Grainger & Allison’s Diagnostic Radiology: A Textbook of Medical Imaging [7th Edition] 0702075248, 9780702075247, 9780702075629, 9780702075612

Table of contents :
Cover......Page 1
Grainger & Allison’s Diagnostic Radiology......Page 3
Copyright Page......Page 4
Table Of Contents......Page 5
Preface......Page 8
List of Section Editors......Page 9
List of Contributors......Page 10
Dedication......Page 19
Portable Chest Radiography......Page 20
From Single-Slice to Multidetector Computed Tomography......Page 21
High-Resolution Computed Tomography......Page 22
Intravenous Contrast Medium Enhancement and Timing of Computed Tomography Acquisition......Page 23
Additional Postprocessing Techniques......Page 25
Computed Tomography Dose Considerations......Page 26
Ultrasound......Page 28
Endoscopic and Endobronchial Ultrasound......Page 29
Ventilation–Perfusion Scintigraphy......Page 30
Positron Emission Tomography......Page 32
Further Reading......Page 33
The Lungs......Page 34
The Central Airways......Page 35
The Lungs Beyond the Hila......Page 37
The Hila......Page 42
Computed Tomography and Magnetic Resonance Imaging......Page 44
Junction Lines......Page 48
Right Mediastinum Above the Azygos Vein......Page 49
Left Cardiac Border Below the Aortic Arch......Page 51
The Diaphragm......Page 53
Further Reading......Page 54
Soft-Tissue Calcification......Page 55
Ribs......Page 56
Soft-Tissue Tumours......Page 57
Free pleural fluid.......Page 60
Loculated (encysted, encapsulated) pleural fluid.......Page 62
Magnetic Resonance Imaging......Page 63
Positron-Emission Tomography/Computed Tomography......Page 64
Exudates and Transudates......Page 65
Chylothorax......Page 66
Typical Signs.......Page 67
Computed Tomography......Page 68
Adhesions......Page 69
Magnetic Resonance Imaging......Page 70
Diffuse Pleural Disease......Page 71
Pleural Fibroma......Page 72
Tumour-Like Conditions of the Pleura......Page 74
Pleural Aspiration......Page 75
Complications of Image-Guided Pleural Intervention......Page 76
Diaphragm......Page 77
Computed Tomography......Page 78
Diaphragmatic Hernias......Page 79
Diaphragmatic Trauma......Page 80
Neoplasms of the Diaphragm......Page 82
Recent Journal Articles......Page 83
Computed Tomography......Page 84
Ultrasound......Page 85
Thyroid Masses......Page 86
Thymomas......Page 87
Thymic Carcinoma......Page 91
Lymphofollicular Thymic Hyperplasia and Rebound Thymic Hyperplasia......Page 92
Teratomas......Page 93
Non-Seminomatous Germ-Cell Tumours......Page 94
Malignant Lymphoma and Leukaemia......Page 95
Low-Attenuation Nodes......Page 96
Lymph Node Enlargement......Page 97
Foregut Duplication Cysts......Page 99
Bronchogenic Cysts......Page 100
Sympathetic Ganglion Tumours......Page 102
Extramedullary Haematopoiesis......Page 105
Lymphangiomas (Cystic Hygromas)......Page 106
Fat-Containing Hernias......Page 107
Fibrosing Mediastinitis......Page 108
Pneumomediastinum......Page 109
Pericardium......Page 112
Pericardial Effusion......Page 113
Pericarditis......Page 115
Constrictive Pericarditis......Page 116
Pericardial Neoplasms......Page 117
Further Reading......Page 119
Clinical Utility and Limitations of Chest Radiography and Computed Tomography......Page 121
Patterns of Pulmonary Infection......Page 122
Complications of Pneumonia......Page 123
Integrating Clinical and Imaging Findings......Page 124
Moraxella catarrhalis.......Page 125
Coxiella burnetii (Rickettsial Pneumonia).......Page 126
Escherichia coli.......Page 127
Influenza A.......Page 128
Epstein–Barr virus (EBV).......Page 129
Cytomegalovirus (CMV).......Page 130
Middle East respiratory syndrome.......Page 131
Changing Spectrum of Human Immunodeficiency Virus Infections: 40 Years Later......Page 132
Pulmonary Non-Tuberculous Mycobacteria (NTMB)......Page 133
Aspergillus Infection......Page 134
Pneumocystis jiroveci......Page 135
Histoplasmosis......Page 136
Strongyloidiasis......Page 137
Echinococcosis (Hydatid Disease)......Page 138
Schistosomiasis......Page 140
Further Reading......Page 141
Primary Malignant Neoplasms......Page 143
Benign Neoplasms......Page 146
Tracheobronchial Amyloidosis......Page 148
Sarcoidosis......Page 149
Tracheobronchomalacia......Page 150
Tracheobronchial Fistula and Dehiscence......Page 152
Bronchiectasis......Page 153
Radiographic Findings......Page 155
Computed Tomographic Findings......Page 156
Cystic Fibrosis......Page 158
Allergic Bronchopulmonary Aspergillosis......Page 160
Radiological Findings......Page 161
Computed Tomography Assessment of Air Trapping......Page 163
Pathological Findings......Page 164
Visually Defined Computed Tomography Pattern of Emphysema......Page 165
Associated Features......Page 167
Quantitative Computed Tomography Image Analysis......Page 168
Quantitative Analysis of Emphysema Extent......Page 173
Quantitative Computed Tomography Analysis of Gas Trapping......Page 174
Quantitative Analysis of Airway Dimensions......Page 175
Computed Tomography Findings......Page 176
Quantitative Computed Tomography Imaging of Airways in Asthma......Page 177
Further Reading......Page 179
Radiographic Considerations......Page 180
Indirect Signs of Volume Loss......Page 181
Ancillary Features of Lobar Collapse......Page 182
Utility......Page 183
Other Imaging Techniques in Lobar Collapse......Page 186
Left Upper Lobe Collapse......Page 187
Right Middle Lobe Collapse......Page 188
Right and Left Lower Lobe Collapse......Page 189
Whole Lung Collapse......Page 193
Combinations of Lobar Collapse......Page 194
Further Reading......Page 195
Treatment-Predictive Biomarkers......Page 196
Re-biopsy......Page 197
Chest Radiographic Screening......Page 198
Radiation Dose Considerations......Page 199
The Future of Screening......Page 200
Management of Small Pulmonary Nodules......Page 201
Nodule Density......Page 202
Subsolid Nodules......Page 203
Positron Emission Tomography/Computed Tomography......Page 204
Lung Cancer Staging—the 8Th Edition of the TNM Staging System......Page 206
Small Cell Lung Cancer......Page 207
The Current Standards of Computer Tomography Technology......Page 208
Calcification......Page 210
Collapse/Consolidation in Association With Central Tumours......Page 211
Chest Wall Invasion......Page 213
Pleural Involvement......Page 215
Pulmonary Sarcoma and Other Primary Malignant Neoplasms......Page 216
Other Benign Pulmonary Neoplasms......Page 217
Follicular Bronchiolitis......Page 218
Other Findings in Pulmonary Lymphoma......Page 219
Metastases......Page 220
Lymphangitic Carcinomatosis......Page 221
Further Reading......Page 222
Nodular Pattern......Page 223
Cystic Pattern......Page 224
Usual Interstitial Pneumonia/Idiopathic Pulmonary Fibrosis......Page 225
Respiratory Bronchiolitis–Interstitial Lung Disease and Desquamative Interstitial Pneumonia......Page 228
Lymphoid Interstitial Pneumonia......Page 230
Idiopathic Pleuroparenchymal Fibroelastosis......Page 231
High-Resolution Computed Tomography Features......Page 232
Hypersensitivity Pneumonitis......Page 233
Langerhans Cell Histiocytosis......Page 234
Lymphangioleiomyomatosis......Page 235
Rheumatoid Disease......Page 236
Sjögren Syndrome......Page 237
Progressive Systemic Sclerosis (Scleroderma)......Page 238
Classical HRCT Findings......Page 239
Eosinophilic Granulomatosis With Polyangiitis (Formerly Churg–Strauss Syndrome)......Page 240
Diffuse Alveolar Damage......Page 241
Eosinophilic Pneumonia......Page 242
The International Labour Office Classification......Page 243
Asbestos-Related Disease......Page 244
Asbestosis......Page 245
Further Reading......Page 246
Anatomical Considerations......Page 248
Pulmonary Oedema......Page 249
Interstitial Oedema......Page 251
Alveolar Oedema......Page 253
Diffuse Pulmonary Haemorrhage......Page 254
Antibasement Membrane Antibody Disease (Goodpasture’s Syndrome)......Page 255
Organising Pneumonia......Page 256
Chronic Eosinophilic Pneumonia......Page 260
Alveolar Microlithiasis......Page 261
Further Reading......Page 262
Cardiac Magnetic Resonance......Page 263
Cardiac Axis Imaging Planes......Page 264
Normal Anatomy on Cardiac Magnetic Resonance Images......Page 266
Computed Tomography Imaging Techniques......Page 267
Coronary Arteries by Computed Tomography......Page 269
Valves......Page 274
Pulmonary Veins......Page 277
Other Structures......Page 278
Echocardiography......Page 279
Valves......Page 283
Further Reading......Page 285
Clinical Presentation......Page 287
Step 1—Atrial Situs......Page 288
Step 3—Ventriculo-Arterial Connection......Page 289
Magnetic Resonance Imaging......Page 290
The Pulmonary Vasculature......Page 291
Atrial Septal Defects......Page 294
Key imaging goals......Page 296
Key imaging goals......Page 297
Abnormalities of the Aortic Arch and Vascular Rings......Page 298
Aortic Valve Disease......Page 299
Anomalous Coronary Arteries......Page 301
Tetralogy of Fallot......Page 302
Transposition of the Great Arteries......Page 303
Key imaging goals......Page 304
Key imaging goals......Page 306
Anomalous Pulmonary Venous Connection/Drainage......Page 307
Key imaging goals following stage 3: post-total cavopulmonary connection (post-TCPC)......Page 309
Hybrid Catheter/Cardiac Magnetic Resonance Imaging Laboratory......Page 310
Conclusion......Page 311
Further Reading......Page 312
Magnetic Resonance Imaging......Page 313
Hypertrophic Pattern......Page 314
Dilated Phenotype......Page 317
Restrictive Phenotype......Page 323
Arrhythmogenic Right Ventricular Cardiomyopathy......Page 327
Unclassified Cardiomyopathy......Page 328
Myocarditis......Page 330
Mitral Valve Disease......Page 332
Chordal Rupture......Page 333
Functional Mitral Regurgitation......Page 334
Mitral Stenosis......Page 335
Tricuspid Valve Disease......Page 336
Aortic Stenosis......Page 337
Aortic Regurgitation......Page 339
New Magnetic Resonance Imaging Techniques......Page 340
Prosthetic Cardiac Valves......Page 341
Infective Endocarditis......Page 344
Tumours of the Heart......Page 346
Cardiac Myxoma......Page 348
Rhabdomyomas......Page 352
Papillary Fibroelastoma......Page 353
Sarcomas......Page 354
Rhabdomyosarcoma.......Page 355
Anatomy......Page 356
Pericardial Effusion......Page 358
Pericardial Inflammation......Page 360
Constrictive Pericarditis......Page 362
Pericardial Masses......Page 364
Further Reading......Page 365
Pulmonary Arteries......Page 366
Pulmonary Circulation Physiology......Page 367
Interstitial Oedema (Grade 2)......Page 368
Alveolar Oedema (Grade 3)......Page 369
Pulmonary Arterial Hypertension......Page 370
Vascular Signs......Page 374
Cardiac Signs......Page 375
Parenchymal Signs......Page 376
Pulmonary Arteriovenous Malformations......Page 378
Clinical (Pre-Test) Probability Estimate and D-Dimer Testing......Page 380
Compression Ultrasound of the Legs......Page 381
Ventilation–Perfusion Scintigraphy......Page 382
Computed tomography pulmonary angiography protocol.......Page 383
Computed tomography perfusion.......Page 384
Chronic Pulmonary Thromboembolism......Page 387
Acknowledgements......Page 388
Further Reading......Page 389
Pathophysiology of Ischaemic Heart Disease......Page 390
Coronary Artery Imaging......Page 392
Functional Imaging......Page 397
Stress Imaging......Page 401
Myocardial Infarct Imaging......Page 405
Myocardial Viability Imaging......Page 411
Imaging of Complications Related to Ischaemic Heart Disease......Page 415
Differential Diagnosis in Ischaemic Heart Disease......Page 416
Further Reading......Page 417
Chest X-Ray and Echocardiography......Page 419
Acute Diseases......Page 420
Chronic Diseases......Page 421
Magnetic resonance imaging.......Page 422
Computed tomography.......Page 424
Intramural Haematoma......Page 426
Magnetic Resonance Imaging......Page 427
Penetrating Atherosclerotic Ulcer......Page 428
Imaging......Page 430
Aortic Aneurysms......Page 432
Atherosclerotic Aortic Aneurysms......Page 433
Computed tomography.......Page 434
Mycotic Aneurysms......Page 435
Aortic Sinus Aneurysms......Page 438
Visceral Malperfusion......Page 439
Preoperative or Preinterventional Evaluation......Page 441
Traumatic Aortic Injury......Page 442
Endovascular treatment of type B dissection.......Page 443
Midaortic Syndrome......Page 444
Granulomatous vasculitis (Takayasu disease).......Page 445
Prognosis and treatment.......Page 446
Chronic Aortic Occlusive Disease......Page 447
Vascular Rings......Page 448
Double Aortic Arch......Page 450
Right aortic arch with mirror-image branching and retro-oesophageal ligamentum arteriosum.......Page 451
Imaging......Page 452
Chest X-Ray......Page 453
Magnetic Resonance Imaging......Page 454
Management......Page 455
Aortic Atresia......Page 456
Further Reading......Page 458
Pneumoperitoneum......Page 459
Gas in Bowel Wall......Page 461
Gas in Other Organs......Page 462
Distinction Between Small- and Large-Bowel Dilatation......Page 463
Small-Bowel Dilatation......Page 464
Pseudo-Obstruction......Page 467
Small-Bowel Ischaemia......Page 468
Pseudomembranous Colitis......Page 470
Ultrasound in Appendicitis......Page 471
Computed Tomography in Appendicitis......Page 472
Imaging the Abdomen With Computed Tomography: Radiation Issues......Page 473
Iterative Reconstruction Algorithms......Page 474
Role of MRI in the Acute Abdomen......Page 475
Radiation Dose Reduction in Clinical Practice......Page 476
Further Reading......Page 477
Plain Radiography......Page 478
Fluoroscopy......Page 479
Computed Tomography......Page 480
Oesophageal Cancer......Page 481
Computed Tomography for Oesophageal Cancer......Page 482
Endoscopic Ultrasound for Oesophageal Cancer......Page 484
Positron-Emission Tomography–Computed Tomography for Oesophageal Cancer......Page 485
Benign Lesions......Page 486
Hiatus Hernia......Page 488
Columnar-lined oesophagus.......Page 489
Motility Disorders......Page 490
Oesophageal Varices......Page 491
Trauma......Page 492
Further Reading......Page 493
Fluoroscopy......Page 494
Multidetector Computed Tomography......Page 496
Gastric Ulcer......Page 498
Atrophic Gastritis......Page 501
Ménétrier Disease......Page 502
Corrosive Ingestion......Page 503
Mucosal Polyps......Page 504
Mesenchymal Tumours......Page 505
Early Gastric Cancer......Page 507
Staging of Gastric Cancer......Page 508
Gastric Lymphoma......Page 510
Hiatus Hernia......Page 511
Gastric Volvulus......Page 512
Gastric Pneumatosis......Page 513
Gastric Distention......Page 514
Gastroparesis......Page 515
Complications of Bariatric Surgery......Page 517
Non-Bariatric Gastric Surgeries......Page 518
Complications of Non-Bariatric Gastric Surgery......Page 520
Further Reading......Page 521
Peptic Ulceration......Page 523
Diverticula......Page 524
Secondary Involvement......Page 525
Intramural Haematoma......Page 526
Radiological Investigation......Page 527
Computed Tomography......Page 528
Nuclear Medicine Studies......Page 529
Radiological Appearances......Page 530
Neoplasms......Page 533
Carcinoid Tumour......Page 534
Adenocarcinoma......Page 535
Benign Neoplasms......Page 536
Tuberculosis......Page 538
Whipple Disease......Page 539
Mechanical Small-Bowel Obstruction......Page 540
Occult Gastrointestinal Bleeding......Page 542
Jejunal Diverticula......Page 543
Ileal Diverticula......Page 544
Eosinophilic Gastroenteritis......Page 545
Peritoneal Spaces......Page 546
The Mesenteries......Page 547
The Omentum......Page 548
Peritonitis......Page 549
Tuberculosis Peritonitis......Page 550
Mesenteric Lymphadenitis......Page 551
Non-Inflammatory Mesenteric Oedema......Page 552
Epiploic Appendagitis......Page 553
Direct Invasion and Along Mesenteric– Ligamentous Attachments......Page 554
Intraperitoneal Seeding and Peritoneal Carcinomatosis......Page 555
Cytoreductive Surgery......Page 557
Mesentery......Page 559
Further Reading......Page 562
Anatomy......Page 564
Radiological Investigation......Page 565
Polyps......Page 567
Familial adenomatous polyposis.......Page 569
Computed tomography colonography.......Page 570
Colorectal Cancer......Page 571
Colon Cancer......Page 572
Rectal Cancer......Page 574
Anal Cancer......Page 575
Appendix Tumours......Page 576
Secondary Cancers......Page 577
Diverticulitis......Page 578
Epiploic Appendagitis......Page 582
Differential Features......Page 583
Ischaemic Colitis......Page 584
Infectious Colitis......Page 585
Parasitic Colitis......Page 586
Acute Fulminant Colitis......Page 587
Large-Bowel Strictures......Page 588
Volvulus......Page 589
Endometriosis......Page 590
Functional Disorders of the Anorectum......Page 591
Anal Fistula......Page 592
Further Reading......Page 593
Gallbladder Anatomical Variants......Page 594
Ultrasound......Page 595
Intraoperative Cholangiography......Page 596
Low Phospholipid-Associated Cholelithiasis......Page 597
Gangrenous Cholecystitis......Page 598
Adenomyomatous Hyperplasia......Page 599
Gallbladder Polyps......Page 600
Gallbladder Carcinoma......Page 601
Gallbladder Metastases and Lymphoma......Page 602
Role of Radiology in Investigation of Jaundice......Page 603
Cholangiography......Page 606
Primary Sclerosing Cholangitis......Page 607
IgG4-Related Disease......Page 608
Mirizzi Syndrome......Page 609
Fascioliasis......Page 610
Cholangiocarcinoma......Page 611
Haemobilia......Page 613
Hilar Strictures—Special Considerations......Page 614
Further Reading......Page 615
Vascular Anatomy Variation (Fig. 23.5)......Page 617
Technique (Fig. 23.9)......Page 618
Intravenous Contrast Agents (Fig. 23.12)......Page 622
Hepatic Steatosis......Page 625
Cirrhosis......Page 626
Haemochromatosis and Iron Overload......Page 627
Malignant Diffuse Disease......Page 628
Calcification......Page 629
Pneumobilia......Page 630
Cysts......Page 631
Haemangioma......Page 632
Focal Nodular Hyperplasia......Page 635
Hepatic Adenoma......Page 637
Focal Fat......Page 638
Biliary Hamartomas (Fig. 23.44)......Page 642
Hepatocellular Carcinoma......Page 643
Fibrolamellar Carcinoma (Fig. 23.53)......Page 648
Metastases......Page 652
Budd–Chiari Syndrome......Page 655
Portal Venous Hypertension......Page 656
Portal Vein Thrombosis......Page 657
Intrahepatic Portosystemic Shunts......Page 659
Hepatic Trauma......Page 660
Living Donor Assessment......Page 661
Devices......Page 662
Practical Procedural Issues......Page 663
Anatomy......Page 664
Infarction (Figs 23.78 and 23.79)......Page 665
Cysts (Fig. 23.80)......Page 669
Hamartomas and Lymphangiomas (Fig. 23.82)......Page 670
Lymphoma (Fig. 23.84)......Page 671
Portal Hypertension/Splenic Vein Thrombosis......Page 672
Further Reading......Page 674
Annular Pancreas......Page 675
Acute Pancreatitis......Page 676
Imaging in Acute Pancreatitis......Page 680
Interstitial Oedematous Pancreatitis......Page 681
Necrotising Pancreatitis......Page 682
Complications of Acute Pancreatitis......Page 684
Chronic Pancreatitis......Page 686
Paraduodenal Pancreatitis......Page 688
Autoimmune Pancreatitis......Page 690
Ductal Adenocarcinoma......Page 694
Imaging Investigations......Page 695
Imaging Appearances......Page 696
Local Staging......Page 697
Follow-Up......Page 699
Functioning Tumours......Page 700
Cystic Masses......Page 703
What to Do With Small Incidental Cystic Lesions?......Page 709
Osler–Weber–Rendu Disease......Page 710
Biopsy of Solid Pancreatic Lesions and Fine-Needle Aspiration of Cystic Pancreatic Lesions......Page 713
Drainage of Pancreatic/Peripancreatic Fluid Collections......Page 714
Further Reading......Page 716
Imaging of Renal and Ureteric Stones......Page 718
Risk Stratification of Patients With Haematuria for Triage to Computed Tomography......Page 719
Imaging Features of Urothelial Tumours on Computed Tomography......Page 720
Computed Tomography......Page 724
Ultrasound......Page 726
Radiography......Page 727
Magnetic Resonance Urography......Page 728
Iodinated Contrast Agents......Page 729
Imaging of Acute Pyelonephritis......Page 730
Computed tomography.......Page 731
Magnetic resonance imaging.......Page 732
Renal scintigraphy.......Page 733
Imaging of Emphysematous Pyelonephritis......Page 734
Xanthogranulomatous Pyelonephritis......Page 735
Imaging of Xanthogranulomatous Pyelonephritis......Page 737
Pyonephrosis......Page 738
Imaging of Pyonephrosis......Page 739
Imaging of Renal Tuberculosis......Page 740
Acute Bacterial Prostatitis and Prostatic Abscess......Page 741
Imaging of acute prostatitis and prostatic abscess.......Page 742
Imaging of chronic prostatitis.......Page 745
Further Reading......Page 746
Normal Urinary Tract Anatomy......Page 747
Renal......Page 751
Ureter and Pelvis......Page 753
Bladder and Urethra......Page 758
Retrograde Urography......Page 760
Retrograde Urethrography and Voiding Cystourethrogram......Page 761
Ultrasound......Page 762
Nuclear Medicine......Page 763
Computed Tomography......Page 765
Magnetic Resonance Imaging......Page 767
Conventional Radiography......Page 769
Further Reading......Page 770
Duplicated Collecting System (Ureteral Duplication)......Page 772
Autosomal Recessive Polycystic Kidney Disease......Page 773
Cysts......Page 774
Bosniak Renal Cyst Classification System......Page 775
Renal Cysts Associated With von Hippel–Lindau Syndrome (vHL)......Page 779
Acquired Cystic Disease of the Kidney......Page 781
Dual-Energy Computed Tomography......Page 784
Nephrocalcinosis......Page 785
Further Reading......Page 786
Magnetic Resonance Imaging......Page 787
Complicated Cysts......Page 788
Adult Polycystic Kidney Disease (ADPKD)......Page 789
Renal Abscesses......Page 790
Intrarenal Vascular Masses......Page 791
Adenoma and Oncocytoma......Page 792
Renal cell carcinoma.......Page 793
Staging of renal cancer.......Page 795
Sarcoma.......Page 796
Squamous cell carcinoma.......Page 797
Further Reading......Page 798
Ultrasound......Page 799
Magnetic Resonance Imaging......Page 800
Catheter-Based Angiography and Fluoroscopy......Page 801
Renal Artery Stenosis......Page 802
Acute Tubular Necrosis......Page 803
Chronic Rejection......Page 804
Lymphocele......Page 805
Ureteric Strictures......Page 806
Cancer and Transplant......Page 807
Renal Transplant Biopsy......Page 808
Radiological Evaluation of Potential Donor Kidneys......Page 809
Further Reading......Page 810
Risk Factors......Page 811
Management......Page 812
Optimisation of Computed Tomography Urography Technique......Page 813
Imaging Findings in Upper Tract Urothelial Carcinoma......Page 814
Magnetic Resonance Urography......Page 815
Conclusions and Summary......Page 816
Clinical Detection......Page 818
Computed Tomography and Magnetic Resonance Urographic Appearance of Bladder Cancers......Page 819
Sensitivity and Specificity of Computed Tomography and Magnetic Resonance Urography in Detecting Bladder Cancers......Page 820
Imaging for Local Staging of Bladder Cancer......Page 823
Upper Tract Evaluation......Page 825
Treatment and Follow-Up......Page 831
Use of Imaging to Identify Tumour Response to Chemotherapy......Page 832
Other Urinary Tract Malignancies......Page 833
Benign Bladder Lesions......Page 834
Recent Journal Articles......Page 838
Anatomy......Page 840
Risk Stratification......Page 841
Prostate Imaging—Reporting and Data System......Page 843
Transition Zone......Page 844
Diffusion-Weighted Imaging......Page 845
Nodal Staging......Page 848
Non-Nodal Metastases......Page 850
Local Disease Stage......Page 851
Post-Treatment Evaluation......Page 852
Prostate Biopsy......Page 853
Further Reading......Page 856
Scrotal Masses......Page 857
Malignant Testicular Pathology......Page 858
Computed Tomography......Page 859
Non-Malignant Focal Testicular Lesions......Page 860
Extratesticular Scrotal Lesions......Page 861
Acute Epididymitis......Page 862
Cryptorchidism......Page 864
Male Infertility......Page 865
Evaluation of the Soft Tissues of the Penis......Page 867
Further Reading......Page 868
Computed Tomography......Page 869
Detection, Diagnosis and Staging......Page 871
Ultrasound......Page 872
Magnetic Resonance Imaging......Page 873
Recommended Imaging Approach......Page 874
Magnetic Resonance Imaging......Page 875
Positron-Emission Tomography/Computed Tomography......Page 876
Recommended Imaging Approach......Page 877
Ovarian Carcinoma......Page 878
Ultrasound......Page 879
Computed Tomography......Page 880
Magnetic Resonance Imaging......Page 881
Conclusion......Page 882
Further Reading......Page 884
Congenital Anomalies of the Female Genital Tract......Page 886
Class II: Unicornuate Uterus......Page 888
Class IV: Bicornuate Uterus......Page 889
Vaginal Anomalies......Page 890
Imaging of Ambiguous Genitalia......Page 891
Fibroids......Page 892
Endometrial Hyperplasia......Page 893
Magnetic Resonance Imaging......Page 897
Pelvic Inflammatory Disease......Page 898
Computed Tomography......Page 901
Ovarian Vein Thrombosis......Page 902
Functional Ovarian Cysts......Page 904
Epithelial Tumours......Page 905
Stromal Cell Tumours......Page 906
Urethral Bulking Agents......Page 909
Mid-Urethral Slings......Page 911
Ectopic Pregnancy......Page 913
Abnormal Placental Positions: Placenta Previa and Low-lying Placenta......Page 914
Placenta Accreta, Percreta and Increta......Page 915
Uterine Arteriovenous Malformation (AVM)......Page 916
Further Reading......Page 918
Renal Trauma......Page 919
Imaging Technique for Renal Injury......Page 920
Computed Tomography Findings of Grade II Renal Injury......Page 921
Computed Tomography Findings of Grade IV Renal Injury......Page 923
Active Bleeding......Page 925
Traumatic Renal Pseudoaneurysm......Page 926
Ureteral Injury......Page 927
Bladder Injury......Page 928
Computed Tomography Findings......Page 929
Imaging Findings of Urethral Injury......Page 930
Imaging Technique and Findings......Page 931
Imaging Technique and Imaging Findings......Page 932
Further Reading......Page 933
Physiology......Page 934
Incidentally Detected Adrenal Mass......Page 935
Lesion Size and Contour......Page 937
Contrast Enhancement and Contrast Washout Characteristics......Page 938
Histogram Analysis Method......Page 939
Positron-Emission Tomography......Page 940
Adrenal Scintigraphy......Page 945
Myelolipoma......Page 946
Imaging Functional Disorders of the Adrenal Gland......Page 948
Cushing’s Syndrome......Page 949
Adrenal carcinoma.......Page 950
ACTH-dependent Cushing’s syndrome.......Page 953
Primary Hyperaldosteronism (Conn’s Syndrome)......Page 954
Congenital adrenal hyperplasia.......Page 955
Phaeochromocytomas......Page 956
Neuroblastoma and Ganglioneuroblastoma......Page 958
Primary Adrenal Hypofunction......Page 959
Secondary Adrenal Hypofunction......Page 960
Further Reading......Page 962
Conventional Radiography......Page 963
Ultrasound......Page 964
Elastography......Page 965
Magnetic Resonance Imaging......Page 966
Metal Artefact Reduction......Page 967
Nuclear Medicine......Page 968
Arthrography......Page 969
Further Reading......Page 970
The Shoulder......Page 971
Rotator Cuff Disease......Page 972
Glenohumeral Joint Instability......Page 974
The Acromioclavicular Joint......Page 976
Tendons......Page 978
Ligaments......Page 980
Hand and Wrist......Page 981
Wrist Ligaments......Page 982
Triangular Fibrocartilage......Page 984
Wrist Tendons......Page 985
Ulnar Collateral Ligament of the Thumb......Page 986
Labrum and Cartilage......Page 988
Muscle and Tendon......Page 989
Bursae......Page 990
The Knee......Page 991
Anterior Cruciate Ligament......Page 992
Medial Collateral Ligament......Page 994
Lateral Collateral Ligament Complex and Posterolateral Corner......Page 995
The Extensor Mechanism and Patellofemoral Joint......Page 996
Bursae......Page 997
Tendons......Page 998
Tarsal Coalition......Page 1000
Other Soft-Tissue Abnormalities......Page 1001
Further Reading......Page 1002
Periosteal Reaction......Page 1003
Computed Tomography and Magnetic Resonance Imaging in Diagnosis and Staging......Page 1004
Osteochondroma......Page 1008
Radiological Features......Page 1009
Enchondromatosis with haemangiomas (Maffucci syndrome).......Page 1010
Chondroblastoma......Page 1012
Radiological Features......Page 1015
Radiological Features......Page 1016
Osteoblastoma......Page 1019
Radiological Features......Page 1020
Non-Ossifying Fibroma......Page 1022
Radiological Features......Page 1023
Radiological Features......Page 1024
Haemangioma......Page 1025
Radiological Features......Page 1026
Radiological Features......Page 1027
Radiological Features......Page 1028
Fibrous Dysplasia......Page 1030
Radiological Features......Page 1031
Radiological Features......Page 1033
Further Reading......Page 1034
Clinical......Page 1035
Radiological Features......Page 1036
Breast.......Page 1038
Melanoma.......Page 1039
Radiological Investigation of Bone Metastases......Page 1040
Bone Metastases in Children......Page 1041
Primary Malignant Neoplasms of Bone......Page 1042
Clinical Presentation......Page 1043
Imaging Features......Page 1045
Periosteal Chondrosarcoma......Page 1046
Clinical presentation.......Page 1047
Imaging features.......Page 1048
Other varieties of central osteosarcoma.......Page 1049
Periosteal osteosarcoma.......Page 1050
Paget sarcoma.......Page 1051
Post-radiation sarcoma.......Page 1052
Ewing Sarcoma and Primitive Neuroectodermal Tumour......Page 1053
Imaging Features......Page 1054
Malignant Vascular Tumours......Page 1055
Adamantinoma......Page 1056
Further Reading......Page 1058
Radiographs......Page 1060
Ultrasound......Page 1061
World Health Organisation Classification of Soft-Tissue Tumours......Page 1062
Magnetic Resonance Imaging......Page 1063
Atypical Lipomatous Tumour/ Well-Differentiated Liposarcoma......Page 1064
Nodular Fasciitis......Page 1065
Desmoid-Type Fibromatosis......Page 1067
Tenosynovial Giant Cell Tumour......Page 1068
Soft-Tissue Chondromas......Page 1070
Benign Nerve Sheath Tumours......Page 1071
Malignant Peripheral Nerve Sheath Tumours......Page 1072
Traumatic Neuroma......Page 1073
Morton’s Neuroma......Page 1074
Myxoma......Page 1075
Non-Neoplastic Tumour Mimics......Page 1077
Traumatic Lesions......Page 1078
Further Reading......Page 1079
Bone Formation and Turnover......Page 1081
Bone Growth and Development......Page 1082
Radiological Features......Page 1084
Spine in Osteoporosis......Page 1085
Vertebroplasty and Kyphoplasty......Page 1086
Osteoporotic Fractures......Page 1087
Generalised Osteoporosis......Page 1088
Idiopathic juvenile osteoporosis.......Page 1089
Osteogenesis Imperfecta......Page 1091
Type II (Lethal Perinatal)......Page 1092
Dual Energy X-Ray Absorptiometry......Page 1093
Quantitative Ultrasound......Page 1095
Radiogrammetry......Page 1096
Primary Hyperparathyroidism......Page 1097
Subperiosteal erosions.......Page 1098
Osteosclerosis.......Page 1099
Osteoporosis.......Page 1100
Pseudohypoparathyroidism......Page 1101
Vitamin D Deficiency......Page 1102
Rickets.......Page 1103
Osteomalacia.......Page 1104
X-linked hypophosphataemia.......Page 1105
Hypophosphatasia......Page 1106
Osteopetrosis......Page 1107
Hypervitaminosis A......Page 1108
Further Reading......Page 1109
Entheseal disease.......Page 1110
Distribution of Joint Involvement......Page 1111
Osteoarthritis......Page 1112
Radiographic Findings......Page 1113
Hands and Wrists......Page 1114
Advanced Imaging......Page 1115
Radiographic Features......Page 1116
Sero-Negative Arthritis......Page 1118
Spinal disease.......Page 1119
Peripheral joint involvement.......Page 1120
Joints of the hands and feet.......Page 1122
Reactive Arthritis......Page 1123
Chronic tophaceous gout.......Page 1124
Pyrophosphate arthropathy.......Page 1125
Intra-articular hydroxyapatite deposition disease.......Page 1126
Systemic Lupus Erythematosus......Page 1127
Other musculoskeletal manifestations of haemophilia.......Page 1128
Neuropathic Arthropathy......Page 1129
Synovial (Osteo)-Chondromatosis......Page 1130
Lipoma Arborescens......Page 1131
Sarcoid......Page 1132
Hypertrophic Pulmonary Osteoarthropathy......Page 1133
Haemochromatosis......Page 1134
Further Reading......Page 1135
Occult-Complete Fractures......Page 1136
Insufficiency, Pathological and Stress Fractures......Page 1138
Growth Plate Injuries......Page 1139
Type III......Page 1140
Anterior Dislocation......Page 1141
Posterior Dislocation......Page 1142
Luxatio Erecta......Page 1143
Lateral epicondyle fractures.......Page 1144
Medial epicondyle fractures.......Page 1145
Adults......Page 1146
Children......Page 1147
Chauffeur Fracture (Hutchinson Fracture)......Page 1148
Scaphoid......Page 1149
Triquetrum......Page 1150
Scapholunate Disassociation......Page 1151
Condylar Fractures......Page 1152
Articular Fractures......Page 1153
Anterior Compression Injuries......Page 1154
Anatomy......Page 1155
Posterior Wall Fractures......Page 1157
Complex Fractures......Page 1159
Pelvic Insufficiency and Stress Fractures......Page 1160
Intertrochanteric and Subtrochanteric Fracture......Page 1161
Atypical Femoral Fractures......Page 1163
Tibial Plateau Fractures......Page 1164
Anterior Cruciate Ligament Avulsion Fracture......Page 1165
Patella Fractures......Page 1166
The Ankle......Page 1167
Supination-Lateral Rotation (Fig. 45.97)......Page 1168
Talar Dome Fractures......Page 1169
Calcaneal Fractures......Page 1170
Foot Injuries......Page 1174
Lisfranc Injury......Page 1175
Further Reading......Page 1176
Further Reading......Page 0
Neural Arch......Page 1217
Facet Joints......Page 1218
Intervertebral Disc—Symphysis......Page 1219
Neural Structures—Spinal Cord, Spinal Nerves, Dura Mater......Page 1220
Imaging Techniques......Page 1221
Plain Radiography......Page 1222
Myelography......Page 1223
Spinal Angiography......Page 1224
Computed Tomography......Page 1226
Gradient-Echo Imaging......Page 1227
Diffusion-Weighted Imaging......Page 1229
Artefacts......Page 1230
Cerebrospinal Fluid Pulsation Artefacts......Page 1231
Single-Photon Emission Computed Tomography......Page 1232
Further Reading......Page 1234
Age-Related Changes in the Intervertebral Disc......Page 1235
Degenerative Disc Disease......Page 1236
Annular Tears......Page 1237
Disc Herniation......Page 1238
Vertebral End Plates and Bone Marrow Changes......Page 1240
Osteoarthritis of the Facet Joints......Page 1247
Degenerative Cysts Arising From the Facet Joints......Page 1249
Cysts of the Ligamentum Flavum......Page 1250
Ligamentum Flavum Hypertrophy......Page 1251
Spinous Process Abnormalities and Associated Ligamentous Changes......Page 1253
Degenerative Spinal Canal Stenosis......Page 1254
Degenerative Scoliosis......Page 1257
Further Reading......Page 1258
Magnetic Resonance Imaging......Page 1259
Bone Scintigraphy......Page 1261
Classification of Spinal Tumours......Page 1262
Myxopapillary Ependymoma......Page 1263
Astrocytoma......Page 1264
Ganglioglioma......Page 1266
Cavernous malformation (cavernoma).......Page 1267
Intradural Extramedullary Tumours......Page 1269
Meningioma......Page 1270
Intradural Extramedullary Tumour Mimics......Page 1271
Metastatic Spine Disease......Page 1272
Osteoid Osteoma/Osteoblastoma......Page 1275
Aneurysmal Bone Cyst......Page 1277
Eosinophilic Granuloma (Unifocal Langerhans Cell Histiocytosis)......Page 1278
Benign Notochordal Cell Tumours......Page 1279
Chordoma.......Page 1280
Multiple Myeloma and Plasmacytoma......Page 1282
Chondrosarcoma......Page 1283
Ewing Sarcoma......Page 1284
Osteosarcoma......Page 1285
Further Reading......Page 1286
Neuromyelitis Optica......Page 1287
Systemic Lupus Erythematosus......Page 1289
Demyelinating Polyneuropathies......Page 1291
Spinal Dural Arteriovenous Fistula......Page 1293
Spinal Arteriovenous Malformations......Page 1295
Spinal Cord Vasculitis......Page 1296
Spinal Cord Infection Can Be Bacterial, Viral, Fungal or Parasitic in Origin......Page 1297
Intramedullary Lipoma......Page 1299
Diastematomyelia......Page 1300
Chiari Type II Malformations......Page 1303
Anterior Thoracic Meningocele With Ventral Herniation of the Spinal Cord......Page 1305
Intraspinal Arachnoid Cyst......Page 1307
Syringomyelia......Page 1308
Subacute Combined Degeneration of the Spinal Cord......Page 1309
Further Reading......Page 1310
Principles of Spinal Surgery......Page 1311
Early Complications......Page 1313
Late Complications......Page 1317
Failed Back Surgery Syndrome......Page 1320
Further Reading......Page 1323
Conventional Radiography......Page 1324
Computed Tomography......Page 1327
Rotatory Subluxation......Page 1328
Odontoid Fractures......Page 1329
Hyperflexion-Rotation Injury......Page 1330
Classification Systems......Page 1331
Flexion-Compression and Flexion-Distraction Injuries......Page 1333
Burst Fractures......Page 1334
The Rigid Spine......Page 1336
Spinal Cord......Page 1337
Brachial Plexus Injury......Page 1338
Further Reading......Page 1339
Cerebral Cortex, Lobar Anatomy and Deep Grey Matter Structures......Page 1340
Limbic System, Hypothalamus and Pituitary Gland......Page 1343
Ventricular System and Subarachnoid Space......Page 1347
Brainstem......Page 1351
Posterior Circulation......Page 1353
Anastomotic Pathways......Page 1354
Intracranial Veins......Page 1356
Introduction......Page 1357
Basic Principles and Scanning Parameters......Page 1358
Image Reconstruction and Display......Page 1359
Image Appearance......Page 1360
Indications, Risks and Benefits......Page 1361
Artefacts and Limitations......Page 1362
Signal generation.......Page 1363
Image appearance and basic applications.......Page 1364
Functional Magnetic Resonance Imaging......Page 1365
Interventional Magnetic Resonance Imaging......Page 1366
Indications, risks and benefits.......Page 1367
MRI artefacts.......Page 1368
Positron Emission Tomography......Page 1369
Digital Subtraction Angiography......Page 1370
Computed Tomography Angiography......Page 1371
Computed tomography perfusion.......Page 1372
Magnetic Resonance Angiography......Page 1373
Further Reading......Page 1374
Modality and Technique......Page 1376
Calvarial/Skull Fractures......Page 1379
Skull Base Fracture......Page 1381
Vascular Injury......Page 1382
Traumatic Intra-Cranial Haemorrhage......Page 1384
Subdural Haemorrhage......Page 1387
Traumatic Subarachnoid Haemorrhage......Page 1392
Diffuse Axonal Injury......Page 1393
Chronic Imaging Findings of Traumatic Parenchymal Injury......Page 1395
Secondary Brain Injury......Page 1396
Further Reading......Page 1399
Structural Magnetic Resonance Imaging......Page 1400
Magnetic Resonance Perfusion-Weighted Imaging......Page 1401
Magnetic Resonance Diffusion Imaging......Page 1402
Positron-Emission Tomography......Page 1403
Gliomas......Page 1404
Oligodendroglioma (IDHmut With 1p19q Codeletion)......Page 1406
Isocitrate Dehydrogenase-Mutant Astrocytoma (IDHmut 1p19q Intact)......Page 1407
Midline Glioma, Histone H3 K27M-Mutant......Page 1408
The Role of Advanced Physiological Magnetic Resonance Imaging in Glial Tumours......Page 1409
Dysembryoplastic Neuroepithelial Tumour......Page 1412
Lymphoma......Page 1413
Metastases......Page 1415
Ependymoma......Page 1416
Choroid Plexus Tumours......Page 1417
Meningioma......Page 1418
Meningiomas......Page 1419
Cranial Nerve Sheath Tumours......Page 1420
Skull Base Tumours......Page 1421
Pineal Region Tumours......Page 1422
Pituitary Adenomas......Page 1424
Craniopharyngiomas......Page 1425
Other Sellar Region Tumours......Page 1426
Further Reading......Page 1428
The Penumbra Model......Page 1430
Stroke Classification......Page 1431
Imaging Strategies and Goals in Acute Stroke......Page 1432
Hyperacute Infarct Imaging Signs on Computed Tomography......Page 1433
Perfusion imaging.......Page 1436
Magnetic Resonance Imaging in Acute Stroke......Page 1437
Angiographic imaging.......Page 1440
Additional advanced imaging techniques.......Page 1441
Computed tomography or magnetic resonance imaging?......Page 1442
Subacute and Chronic Infarct Imaging Signs......Page 1444
Imaging Options for Carotid Stenosis......Page 1447
Imaging Signs......Page 1448
Arterial Dissection......Page 1449
Ischaemic Microangiopathy......Page 1450
Vasculitis......Page 1452
Cerebral Venous Thrombosis......Page 1453
Non-Traumatic Intracranial Haemorrhage......Page 1454
Subarachnoid Haemorrhage......Page 1455
Initial Investigation of Acute Subarachnoid Haemorrhage......Page 1456
Aneurysmal Subarachnoid Haemorrhage......Page 1459
Imaging of Incidental Intracranial Aneurysms......Page 1460
Appearance on Computed Tomography and Magnetic Resonance Imaging......Page 1462
Angiography in Intracerebral Haemorrhage......Page 1464
Subdural and Extradural Haemorrhage......Page 1465
Further Reading......Page 1469
Cerebritis and Brain Abscess......Page 1470
Ventriculitis......Page 1471
Tuberculosis......Page 1472
Neurosyphilis......Page 1473
Herpes Simplex Encephalitis......Page 1475
Other Viral Encephalitides......Page 1476
Human Immunodeficiency Virus Encephalopathy......Page 1477
Immune Reconstitution Inflammatory Syndrome......Page 1479
Cysticercosis......Page 1481
Echinococcus (Hydatid Disease)......Page 1483
Malaria......Page 1484
Further Reading......Page 1486
Indications for Imaging......Page 1487
Alheimer Disease......Page 1488
Vascular Dementia......Page 1489
Systemic Causes of Vascular Dementia......Page 1490
Autoimmune Limbic Encephalitis......Page 1491
Introduction and Clinical Overview......Page 1492
Indications for Imaging......Page 1493
Epilepsy—Infectious Aetiology......Page 1494
Epilepsy—Structural Brain Alterations— Genetic Syndromes......Page 1495
Neoplasms......Page 1496
Epilepsy—Infectious Aetiology......Page 1497
Further Reading......Page 1498
Multiple Sclerosis......Page 1499
Magnetic Resonance Imaging......Page 1500
Baló Concentric Sclerosis......Page 1508
Pseudotumoural Lesions......Page 1509
Neuromyelitis Optica Spectrum Disorders......Page 1511
Acute Disseminated Encephalomyelitis......Page 1513
Acute haemorrhagic leukoencephalitis.......Page 1518
Primary Angiitis of the Central Nervous System......Page 1519
Neurosarcoidosis......Page 1522
Behçet Disease......Page 1523
Systemic Lupus Erythematosus......Page 1524
Postradiotherapy Leukoencephalopathy......Page 1525
Radionecrosis......Page 1526
Stroke-Like Migraine Attacks After Radiation Therapy Syndrome......Page 1527
Posterior Reversible Encephalopathy Syndrome......Page 1528
Brain Gadolinium Deposition......Page 1531
Cocaine......Page 1533
Diethylene Glycol......Page 1535
Organic Solvent Poisoning......Page 1536
Krabbe Disease......Page 1538
Wernicke Encephalopathy......Page 1540
Osmotic Demyelination......Page 1542
Hepatic Encephalopathy......Page 1543
Acknowledgements......Page 1545
Further Reading......Page 1550
The Globe......Page 1551
Vascular and Nervous Supply of the Orbit......Page 1552
Persistent Hypertrophic Primary Vitreous......Page 1553
Coloboma......Page 1554
Thyroid Orbitopathy......Page 1555
Idiopathic Orbital Inflammation......Page 1556
Systemic Inflammatory Diseases With Orbital Involvement......Page 1557
Benign Lacrimal Fossa Masses......Page 1558
Nerve Sheath Tumour......Page 1563
Cavernous Haemangioma......Page 1564
Arteriovenous Malformations......Page 1567
Lymphoproliferative Malignancy......Page 1568
Retinoblastoma......Page 1571
Rhabdomyosarcoma......Page 1572
Orbital Trauma......Page 1573
Anatomy......Page 1576
Congenital......Page 1577
Inflammatory/Demyelinating Lesions......Page 1578
Pathologies of the Posterior Visual Pathway (Lateral Geniculate Nucleus, Optic Radiation and Visual Cortex)......Page 1582
Central Gaze Palsies......Page 1584
Further Reading......Page 1586
Keratosis Obturans.......Page 1588
Necrotising Otitis Externa.......Page 1589
Neoplasia of the Auricle and External Auditory Canal.......Page 1590
Cholesteatoma.......Page 1591
Retrofenestral or pericochlear.......Page 1593
Ossicular disruption.......Page 1594
Anatomy and Physiology......Page 1595
Trauma.......Page 1596
Congenital malformations.......Page 1597
Glomus tumours (paragangliomas).......Page 1598
Anatomy and Physiology......Page 1599
Rhinosinusitis......Page 1601
Mucocoeles......Page 1602
Epistaxis......Page 1603
Juvenile Angiofibroma......Page 1604
Sinonasal Malignancy......Page 1605
Anatomy......Page 1606
Anatomy......Page 1607
Sialadenitis......Page 1608
Fractures of the Facial Skeleton......Page 1609
Orbital Blow-Out Fractures......Page 1611
Developmental Abnormalities......Page 1612
Temporomandibular Joint Dysfunction......Page 1613
Arthritides......Page 1614
Injury......Page 1615
Pathology.......Page 1616
Anatomy.......Page 1617
Anatomy.......Page 1618
Radiology and pathology.......Page 1619
Anatomy......Page 1620
Radiology and Pathology......Page 1621
Anatomy......Page 1622
Nuclear medicine.......Page 1623
Parathyroid Pathology......Page 1624
Acknowledgements......Page 1626
Further Reading......Page 1627
Chapter Outline......Page 1628
Computed Tomography......Page 1629
Confirmation of Diagnosis......Page 1630
Staging Systems......Page 1631
Prostate Cancer......Page 1632
Assessment of Treatment Response......Page 1633
Objective Response Assessment......Page 1634
Treatment Toxicity......Page 1635
Lung......Page 1636
Hepatic Toxicity......Page 1637
Surveillance of Asymptomatic Patients......Page 1638
Further Reading......Page 1639
Standard Projections......Page 1641
Additional Projections......Page 1642
Breast Compression......Page 1643
Digital Breast Tomosynthesis in Breast Cancer Screening......Page 1644
Contrast Enhanced Spectral Mammography......Page 1645
Ultrasound......Page 1646
Normal Breast Anatomy......Page 1647
Papilloma......Page 1648
Classification of Invasive Breast Cancer......Page 1649
Mammography.......Page 1650
Ultrasound.......Page 1651
The Differential Diagnosis of Malignancy......Page 1652
Benign Microcalcifications......Page 1653
Malignant Microcalcifications......Page 1654
Lesion Characterisation......Page 1656
Controversies Surrounding the Use of Breast Magnetic Resonance Imaging......Page 1657
Interventional Breast Radiology......Page 1659
Guidance Methods for Breast Needle Biopsy......Page 1660
Magnetic Resonance Imaging-Guided Biopsy......Page 1661
Introduction......Page 1662
Which Age Groups Should Be Screened?......Page 1663
Conclusion......Page 1664
Further Reading......Page 1665
Genetic Factors......Page 1666
Prognosis and Treatment......Page 1667
Prognosis and Treatment......Page 1669
Choice of Imaging Technique......Page 1670
Abdomen and Pelvis......Page 1671
Extranodal Disease in Lymphoma......Page 1672
Primary Pulmonary Lymphoma......Page 1674
Liver......Page 1675
Gastrointestinal Tract......Page 1676
Small Bowel......Page 1677
Kidneys......Page 1678
Testis......Page 1679
Bone Marrow......Page 1680
Primary......Page 1681
Waldeyer’s Ring......Page 1682
Salivary Glands......Page 1683
Lymphomas Associated With HIV (AIDS-Related Lymphomas)......Page 1684
Monitoring Response to Therapy......Page 1685
Magnetic Resonance Imaging......Page 1686
Response Criteria......Page 1687
FDG-PET/CT in Response Assessment......Page 1688
Further Reading......Page 1689
Clinical Features......Page 1690
Imaging Features......Page 1691
Imaging Features......Page 1692
Lymphoma......Page 1693
Imaging Features......Page 1694
Imaging Features......Page 1696
Plasma Cell Dyscrasias......Page 1697
Imaging Features......Page 1698
Clinical Features......Page 1699
Imaging Features......Page 1700
Clinical Features......Page 1701
Langerhans Cell Histiocytosis......Page 1702
Imaging Features......Page 1704
Imaging Findings......Page 1707
Further Reading......Page 1709
The Haemoglobinopathies......Page 1711
Thalassaemia......Page 1712
Iron chelation therapy.......Page 1714
Sickle Cell Disease......Page 1715
Clinical Features......Page 1716
Bone infarction.......Page 1717
Osteomyelitis.......Page 1718
Radiological Features......Page 1720
Further Reading......Page 1723
Three-Dimensional Conformal Radiotherapy......Page 1724
Brachytherapy......Page 1725
Particle Therapy......Page 1727
Target Volume Definition......Page 1728
Computed Tomography Simulation......Page 1729
Delivery and Verification......Page 1730
Four-Dimensional Imaging......Page 1732
Adaptive Radiotherapy......Page 1734
Inclusion of Biological Information to the Treatment Process......Page 1735
Further Reading......Page 1737
Dynamic Contrast Enhanced- Computed Tomography......Page 1739
Contrast Agent Kinetics......Page 1741
Dynamic Contrast-Enhanced Magnetic Resonance Imaging......Page 1743
Diffusion-Weighted Imaging......Page 1744
Magnetic Resonance Spectroscopy......Page 1745
Non-Fluorodeoxyglucose Positron Emission Tomography Tracers......Page 1748
Optical Imaging......Page 1750
Conclusion: Role of Functional and Molecular Imaging in Oncology......Page 1751
Further Reading......Page 1752
Fluoroscopy......Page 1753
320-Row Multidetector Computed Tomography......Page 1754
Radiation Dose Considerations......Page 1755
Dental Cone Beam Computed Tomography......Page 1756
Motion Artefact Reduction......Page 1757
Imaging Planes......Page 1758
Further Reading......Page 1759
Idiopathic Respiratory Distress Syndrome......Page 1760
Transient Tachypnoea of the Newborn......Page 1763
Meconium Aspiration Syndrome......Page 1764
Neonatal Pneumonia......Page 1765
Lines and Tubes......Page 1766
The Chest Radiograph......Page 1767
Cardiac or Respiratory?......Page 1768
Bacterial Versus Viral......Page 1769
Tuberculosis......Page 1770
Mycoplasma pneumoniae......Page 1771
Inhaled Foreign Bodies......Page 1773
Congenital Thoracic Cysts......Page 1774
Congenital Pulmonary Airway Malformations......Page 1776
Congenital Lobar Overinflation......Page 1777
Lung Agenesis–Hypoplasia Complex......Page 1778
Mediastinal Masses......Page 1779
Middle mediastinum.......Page 1780
Pulmonary and Endobronchial Tumours......Page 1781
Pneumocystis jiroveci (carinii) pneumonia.......Page 1782
Lymphoproliferative disease and lymphocytic interstitial pneumonia.......Page 1783
Lung ultrasound......Page 1784
Lung MRI......Page 1785
Further Reading......Page 1786
Omphalocele......Page 1787
Oesophageal Atresia and Tracheo-Oesophageal Fistula......Page 1788
Post-surgery imaging.......Page 1789
The Vomiting Neonate......Page 1790
Small Bowel Malrotation and Volvulus......Page 1791
Duodenal Atresia and Stenosis......Page 1793
Necrotising Enterocolitis......Page 1794
Megacystis–Microcolon–Intestinal Hypoperistalsis (Berdon) Syndrome......Page 1796
Functional Immaturity of the Colon and Meconium Plug Syndrome......Page 1797
Meconium Ileus......Page 1798
Distal Ileal Atresia......Page 1799
Anorectal and Cloacal Malformations......Page 1800
Mesenteric Lymphadenitis......Page 1801
Magnetic resonance imaging.......Page 1802
Computed tomography.......Page 1803
Intussusception......Page 1804
Intestinal Motility Disorders......Page 1806
Henoch–Schönlein Purpura......Page 1807
Enteric Duplication Cysts......Page 1808
Non-Bilious Vomiting......Page 1809
Hypertrophic Pyloric Stenosis......Page 1810
Meckel Diverticulum......Page 1812
Gastrointestinal Malignancies......Page 1814
Abdominal Manifestations of Cystic Fibrosis......Page 1815
Abdominal Trauma......Page 1816
Magnetic Resonance Imaging......Page 1817
Imaging Anatomy......Page 1818
Fibrosis......Page 1819
Preduodenal Portal Vein......Page 1820
Transplant......Page 1821
Biliary Atresia (Fig. 71.50)......Page 1822
Choledochal Malformation (Choledochal Cyst; Fig. 71.51)......Page 1823
Other......Page 1824
Congenital Hyperinsulinism......Page 1825
Lateralisation Disorders (Fig. 71.59)......Page 1826
Neoplasia......Page 1827
Further Reading......Page 1828
Plain Radiography......Page 1830
Standard Technique......Page 1831
Cystography......Page 1832
Indications.......Page 1833
Contrast-Enhanced Ultrasonography......Page 1834
Dynamic Renography......Page 1835
Technique.......Page 1836
Magnetic Resonance Imaging......Page 1837
Angiography......Page 1838
Abnormalities With Renal Fusion......Page 1839
Imaging......Page 1840
Imaging......Page 1842
Functional Bladder Disturbance and Neurogenic Bladder......Page 1843
Posterior Urethral Valves......Page 1844
Anterior Urethral Abnormalities......Page 1845
Undescended Testis......Page 1846
Bilateral Renal Pelvic Dilatation......Page 1847
Unilateral Renal Pelvic Dilatation......Page 1848
Urinary Tract Infection and Vesicoureteric Reflux......Page 1849
Imaging......Page 1850
Xanthogranulomatous Pyelonephritis......Page 1851
Cystic Dysplasia......Page 1853
Localised Cystic Disease of the Kidney......Page 1854
Autosomal Recessive Polycystic Kidney Disease......Page 1855
Renal Calculi......Page 1857
Nephroblastomatosis......Page 1859
Wilms Tumour......Page 1860
Renal Cell Carcinoma......Page 1861
Scrotal Masses......Page 1862
Presacral Masses......Page 1863
Hypertension......Page 1864
Renal Transplantation......Page 1867
Post-Transplantation......Page 1868
Further Reading......Page 1869
Diagnosis......Page 1870
Imaging for Complications......Page 1894
Genetic Counselling......Page 1895
Trisomy 21 (Down Syndrome)......Page 1896
Idiopathic Avascular Necrosis of the Femoral Head (Perthes Disease)......Page 1897
Slipped Capital Femoral Epiphysis......Page 1899
Tibia Vara and Tibial Bowing......Page 1900
Madelung Deformity......Page 1901
Sprengel Deformity (Congenital Elevation of the Scapula)......Page 1902
Juvenile Idiopathic Arthritis......Page 1903
Chronic Recurrent Multifocal Osteomyelitis......Page 1906
Haemophilia......Page 1908
Rickets......Page 1909
Vitamin D-Dependent Rickets......Page 1910
Lead Poisoning......Page 1911
Hypoparathyroidism......Page 1912
Hypothyroidism......Page 1913
Thalassaemia......Page 1914
Infective Arthritis......Page 1915
Infection of the Spine (Discitis and Osteomyelitis)......Page 1916
Further Reading......Page 1917
Physeal Injuries......Page 1919
Shoulder/Humerus......Page 1920
Elbow......Page 1921
Forearm/Wrist/Hand......Page 1927
Pelvis......Page 1928
Knee......Page 1929
Foot......Page 1931
Cervical Spinal Injuries......Page 1932
Clinical Presentation and the Role of the Radiologist......Page 1933
Injury Patterns......Page 1934
General Imaging Strategies......Page 1935
Fracture Patterns in Accidental Versus Nonaccidental Injury......Page 1936
Rib Fractures......Page 1937
Long Bone Fractures......Page 1938
Unusual Fractures......Page 1939
Fracture Healing......Page 1940
Birth Trauma......Page 1941
Brain Injuries......Page 1942
Extra-Axial Haemorrhages......Page 1943
Further Reading......Page 1950
The Ewing Sarcoma Family of Tumours......Page 1952
Rare Malignant Bone Tumours in Children......Page 1954
Tumours of Fibrous Tissue Origin......Page 1955
Osteochondroma (Exostosis)......Page 1956
Locally Aggressive Tumours......Page 1957
Aneurysmal Bone Cyst......Page 1958
Neuroblastoma......Page 1960
Computed Tomography and Magnetic Resonance Imaging......Page 1962
Radionuclide Radiology......Page 1963
Neuroblastoma Staging......Page 1964
Further Reading......Page 1967
Computed Tomography Angiography in Clinical Practice......Page 2030
Computed Tomography Angiography or Magnetic Resonance Angiography: Safety, Preference and Ease of Use......Page 2031
Abdominal Aortic Aneurysm Assessment for Suitability for Endovascular Repair and Postintervention Follow-Up......Page 2032
Assessing the Aorta for Other Pathological Conditions......Page 2033
Acute Mesenteric Ischaemia......Page 2035
Computed Tomography Angiography for Acute Gastrointestinal Bleeding......Page 2036
Other Pulmonary Conditions......Page 2037
Diagnosis of Stroke: Computed Tomography and Magnetic Resonance......Page 2038
Elucidating the Source of Embolic Stroke......Page 2039
Peripheral Vascular Disease......Page 2041
Magnetic Resonance Angiography......Page 2043
Arteriovenous Malformations......Page 2044
Further Reading......Page 2045
Arterial Puncture......Page 2046
Sheaths......Page 2047
Closure Devices......Page 2048
False Aneurysm......Page 2049
Further Reading......Page 2050
Stent-Grafts and Basic Principles of Stent-Grafting......Page 2051
Anatomical Considerations......Page 2057
Outcomes of Endovascular Repair and Comparison With Surgery......Page 2059
Acute Aortic Syndrome......Page 2060
Complicated Type B Dissection......Page 2061
Chronic Dissection and Aneurysmal Development......Page 2062
Aortic Coarctation......Page 2063
Outcomes of Endovascular Repair and Comparison With Surgery for Elective Abdominal Aortic Aneurysm Repair......Page 2066
Thoraco-Abdominal Aneurysms......Page 2067
Surveillance Imaging and Complications......Page 2070
Graft Infection (See Fig. 79.33)......Page 2071
Summary and Conclusion......Page 2073
Further Reading......Page 2074
Embolisation......Page 2075
Thrombolysis......Page 2077
Diagnosis......Page 2078
Occlusions.......Page 2079
Popliteal Artery......Page 2081
Treatment of Acute Lower Limb Ischaemia......Page 2082
Gastrointestinal System......Page 2083
Mesenteric Haemorrhage......Page 2084
Visceral Artery Aneurysms......Page 2086
Occlusive Mesenteric Vascular Disease......Page 2088
Bronchial Artery Embolisation......Page 2089
Lower Extremity Venous System......Page 2090
Complications of Endovascular Procedures......Page 2091
Further Reading......Page 2093
Preprocedural Assessment......Page 2094
Semiautomatic, For Example Temno, SuperCore (Fig. 81.1)......Page 2095
Coaxial Technique......Page 2096
Magnetic Resonance Imaging......Page 2097
Improving Needle Tip Visualisation in Ultrasound-Guided Biopsy......Page 2098
The Case for Tumour Ablation......Page 2099
Microwave Ablation......Page 2100
Preprocedural Planning......Page 2101
Postprocedural Imaging......Page 2102
Understanding and Modifying Tumour Pathophysiology......Page 2103
Lung Tumour Ablation......Page 2105
Bone Tumour Ablation......Page 2106
Further Reading......Page 2108
Indications and Contraindications......Page 2109
Management of Anticoagulant Medication Prior to Percutaneous Drainage......Page 2110
Imaging Guidance......Page 2111
Patient Preparation and Care......Page 2113
Catheter Insertion......Page 2114
Catheter Management......Page 2115
Chest......Page 2116
Gallbladder......Page 2117
Spleen......Page 2118
Peritoneum......Page 2119
Deep Pelvic Collections......Page 2122
Organ Traversal......Page 2124
Complications......Page 2125
Conclusion......Page 2126
Further Reading......Page 2127
Percutaneous Transhepatic Cholangiography......Page 2128
Benign Strictures......Page 2129
Vascular Interventional Techniques in the Liver......Page 2132
Performing the Procedure......Page 2133
Background......Page 2134
Performing the Procedure: Planning......Page 2135
Hepatic Arterial Embolisation for Haemorrhage......Page 2136
Performing the Procedure......Page 2138
Background......Page 2139
Background......Page 2140
Diagnosis......Page 2141
Further Reading......Page 2142
Renal revascularisation trials.......Page 2143
Renal artery angioplasty.......Page 2144
Takayasu arteritis.......Page 2145
Results......Page 2146
Malignant......Page 2147
Renal Arteriovenous Malformation......Page 2148
Ureter......Page 2149
Treatment Options......Page 2150
Imaging......Page 2151
Safety and Efficacy......Page 2152
Management of abnormal placentation.......Page 2153
Pelvic Congestion Syndrome......Page 2155
Treatment......Page 2156
Further Reading......Page 2160
Other Anatomical Factors Important for Renal Access......Page 2161
Access Needle......Page 2162
Percutaneous Nephrostomy (PCN)......Page 2163
Patient Preparation and Procedure......Page 2164
Transplant Kidney......Page 2165
Complications of Percutaneous Nephrostomy (PCN) and Management......Page 2168
Tract Planning......Page 2169
Tract Dilatation......Page 2170
Complications of Percutaneous Nephrolithotomy and Management......Page 2171
Technique of Antegrade Stenting......Page 2172
Insertion of a Metal Ureteric Stent......Page 2173
Tortuous Ureter......Page 2174
Monitoring Ureteric Stents......Page 2175
Suprapubic Bladder Catherisation......Page 2176
Interventional Procedures in the Prostate Gland and Seminal Vesicles......Page 2178
Insertion of Prostate Fiducials and Absorbable Hydrogel Spacer......Page 2179
Further Reading......Page 2180
General Patient and Interventional Suite Preparation......Page 2181
Tunnelled Central Venous Catheter: Hickman Line......Page 2182
Tunnelled Central Venous Catheter: Totally Implantable Vascular Access Device......Page 2183
Translumbar Dialysis Catheter Insertion......Page 2184
Complications and Their Management......Page 2185
Inadvertent Arterial Puncture......Page 2186
Air Embolism......Page 2187
Laceration of Central Veins......Page 2188
Catheter Kinks/Fractures......Page 2189
Diagnostic Evaluation......Page 2190
Further Reading......Page 2191
Bone Biopsy Techniques......Page 2192
Performing the Procedure......Page 2194
Technique......Page 2195
Complications......Page 2197
Pulsed Radiofrequency Ablation......Page 2199
Facet Joint Syndrome......Page 2201
Cervical Spine Procedures......Page 2202
Technique......Page 2205
Percutaneous Vertebral Augmentation......Page 2206
Osteoplasty and Osteosynthesis......Page 2208
Ablation of Bone Tumours......Page 2211
Complications......Page 2213
Osteoid Osteomas......Page 2214
Neuromodulation and Neurostimulation......Page 2215
Further Reading......Page 2218

Citation preview

https://afkebooks.com

Grainger & Allison’s

DIAGNOSTIC RADIOLOGY A Textbook of Medical Imaging

https://t.me/afkebooks

SEVENTH EDITION

VOLUM E ONE

Grainger & Allison’s

DIAGNOSTIC RADIOLOGY A Textbook of Medical Imaging EDITED BY

Andreas Adam, CBE, MB, BS(Hons), PhD, PhD (hon caus), DSc (hon caus) FRCP, FRCR, FRCS, FFRRCSI(Hon), FRANZCR(Hon), FACR(Hon), FMedSci Professor of Interventional Radiology, King’s College London, London, UK

Adrian K. Dixon, MD, MD (hon caus), FRCP, FRCR, FRCS, FFRRCSI(Hon), FRANZCR(Hon), FACR(Hon), FMedSci Professor Emeritus of Radiology, Cambridge, UK

Jonathan H. Gillard, BSc, MA, MD, FRCP, FRCR, MBA Professor of Neuroradiology, Christ’s College, Cambridge, UK

Cornelia M. Schaefer-Prokop, MD, PhD Professor of Radiology, Meander Medical Centre, Amersfoort, The Netherlands

https://afkebooks.com

© 2021, Elsevier Limited. All rights reserved. First edition 1986 Second edition 1991 Third edition 1997 Fourth edition 2001 Fifth edition 2008 Sixth edition 2015 No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds or experiments described herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. To the fullest extent of the law, no responsibility is assumed by Elsevier, authors, editors or contributors for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. ISBN: 9780702075247 Executive Content Strategist: Michael Houston Content Development Specialists: Martin Mellor Publishing Services / Louise Cook Project Manager: Andrew Riley Design: Renee Duenow Illustration Manager: Paula Catalano Marketing Manager: Claire McKenzie

Printed in Poland Last digit is the print number: 9 8 7 6 5 4 3 2 1

https://t.me/afkebooks

CONTENTS 19 The Oesophagus, 482

Preface, viii List of Section Editors, ix List of Contributors, x Dedication, xix

Edmund M. Godfrey, Sibu Varghese, Alan H. Freeman

20 The Stomach, 498

Omar Agosto, Chandra Dass, Dina F. Caroline

21 The Small Intestine, Mesentery and Peritoneal Cavity, 527

VOLUME 1

Abdullah A. Al Sarraf, Patrick D. McLaughlin, Michael M. Maher

SECTION A The Chest and Cardiovascular System Cornelia M. Schaefer-Prokop, Simon P.G. Padley

22 The Large Bowel, 568

Darren Boone, Andrew Plumb, Stuart A. Taylor

23 The Liver and Spleen, 598

Lorenzo Mannelli, David J. Lomas, Orpheus Kolokythas

24 The Biliary System, 656

1 Current Status of Thoracic Imaging, 3

Robert N. Gibson, Tom R. Sutherland

Arjun Nair, Joseph L. Barnett, Thomas R. Semple

25 The Pancreas, 679

Simon P.G. Padley, Bhavin Rawal

26 Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection, 722

2 The Normal Chest, 17

Wolfgang Schima, Helmut Kopf

3 The Chest Wall, Pleura, Diaphragm and Intervention, 38 Johny A. Verschakelen, Fergus Gleeson, Maria Tsakok

4 The Mediastinum, Including the Pericardium, 67 Nadeem Parkar, Cylen Javidan-Nejad, Sanjeev Bhalla

5 Pulmonary Infection in Adults, 104 Tomás Franquet, Gustavo Meirelles

6 Large Airway Disease and Chronic Airflow Obstruction, 126 Philippe A. Grenier, Catherine Beigelman-Aubry, Pierre-Yves Brillet, Catalin I. Fetita

7 Pulmonary Lobar Collapse: Essential Considerations, 163

Richard G. Kavanagh, Michael M. Maher, Owen J. O’Connor

27 Current Status of Imaging of the Urinary Tract: Imaging Techniques, Overview of Anatomy and Radiation Issues, 751 Stephen P. Power, Michael M. Maher, Owen J. O’Connor

28 Benign Upper Urinary Tract Conditions: Congenital Anomalies, Cysts, Calculi, Nephrocalcinosis, 776 Richard G. Kavanagh, Tristan Barrett, Brian Carey, Michael M. Maher

Susan J. Copley

29 Renal Masses: Imaging and Biopsy, 791

Simon P.G. Padley, Sanjay Popat, Anand Devaraj

30 Renal Transplantation: Imaging, 803

8 Pulmonary Neoplasms, 179

Giles Rottenberg, Hema Verma

9 High-Resolution Computed Tomography of Interstitial and Occupational Lung Disease, 206 Mario Silva, Nicola Sverzellati

Giles Rottenberg, Eamon Lagha

31 Urothelial Cell Cancer, Upper Tract and Lower Tract, 815 Richard G. Kavanagh, Kevin O’Regan, Michael M. Maher

32 Prostate, 844

10 Thoracic Trauma and Related Topics, 231

Maarten de Rooij, Joyce G.R. Bomers, Geert M. Villeirs, Jelle O. Barentsz

Hefin Jones, John H. Reynolds

11 Airspace Diseases, 250

Gurinder S. Nandra, Sujal R. Desai

33 The Male Reproductive Structures, 861

Hans-Marc J. Siebelink, Jos J.M. Westenberg, Lucia J.M. Kroft, Albert de Roos

34 Gynaecological Imaging in Oncology, 873

Michael A. Quail, Andrew M. Taylor

36 Genitourinary Tract Trauma, 923

Luigi Natale, Veronica Bordonaro

37 Adrenal Imaging, 938

12 Cardiac Anatomy and Imaging Techniques, 265

Nadeem Shaida, Tristan Barrett

13 Congenital Heart Disease: General Principles and Imaging, 289 14 Nonischaemic Acquired Heart Disease, 315

Susan Freeman, Amreen Shakur, Evis Sala

35 Benign Gynaecological Disease, 890 Jan Smith, Fleur Kilburn-Toppin, Helen Addley Peter Beddy, Roisin M. Heaney

15 Ischaemic Heart Disease, 368

Anju Sahdev, Richard G. Kavanagh, Rodney H. Reznek

Jan Bogaert, Rolf Symons

SECTION C The Musculoskeletal System

16 Pulmonary Circulation and Pulmonary Thromboembolism, 397

Andrew J. Grainger, Philip O’Connor

Ieneke J.C. Hartmann, Nicholas J. Screaton

17 The Thoracic Aorta: Diagnostic Aspects, 421

38 Current Status of Imaging of The Musculoskeletal System, 969

Rossella Fattori, Luigi Lovato, Vincenzo Russo

Geeta Kapoor, Andoni P. Toms

SECTION B Abdominal Imaging

39 Internal Derangements of Joints: Upper and Lower Limbs, 977

Michael M. Maher, Adrian K. Dixon

18 Current Status of Imaging of the Gastrointestinal Tract, 463 Abdullah A. Al Sarraf, Patrick D. McLaughlin, Michael M. Maher

Robert S.D. Campbell, Alpesh Mistry, Eugene McNally, Edward Sellon

40 Bone Tumours (1): Radiological Approach, Benign Tumours and Tumour-Like Lesions of Bone, 1009 Asif Saifuddin, Philippa Tyler

v https://afkebooks.com

vi

Contents

41 Bone Tumours (2): Radiological Approach, Malignant Bone Tumours, 1041 Anish Patel, Steven L.J. James, A. Mark Davies

42 Soft-Tissue Tumours, 1066

SECTION F Oncological Imaging Viky Goh, Andreas Adam

62 Introduction to Oncological Imaging, 1641

Paul O’Donnell

43 Metabolic and Endocrine Skeletal Disease, 1087 Janina M. Patsch, Christian R. Krestan

Christian Kelly-Morland, Davide Prezzi, David MacVicar, Vicky Goh

63 The Breast, 1654

Jonathan J. James, Andrew J. Evans

44 Arthritis, 1116

64 Reticuloendothelial Disorders: Lymphoma, 1679

Emma L. Rowbotham, Andrew J. Grainger, Philip O’Connor

Sarah J. Vinnicombe, Victoria S. Warbey, Gary J.R. Cook

65 Bone Marrow Disorders: Haematological Neoplasms, 1703

45 Appendicular and Pelvic Trauma, 1142 Philip M. Hughes, Alun Davies

Asif Saifuddin

46 Bone, Joint, and Spinal Infections, 1184

66 Bone Marrow Disorders: Non-Neoplastic Conditions, 1724

Jaspreet Singh, Radhesh Lalam

Asif Saifuddin

67 Imaging for Radiotherapy Planning, 1737

VOLUME 2

Peter Hoskin, Ananya Choudhury, Lizbeth M. Kenny

68 Functional and Molecular Imaging for Personalized Medicine in Oncology, 1752

SECTION D The Spine

Eva M. Serrao, Avnesh S. Thakor, Vicky Goh, Ferdia A. Gallagher

Jonathan H. Gillard, H. Rolf Jäger

47 Current Status of Imaging of the Spine and Anatomical Features, 1225 Thomas Van Thielen, Luc van den Hauwe, Johan W. Van Goethem, Paul M. Parizel

48 Degenerative Disease of the Spine, 1243

SECTION G Paediatric Imaging Catherine M. Owens, Jonathan H. Gillard

69 Current Status of Paediatric Imaging, 1769 Thomas R. Semple, Kristian H. Mortensen, Øystein E. Olsen, Catherine M. Owens

Paul M. Parizel, Thomas Van Thielen, Luc van den Hauwe, Johan W. Van Goethem

70 The Neonatal and Paediatric Chest, 1776

49 Spinal Tumours, 1267

Luc van den Hauwe, Johan W. Van Goethem, Danielle Balériaux, Arthur M. De Schepper

50 Non-Tumoural Spinal Cord Lesions, 1295

Thomas R. Semple, Kristian H. Mortensen, Tom A. Watson, Catherine M. Owens

71 Paediatric Abdominal Imaging, 1803

Tom A. Watson, Øystein E. Olsen, Lil-Sofie Ording Müller

72 Imaging of the Kidneys, Urinary Tract and Pelvis in Children, 1846

Farah Alobeidi, Majda M. Thurnher, H. Rolf Jäger

51 Postoperative Spine, 1319

Tomasz Matys, Nasim Sheikh-Bahaei, Jonathan H. Gillard

Owen Arthurs, Marina Easty, Michael Riccabona

52 Spinal Trauma, 1332

73 Skeletal Radiology in Children: Non-Traumatic and Non-Malignant, 1886

James J. Rankine

Joy L. Barber, Amaka C. Offiah

SECTION E Neuroimaging H. Rolf Jäger, Jonathan H. Gillard

53 Current Status of Imaging of the Brain and Anatomical Features, 1351

74 Paediatric Musculoskeletal Trauma and the Radiology of Nonaccidental Injury and Paediatric Factures, 1935 Karen Rosendahl, Jean-François Chateil, Karl Johnson, Rachel M. Martin

Christen D. Barras, Joti Jonathan Bhattacharya

75 Bone Tumours and Neuroblastoma in Children, 1968

Adam Kenji Yamamoto, Ashok Adams

76 Paediatric Neuroradiology, 1984

54 Imaging of Head Trauma, 1387

55 Benign and Malignant Intracranial Tumours in Adults, 1411

Paul D. Humphries, Claudio Granata

Roxana S. Gunny, Dawn E Saunders, Maria I. Argyropoulou

Steffi Thust, Caroline Micallef, H. Rolf Jäger

SECTION H Interventional Radiology

Brynmor P. Jones, Charles B.O. Hall, Amrish Mehta

Robert A. Morgan, Michael J. Lee, Andreas Adam

56 Neurovascular Diseases, 1441 57 Intracranial Infections, 1481

Daniel J. Scoffings, Majda M. Thurnher, H. Rolf Jäger

58 Inflammatory and Metabolic Disease, 1498 Àlex Rovira, Stéphane Kremer, Pia C. Sundgren

59 Neurodegenerative Diseases and Epilepsy, 1550

77 Current Status of Imaging for Interventional Procedures, 2049 James F.M. Meaney, Simon P. Doran

78 Angiography: Principles, Techniques and Complications, 2065

Beatriz Gomez Anson, Frederik Barkhof, Francesca Benedetta Pizzini

79 Aortic Intervention, 2070

Tom Campion, Katherine Miszkiel, Indran Davagnanam

80 Peripheral Vascular Disease Intervention, 2094

60 The Orbit, 1562

61 Head and Neck Radiology, 1599 Timothy Beale, Susan Jawad

Jim A. Reekers

Raman Uberoi, Mohammed Hadi

Seyed Ameli-Renani, Anna-Maria Belli, Joo-Young Chun, Robert A. Morgan

https://t.me/afkebooks

Contents

81 Image-Guided Biopsy and Ablation Techniques, 2113

85 Non-Vascular Genitourinary Tract Intervention, 2180 Uday Patel, Lakshmi Ratnam

David J. Breen, Elizabeth E. Rutherford, Beth Shepherd

86 Venous Access and Interventions, 2200

Stephen P. Power, Michael M. Maher, Owen J. O’Connor

87 Skeletal Interventions, 2211

82 Image-Guided Drainage Techniques, 2128

Mohamad Hamady, Wasim Hakim

83 Hepatobiliary Intervention, 2147

Timothy E. Murray, Aoife N. Keeling, Michael J. Lee

84 Vascular Genitourinary Tract Intervention, 2162 Jonathan G. Moss, Reddi Prasad Yadavali, Ram S. Kasthuri

Konstantinos Katsanos, Tarun Sabharwal, Roberto Luigi Cazzato, Afshin Gangi

Index, I-1

https://afkebooks.com

vii

P R E FA C E This seventh edition of the landmark Grainger and Allison textbook Diagnostic Radiology is once again a truly cooperative venture. The four Lead Editors and the very energetic Section Editors (Vicky Goh, Andrew Grainger, Rolf Jäger, Michael Lee, Michael Maher, Robert Morgan, Phil O’Connor, Cathy Owens and Simon Padley) have recruited outstanding authors who are all international experts within their field. We hope that this new edition will help to maintain the role of this book as the leading educational text for those pursuing radiological training in the UK, mainland Europe, Asia, Africa, Australia and New Zealand; certainly it has been written very much with qualifying examinations in mind, such as the FRCR, FFRCSI, EDiR, FRANZCR, DNB, etc. Radiologists in North America have also found previous editions helpful when preparing for their board examinations. We hope it will remain as a ready source of reference for most radiological queries—even in the days of the internet. One major change is a reduction in the amount of material about physics, equipment, techniques and basic anatomy, much of which can be found elsewhere. Indeed the whole introductory section in the previous edition has been removed. Instead, authors have been encouraged to include important aspects of radiological techniques and anatomy within their individual chapters. Furthermore the first chapter in each section entitled ‘Current status of Imaging in the …’ outlines the imaging techniques relevant for that subspecialty. Approximately half of the images are ‘new’ and the text has been updated to reflect recent changes in practice. Once again we gratefully acknowledge the work and material of previous authors which ‘live on’ in this edition.

The support given to the editors and authors by the team at Elsevier has been exemplary. Michael Houston has wholeheartedly supported this project for several decades. Martin Mellor has been the lynchpin in keeping authors (and editors!) up to the mark. The Production Manager, Andrew Riley, has been instrumental in the day to day stages of typesetting and proofreading. Thank you one and all: we could not have done it without you! Nor could we have done it without the continued support of the cast of nearly 200 authors from around the world who have generously given of their time and expertise to this ‘living’ textbook, not only those who have helped in this edition but also those who have contributed to previous editions. Again we are very grateful. Finally, a note of sadness. Many readers will know that Ronald Grainger, who launched the first edition of this textbook with David Allison back in 1986, sadly passed away during 2014 aged 91. He was a tower of strength within British and international radiology and led a remarkable training department in Sheffield which spawned many future leaders of the specialty. Ronald himself was a pioneer in cardiovascular radiology but this book is perhaps one of his greatest testaments and he was very proud of its continuing success. Ronald, this seventh edition is dedicated to your vision.

viii https://t.me/afkebooks

Andreas Adam, Adrian K. Dixon, Jonathan H. Gillard, Cornelia M. Schaefer-Prokop

LIST OF SECTION EDITORS Andreas Adam, CBE, MB, BS(Hons), PhD, FRCP, FRCR, FRCS, FFRRCSI(Hon), FRANZCR(Hon), FACR(Hon), FMedSci Professor of Interventional Radiology King’s College London London, UK (Section F: Oncological Imaging; Section H: Interventional Radiology)

Adrian K. Dixon, MD, MD(Hon caus), FRCP, FRCR, FRCS, FFRRCSI(Hon), FRANZCR(Hon), FACR(Hon), FMedSci Professor Emeritus of Radiology Cambridge, UK (Section B: Abdominal Imaging)

Jonathan H. Gillard, BSc, MA, MD, FRCP, FRCR, MBA Professor of Neuroradiology University of Cambridge, Addenbrooke’s Hospital Cambridge, UK (Section D: The Spine; Section E: Neuroimaging; Section G: Paediatric Imaging)

Vicky Goh, MA, MBBChir, MD, MRCP, FRCR Chair of Clinical Cancer Imaging Division of Imaging Sciences and Biomedical Engineering, King’s College London; Honorary Consultant Radiologist Guy’s and St Thomas’ Hospitals London, UK (Section F: Oncological Imaging)

Andrew J. Grainger, BM, BS, MRCP, FRCR Consultant Musculoskeletal Radiologist Addenbrooke’s Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK (Section C: The Musculoskeletal System)

H. Rolf Jäger, MD, FRCR Professor in Neuroradiology Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, UCL Faculty of Brain Sciences; Consultant Neuroradiologist Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, and Department of Imaging, University College London Hospitals London, UK (Section D: The Spine; Section E: Neuroimaging)

Michael J. Lee, MSc, FRCPI, FRCR, FFR (RCSI), FSIR, EBIR Professor of Radiology Royal College of Surgeons in Ireland; Consultant Interventional Radiologist and Clinical Director Beaumont Hospital Dublin, Ireland (Section H: Interventional Radiology)

Michael M. Maher, MD, FRCSI, FRCR, FFR (RCSI) Professor of Radiology University College Cork; Consultant Radiologist Cork University Hospital and Mercy University Hospital Cork, Ireland (Section B: Abdominal Imaging)

Philip O’Connor, MBBS, MRCP, FRCR, FFSEM(UK) Musculoskeletal Radiologist Clinical Radiology, Leeds Teaching Hospitals Trust, The University of Leeds West Yorkshire, UK (Section C: The Musculoskeletal System)

Catherine M. Owens, BSc, MBBS, MRCP, FRCR Consultant Radiologist and Honorary Reader Department of Imaging, Great Ormond Street Hospital London, UK (Section G: Paediatric Imaging)

Simon P.G. Padley, BSc, MBBS, FRCP, FRCR Professor of Practice (Diagnostic and Interventional Radiology) Imperial College School of Medicine; Consultant Radiologist, Cross-site Director of Radiology Royal Brompton and Harefield Hospital London, UK (Section A: The Chest and Cardiovascular System)

Cornelia M. Schaefer-Prokop, MD, PhD Professor of Radiology Meander Medical Centre Amersfoort, The Netherlands (Section A: The Chest and Cardiovascular System)

Robert A. Morgan, MB ChB, MRCP, FRCR, EBIR Consultant Vascular and Interventional Radiologist Radiology Department, St George’s NHS Trust London, UK (Section H: Interventional Radiology)

ix https://afkebooks.com

LIST OF CONTRIBUTORS The editor(s) would like to acknowledge and offer grateful thanks for the input of all previous editions’ contributors, without whom this new edition would not have been possible.

Ashok Adams, BSc (Hons), MRCP, FRCR

Owen Arthurs, MB BChir, PhD, FRCR

Tristan Barrett, MBBS, MD, BSc, FRCR

Consultant Neuroradiologist Department of Neuroradiology BartsHealth NHS Trust London, UK

Consultant Paediatric Radiologist Radiology Great Ormond Street Hospital London, UK

Department of Radiology Cambridge University Hospitals Cambridge, UK

Helen Addley, BM, BCh, MRCP, FRCR

Danielle Balériaux, MD

Consultant Radiologist Department of Radiology Cambridge University Hospitals Cambridge, UK

Professor Emeritus Neuroradiology Hôpital Erasme ULB Brussels, Belgium

Omar Agosto, MD

Joy L. Barber, MBBS, MA, FRCR

Clinical Assistant Professor Diagnostic Radiology Lewis Katz School of Medicine at Temple University; Director of Body MRI and Body CT Diagnostic Radiology Jeanes Hospital and Northeastern Ambulatory Care Center-Temple University Healthcare System Philadelphia, PA, USA

Consultant Radiologist Department of Radiology St George’s University Hospitals NHS Foundation Trust London, UK

Timothy Beale, MBBS, FRCS, FRCR

Abdullah A. Al Sarraf, MB BCh BAO BMedSc (Hons), FFR RCSI Consultant Radiologist – Cardiothoracic & Oncology imaging. Dar Al Shifa Hospital Mubarak Al Kabeer (MOH) Hospital

Farah Alobeidi, MA, MB BChir, MRCS, FRCR Consultant Neuroradiologist Radiology Imperial College Healthcare NHS Trust London, UK

Seyed Ameli-Renani, MBBS, FRCR, EBIR Consultant Diagnostic and Interventional Radiologist Radiology Department St George’s University Hospitals NHS Foundation Trust London, UK

Maria I. Argyropoulou, MD, PhD Professor of Radiology Department of Radiology University of Ioannina Greece

Consultant Head and Neck Radiologist University College Hospital Royal National Throat Nose and Ear Hospital London, UK

Peter Beddy, MB, MRCS, FFR RCSI, FRCR Department of Radiology St James Hospital and Trinity College Dublin Dublin, Ireland

Jelle O. Barentsz, MD, PhD

Catherine Beigelman-Aubry, MD

Professor of Radiology Department of Radiology Prostate MR Reference Center Radboud University Nijmegen Medical Center Nijmegen, The Netherlands

Priva Docent-Maitre d’enseignement et recherche Radiodiagnostic and Interventional Radiology Centre Hospitalier Universitaire Vaudois Lausanne, Switzerland

Frederik Barkhof, MD, PhD

Anna-Maria Belli, FRCR, EBIR

Professor of Neuroradiology Alzheimer Center and Department of Radiology VU University Medical Center Amsterdam, The Netherlands

Professor of Interventional Radiology; Consultant Radiologist Radiology Department St George’s Healthcare NHS Trust London, UK

Joseph L. Barnett, MBBS, MA (Cantab), FRCR

Sanjeev Bhalla, MD

National Thoracic Imaging Fellow Department of Radiology Royal Brompton Hospital London, UK

Christen D. Barras, MBBS (Hons), BMedSci (Hons), MMed, PhD, DipSurgAnat, DipOMS, FRANZCR Consultant Neuroradiologist Department of Radiology Royal Adelaide Hospital South Australian Health and Medical Research Institute The University of Adelaide Adelaide, Australia

x https://t.me/afkebooks

Chief Cardiothoracic Radiology Mallinckrodt Institute of Radiology Washington University in St Louis St Louis, MO, USA

Joti Jonathan Bhattacharya, MBBS, MSc, FRCR Consultant Neuroradiologist Department of Neuroradiology Institute of Neurological Sciences Southern General Hospital Glasgow, UK

List of Contributors

Jan Bogaert, MD, PhD

Dina F. Caroline, MD, PhD

Professor of Medicine Department of Radiology University Hospital Gasthuisberg University of Leuven Leuven, Belgium

Professor Emerita Department of Radiology Lewis Katz School of Medicine at Temple University Philadelphia, PA, USA

Joyce G.R. Bomers, MSc, PhD

Roberto Luigi Cazzato, Cazzato MD, PhD

Technical Physician Radiology and Nuclear Medicine Radboud University Nijmegen Medical Centre Nijmegen, The Netherlands

Associate Professor Repartment of Interventional Radiology University Hospital of Strasbourg Strasbourg, France

xi

Indran Davagnanam, MB BCh, BMedSci, FRCR Consultant Neuroradiologist National Hospital for Neurology and Neurosurgery Lysholm Radiological Department; Radiology Moorfields Eye Hospital; Honorary Research Associate The Brain Repair and Rehabilitation Unit Institute of Neurology London, UK

A. Mark Davies, MBChB, DMRD, FRCR Darren Boone, MBBS BSc MD MRCS FRCR Consultant Radiologist Department of Specialist Imaging University College Hospital London, UK

Veronica Bordonaro, MD Researcher Department of Diagnostic Imaging Bambino Gesù Children’s Hospital; Rome, Italy

Jean-François Chateil, MD, PhD Professor CHU de Bordeaux Service d’imagerie anténatale de l’enfant et de la femme; University of Bordeaux Bordeaux, France

Ananya Choudhury, MA, PhD, MRCP, FRCR Chair in Clinical Oncology Division of Cancer Sciences University of Manchester Manchester, UK

David J. Breen, MRCP, FRCR Consultant Abdominal Radiologist; Honorary Senior Lecturer in Radiology Department of Radiology Southampton University Hospitals Southampton, UK

Joo-Young Chun, MBBS, MSc, MRCS, FRCR Consultant Interventional Radiologist Royal London Hospital Barts Health NHS Trust London, UK

Pierre-Yves Brillet, MD, PhD

Consultant Radiologist MRI Centre Royal Orthopaedic Hospital NHS Foundation Trust Birmingham, UK

Alun Davies, MA, BM, BCh, MRCS, FRCR Consultant Musculoskeletal Radiologist Department of Radiology University Hospitals Plymouth NHS trust Plymouth, UK

Maarten de Rooij, MD PhD Radiology Resident Department of Radiology and Nuclear Medicine Radboud University Medical Center Nijmegen, The Netherlands

Albert de Roos, MD

Professor Department of Radiology Hôpital Avicenne Assistance Publique des Hôpitaux de Paris Bobigny, France

Gary J.R. Cook, MBBS, MSc, MD

Robert S.D. Campbell, MB, ChB, FRCR

Susan J. Copley, MBBS, FRCP, FRCR, MD (Res)

Professor Cancer Imaging King’s College London London, UK

Professor of Radiology Department of Radiology Leiden University Medical Center Leiden, The Netherlands

Sujal R. Desai, MD, FRCP, FRCR Consultant Musculoskeletal Radiologist Department of Radiology Royal Liverpool University Hospital Liverpool, UK

Consultant Radiologist Imperial College NHS Healthcare Trust London, UK

Tom Campion, BA, BMBCh, MSc, FRCR

Chandra Dass, MBBS, DMRD

Neuroradiology Fellow Lysholm Department of Neuroradiology National Hospital for Neurology and Neurosurgery London, UK

Professor of Clinical Radiology Department of Radiology Lewis Katz School of Medicine at Temple University Philadelphia, PA, USA

Brian Carey, MB BCh, BAO, MCh, MRCS Clinical Lecturer Department of Radiology Cork University Hospital Cork, Ireland

Consultant Radiologist Royal Brompton and Harefield NHS Foundation Trust London; Professor of Practice (Thoracic Imaging) National Heart and Lung Institute Imperial College London, UK

Anand Devaraj, MD, FRCR, MRCP Consultant Thoracic Radiologist Department of Radiology Royal Brompton and Harefield NHS Trust London, UK

Simon P. Doran, MB BCh BAO, MRCPI Radiology Registrar and Lecturer Centre For Advanced Medical Imaging St James Hospital Dublin, Ireland

https://afkebooks.com

xii

List of Contributors

Marina Easty, MBBS, BSc, MRCP, FRCR, PGCertNucMed Consultant Paediatric Radiologist Radiology Department Great Ormond Street Hospital London, UK

Andrew J. Evans, MB ChB, MRCP, FRCR Professor of Breast Imaging University of Dundee; Honorary Consultant Radiologist NHS Tayside Ninewells Hospital and Medical School Dundee, UK

Afshin Gangi, Professor of Radiology University of Strasbourg France; Guy’s and St Thomas’ Hospitals London, UK

Robert N. Gibson, MBBS, MD, FRANZCR, DDU Professor Department of Radiology Royal Melbourne Hospital Parkville Victoria, Australia

Andrew J. Grainger, BM, BS, MRCP, FRCR Consultant Musculoskeletal Radiologist Addenbrooke’s Hospital Cambridge University Hospitals NHS Foundation Trust Cambridge, UK

Claudio Granata, MD Consultant Radiologist Department of Radiology IRCCS Giannina Gaslini Italy

Philippe A. Grenier, MD Rossella Fattori, MD, PhD Professor of Radiology; University of Bologna Italy

Catalin I. Fetita, PhD Professor Department ARTEMIS Telecom SudParis Institut Mines-Telecom CNRS Evry, France

Jonathan H. Gillard, BSc, MA, MD, FRCP, FRCR, MBA Professor of Neuroradiology University of Cambridge Addenbrooke’s Hospital Cambridge, UK

Fergus Gleeson, MD, FRCP, FRCR Radiologist Department of Radiology Churchill Hospital Oxford, UK

Tomás Franquet, MD

Edmund M. Godfrey, MA MRCS FRCR

Chief Section of Thoracic Imaging Department of Radiology Hospital de Sant Pau; Associate Professor of Radiology Universitat Autonoma de Barcelona Barcelona, Spain

Consultant Radiologist Department of Radiology Cambridge University Hospitals NHS Foundation Trust Cambridge, UK

Alan H. Freeman, MBBS, FRCR Consultant Radiologist Department of Radiology Addenbrooke’s Hospital Cambridge University Hospitals NHS Foundation Trust Cambridge, UK

Vicky Goh, MA, MB BChir, MD MRCP, FRCR Chair of Clinical Cancer Imaging Division of Imaging Sciences and Biomedical Engineering King’s College London; Honorary Consultant Radiologist Guy’s and St Thomas’ Hospitals London, UK

Susan Freeman, MRCP, FRCR Consultant Radiologist Department of Radiology Cambridge University Hospitals NHS FoundationTrust Cambridge, UK

Ferdia A. Gallagher, PhD, MRCP, FRCR

Beatriz Gomez Anson, MD, PhD, FRCR Clinical Head of Neuroradiology Unit of Neuroradiology Department of Radiology Hospital Santa Creu i Sant Pau Universitat Autonoma Barcelona, Spain

Cancer Research UK Senior Cancer Research Fellow and Reader in Molecular Imaging University of Cambridge; Honorary Consultant Radiologist Department of Radiology Addenbrooke’s Hospital Cambridge, UK

Professor Department of Radiology Service d’imagerie Hopital Foch Suresnes, France

Roxana S. Gunny, MBBS, MRCP, FRCR Consultant Neuroradiologist Department of Radiology Great Ormond Street Hospital London, UK

Mohammed Hadi, MD, FRCR, MRCS, MBChB Specialist Trainee Interventional Radiology John Radcliffe Hospital Oxford, UK

Wasim Hakim, BSc, MBBS, MRCS, FRCR Interventional Radiology Fellow Department of Radiology St Mary’s Hospital London, UK

Charles B.O. Hall, BSc, MBBS, FRCR Interventional Neuroradiology Fellow Imperial College Healthcare NHS Trust London, UK

Mohamad Hamady, MBChB, FRCR, EBIR Consultant Interventional Radiologist Department of Clinical Radiology Imperial College NHS Trust London, UK

Ieneke J.C. Hartmann, MD, PhD Radiologist Radiology Maasstad Ziekenhuis Rotterdam, The Netherlands

https://t.me/afkebooks

List of Contributors

xiii

Roisin M. Heaney, MB BCh, BAO, BA

Cylen Javidan-Nejad, MD

Aoife N. Keeling, FFRRCSI, MRCPI, MSc

Department of Radiology St James’s Hospital Dublin, Ireland

Associate Professor Diagnostic Radiology Mallinckrodt Institute of Radiology Washington University in St Louis St Louis, MO, USA

Consultant Interventional Radiologist Beaumont Hospital Dublin, Ireland

Susan Jawad, MBBS, BSc(Hons), FRCR

Consultant Radiologist School of Biomedical Engineering and Imaging Sciences King’s College London and Department of Radiology Guy’s and St Thomas’ Hospitals London, UK

Peter Hoskin, MD, FRCP, FRCR Consultant in Clinical Oncology Cancer Centre Mount Vernon Hospital Northwood; Professor in Clinical Oncology University of Manchester Manchester; Honorary Consultant in Clinical Oncology Christie Hospital Manchester; Honorary Consultant in Clinical Oncology University College London Hospital London, UK

Consultant Head and Neck Radiologist Department of Radiology University College Hospitals NHS Fountation Trust London, UK

Karl Johnson, BSC MB ChB, MRCP, FRCR Consultant Paediatric Radiologist Birmingham Children’s Hospital Birmingham, UK

Philip M. Hughes, MBBS, MRCP, FRCR

Hefin Jones, MBBS BSc, FRCR

Consultant Radiologist Radiology Plymouth Hospitals Trust Plymouth, Cornwall, UK

Consultant Cardiothoracic Radiologist Department of Imaging University Hospitals North Midlands Stoke-on-Trent, UK

Paul D. Humphries, BSc, MBBS, MRCP, FRCR

Brynmor P. Jones, BSc(Hons), MBBS, MRCP, FRCR

Consultant Paediatric Radiologist University College London Hospital NHS trust and Great Ormond Street Hospital London, UK

Consultant Neuroradiologist Imperial College Healthcare NHS Trust Charing Cross Hospital London, UK

H. Rolf Jäger, MD, FRCR

Geeta Kapoor, MBBS, FRCR

Professor in Neuroradiology Department of Brain Repair and Rehabilitation UCL Institute of Neurology UCL Faculty of Brain Sciences; Consultant Neuroradiologist Lysholm Department of Neuroradiology National Hospital for Neurology and Neurosurgery and Department of Imaging University College London Hospitals London, UK

Radiology Registrar Department of Radiology Norfolk and Norwich University Hospital Norwich, UK

Steven L.J. James, MB, ChB, FRCR

Assistant Professor Department of Interventional Radiology Patras University Hospital Patras, Greece

Consultant Musculoskeletal Radiologist The Royal Orthopaedic Hospital NHS Foundation Trust Birmingham, UK

Ram S. Kasthuri, MRCS, FRCR, EBIR Consultant Interventional Radiologist Interventional Radiology Queen Elizabeth University Hospital Glasgow, UK

Christian Kelly-Morland, FRCR, MRCS, MBBS, BSc

Lizbeth M. Kenny, MBBS, FRANZCR, FACR(Hon), FBIR(Hon), FRCR(Hon), FCIRSE Professor Cancer Care Services Royal Brisbane and Women’s Hospital Herston, Queensland, Australia

Fleur Kilburn-Toppin, MA, MB BChir, FRCR Consultant Radiologist Addenbrooke’s Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

Orpheus Kolokythas, MD, FSAR Associate Professor Department of Radiology University of Washington Seattle, WA, USA

Helmut Kopf, MD, MSc Radiologist Institute for Diagnostic and Interventional Radiology Göttlicher Heiland Hospital Vienna, Austria

Stéphane Kremer, MD, PhD Konstantinos Katsanos, MSc, MD, PhD EBIR

Jonathan J. James, BMBS, FRCR

Richard G. Kavanagh, MB BCh, BAO, BSc, MCh, FFRRCSI

Consultant Radiologist Nottingham Breast Institute City Hospital Nottingham, UK

Radiology Resident Radiology Cork University Hospital Cork, Ireland

https://afkebooks.com

Professor Imagerie 2 CHU de Strasbourg Strasbourg, France

Christian R. Krestan, MD Associate Professor of Radiology; Head Trauma Radiology Section Department of Biomedical Imaging und Image-guided Therapy Division of General and Pediatric Radiology Medical University of Vienna - Vienna General Hospital Vienna, Austria

xiv

List of Contributors

Lucia J.M. Kroft, MD, PhD

Lorenzo Mannelli, MD, PhD

Alpesh Mistry, MBChB, BSc, FRCR

Radiologist Department of Radiology Leiden University Medical Centre Leiden, The Netherlands

Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, USA

Consultant Musculoskeletal Radiologist Department of Radiology Royal Liverpool and Broadgreen University Hospitals Liverpool, UK

Eamon Lagha, BEng, MBChB

Rachel M. Martin, MB BCh, BAO, MA, SEM

Specialist Registrar in Radiology Department of Radiology Guy’s & St Thomas’ NHS Foundation Trust London, UK

Consultant Radiologist Department of Radiology Craigavon Area Hospital Craigavon, UK

Radhesh Lalam, MBBS, MRCS, FRCR

Tomasz Matys, PhD, FRCR

Consultant Radiologist Department of Diagnostic Radiology The Robert Jones and Agnes Hunt Orthopaedic and District Hospital NHS Foundation Trust Oswestry, UK

University Lecturer and Honorary Consultant Radiologist Department of Radiology University of Cambridge UK

Michael J. Lee, MSc, FRCPI, FRCR, FFR (RCSI), FSIR, EBIR Professor of Radiology Royal College of Surgeons in Ireland; Consultant Interventional Radiologist and Clinical Director Beaumont Hospital Dublin, Ireland

David J. Lomas, MA, MB BChir, FRCR, FRCP Professor of Clinical MRI University Radiology Department Addenbrooke’s Hospital Cambridge, UK

Luigi Lovato, MD Radiologist Cardiovascular Radiology Unit Cardiothoracic Radiology Cardiovascular-Thoracic Department S.Orsola-Malpighi Hospital Bologna, Italy

David MacVicar, MA, FRCP, FRCR, FBIR Consultant Radiologist Department of Diagnostic Radiology Royal Marsden Hospital Sutton, Surrey UK

Michael M. Maher, MD, FRCSI, FFR(RCSI), FRCR Professor of Radiology University College Cork; Consultant Radiologist Cork University Hospital and Mercy University Hospital Cork, Ireland

Patrick D. McLaughlin, MB BCh, BAO, BMedSc, FFR, RCSI Lecturer in Radiology Department of Radiology University College Cork Cork, Ireland

Katherine Miszkiel, BM(Hons), MRCP, FRCR Consultant Neuroradiologist Lysholm Department of Neuroradiology The National Hospital for Neuroradiology and Neurosurgery; Honorary Consultant Neuroradiologist Moorfields Eye Hospital London, UK

Robert A. Morgan, MB ChB, MRCP, FRCR, EBIR Consultant Vascular and Interventional Radiologist Radiology Department St George’s NHS Trust London, UK

Kristian H. Mortensen, MD, PhD, FRCR Eugene McNally, MB Bch, BAO, FRCPI, FRCR Consultant Musculoskeletal Radiologist Department of Radiology Oxford Musculoskeletal Radiology Oxford, UK

James F.M. Meaney, FRCR, FFRRCSI, FISMRM Professor Department of Radiology St James Hospital Dublin and Trinity College Dublin Ireland

Amrish Mehta, MBBS, BSc(Hons), FRCR

Consultant Imager Cardiorespiratory Unit Great Ormond Street Hospital for Children London, UK

Jonathan G. Moss, MB ChB, FRCS(Ed), FRCR Professor of Interventional Radiology Department of Radiology North Glasgow University Hospitals Gartnavel General Hospital Glasgow, UK

Timothy E. Murray, MB MCh, MRCS, FFRRCSI

Consultant Neuroradiologist Department of Imaging Imperial College Healthcare NHS Trust London, UK

Vascular and Interventional Radiology Fellow Department of Radiology Beaumont Hopsital Dublin, Ireland

Gustavo Meirelles, MD, PhD

Arjun Nair, MD MRCP FRCR

Medical Manager and Head of Thoracic Imaging Radiology Fleury Group São Paulo, Brazil

Consultant Cardiothoracic Radiologist Department of Radiology University College London Hospitals NHS Foundation Trust London, UK

Caroline Micallef, MD, FRCR

Gurinder S. Nandra, MBChB, BSc, FRCR

Consultant Neuroradiologist National Hospital for Neurology and Neurosurgery University College London Hospitals London, UK

Radiology Registrar St George’s University Hospitals NHS Foundation Trust London, UK

https://t.me/afkebooks

List of Contributors

xv

Luigi Natale, MD

Kevin O’Regan, MB BCh, BAO, FFRRCSI

Andrew Plumb, BMBCh, MRCP, FRCR

Researcher and Aggregate Professor of Radiology Radiological Sciences Department Catholic University of Sacred Heart Polyclinic Gemelli Foundation IRCCS Rome, Italy

Consultant Radiologist Department of Radiology Cork University Hospital Cork, Ireland

Research Fellow in Gastrointestinal Radiology University College London Hospitals London, UK

Catherine M. Owens, BSc, MBBS, MRCP, FRCR

Sanjay Popat, FRCP, PhD

Owen J. O’Connor, MD, FFR(RCSI) MB BCh, BAO, BMedSci

Consultant Radiologist and Honorary Reader Department of Imaging Great Ormond Street Hospital London, UK

Consultant Radiologist Department of Radiology Cork University Hospital and Mercy University Hospital Cork; Senior Lecturer Department of Radiology University College Cork Cork, Ireland

Philip O’Connor, MBBS, MRCP, FRCR, FFSEM(UK) Musculoskeletal Radiologist Clinical Radiology Leeds Teaching Hospitals Trust The University of Leeds West Yorkshire, UK

Paul O’Donnell, MBBS, MRCP, FRCR Consultant Radiologist Royal National Orthopaedic Hospital; Honorary Senior Lecturer University College London London, UK

Amaka C. Offiah, BS, MBBS, MRCP, FRCR, PhD HEFCE Clinical Senior Lecturer Academic Unit of Child Health University of Sheffield; Consultant Paediatric Radiologist Academic Unit of Child Health Sheffield Children’s NHS Foundation Trust Sheffield, UK

Øystein E. Olsen, PhD Consultant Radiologist Radiology Department Great Ormond Street Hospital for Children London, UK

Lil-Sofie Ording Müller, MD, PhD Consultant Paediatric Radiologist Section for Paediatric Radiology Division of Diagnostics and Intervention Oslo University Hospital Ullevål Oslo, Norway

Simon P.G. Padley, BSc, MBBS, FRCP, FRCR Professor of Practice (Diagnostic and Interventional Radiology) Imperial College School of Medicine; Consultant Radiologist Cross-site Director of Radiology Royal Brompton and Harefield Hospital London, UK

Paul M. Parizel, MD, PhD Professor and Chair Department of Radiology Antwerp University Hospital University of Antwerp Edegem, Belgium

Nadeem Parkar, MD Department of Diagnostic Radiology Cleveland Clinic Main Campus Cleveland, OH, USA

Anish Patel, MBChB, MRCP, FRCR Consultant Radiologist Clinical Imaging Royal Orthopaedic Hospital Birmingham, UK

Uday Patel, MB ChB, MRCP, FRCR Consultant Radiologist Department of Radiology St George’s Hospital London, UK

Janina M. Patsch, MD, PhD Assistant Professor Diagnostic Imaging and Image-Guided Therapy Medical University of Vienna Vienna, Austria

Francesca Benedetta Pizzini, MD, PhD Neuroradiologist Department of Diagnostics and Pathology Verona University Hospital Verona, Italy

https://afkebooks.com

Consultant Medical Oncologist Royal Marsden Hospital London, UK

Stephen P. Power, MB BCh, BAO, BPHarm, MRCPI, FFRCSI Specialist Registrar Department of Radiology Cork University Hospital Cork, Ireland

Davide Prezzi, FRCR Consultant Radiologist School of Biomedical Engineering and Imaging Sciences King’s College London; Department of Radiology Guy’s and St Thomas’ Hospitals London, UK

Michael A. Quail, MSc, MB ChB(Hons), MRCPCH British Heart Foundation Clinical Research Training Fellow Centre for Cardiovascular Imaging UCL Institute of Cardiovascular Science; Paediatric Cardiology Academic Clinical Fellow Department of Cardiology Great Ormond Street Hospital London, UK

James J. Rankine, MB ChB, MRCP, MRaD, FRCR, MD Consultant Radiologist and Honorary Clinical Associate Professor Department of Radiology Leeds General Infirmary Leeds, West Yorkshire, UK

Lakshmi Ratnam, MBChB, MRCP, FRCR Consultant Interventional Radiologist Radiology St George’s Hospital London, UK

Bhavin Rawal, MBBS, FRCR Clinical Research Fellow Department of Radiology Royal Brompton Hospital London, UK

xvi

List of Contributors

Jim A. Reekers, MD, PhD, EBIR

Vincenzo Russo, MD, PhD

Professor of Radiology Department of Radiology Academic Medical Centre Amsterdam University of Amsterdam The Netherlands

Radiologist CardioVascular Radiology Unit Cardio-Thoracic Radiology S.Orsola-Malpighi Hospital Bologna, Italy

John H. Reynolds, MMedSci, FRCR, DMRD

Elizabeth E. Rutherford, BMedSci, MBBS, MRCS, FRCR

Consultant Radiologist Department of Radiology Birmingham Heartlands Hospital Birmingham, UK

Consultant Radiologist University Hospitals Southampton NHS Trust Southampton, UK

Rodney H. Reznek, MA, FRANZCR(Hon), FFR RCSI(Hon), FRCP, FRCR

Tarun Sabharwal, FRCR, FRCRI, EBIR, FRSIR, FCIRSE

Emeritus Professor of Cancer Imaging Cancer Institute Queen Mary’s University London St Bartholomew’s Hospital West Smithfield London, UK

Michael Riccabona, OA Professor Department of Radiology Division of Pediatric Radiology Universitätsklinikum- LKH Graz, Austria

Karen Rosendahl, MD, PhD Consultant Paediatric Radiologist Haukeland University Hospital; Professor Department of Clinical Medicine University of Bergen Norway

Giles Rottenberg, MBBS, MRCP, FRCR Consultant Radiologist Department of Radiology Guy’s and St Thomas’ NHS Foundation Trust London, UK

Àlex Rovira, MD Head of Magnetic Resonance Unit (IDI) Department of Radiology Vall d’Hebron University Hospital Barcelona, Spain

Emma L. Rowbotham, MB BChir, MRCS, FRCR Musculoskeletal Radiology Consultant Department of Radiology Leeds Teaching Hospitals Leeds, UK

Consultant Interventional Radiologist Clinical Lead Interventional Radiology Guy’s and St Thomas’ Hospitals London, UK

Anju Sahdev, MBBS, MRCP, FRCR Consultant Uro-Gynae Radiologist Department of Imaging St Bartholomew’s Hospital Barts Health West Smithfield London, UK

Asif Saifuddin, BSc(Hons), MB ChB, MRCP, FRCR Consultant Musculoskeletal Radiologist Imaging Department The Royal National Orthopaedic Hospital Stanmore Middlesex, UK

Evis Sala, MD, PhD, FRCR Professor of Oncological Imaging Department of Radiology University of Cambridge Cambridge, UK

Dawn E. Saunders, MBBS, MD, MRCP, FRCR Honorary Senior Lecturer Radiology & Physics UCL Institute of Child Health London, UK

Wolfgang Schima, MD, MSc Professor Department of Diagnostic and Interventional Radiology KH Göttlicher Heiland Hospital; KH der Barmherzigen Schwestern Wien; Sankt Josef-Krankenhaus Vienna, Austria

https://t.me/afkebooks

Daniel J. Scoffings, BSc(Hons), MBBS, MRCP(UK), FRCR Consultant Neuroradiologist Department of Radiology Addenbrooke’s Hospital Cambridge, UK

Nicholas J. Screaton, FRCR, FRCP Consultant Radiologist Department of Radiology Royal Papworth Hospital Cambridge, UK

Edward Sellon, MRCS, MSc (SEM), DipESSR, FRCR, RAMC Consultant Radiologist Department of Radiology Oxford University Hospitals Oxford, UK

Thomas R. Semple, FRCR MBBS BSc(Hons) The Royal Brompton Hospital London, UK

Eva M. Serrao, MD, PhD Academic Clinical Lecturer Department of Radiology University of Cambridge School of Clinical Medicine Cambridge, UK

Nadeem Shaida, MBBS, FRCR, FHEA Consultant Radiologist Department of Radiology Cambridge University Hospitals NHS Foundation Trust Cambridge, UK

Amreen Shakur, MBBS, BMedSci (Hons) Department of Radiology Addenbrooke’s Hospital Cambridge, UK

Nasim Sheikh-Bahaei, MD, MRCP, FRCR, PhD Assistant Professor of Radiology and Neurology Keck School of Medicine of USC University of Southern California Los Angeles, CA, USA

Beth Shepherd, MBBS, MA(cantab), MRCS, FRCR Consultant Oncological Radiologist Department of Radiology University Hospitals Southampton Southampton, UK

List of Contributors

Hans-Marc J. Siebelink, MD, PhD Cardiologist Non-Invasive Imaging Department of Cardiology Leiden University Medical Center Leiden, The Netherlands

Mario Silva, MD, PhD Assistant Professor Section of Radiology Unit of Surgical Sciences Department of Medicine and Surgery (DiMeC) University of Parma Parma, Italy

Jaspreet Singh, MBBS, MRCP, FRCR Consultant Musculoskeletal Radiologist Department of Radiology Robert Jones and Agnes Hunt Orthopaedic Hospital Oswestry, UK

Jan Smith, MB BChir, PhD, FRCR Specialist Registrar in Radiology Department of Radiology Addenbrooke’s Hospital Cambridge, UK

Pia C. Sundgren, MD, PhD Professor of Radiology; Head Department of Diagnostic Radiology Clinical Sciences Lund Lund University; Center for Medical Imaging and Physiology Skåne University Hospital Lund, Sweden

Tom R. Sutherland, MBBS, MMed, FRANZCR Radiologist and Director of Ultrasound St Vincents Hospital Melbourne, Victoria, Australia

Andrew M. Taylor, BA(Hons), BM BCh, MRCP, FRCR Professor of Cardiovascular Imaging Centre for Cardiovascular Imaging UCL Institute of Cardiovascular Science and Great Ormond Street Hospital for Children London, UK

Stuart A. Taylor, MBBS, BSc, MD, MRCP, FRCR Professor of Medical Imaging Centre for Medical Imaging University College London London, UK

Avnesh S. Thakor, BA, MA, MSc, MD, PhD, MB BChir, FHEA, FRCR Fellow in Interventional Radiology University of Cambridge UK; Visiting Scholar Molecular Imaging Program Stanford University, CA, USA

Majda M. Thurnher, MD Associate Professor of Radiology Medical University Vienna Department of Biomedical Imaging and Image-guided Therapy Vienna, Austria

Professor Unit of Radiology Department of Medicine and Surgery (DiMeC) University of Parma Parma, Italy

Rolf Symons, MD, PhD Radiologist Department of Imaging and Pathology University Hospitals Leuven Leuven, Belgium

Philippa Tyler, MBBS, BSc, MRCS, FRCR Consultant Musculoskeletal Radiologist Imaging Department The Royal National Orthopaedic Hospital Stanmore Middlesex, UK

Raman Uberoi, BMSCPath, MBBchir, MRCP, FRCR, EBIR Consultant Interventional Radiologist Oxford University Hospitals Oxford, UK

Luc van den Hauwe, MD Consultant Radiologist Department of Radiology University Hospital Antwerp Edegem; Department of Radiology AZ KLINA Brasschaat, Belgium

Johan W. van Goethem, MD, PhD Professor Department of Radiology University Hospital Antwerp Antwerp Edegem; Radiologist Department of Radiology AZ Nikolaas Sint-Niklaas, Belgium

Steffie Thust, MD, FRCR “Consultant Neuroradiologist Lysholm Department of Neuroradiology National Hospital of Neurorology and Neurosurgery London; Honorary Senior Lecturer Department of Brain Rehabilitation and Repair UCL Institute of Neurology London, UK

Andoni P. Toms, MBBS, PhD, FRCR Nicola Sverzellati, MD, PhD

xvii

Professor Department of Radiology Norfolk and Norwich University Hospital Norwich, UK

Maria Tsakok, BM, BCh, BA(Hons) Academic Clinical Fellow in Clinical Radiology Department of Radiology Oxford University Hospitals NHS Trust Oxford, UK

https://afkebooks.com

Thomas Van Thielen, MD Radiologist Department of Radiology AZ Klina Brasschaat, Belgium; Antwerp University Hospital Edegem, Belgium

Sibu Varghese, MMBS, MRCP, PhD Consultant Gastroenterologist Department of Gastroenterology Cambridge University Hospitals NHS Trust Cambridge, UK

Hema Verma, MB, BS, MRCP, FRCR Consultant Radiologist Department of Radiology Guy’s and St Thomas’ NHS Foundation Trust London, UK

Johny A. Verschakelen, MD, PhD Director Chest Radiology Department of Radiology University Hospitals Leuven Leuven, Belgium

xviii

List of Contributors

Geert M. Villeirs, MD, PhD Clinical Head of Department Division of Genitourinary Radiology Ghent University Hospital Ghent, Belgium

Sarah J. Vinnicombe, BSc(Hons), MRCP, FRCR Consultant Radiologist Thirlestaine Breast Unit Cheltenham General Hospital Cheltenham; Honorary Senior Lecturer Cancer Research Ninewells Hospital and Medical School University of Dundee Dundee, UK

Victoria S. Warbey, MA, MBBS, FRCR, MSc, FRCP

Reddi Prasad Yadavali, MBBS, MRCS, FRCR, EBIR

King’s College London & Guy’s and St Thomas’ PET Centre St Thomas’ Hospital London, UK

Consultant Interventional Radiologist Department of Radiology Manipal Hospitals Bengaluru, India; Adjunct Professor of Radiology Manipal University Manipal Academy of Higher Education Manipal, India

Tom A. Watson, MBChB, FRCR Consultant Paediatric Radiologist Department of Imaging Great Ormond Street Hospital London, UK

Jos J.M. Westenberg, PhD Associate Professor Department of Radiology Leiden University Medical Center Leiden, The Netherlands

https://t.me/afkebooks

Adam Kenji Yamamoto, MBBS, BSc (Hons), MRCS, FRCR Consultant Neuroradiologist Lysholm Department of Neuroradiology National Hospital for Neurology and Neurosurgery London, UK

In Memoriam Ronald Grainger (1922–2014)

https://afkebooks.com

1  Current Status of Thoracic Imaging Arjun Nair, Joseph L. Barnett, Thomas R. Semple

CHAPTER OUTLINE Chest Radiography, 3 Computed Tomography of the Thorax, 4 Ultrasound, 11

Magnetic Resonance Imaging, 13 Radionuclide Imaging, 13

Chest radiography and computed tomography (CT) remain the stalwarts of thoracic imaging. The basic technique of chest radiography has remained largely unchanged since its inception, but continuing developments in image receptor technology have resulted in techniques which are simultaneously efficient and radiation dose-optimised. Radiographs are now mainly produced in digital format, thus facilitating their incorporation into picture archiving and communications systems (PACS). Evolving CT technology has meant that multidetector row CT (MDCT) systems have largely replaced single-detector CT. Newer dual-energy CT (DECT) systems may provide new applications, and these continue to be validated. The latest generation of iterative reconstruction techniques can be manipulated to achieve minimal image noise and lower radiation doses. MDCT protocols integrating these various aspects of novel technology are being continuously refined to strike a balance between obtaining diagnostically adequate information and dose minimisation. Ultrasound and magnetic resonance imaging (MRI) are increasingly being applied in the investigation of specific thoracic diseases. Positron emission tomography (PET) fused with CT (PET–CT) now has an established role in the investigation of neoplastic disease, enabling the simultaneous assessment of metabolic function, anatomical location, and unsuspected extrathoracic metastatic disease.

CHEST RADIOGRAPHY

consequent high retake rate, as well as inflexibility in image display and manipulation. The exponential advances in computational power, storage capacity and detector technology have led to the replacement of filmscreen radiography by PACS and digital imaging systems in most modern imaging departments. Early digital imaging systems introduced over 30 years ago—generally termed ‘computed radiography’ (CR)—used a photostimulable phosphor image receptor plate, and continue to be used in some departments because of their compatibility with conventional radiography equipment. However, CR systems have largely been superseded by direct radiography (DR) systems (Table 1.1). DR systems employ either direct or indirect methods for converting x-ray photons into electrical charges, thereby generating an electrical signal that can be read directly. Direct conversion may be achieved by photoconductors within flat-panel detectors (FPDs) (most commonly amorphous selenium), or using a selenium drum. Indirect conversion involves the use of a scintillator associated with either a charge-coupled device (CCD) or FPD. Scintillators most commonly use thallium-doped caesium iodide-based or, more recently, gadolinium-based compounds. Both CR and DR systems offer many advantages over conventional film-screen radiography, including wider latitude (hence reducing error and repeat examination rates), reusable detectors, almost seamless integration with PACS, and (definitely in the case of DR) superior image quality.

Radiographic Projections

Equipment Considerations Chest radiography remains the commonest diagnostic radiographic procedure. Chest radiographs were traditionally acquired with conventional film-screen radiography systems that provide, at low cost, good image quality and high spatial resolution. However, film-screen radiography is limited by a relatively narrow exposure range and

SUMMARY BOX • Chest radiography remains the mainstay of thoracic imaging • Novel techniques such as digital tomosynthesis, while useful, are still expensive and limited in availability

Frontal and lateral projections of the chest are adequate for most purposes. Other radiographic views (e.g. lordotic, expiratory, or portable lateral shoot-through radiographs) are now usually performed only when CT is not readily available. The lateral decubitus view, taken as a frontal projection with a horizontal beam and the patient lying on his or her side, is still sometimes used to identify an effusion that is not visible on an erect chest radiograph to demonstrate the movement of fluid in the pleural space or to localise or confirm an equivocal opacity seen on a frontal projection.

Portable Chest Radiography CR systems remain the most widely used portable radiography system. They overcome some of the limitations of portable chest radiography

3 https://t.me/afkebooks

4

SECTION A  The Chest and Cardiovascular System

TABLE 1.1  Methods of Computed Radiography and Direct Radiography X-Ray Conversion Method Devices Image Readout

Computed Radiography

Direct Radiography

Indirect conversion

Indirect conversion

Direct conversion

Removable image plates using storage phosphors (analogous to conventional film cassettes) Separate readout process: detector layer must be analysed by laser (point scan or line-scan); resulting output is converted into electrical signal

Scintillator-thin-film transistor array Scintillator-charge-coupled device Direct readout process: x-rays are converted immediately into electrical signal and read

Selenium drum Photoconductor flat-panel detector

BOX 1.1  Limitations of Portable Chest

Radiography

• Scattered radiation • Inability of the radiograph to capture all relevant information • Significant underexposure leads to unacceptable image noise and low contrast resolution. • Shorter focus–detector distance results in undesirable, and sometimes misleading, magnification of structures • High kilovoltage techniques cannot be used because portable machines are unable to deliver a sufficiently high kilovoltage, and as the maximum current is limited, long exposure times are needed, increasing movement artefact

(Box 1.1) by controlling optical density and contrast, but are unable to overcome the problem of scatter inherent to all portable systems. An increasing array of portable DR detectors is now being made available, including FPDs, those that can be integrated into existing CR cassettes for flexible positioning, and those with wireless transmission capabilities to allow immediate transfer to PACS, which may substantially improve technologists’ workflow. Such developments are likely to encourage the uptake of portable DR systems in the near future, but at present they still remain relatively costly.

Novel Radiographic Techniques The techniques of digital tomosynthesis, dual-energy subtraction radiography and temporal subtraction radiography, although still confined mainly to research studies, have potential clinical application, chiefly for nodule detection. The decreasing cost of these systems—in particular, of dual-energy subtraction radiography—has enabled their use as a routine chest radiographic technique in some centres. Dual-energy subtraction radiography is essentially a bone suppression technique that takes advantage of the differential attenuation of x-ray photons of high atomic number materials (such as calcium and iodine) at different photon energies. Such differential attenuation causes the contrast from calcium and bone in a high kVp image to be reduced. As such, subtraction of the low-energy from the high-energy image allows subtraction of obscuring bony structures, potentially increasing pulmonary nodule conspicuity (Fig. 1.1).

COMPUTED TOMOGRAPHY OF THE THORAX Principles The fundamental principles that apply to CT of any anatomical part, including the thorax, have remained essentially unchanged since CT’s inception by Sir Godfrey Hounsfield. The components required are: (1) an x-ray source emitting a fan beam; (2) a rotating gantry housing the x-ray source; and (3) a ring composed of a detector array, located diametrically opposite the x-ray source. The patient’s thorax is interpolated

in either craniocaudal or caudocranial directions through the ring. The x-rays penetrate the thorax and reach the detector; signals emitted from the detector are reconstructed by a computer, and the resulting image is an anatomical ‘map’ quantifying the different densities within the scanned volume, displayed by three-dimensional (3D) pixels (termed voxels). Depending on the type of CT system, thicker or thinner CT image slices can then be reconstructed by the computer as necessary. SUMMARY BOX • With modern multidetector computed tomography (CT), high resolution CT reconstructions of the entire volume of the thorax can be obtained from virtually any thoracic CT acquisition • Visualisation in the axial (or transverse) plane remains the main method of CT review, but multiple three-dimensional post-processing techniques complement this review • Multiple dose reduction strategies are increasingly available

Reconstruction algorithms, sometimes also called ‘kernels’ or ‘filters’ by different manufacturers, are applied to the CT data set during reconstruction to optimise visualisation of different structures. Low spatial resolution algorithms are used to reduce image noise and improve contrast, at the expense of spatial resolution, and as such are used to optimise soft tissue and vascular visualisation. Conversely, high spatial resolution reconstruction algorithms (so-called ‘lung’ or ‘sharp’ algorithms) enhance fine structural detail with the penalty of increased image noise. Finally, it is necessary to manipulate the available greyscale on a standard viewing monitor to optimally display structures of particular interest within the thorax. A particular greyscale setting is termed a window. Different window settings are required for the lung parenchyma (requiring wider windows) compared with those for the mediastinum (requiring narrower windows). As a general rule, the window centre level is usually halfway between the density of the structure to be measured and the density of the surrounding tissue. Default window settings are available on PACS and CT viewing systems, but manual adjustment is often required: for instance, when administered contrast agents appear too dense and may obscure abnormalities such as a pulmonary embolus.

From Single-Slice to Multidetector Computed Tomography The introduction of spiral (helical) CT in the early 1990s fundamentally altered thoracic CT imaging, in particular the discontinuous acquisition of data in conventional, sequential single-slice CT could now be replaced with volumetric data acquisition. In 1998 several CT manufacturers introduced multidetector systems, which provided considerable improvement in acquisition speed, coverage, and temporal and spatial resolution. Since then, there has been rapid improvement in CT performance with increased numbers of detector rows and faster tube rotation; currently,

https://afkebooks.com

CHAPTER 1  Current Status of Thoracic Imaging

A

5

B

Fig. 1.1  Series of Dual-Energy Subtraction Chest Radiographs in a Man With a Smoking History. A possible right apical opacity is seen on a conventional posteroanterior radiograph (A), with a more dubious left apical opacity projected over the left first costochondral junction. The right apical opacity is made conspicuous, while the left apical opacity disappears on a bone-subtracted image (B); conversely, the right apical opacity is absent while the left apical opacity becomes more apparent on the soft-tissue subtracted image (C). Thus, the right apical opacity is an actual soft-tissue pulmonary nodule (confirmed on subsequent computed tomography), while the left apical opacity is actually calcification of the first costochondral junction. (Images courtesy Dr Teodora Wetscherek, Consultant Thoracic Radiologist, Addenbrooke’s Hospital, Cambridge, UK.)

systems with up to 320 active detector rows are available. Rotation times of the x-ray tubes are now as low as 0.33 seconds per rotation. Recent developments in CT technology have not decreased this rotation time further; however, novel methods are used for increasing temporal and spatial resolution, with either dual-source technology (more recently augmented with high-pitch acquisition), a ‘flying focal spot’ (converting a 128-detector row array into a virtual 256-detector row array), or wide-area detectors (typically 16 cm wide in the z-axis), effectively providing 320-detector row coverage. Three major benefits arise from these technological advances: (1) single breath-hold imaging, (2) reduction in motion artefacts, which in paediatric practice has lessened the requirement for sedation; and (3) improved spatial resolution. With respect to the latter, MDCT systems permit reconstructions of varying slice thickness by collimating and adding together the signals of neighbouring detector rows, with overlap between these rows if necessary. Hence, from the same data set, both

C

narrow sections (0.6 to 1.25 mm thickness) for high spatial resolution detail or 3D postprocessing and wide sections (2.5 to 5 mm) can be produced for better contrast resolution or quick review. The convenience of a single protocol is particularly useful for patients with a suspected focal lesion as well as diffuse interstitial lung disease. Thin-section reconstructions are recommended for volumetric assessment and characterisation of pulmonary nodules, the evaluation of interstitial lung disease and the evaluation of pulmonary embolism. For the initial assessment of mediastinal masses and for lung cancer staging studies, 3- to 5-mm reconstructions are usually adequate.

High-Resolution Computed Tomography The essential requirements for a high-resolution CT (HRCT) technique are: (1) thin collimation, usually 1 to 2 mm, (2) reconstructions using a high spatial frequency (‘sharp’ or ‘lung’) algorithm and (3) single breath-hold imaging, to obtain images undistorted by respiratory motion

https://t.me/afkebooks

6

SECTION A  The Chest and Cardiovascular System

artefact. The improved spatial resolution afforded by the thin collimation and the ‘sharp’ algorithm enhances the detection of key morphological features in HRCT interpretation, such as thickened interlobular septa, ground-glass opacification and abnormal subsegmental airways. Reducing the section thickness below 1 mm will not yield any significant further improvement in spatial resolution and at the same time will reduce the signal-to-noise ratio of the image. A sharp reconstruction algorithm reduces image smoothing and makes structures visibly sharper, although image noise becomes more obvious. Intravenous (IV) contrast medium should be avoided unless there is another clinical indication necessitating its use (such as pulmonary embolism) because it can spuriously increase parenchymal opacification and interfere with interpretation, especially in comparison examinations (Fig. 1.2). Images are usually obtained in the supine position from the apices to the lung bases at full inspiration. When early interstitial fibrosis is suspected, HRCT is often performed in the prone position to prevent confusion with the increased opacification often seen in the dependent posterobasal segments in the usual supine position. However, there is no advantage in prone CT if there is obvious diffuse lung disease on a contemporary chest radiograph.

The necessity of expiratory CT sections is somewhat controversial. Although images at end-expiration can reveal small or subtle areas of air trapping, the mosaic attenuation pattern attributable to small airways disease is usually apparent, albeit less conspicuous, on inspiratory images in most patients with clinically significant small airways disease, and can be readily accentuated on minimum intensity projection (minIP) reconstructions (Fig. 1.3). However, paired inspiratory and expiratory CT can be useful in the non-invasive confirmation of tracheobronchomalacia or excessive dynamic airways collapse (Fig. 1.4). Historically, thin-section volumetric acquisition in a single breath-hold was not possible on conventional or spiral single-slice CT systems. Volumetric thoracic CT would be performed with thicker collimation, and thus acquired more quickly, with focussed thinner sections only performed (targeted repeat acquisitions) on areas identified on the initial thick section acquisition as requiring more detailed visualisation, such as suspected nodules or masses. Alternatively, a ‘sampling’ technique to assess diffuse lung disease was used by acquiring non-contiguous thin sections at 10- to 20-mm intervals (and acquiring 3 to 4 slices per breath-hold), allowing the patient to inbreathe between acquisitions. However, with MDCT, volumetric acquisitions in a single breath-hold, even in tachypnoeic patients, are now possible. ‘Sharp’ reconstruction algorithms can be performed to obtain images that meet the spatial resolution requirements of HRCT. Such reconstructions can even be performed retrospectively, so long as the raw acquisition information is saved; in this way HRCT images can be obtained from any thoracic CT acquisition. This has meant that the non-contiguous acquisition method historically required for, and still often misconstrued as ‘essential’ for, HRCT is no longer mandatory.

Intravenous Contrast Medium Enhancement and Timing of Computed Tomography Acquisition

A

B Fig. 1.2  (A) Unenhanced and (B) intravenously enhanced volumetric 1-mm section high-resolution computed tomography images in a patient with biopsy-proven non-specific interstitial pneumonia, taken 1 week apart. Generally, increased ground-glass opacity is seen in both lungs, but it is difficult to determine whether this represents new parenchymal opacification, or whether it is purely the consequence of contrast enhancement.

IV enhancement is only required in thoracic CT for certain indications, most frequently for lung cancer staging, CT pulmonary angiography (CTPA), CT coronary angiography (CTCA) and aortic evaluation. IV enhancement in general is influenced by several factors. • patient factors, such as body size (as measured by body mass index and body surface area), and cardiac output; • contrast medium factors, including the iodine delivery rate (itself a product of iodine concentration, injected volume and the rate of injection); • timing of acquisition, depending on whether automated bolus triggering, test bolus or a set delay is used to initiate the CT data acquisition; and • changes in respiration, e.g. suboptimal pulmonary artery opacification in CTPA on deep inspiration, due to a variety of suggested mechanisms. With single-detector CT, a volume of 100 mL of 150 mg mL−1 of iodine injected at a rate of 2.5 mL s−1 after a 25-second delay was recommended for general thoracic work, while 120 to 140 mL of 240 to 300 mg mL−1 of iodine injected at a rate of 3 to 4 mL s–1, with either a fixed delay or the use of automated triggering mechanisms, was recommended for CTPA. However, it has been necessary to redesign contrast medium administration protocols and the timing of acquisition, caused by: (A) the reduced acquisition time brought about by MDCT; (B) newer technology, such as DECT; (C) the need to reduce the dose of contrast agent to minimise potential nephrotoxicity; and (D) the increasing feasibility of ‘triple-rule-out’ CT to provide simultaneous evaluation of the coronary arteries, pulmonary arteries and aorta, as well as other intrathoracic abnormalities in patients presenting with acute chest pain. In general, the faster acquisition times of MDCT require a higher iodine delivery rate to achieve faster peak arterial contrast medium enhancement. If contrast material volume is to be reduced, a higher rate

https://afkebooks.com

CHAPTER 1  Current Status of Thoracic Imaging

B

A

Fig. 1.3  A Patient With Confirmed Obstructive Small Airways Disease on Pulmonary Function Testing. On a 1-mm-thin highresolution image obtained at end-inspiration (A), subtle mosaic attenuation, with decreased centrilobular pulmonary arterial conspicuity in the lobules of decreased attenuation, is seen. This is made more conspicuous on 8-mm minimum-intensity projection (minIP) reconstructions (B), when viewed on appropriate narrow windows, and is confirmed on a subsequent end-expiratory acquisition (C), representing air-trapping in this context.

A

C

B Fig. 1.4  A 57-Year-Old Woman With Relapsing Polychondritis. Paired end-inspiratory (A) and end-expiratory (B) imaging demonstrates excessive collapse of the trachea, with the tracheal area decreasing by 52%.

https://t.me/afkebooks

7

8

SECTION A  The Chest and Cardiovascular System

can be achieved by a faster injection rate and a higher concentration of contrast medium. In addition, biphasic injection protocols are now preferred. Single-bolus (i.e. monophasic) contrast medium administration can cause thoracic MDCT acquisitions to suffer from streak and beamhardening artefacts, due to the dense contrast medium within the brachiocephalic veins during CT data acquisition. To overcome this, biphasic injection protocols are now the norm, using dual-headed power injectors to deliver both contrast medium and a saline chaser to dilute the contrast material density in the peripheral veins, thus overcoming the artefacts, as well as providing a more homogeneous enhancement profile. For 64-slice thoracic MDCT acquisitions, typical injection parameters are 60 to 120 mL of 320 to 400 mg mL−1 of iodine injected at 3.5 to 5 mL s−1, followed by 20 to 40 mL of normal saline injected at the same rate. For triple-rule-out studies, variations in the number of phases, timing, volume and composition of the chaser (e.g. using a mixture of 50 : 50 contrast material and saline) are employed. For example, a triphasic

protocol comprising an initial injection of undiluted contrast medium, followed by a second phase with a mixture of contrast media and saline and, finally, a third pure saline flush may be used.

Additional Postprocessing Techniques The introduction of MDCT systems has allowed truly isotropic imaging: that is, each voxel is of equal dimension in all three axes, thereby providing display in any arbitrarily chosen imaging plane (both orthogonal and non-orthogonal). The acquisition of volumetric high-resolution data has permitted new methods of two-dimensional (2D) and 3D reconstruction that can complement conventional axial image review, particularly in the display of airways and vascular structures. Furthermore, isotropic imaging has facilitated the development of computer-aided detection and diagnosis systems for the detection and evaluation of pulmonary nodules and pulmonary emboli, as well as automated quantification of disease processes, most notably emphysema. Table 1.2 summarises the

TABLE 1.2  Postprocessing Techniques and Examples of Clinical Application Technique

Technical Considerations

Examples of Use

Multiplanar and curved multiplanar reconstructions (MPR and CMPR)

2D techniques that provide alternate viewing perspectives, usually with conventional window settings. Images are obtained by a reordering of the voxels into 1-voxel-thick tomographic sections, excluding those voxels outside the imaging plane A ray is cast through the computed tomography (CT) data and only data above an assigned value are displayed, thus reducing all data in the line of the ray to a single plane. Sliding slabs of 5n ability to differentiate Similar to MIP, but only data below an assigned value are displayed and thus it is best suited for showing areas of low density Data reformatted around a threshold that defines the interface of tissues. SSD does not reveal any internal detail Histogram-based classification is applied to attenuation values in the entire CT data set. CT attenuation can be mapped to brightness, opacity and colour to display a structure of interest. Voxels partially filled with a density of interest are also included. The resultant images contain depth information whilst maintaining 3D spatial relationships Surface rendering and volume rendering are used to produce endoscopic simulations of the airway

Evaluation of the large airways and pulmonary emboli, particularly for interpretative difficulties on axial sections due to either partial volume averaging or the inability to differentiate periarterial from endoluminal abnormalities

Maximum intensity projection (MIP)

Minimum intensity projection (MinIP) Shaded surface display (SSD)

Volume rendering

Virtual bronchoscopy

Computer-aided detection

Quantitative lung parenchymal assessment Dual-energy CT pulmonary blood volume assessment Dual-energy CT virtual unenhanced imaging

Computerised complex pattern recognition employing a combination of image processing, segmentation and pattern classifiers to identify lesions of interest Quantification of parenchymal lung density using techniques such as density masks and histogram analysis to allow objective parenchymal assessment Generation of iodine maps that act as a surrogate of lung perfusion and therefore of pulmonary blood volume on dual-energy CT Removal of iodine from post-contrast dual-energy acquisitions to generate virtual unenhanced images

Mainly used in vascular imaging and in the evaluation of micronodular disease (more accurate identification of nodules vs vessels, and more precise characterisation of nodule distribution) May improve conspicuity of subtle density differences of lung parenchyma and therefore highlight regions of emphysema or air trapping (see Fig. 1.3) Evaluation of airway abnormalities

Used in angiographic examinations and also to evaluate large airway abnormalities

Virtual endoscopic or perspective volume rendering images are not widely applied as they seldom give information that cannot be obtained by MPR. However, virtual CT bronchoscopy can provide a view ‘through’ an obstructing lesion to visualise the airway distal to it, which may not be possible with conventional bronchoscopy Detection and volumetric assessment of pulmonary nodules and pulmonary emboli Emphysema quantification (see Fig. 1.5)

Assessment of perfusion defects in acute and chronic pulmonary thromboembolism Evaluation of apparent enhancement in pulmonary nodules without need for two separate pre- and post-contrast CT acquisitions

https://afkebooks.com

CHAPTER 1  Current Status of Thoracic Imaging

9

B

A

Fig. 1.5  Emphysema Quantification on Computed Tomography. Coronal thin-section computed tomography image in a patient with emphysema demonstrates visually upper lobe predominant emphysema, but densitometric evaluation using a threshold of –950 HU allows assessment of more subtle foci of emphysema in the lower lobes, as well as proportional lobar (or even segmental) quantification as well.

various postprocessing techniques used in evaluating thoracic disease. The increasing integration of quantitative assessment of both focal (e.g. pulmonary nodule sizing) and diffuse (e.g. emphysema and diffuse lung disease quantification) (Fig. 1.5) into clinical guidelines will mandate standardisation of thoracic CT acquisition protocols, reconstruction kernels, and slice thickness both within and between institutions. Also, with the expanding storage power and multiplanar reconstruction capability available on networked PACS solutions, storing the thinnest sections possible is increasingly regarded as good practice to ensure that such postprocessing can be performed should it be required (once raw CT data are deleted, such thin sections can no longer be reconstructed).

unenhanced image data set. This may have potential clinical application in pulmonary nodule and nodule characterisation: for example, where malignant and inflammatory nodules may show relatively different iodine content or enhancement. Alternatively, by creating an ‘iodine-only’ image data set, a map of pulmonary blood volume can be generated from a contrast-enhanced thoracic DECT acquisition. The latter may demonstrate perfusion defects in acute and chronic pulmonary thromboembolism analogous to, and potentially comparable with, that depicted by perfusion scintigraphy (Fig. 1.6). While promising, these potential clinical applications of DECT in thoracic imaging are still undergoing validation.

Dual-Energy Computed Tomography

Computed Tomography Dose Considerations

The concept of DECT was first explored in the 1970s, although initial efforts were hampered by limitations in the CT hardware and computational power available at the time. New CT technology has permitted the development of DECT systems over the past 7 years. The principles underpinning DECT are the same as for dual-energy subtraction radiography (discussed earlier). DECT allows materials to be differentiated by analysing their attenuation properties at different photon energies using the material decomposition theory. This theory is particularly applicable to high atomic number materials, such as iodine or calcium, owing to the photoelectric effect, as they exhibit different degrees of attenuation at different energies. Different methods for achieving DECT acquisitions are currently used. A dual-source CT system with two X-ray tubes mounted at 90 degrees to each other can provide a dual-energy acquisition by operating the tubes at different kilovoltages (typically 80 or 100 kVp and 140 kVp, respectively). DECT imaging with single-source CT currently uses either rapid switching between two kilovoltages or a dual-layer (‘sandwich’) detector, with different detector layers absorbing the different energy spectra. Regardless of the acquisition technology employed, these DECT systems are all able to generate material-specific image data sets, or subtraction images, from a single CT acquisition. In doing so, the need for pre- and post-contrast imaging can be obviated, thus reducing dose while simultaneously avoiding problems arising from misregistration between acquisitions before and after IV contrast agents. For example, material differentiation of iodine makes it possible to create a virtual

It is important to appreciate some fundamental principles of CT acquisition, and factors affecting radiation dose, before considering strategies for dose reduction. In a CT x-ray tube, a small area on the anode plate emits x-rays that penetrate the patient and are registered by the detector. A collimator between the x-ray tube and the patient, the pre-patient collimator, is used to shape the beam and establish the dose profile. In general, the collimated dose profile is a trapezoid in the longitudinal direction, resulting in umbral and penumbral regions within the area of coverage. In the umbral region, x-rays emitted from the entire area of the focal spot fall on the detector; however, in the penumbral regions at the edge of the beam, only a part of the focal spot illuminates the regions within the area of coverage. Despite its undoubted clinical benefit, MDCT carries an increased radiation exposure burden compared with conventional single-slice CT. The increase in radiation arises primarily as a result of wasted radiation dose, due to decreased geometric efficiency of MDCT. Geometric efficiency indicates the proportion of the x-ray beam used in image formation. Lower efficiency thus means that increased doses will be required to maintain similar image quality. With MDCT, only the plateau (umbral) region of the dose profile is used to ensure an equal signal level for all detector elements—the penumbral region is discarded, either by a post-patient collimator or by the intrinsic selfcollimation of the MDCT, and represents ‘wasted’ dose (Fig. 1.7) The relative contribution of the penumbral region decreases with increasing section width (due to a decrease in the incident beam width) and with

https://t.me/afkebooks

10

SECTION A  The Chest and Cardiovascular System

TABLE 1.3  Dose Reduction Strategies in

Thoracic Computed Tomography

Tube current modulation using: Automatic exposure control (AEC) Weight-based modulation Size-based modulation Gated modulation, e.g. prospective ECG gating in CT coronary angiography Tube current reduction in low-dose examinations Tube potential reduction Beam-shaping filters (e.g. bowtie filters) Restricting length of coverage to area of interest Higher pitch (by increasing table speed); wider collimation where possible Faster gantry rotation time New detectors with higher radiation sensitivity to decrease exposure time Patient shielding Iterative reconstruction techniques (especially model-based iterative reconstruction)

A

CT, Computed tomography; ECG, electrocardiographic.

B Fig. 1.6  Dual-Energy Computed Tomography Pulmonary Angiogram in a 32-Year-Old Patient With Acute on Chronic Left Pulmonary Arterial Occlusion. The conventional computed tomography image can be fused with the derived pulmonary iodine map (A) to demonstrate both the left pulmonary arterial occlusion and the almost total absence of perfusion to the left lung, correlating very well with perfusion scintigraphy (B).

U

P

U

P

P

P

U

U

P

P

Fig. 1.7  Geometry and Dose Profile for Spiral, 4-, 16- and 64-Slice Computed Tomography. In spiral computed tomography (CT), the whole dose within the umbral region (U) contributes to image reconstruction with no wastage. In 4-slice CT, wastage occurs within the penumbral regions (P). The relative contribution of the penumbral region decreases with an increasing number of simultaneously acquired sections. The effect of this wastage is minimised in 64-slice CT.

an increasing number of simultaneously acquired images, in 4- to 16-section MDCT systems. However, with 64-channel MDCT, the incident beam width remains constant over both narrow and wide collimation acquisitions; therefore, geometric efficiency of 64-MDCT is high, with little consequent penalty by way of increased dose. Another factor that decreases geometric efficiency in MDCT arises from gaps between detector elements in the multidetector array. Photons incident on these regions do not contribute to image signal and are another form of wasted dose. In general, the number of gaps increases with increasing numbers of sections, thus decreasing the efficiency. Dose reduction strategies are summarised in Table 1.3. The CT parameters that directly affect radiation dose include gantry geometry, rotation time, tube current and voltage, acquisition modes, z-axis coverage, pitch, section collimation and section overlap or interval. Factors that indirectly affect dose include reconstruction methods and image filters. Reduction in tube current is the most practical means of decreasing CT radiation dose, provided this does not compromise image quality due to increased noise. A 50% reduction in tube current can halve effective radiation dose. Authors of several studies using MDCT have suggested that it is possible to reduce tube current markedly (to tube current–time products of between 40 and 70 milliampere seconds (mAs)) in chest examinations without significantly affecting image quality. In the paediatric population, some institutions favour the use of a tube current manually tailored to body weight for imaging the thorax, an approach that significantly reduces radiation dose. Tube potential (peak voltage) determines the incident x-ray mean energy, and variation in tube potential causes a substantial change in CT radiation dose. The effect of tube voltage on image quality is complex, because it affects both image noise and tissue contrast. Thus, the image quality ramifications of a decrease in tube voltage to reduce radiation exposure must be carefully examined before being implemented. For chest examinations, 120 kVp is commonly used. In thin patients (15,000 leucocytes per mL, a fibrinopurulent stage with prominent adhesion formation and finally an organising stage with development of a thick pleural peel. An empyema becomes more difficult to drain as it evolves and forms thick adhesions and fibrous bands, which can be easily identified on ultrasound. On chest radiography, an empyema may be suggested by a loculated pleural effusion, whilst CT shows the ‘split pleura sign’ and thickened pleura, often alongside stranding and widening of the extrapleural space (Fig. 3.16). Differentiating empyema from lung abscess may be difficult—a thicker and more irregular wall and destruction rather than compression of the underlying lung suggest abscess rather than empyema. Empyema less commonly is caused by transdiaphragmatic extension of a liver abscess or by bronchopleural fistula the latter of which may demonstrate a gas–liquid level. Chronic tuberculous empyema may demonstrate a large, loculated pleural effusion with pleural calcification and enlargement of the overlying ribs.

Chylothorax

Fig. 3.15  Computed Tomography of Malignant Pleural Disease. In this right pleural effusion, computed tomography identifies the extensive and irregular pleural thickening characteristic of a malignant process (pleural metastases). Note also the primary tumour in the right breast.

A

Chylothorax is caused by disruption or obstruction of the thoracic duct or collaterals, with an equal incidence of nontraumatic and traumatic causes. The leading nontraumatic cause is malignancy, with a particular association with advanced lymphoma, whilst the leading traumatic cause is from accidental damage from surgical procedures. Traumatic rightsided chylothoraces suggest injury to the lower third of the thoracic duct, whilst left-sided chylothoraces suggest injury to the upper two-thirds of the thoracic duct. Rarely, chylothorax results from the abdomen in the presence of chylous ascites crossing the diaphragm. Chylous effusions are commonly milky because they contain triglycerides in the form of chylomicrons. Chylous and nonchylous pleural effusions are indistinguishable on the chest radiograph. In addition, despite its high fat

B

C

Fig. 3.16  Empyema. (A) Chest x-ray shows an encapsulated pleural effusion on the right and a free pleural effusion on the left. (B and C) An enhanced computed tomography confirms this bilateral fluid collection. However, the pleura on the right is thickened but smooth and enhancing while subpleural fat is infiltrated and widened, which is the result of oedema. The empyema followed pneumonia, which can be seen in the middle lobe (C). Compare with noncomplicated left pleural effusion.

https://t.me/afkebooks

50

SECTION A  The Chest and Cardiovascular System

content, the increased protein level of a chylothorax gives it an attenuation on CT similar to that of other pleural effusions. The exception where a presumptive diagnosis of chylothorax can be made is when associated with lymphangioleiomyomatosis. Chylous effusion may be of high signal intensity on T1 weighted images similar to subcutaneous fat and therefore MRI can be used in diagnostic difficulty. T2 weighted sequences may also show morphological changes of the thoracic duct.

Haemothorax On the plain chest radiograph an acute haemothorax is indistinguishable from other pleural fluid collections. Once the blood clots, there is a tendency for loculation and occasionally a fibrin body will form. Pleural thickening and calcification are recognised sequelae. On CT a haemothorax may show areas of hyperdensity, which has a varying appearance dependent on age. Fresh blood has an attenuation of above 35 HU, clotted blood has an attenuation of 70 HU and a haematocrit effect with layering of blood may be seen in a subacute stage. MRI can be used to identify blood and estimate the age of haemorrhage, which in the subacute or chronic stage will usually appear on MRI as a high signal on T1 and T2 weighted images, possibly with a low signal rim caused by haemosiderin. The most common cause of haemothorax is trauma, but it is seen in a number of other conditions, including pulmonary embolic disease, ruptured aortic aneurysm, pneumothorax, extramedullary haematopoiesis and coagulopathies.

A

Hepatic Hydrothorax Pleural effusion is a recognised complication of hepatic cirrhosis. The principal mechanism of its production is the transdiaphragmatic passage of ascites, although other factors such as hypoalbuminaemia may contribute in a small number of cases. In other cases a large effusion may accumulate and is termed hepatic hydrothorax. Hepatic hydrothorax is a transudate and is defined as a pleural effusion, usually greater than 500 mL, in patients with cirrhosis but without other cause; 85% occur in the right hemithorax, and chest CT, echocardiogram and thoracocentesis should be performed to confirm the diagnosis and exclude infection or alternate causes. Severely symptomatic patients should undergo therapeutic thoracocentesis, and those refractory to medical management may undergo transjugular intrahepatic portosystemic shunt (TIPS) placement. Importantly, long-term chest drains should be avoided as such drainage may result in massive protein and electrolyte depletion, infection, renal failure and bleeding.

Bronchopleural Fistula

B

Bronchopleural fistula differs from a pneumothorax in that the communication with the pleural space is via airways rather than distal air spaces. It occurs in two main settings: following partial or complete lung resection and in association with necrotising infections.

Fig. 3.17  Left Primary Spontaneous Pneumothorax. Chest radiograph (A) at deep inspiration and (B) at deep expiration. The left lung has partially collapsed and an area of extreme low density without vascular markings becomes visible. The pneumothorax is accentuated on the chest radiograph at suspended deep expiration (B).

PNEUMOTHORAX

Chest Radiography

Air in the pleural space is a pneumothorax. When air and liquid are present the nomenclature depends on their relative volumes and the type of liquid. Small amounts of liquid are disregarded, and the condition is still called a pneumothorax; otherwise, the prefix hydro-, haemo-, pyo- or chylo- is added, depending on the nature of the liquid.

Imaging Pneumothorax The diagnosis of pneumothorax is mostly made with the chest radiograph, although other techniques (e.g. CT) may be used for smaller pneumothoraces and detecting complications and predisposing conditions, as well as helping in management (Fig. 3.17).

Typical Signs.  These are seen on erect radiographs in which the pleural air rises to the lung apex. Under these conditions the visceral pleural line at the apex becomes separated from the chest wall by a transradiant zone devoid of vessels. Although this sounds a straightforward sign to assess, difficulties of interpretation can arise with avascular lung apices, as in bullous disease and when linear shadows are created by clothing or dressing artefacts, tubes and skin folds. Skin folds cause problems particularly in neonates and in old people radiographed slumped against a cassette in the Anteroposterior (AP) projection (Fig. 3.18). Features that help identify artefacts and skin folds include extension of the ‘pneumothorax’ line beyond the margin of the chest

https://afkebooks.com

CHAPTER 3  The Chest Wall, Pleura, Diaphragm and Intervention

Fig. 3.18  Skin Folds Mimicking a Right Pneumothorax (arrows). The laterally located blood vessels, the wide margin of the lines, and the orientation of the lines that is inconsistent with the edge of a slightly collapsed lung help to differentiate them from a real pneumothorax.

cavity, laterally located vessels and an orientation of a line that is inconsistent with the edge of a slightly collapsed lung. In addition, the margin of skin folds tends to be much wider than the normally thin visceral pleural line. In indeterminate circumstances a repeat chest radiograph, an expiratory radiograph (see Fig. 3.17B) or one taken with the patient decubitus may clarify the situation. Should doubt still remain, then CT is particularly helpful in distinguishing between bullae and a pneumothorax. Ultrasound has also been reported to be useful in this instance.

Atypical Signs These arise when the patient is supine or when the pleural space is partly obliterated. In the supine position, pleural air rises and collects anteriorly, particularly medially and basally, and may not extend far enough posteriorly to separate lung from the chest wall at the apex or laterally. Signs that suggest a pneumothorax under these conditions are (Fig. 3.19): • Ipsilateral transradiancy, either generalised or hypochondrial. • A deep, finger-like costophrenic sulcus laterally. • A visible anterior costophrenic recess seen as an oblique line or interface in the hypochondrium; when the recess is manifest as an interface it mimics the adjacent diaphragm (‘double diaphragm sign’). • A transradiant band parallel to the diaphragm and/or mediastinum with undue clarity of the mediastinal border. • Visualisation of the undersurface of the heart, and of the cardiac fat pads as rounded opacities suggesting masses. • Diaphragm depression. In a patient who cannot stand, the presence of a pneumothorax can be confirmed with a lateral decubitus view or a supine decubitus projection with the cassette placed dorsolaterally at 45 degrees and the x-ray tube angled perpendicular to the cassette. When the pleural space is partly obliterated a pneumothorax may be loculated and must be differentiated from other localised transradiancies. These include cysts, bullae, pneumatoceles, pneumomediastinum and local emphysema. These cannot always be differentiated by plain radiographs but can be by CT.

51

Fig. 3.19  Supine Pneumothorax. Portable chest radiograph after development of a pneumothorax in a patient with a bilateral pneumonia. There is an increase of transradiancy at the left lung base and the costophrenic sulcus laterally is more pronounced (‘deep sulcus sign’).

Ultrasound Normal pleura is seen as an echogenic line, the ‘pleural stripe’, with the majority of acoustic energy of the ultrasound (US) beam reflected at the air–pleural interface. Beyond the pleural stripe, distal reverberation artefacts are seen as vertical echogenic bands (‘comet tails’) (Fig. 3.20). The pleural stripe tends to shimmer at the pleural surface and is described as ‘lung sliding’. These features disappear in a pneumothorax. The position of the border between sliding lung and pneumothorax is known as the ‘lung point’ sign and can help determine the size of the pneumothorax, although this relies on at least part of the lung being in contact with the chest wall and therefore is not present in large pneumothoraces. ‘A lines’ are horizontal reverberation artefacts that are equally spaced lines caused by reflection from the pleura in the presence of pneumothorax. Ultrasound is useful in detecting pneumothorax immediately after central line insertion, and post-lung biopsy, and potentially for monitoring the resolution of pneumothorax during drainage and residual pneumothorax at follow-up. However, there is a risk of falsepositives, particularly in the inexperienced practitioner, as patients with hyperinflation, bullous lung disease, air trapping or previous pleurodesis may not demonstrate lung sliding and comet tails.

Computed Tomography CT is the most sensitive investigation for the detection of pneumothoraces, particularly if they are small and when the patient must remain supine. CT is useful in patients with widespread subcutaneous emphysema or consolidation, when a pneumothorax on chest radiography may not be readily apparent. It may also detect ancillary relevant findings (e.g. lung contusions in the context of trauma). CT can help guide chest drain insertion, particularly in the context of severe emphysema when bullous disease is to be avoided. Furthermore, a significant proportion of pneumothoraces post-lung biopsy are not detectable on chest radiography (25%–40%).

https://t.me/afkebooks

52

SECTION A  The Chest and Cardiovascular System

TABLE 3.2  Causes of Adult Pneumothorax Spontaneous, Primary Spontaneous, Secondary Airflow obstruction

Pulmonary infection

Pulmonary infarction Neoplasm Diffuse lung disease

Hereditable disorders of fibrous connective tissue Endometriosis (catamenial pneumothorax)

Asthma Chronic obstructive pulmonary disease (COPD) Cystic fibrosis Cavitating pneumonia Tuberculosis Fungal disease AIDS Pneumatocele Metastatic sarcoma Histiocytosis X Lymphangioleiomyomatosis Fibrosing alveolitis Other diffuse fibroses Marfan syndrome

Traumatic, Noniatrogenic Ruptured oesophagus/trachea Closed chest trauma (± rib fracture) Penetrating chest trauma

Fig. 3.20  ‘Comet Tails’ of Pleural Ultrasound. Vertical echogenic bands caused by reverberation artefacts (horizontal arrows). The echogenic ‘pleural stripe’ represents normal pleura (vertical arrows).

Primary Spontaneous Pneumothorax Iatrogenic causes apart, the most common type of pneumothorax in the adult is the so-called primary spontaneous pneumothorax (PSP). A pneumothorax occurring without an obvious precipitating event is spontaneous, and if the patient has essentially normal lungs it is termed primary. PSP occurs predominantly in young adults (65% are between 20 and 40 years of age), and it is five times more common in men than women. Untreated, at least one-third of patients will have a recurrence, most commonly within a few years and on the ipsilateral side. PSP is nearly always caused by the rupture of an apical pleural bleb. Although not detectable on interval chest radiographs, one taken at the time of the pneumothorax will show one or more blebs projecting from the apical lung margin in 20% of patients; such abnormal apical airspaces are much more commonly shown by interval CT.

Traumatic, Iatrogenic Thoracotomy/thoracocentesis Percutaneous biopsy Tracheostomy Central venous catheterisation

Tension Pneumothorax

Secondary Spontaneous Pneumothorax

This life-threatening complication is present when intrapleural pressure becomes positive relative to atmospheric pressure for a significant part of the respiratory cycle. Tension has an adverse effect on gas exchange and cardiovascular performance, causing a rapid deterioration in the patient’s clinical condition. The diagnosis is usually made clinically and treatment instituted without a radiograph. Should a chest radiograph be taken, it will show contralateral mediastinal shift and ipsilateral diaphragm depression. Mild degrees of contralateral mediastinal shift are not unusual with a nontension pneumothorax because of the negative pressure in the normal pleural space. However, moderate or gross mediastinal shift should be taken as indicating tension, particularly if the ipsilateral hemidiaphragm is depressed. This latter sign is the more reliable and is almost invariably present with significant tension.

A large number of conditions predispose to pneumothorax (Table 3.2). In a number of these disorders, pneumothorax occurs frequently.

Pyopneumothorax

Complications

This unusual complication is seen most commonly following necrotising pneumonia or oesophageal perforation.

Haemopneumothorax This is a common complication of traumatic pneumothorax. Small amounts of serous or bloody fluid may also occur with a spontaneous pneumothorax, but only 2% of individuals develop a clinically significant haemothorax in these circumstances. Blood may clot in the pleural space, producing a mass which can mimic a pleural tumour.

Adhesions These generate straight band shadows extending from the lung margin to the chest wall. They limit collapse but at the same time may account for continued air leakage from the lung surface, and if they tear they may bleed. They can be identified with CT.

https://afkebooks.com

CHAPTER 3  The Chest Wall, Pleura, Diaphragm and Intervention

53

Reexpansion Oedema This unusual complication is sometimes seen following the rapid therapeutic reexpansion of a lung that has been markedly collapsed for several days or more. Oedema comes on within hours of drainage, may progress for a day or two and clears within a week. It usually causes only mild morbidity.

PLEURAL THICKENING AND TUMOURS Pleural thickening may be benign or malignant and, if benign, usually represents the organised end stage of various active processes such as infective and noninfective inflammation (including asbestos exposure and pneumothorax) and haemothorax. Malignant processes include mesothelioma and metastatic deposits. Significant advances have been made in the use of different diagnostic imaging techniques to evaluate pleural abnormalities:

A

Imaging Pleural Thickening and Tumours Chest Radiography

The chest radiograph remains the initial investigation of pleural thickening and provides a starting assessment of whether localised or diffuse, smooth or nodular, the presence of any calcification and whether associated with volume loss or pleural effusion. Pleural thickening appears as fixed opacification of water/soft-tissue density, which if viewed en profile, appears more or less parallel to the chest wall and with a sharp lung interface. En face, it causes ill-defined, veil-like shadowing. Benign pleural thickening tends to be smooth and located in specific areas on the chest x-ray (CXR), such as the apex of the lungs, seen as smooth apical pleural thickening secondary to ageing or prior tuberculosis or fibrotic pulmonary disease, or at the costophrenic angles secondary to prior empyema or asbestos exposure causing diffuse visceral pleural thickening. Malignant pleural thickening is often lobulated and may involve the whole hemithorax and extend into and along fissures and be associated with rib destruction, lymphadenopathy, volume loss and a pleural effusion.

B Fig. 3.21  Malignant Mesothelioma. (A) Axial and (B) coronal computed tomography. Diffuse lobulated and nodular thickening of the pleura with tumour extension into the lobar fissure (arrows). Note the metastatic enlargement of some hilar and mediastinal lymph nodes.

Ultrasound Normal pleura is seen as a bright echogenic line (the ‘pleural stripe’) that represents the air–visceral pleural interface. Diffuse or smooth thickening is often challenging to visualise on ultrasound, having a variable echogenicity and only appreciated when >1 cm in depth. The presence of pleural fluid aids visualisation, although in some cases small loculated effusions can themselves be mistaken for pleural thickening. In such cases, colour Doppler may demonstrate fluid movement in effusions with cardiac pulsation. Where visualised, parietal, visceral or diaphragmatic nodularity is strongly suggestive of pleural malignancy.

Computed Tomography CT is very sensitive in the detection of pleural thickening, which is most easily assessed on the inside of the ribs, where there should normally be no soft-tissue opacity. CT is the workhorse of the diagnostic evaluation of pleural thickening and tumours, being able to define its exact location, the degree and extent of thickening, the presence of calcification and enhancement, and also with the visualisation of any ancillary finding such as pleuroparenchymal fibrous bands (‘crow’s feet’) or rounded atelectasis with asbestos exposure, or lymphadenopathy or rib destruction and extrapleural extension in cases of malignancy. Differentiating between generalised, postinflammatory pleural thickening and diffuse pleural malignancy caused by mesothelioma, metastatic disease (particularly adenocarcinoma), lymphoma and leukaemia may be challenging. The

most useful signs on CT that indicate malignant as opposed to benign pleural thickening are circumferential thickening, nodularity, parietal thickening of more than 1 cm and involvement of the mediastinal pleura (Figs 3.21 and 3.22).

Magnetic Resonance Imaging MRI is particularly useful to help problem-solve in cases where there is uncertainty between benign and malignant disease and tumour stage. It may also have a role in treatment evaluation. Signal hypointensity with long repetition time (TR) sequences has been shown to be a reliable predictive sign of benign pleural disease. On morphological sequences, the superior soft-tissue contrast helps demonstrate infiltration of the chest wall or other adjacent tissues in cases of malignant disease, differentiating between T2 and T3 tumour stage. Particular benefit over CT has been seen in assessing diaphragmatic invasion and assessment of resectability of solitary tumour foci. The diagnosis of thoracic endometriosis may also be strongly suggested by detecting T1 and T2 hyperintensity and susceptibility artefact from degradation blood products on implants at the diaphragm or in the pleural cavity. Meanwhile, functional sequences may aid differentiation between benign and malignant pleural disease—the presence of multiple hyperintense pleural spots on high b-value diffusion-weighted MRI (DW-MRI; ‘pleural pointillism sign’) has been shown to potentially be a useful marker of malignant disease and may help to guide biopsy. DW-MRI may also

https://t.me/afkebooks

54

SECTION A  The Chest and Cardiovascular System as determining residual disease or tumour relapse in the apex post radiotherapy.

Pleural Plaques

Fig. 3.22  Malignant Pleural Thickening Caused by Metastatic Disease. Malignant pleural thickening was caused by pleural metastases. Note the compression on the right hemidiaphragm and the extension of the tumour into the liver (arrows).

Fibrous pleural thickening can be induced by asbestos exposure (Fig. 3.23). When focal, they are termed pleural plaques and can undergo hyaline transformation, calcify or ossify. They almost always involve the parietal pleura alone and are most commonly found along the lower thorax and on the diaphragmatic pleura. In extensive disease the anterior and ventral part of the thorax may also be involved. Pleural plaques need to be large before they become visible on a chest x-ray, where they may be unilateral (more frequently on the left) and asymmetrical if bilateral (see Fig. 3.23E). When calcified, pleural plaques are visualised as opaque lines on tangential views parallel to the chest wall, diaphragm or cardiac border, and when seen en face, they produce irregular linear or stippled uneven calcifications (‘holly leaf ’ calcification). On CT they are visualised much earlier (see Fig. 3.23A–D) and appear as circumscribed areas of pleural thickening separated from the underlying rib and extrapleural soft tissues by a thin layer of fat. Because of their higher density they can easily be differentiated from a circumscribed increase in extrapleural fat, as may be seen in obese patients. Plaques may have ‘rolled’ edges in which they are thicker at the edges than centrally, and when associated with interstitial lines are termed ‘hairy plaques’. On MRI, pleural plaques are low signal on T1 and T2 weighted sequences with areas of signal void where there is calcification.

Diffuse Pleural Disease suggest the tumour subtype in mesothelioma because the sarcomatoid subtype of malignant pleural disease has been shown to have significantly lower ADC values than the epithelioid subtype, although biopsy is still required because of the considerable overlap of the biphasic subtype. Dynamic contrast-enhanced weighted MRI can provide extra information about tumour perfusion, vascularity and vascular permeability, which can be correlated with tumour response to chemotherapy and therefore prognosis.

Positron-Emission Tomography/Computed Tomography FDG PET/CT currently has a significant role in both the diagnosis and management of malignant pleural disease, particularly in patients with suspected or proven mesothelioma, with increased FDG activity on PET/CT helping to differentiate benign from malignant pleural disease. However, false-positive FDG activity can occur due to inflammation of the pleura in infection, inflammatory pleurisy or talc pleurodesis. Talc pleurodesis presents a particular challenge, resulting in intense FDG uptake which can persist for several years, with FDG uptake mostly correlating with areas of high density on CT. False negatives may occur due to small tumour size or depth, or metabolically inactive tumours, with epithelioid subtype mesothelioma being less FDG-avid than sarcomatoid mesothelioma. The added value of PET/CT in pleural malignancy tumour staging is limited for the local staging due to its poor spatial resolution; however, it has considerable value in determining nodal status and the detection of unsuspected extrathoracic metastases, which may alter treatment planning. In terms of prognosis and determining response to chemotherapy, a reduction in metabolic activity after chemotherapy may correlate with increased time to progression and longer survival. PET/CT may also be of value in determining the optimal target site for image-guided pleural biopsy and in detecting tract seeding following percutaneous biopsy. PET/CT can provide extra diagnostic information when differentiating between benign apical pleural caps from a Pancoast tumour, as well

The radiographic definition of diffuse pleural thickening is somewhat arbitrary. It has been suggested that to qualify it should appear as a smooth uninterrupted pleural density that extends over at least one-quarter of the chest wall. On CT, it has been defined as a continuous sheet of pleural thickening that extends more than 8 cm in the craniocaudal direction and 5 cm laterally and with a thickness of more than 3 mm. It may be caused by asbestos exposure or other causes, as described as follows. Asbestos-related plaques may cause diffuse pleural thickening. Involvement only of the parietal surface is not associated with changes in the adjacent lung, while diffuse visceral pleural thickening is commonly associated with pulmonary changes. Distinction between the two, although often challenging, is important medicolegally as it is only visceral pleural disease that is thought to result in respiratory compromise. CT may help differentiate extensive plaques from diffuse visceral pleural thickening: plaques tend to have steep shoulders rather than the sloping shoulders of diffuse visceral pleural thickening. Diffuse pleural thickening is more rare than pleural plaques; it is most commonly bilateral and often associated with a thin rim of fluid on CT; it is rarely densely calcified but may contain flecks of calcification. Parenchymal changes associated with visceral pleural involvement are rounded atelectasis and pleuroparenchymal fibrous bands (crow’s feet). Common causes of non–asbestos-related diffuse pleural thickening are empyema, tuberculosis and haemorrhagic effusion. The CXR changes are generally unilateral, affecting the lateral and posterior costophrenic recesses and appear as smooth, often angular thickening. A veil-like increase in radio-opacity may occur en face, extending across fissures, whilst tangentially a thickened pleural rim may occur. CT is helpful for clues as to its aetiology—extensive unilateral pleural calcification, volume loss, thickening of the extrapleural fat and evidence of prior pulmonary parenchymal disease may suggest it is due to a previous empyema, particularly tuberculosis (TB), whilst rib deformity with normal lung parenchyma may suggestive previous traumatic haemothorax. In post-talc pleurodesis there is a ‘sandwich’ of parietal pleural thickening, high attenuation talc and visceral pleural thickening (Fig. 3.24).

https://afkebooks.com

CHAPTER 3  The Chest Wall, Pleura, Diaphragm and Intervention

A

55

B

C

D

Fig. 3.23  Pleural Plaques Caused by Asbestos Exposure. (A–E) Pleural plaques are most commonly found along the lower thorax, on the diaphragmatic pleura and, when involvement is extensive, also along the lateral and anterior thorax (arrows). They can partially or completely calcify or ossify. In this situation, and when large, they can be seen on a chest radiograph (E).

E

Pleural Tumours These may be localised or diffuse; benign or malignant, in which they may be primary or secondary in origin. Localised tumours are relatively uncommon, with diffuse malignant disease accounting for the vast majority of pleural tumours.

Pleural Fibroma Also known as solitary fibrous tumour of the pleura (SFTP) (Fig. 3.25), these lesions most commonly present in middle age, approximately half the patients being asymptomatic. Hypertrophic osteoarthropathy (HPOA) is a well-recognised complication (10%–30% of patients), and rarely the tumour is associated with hypoglycaemia. Microscopically, two-thirds

are benign and one-third are malignant. The CXR findings are of a pleurally based, well-demarcated, rounded and often slightly lobulated mass (2–20 cm diameter) which may, because of pedunculation, show marked positional variation with changes in posture and respiration. Pleural fibromas usually make an obtuse angle with the chest wall and may reach enormous sizes (see Fig. 3.25A and B). They are usually seen in the lower third of the chest and may be in a fissure (30%), along the mediastinal pleura (18%), the thoracic pleura (46%) or adjacent to the hemidiaphragm (6%). CT findings are similar to those observed on plain radiography: a mass, which is often heterogeneous because of necrosis and haemorrhage, frequently enhancing after contrast medium administration with a characteristic swirling pattern, with calcification

https://t.me/afkebooks

56

SECTION A  The Chest and Cardiovascular System

A

B

C

Fig. 3.24  Pleural Calcification. (A–C) On the chest radiograph (A) an extensive sheet-like calcification of the left pleura and a smaller localised calcification of the right pleura is seen together with focal areas of calcification of the diaphragmatic pleura. (B and C) Computed tomography demonstrates the extent and thickness of the pleural calcification.

A B

C

Fig. 3.25  Benign Pleural Fibroma. (A) Frontal chest radiograph, (B) coronal computed tomography and (C) axial magnetic resonance imaging show a well-demarcated and homogeneous mass abutting the diaphragm. Note the obtuse angle between the mass and the chest wall, suggesting the extrapulmonary origin of the mass.

https://afkebooks.com

CHAPTER 3  The Chest Wall, Pleura, Diaphragm and Intervention most commonly seen when they are large. The adjacent lung parenchyma may be displaced with atelectasis, with bowing of the bronchi and pulmonary vessels around the mass. They also may collaterise adjacent blood supply, which is helpful diagnostically and useful information prior to surgical resection. Malignant types are usually larger than 10 cm and may invade the chest wall or have associated pleural or pulmonary metastases, or an effusion. Typically, these tumours show low signal on T1 weighted images and heterogeneous high signal on T2 weighted images, occasionally with a peripheral rim of low intensity seen on T1 weighted images. PET/CT may be of value in identifying sarcomatous change within a pleural fibroma, by demonstrating significantly increased FDG avidity.

Pleural Lipoma Lipomas are asymptomatic benign tumours usually discovered incidentally on chest radiographs as sharply defined pleural masses. Diagnosis is straightforward with CT delineating its pleural origin and fatty composition, which is usually homogeneous. If it appears heterogeneous with areas of soft-tissue attenuation, a liposarcoma or an area of tumour infarction should be suspected. Pleural lipomas have high signal intensity on T1 weighted images. On T2 weighted images the signal is moderately increased. PET/CT is rarely used to image them but may be used to confirm suspected sarcomatous change, as they usually demonstrate FDG avidity less than that of background activity when benign.

Malignant Mesothelioma (Primary Pleural Malignancy) Diffuse malignant mesothelioma is an uncommon primary neoplasm, and its development is strongly related to asbestos exposure. It is usually indistinguishable from metastatic disease on imaging, although the presence of calcified or noncalcified pleural plaques, alongside interstitial disease or asbestosis, may provide evidence of prior asbestos exposure. It most commonly presents on a chest radiograph as irregular and nodular pleural thickening with an associated pleural effusion. Tumour extension into the interlobular fissures, accompanying the pleural effusion, and invasion into the chest wall are better appreciated with CT (see Fig. 3.22). On CT, malignant mesothelioma presents as a nodular soft-tissue mass sometimes with hypodense areas corresponding with necrosis. Metastatic enlargement of hilar and mediastinal nodes is seen in up to 50% of patients. The epithelioid subtype has a better prognosis than sarcomatoid. Malignant mesothelioma has minimally increased signal on T1 and moderately increased signal on T2. MRI may be superior to CT in determining the extent of local disease because it allows better evaluation of the relationship of the tumour to the structures of the chest wall, mediastinum and diaphragm. However, in most cases, CT and MRI provide similar morphological information. Ultrasound may be a supplementary method for biopsy and surgery planning. Advances in the use of functional MRI and PET/CT to differentiate benign from malignant disease, to provide clues as to the subtype, stage, assessment of treatment response and prognosis and to plan biopsy have already been discussed (Fig. 3.26).

Pleural Metastases (Secondary Pleural Malignancy) Pleural metastases are the most common pleural neoplasm. They are usually adenocarcinoma with common sites of origin including the ovary, stomach, breast and lung. Pleural metastatic disease can present as a solitary mass but is more often seen as multiple pleural lesions or diffuse pleural thickening (see Fig. 3.22). Pleural metastases are very often accompanied by a pleural effusion, which may be the only finding on a chest radiograph. CT, MRI and ultrasound are more sensitive in demonstrating pleural metastases as the cause of a pleural effusion. Bronchogenic carcinoma may directly invade the adjacent pleural surface and chest wall or may have metastasised to it. The diagnosis

57

of pleural and chest wall involvement of bronchogenic carcinoma on a chest radiograph can be very difficult. The only highly specific indicator is rib destruction. The diagnosis may also be difficult on CT and MRI. Features such as a large contact (>3 cm) between the mass and the pleura, an obtuse angle between the tumour and the chest wall and associated pleural thickening are usually considered signs of chest wall invasion but may also be seen in benign disease. The diagnostic accuracy of CT can be increased by performing and carefully reviewing 2D and 3D reconstructions of thin sections. In cases where tumour invasion is obvious, 2D sagittal or coronal reconstructions can be helpful in ascertaining the extent of the mass. MRI was previously thought to have a slight advantage over CT in the evaluation of chest wall and pleura invasion due to its increased soft-tissue contrast resolution. However, studies have shown that spiral CT and MRI have comparable sensitivity but that spiral CT has a higher specificity. CT is superior in the detection of pleural calcifications and bone involvement (see Fig. 3.4).

Tumour-Like Conditions of the Pleura There are a few rare causes of diffuse pleural thickening whose aetiology may be determined because of accompanying nonpleural abnormalities. Erdheim–Chester disease should be considered in patients with pleural or pericardial thickening and retroperitoneal/renal abnormalities along with musculoskeletal features such as osteosclerosis; diffuse pulmonary lymphangiomatosis has accompanying pulmonary findings predominantly affecting the upper lobes, including smooth symmetrical interlobular septal and peribronchovascular interstitial thickening, pericardial and pleural effusion and mediastinal fat infiltration. Thoracic splenosis, thoracic endometriosis and extrapleural hamartomas are rare tumour-like focal lesions of the pleura, which again may be suggested as a diagnosis because of an appropriate clinical history and possible extrapleural abnormalities.

SUMMARY BOX: Pleura • Pleural effusion—accumulation of fluid in the pleural space. Different techniques, including chest radiograph, ultrasound, CT, MRI and PET, may add different value in the assessment of pleural fluid: for example, allowing differentiation between empyema, haemothorax, hepatic hydrothorax and bronchopleural fistula. • Pneumothorax—accumulation of air in the pleural space. CXR is the most commonly used investigation, but CT adds further sensitivity, particularly for supine patients. Ultrasound may also have utility, particularly as a bedside investigation. Pneumothorax may be primary (occurring in normal lungs without precipitant) or secondary (abnormal lungs) and may be complicated by blood (haemopneumothorax), tension, pus/infection, adhesions and reexpansion oedema following drainage. • Pleural thickening—multiple imaging techniques are often used to assess, for pleural thickening, with FDG-PET/CT currently of particular use in attempting to differentiate benign from malignant disease, as well as targeting biopsy. Pleural thickening may be localised or diffuse and may relate to asbestos exposure, previous empyema, tuberculosis or haemorrhagic effusion. • Pleural tumours include benign causes (fibroma, lipoma) or malignant causes, which may be primary (malignant mesothelioma) or more commonly metastatic (usually adenocarcinoma with common sites of origin being ovary, stomach, breast and lung). • Tumour-like conditions of the pleura are rare but include Erdheim–Chester disease, thoracic splenosis and thoracic endometriosis. CT, Computed tomography; CXR, chest x-ray; FDG, fluorodeoxyglucose; MRI, magnetic resonance imaging; PET, positron-emission tomography.

https://t.me/afkebooks

58

SECTION A  The Chest and Cardiovascular System

A

B

D

Fig. 3.26  Malignant Mesothelioma. FDG-PET/CT of a patient with malignant mesothelioma. CT (A), PET (B and C) and PET/CT fusion image (D) showing extent of the tumour. CT, Computed tomography; FDG, fluorodeoxyglucose; PET, positron-emission tomography.

C

Pleural Calcification Pleural calcification is most commonly seen following asbestos exposure, empyema (usually tuberculous) and haemothorax (see Fig. 3.17). In the last two conditions, calcification is irregular, resembles a plaque or sheet and is contained within thickened pleura. It may occur anywhere but is most common in the lower posterior half of the chest and is usually unilateral. However, in silicosis, particularly of the asbestos-related type, calcification occurs as more discrete collections within plaques and is usually bilateral.

efficacious. Soft-tissue lesions are readily identified in most patients using ultrasound, with previous diagnostic CT or MRI used to identify the site of impalpable abnormalities (Fig. 3.27). Rib lesions associated with a cortical break may be readily identified and biopsied, and this technique has been shown to be very useful in providing histological confirmation of primary and secondary malignancies when needed. Haemorrhagic diatheses are the main contraindication to pleural intervention. Patients unable to control their breathing and/or coughing pose relative contraindications.

Pleural Intervention

INTERVENTION

Pleural Aspiration

Interventional procedures of the chest wall and pleura may be performed for both diagnostic and therapeutic reasons. Ultrasound or CT guidance is most commonly used, although fluoroscopy and, rarely, MRI have also been used. PET/CT is now increasingly performed before chest wall or pleural biopsy to guide the procedure to the most metabolically active and therefore most likely diagnostic site.

Chest Wall Intervention It is usual to biopsy chest wall lesions, soft tissue and rib, using realtime ultrasound guidance. This has been shown to be safe and highly

This is one of the commonest interventional procedures performed in hospital practice. The main indications are in malignant pleural effusion, in which pleural fluid is sampled for diagnosis or larger-volume aspiration is performed to relieve symptoms of dyspnoea, and in pleural effusion associated with sepsis (suspected empyema), where sampling guides the diagnostic decision to drain. All pleural aspirations should be guided by ultrasound, which has been shown to reduce the incidence of complications (see later for specific complications). Pleural aspiration should be performed at the time of ultrasound, rather than marking the site for a subsequent procedure, as pleural fluid may move in the

https://afkebooks.com

CHAPTER 3  The Chest Wall, Pleura, Diaphragm and Intervention

59

considered are observation, therapeutic pleural aspiration, thoracoscopy and pleurodesis or indwelling pleural catheter placement. Factors that determine suitability include symptoms and patient performance status, the primary tumour type and the degree of lung reexpansion following pleural fluid aspiration. Drainage of infected pleural collections should occur following aspiration of frank pus or turbid/cloudy fluid and/or the presence of organisms identified by positive Gram stain. However, 40% of pleural fluid cultures are negative despite a high clinical suspicion of infection, and it is currently recommended that the aspirates for suspected infection should be sent for culture in blood culture bottles and not sterile containers. A pleural fluid pH 90%) than in women. They are usually more symptomatic than teratomas, either from mass effect or invasion of adjacent structures. Their plain radiographic findings are similar, except that the malignant tumours are more often lobular in outline. Fat density and visible calcifications are rare. Because these are malignant tumours, they grow rapidly and metastasise readily to the lungs, bones or, less often, pleura (Fig. 4.12). CT often shows a lobular, asymmetric mass. The adjacent mediastinal fat planes may be obliterated, and the tumours either are of homogeneous soft-tissue density or show multiple areas of contrast media enhancement interspersed with rounded areas of decreased attenuation caused by necrosis and haemorrhage. On MRI, these tumours may demonstrate heterogeneous intensities with areas of high T2 signal intensity corresponding to degenerative cystic changes.

https://t.me/afkebooks

78

SECTION A  The Chest and Cardiovascular System

B

A

D C Fig. 4.9  Thymic Cyst. Contrast media-enhanced thoracic computed tomography of a 42-year-old woman with dysphagia shows (A) a well-circumscribed, low-attenuation anterior mediastinal mass causing marked mass effect upon the trachea and oesophagus. Magnetic resonance imaging demonstrates a homogeneous low T1 signal intensity (B) and high T2 signal intensity (C) without evidence of enhancement following intravenous contrast media administration (D).

Mediastinal Lymphadenopathy

Malignant Lymphoma and Leukaemia Malignant lymphoma often involves mediastinal and hilar lymph nodes, multiple nodal groups usually being involved, particularly in Hodgkin disease. Hodgkin disease is the most common primary mediastinal lymphoma. In non-Hodgkin lymphoma, the two most common forms of primary mediastinal lymphomas are lymphoblastic lymphoma and diffuse large B-cell lymphoma. Any intrathoracic nodal group may be enlarged and the following generalisations regarding plain radiograph, CT and MRI findings can be made: 1. The prevascular, para-aortic and paratracheal nodes are the groups most frequently involved (Fig. 4.13). The tracheobronchial and

subcarinal nodes also may be enlarged in many cases. In most cases, the lymphadenopathy is bilateral but asymmetric. Hodgkin disease, particularly the nodular sclerosing form, has a propensity to involve the prevascular, para-aortic and paratracheal nodes. 2. Hilar node enlargement is usually seen with mediastinal node enlargement. Hilar lymphadenopathy is rare without accompanying mediastinal node enlargement, particularly in Hodgkin disease. 3. The posterior mediastinal nodes are infrequently involved—the enlarged nodes are often low in the mediastinum and contiguous retroperitoneal disease is likely. 4. The paracardiac nodes are rarely involved but become important as sites of recurrent disease because they may not be included in the initial radiation therapy fields.

https://afkebooks.com

CHAPTER 4  The Mediastinum, Including the Pericardium

79

B

A

Fig. 4.10  Cystic Teratoma. An 8-year-old girl with intermittent chest pain for 2 years was found to have an anterior mediastinal mass on a chest radiograph (A). Contrast media-enhanced computed tomography images in (B) coronal and (C) sagittal projections show a mostly cystic lesion in the anterior mediastinum, with a thick, nodular, and enhancing wall. Pathological examination after surgical excision of the mass showed typical features of a mature teratoma.

C

Lymph node enlargement is also seen occasionally with leukaemia, the pattern being the same as with lymphoma. The lymph node enlargement in both lymphoma and leukaemia may resolve remarkably rapidly with therapy.

Lymph Node Calcification Extensive lymph node calcification is common following tuberculosis and fungal infection, and is occasionally seen with other infections. It may also be encountered in a variety of other conditions, notably sarcoidosis, silicosis and amyloidosis. Although it may be seen in lymph node metastases from primary malignancies, such as osteosarcoma (Fig. 4.14), chondrosarcoma and mucinous colorectal and ovarian tumours, lymph node calcification is rare in metastatic neoplasm. It is virtually unknown in untreated lymphoma, though it is occasionally seen in nodes following therapy for Hodgkin disease.

CT is the most sensitive technique for the detection of lymph node calcification. Two common patterns of calcification are differentiated: coarse, irregularly distributed clumps within the node and homogeneous calcification of the whole node. A strikingly foamy appearance is rarely seen with Pneumocystis jiroveci (previously Pneumocystis carinii) infection in patients with acquired immune deficiency syndrome (AIDS) and in some cases of metastatic mucinous neoplasms. Sometimes there is a ring of calcification at the periphery of the node—so-called ‘eggshell calcification’—which is a particular feature of sarcoidosis and of prolonged dust exposure (e.g. mining).

Low-Attenuation Nodes On CT, areas of low attenuation within enlarged nodes, corresponding to necrosis, may be seen in a variety of conditions. This is particularly seen in tuberculosis and occasionally in fungal disease, infections in

https://t.me/afkebooks

80

SECTION A  The Chest and Cardiovascular System

A A

B

B Fig. 4.11  Seminoma. A 32-year-old man with 2 months of cough, not improving with medical treatment. He was found to have a large anterior mediastinal mass on contrast media-enhanced computed tomography of the chest. The mass has a heterogeneous attenuation, exerting significant mass effect upon the aorta and the main stem bronchi (A), as well as the right atrium and hilar vasculature (B).

immunocompromised patients, metastatic neoplasm (notably from testicular tumours) and lymphoma (Fig. 4.15). Necrotic lymph nodes are common in patients with active tuberculosis. They demonstrate central areas of low attenuation with peripheral enhancement when IV contrast medium has been administered. Attenuation values below that of water are seen in fatty replacement of inflammatory nodes and have been described in Whipple disease.

Enhancing Lymph Nodes Castleman disease is a rare cause of strikingly uniform enhancing lymph nodes. Castleman disease (also referred to as angiofollicular lymph node hyperplasia) is a specific type of lymph node hyperplasia of uncertain aetiology which can cause substantial lymph node enlargement in many sites in the body. The lymph node mass is often localised to one area, can be huge and may be very vascular. The nodes may calcify and may show striking contrast media enhancement on both CT and MRI (Fig. 4.16). Histologically, there are two types: the hyaline vascular type and the plasma cell type. In addition to Castleman disease, marked lymph node enhancement may occur in hypervascular metastases from melanoma, renal cell carcinoma, carcinoid tumours, papillary thyroid cancer and Kaposi sarcoma.

Fig. 4.12  Metastatic Choriocarcinoma. A 23-year-old man with hae­ moptysis for 2 weeks. Chest computed tomography (CT) in soft tissuewindow setting (A) demonstrates an anterior mediastinal mass of soft-tissue attenuation with lobular contours and mediastinal lymphadenopathy with enhancing margins and multiple solid pulmonary nodules and masses. (B) CT image in lung window setting demonstrates central cavitation of a few of these pulmonary metastases.

Contrast media enhancement of enlarged nodes, when moderate in degree, is non-specific, being seen with inflammatory disorders, particularly tuberculosis, fungal disease, sarcoidosis and neoplasm. A low-density centre with rim enhancement of the enlarged node is a useful pointer towards the diagnosis of tuberculous or other granulomatous infections.

Lymph Node Enlargement Normal-sized nodes are demonstrable at CT/MRI but are not visible on CXR. The ease with which enlarged nodes can be recognised at CXR varies according to their location. Nodes in the right paratracheal group are readily identified: they show uniform or lobular widening of the right paratracheal stripe. Enlarged azygos nodes displace the azygos vein laterally and enlarge the shadow that normally represents the azygos vein to over 10 mm in its short-axis diameter. If the lymph nodes beneath the aortic arch become large enough to project beyond the aortopulmonary window they cause a local bulge in the angle between the aortic arch and the main pulmonary artery. Hilar lymph node enlargement causes enlargement and/or lobulation of the outline of the hilar shadows. The diagnosis of lymph node enlargement on plain radiography depends on the recognition of the

https://afkebooks.com

CHAPTER 4  The Mediastinum, Including the Pericardium

A B

C

A

Fig. 4.13  Large B-Cell Lymphoma. A 44-year-old woman with 2 weeks of dysphonia, facial swelling and dyspnoea. Thoracic computed tomography images show an infiltrative anterior mediastinal mass exerting mass effect upon the mediastinal structures, markedly narrowing the trachea and superior vena cava (A) and causing collateral venous flow into the anterior chest wall (B). The heart is displaced into the left hemithorax by the mass (C).

B Fig. 4.14  High-Attenuation Lymph Nodes. Transaxial chest computed tomography images show densely calcified mediastinal lymphadenopathy, calcified pulmonary nodules and sclerotic osseous lesions in a woman with stage IV osteosarcoma (A and B). Dystrophic calcification is a common sequela of Histoplasma capsulatum infection; it can also be seen with metastatic lymphadenopathy of mucinous adenocarcinomas.

https://t.me/afkebooks

81

82

SECTION A  The Chest and Cardiovascular System

A

B Fig. 4.15  Low-Attenuation Lymph Nodes. Transaxial contrast mediaenhanced computed tomography images of the chest in a 45-year-old woman with fever and chest pain show rim-enhancing para-aortic (A) and subcarinal lymph nodes (B) with central low-attenuation caused by necrosis. Cultures of fine-needle aspiration of the lymphadenopathy revealed Histoplasma capsulatum.

edge of a round or oval hilar mass, an analysis that requires a detailed understanding of the normal anatomy of the hilar blood vessels. Subcarinal lymph node enlargement widens the carinal angle and displaces the azygo-oesophageal line, so that the subcarinal portion of the azygo-oesophageal line, which is normally concave towards the lung, flattens or becomes convex towards the lung. This appearance may be confused with left atrial enlargement. Subcarinal lymphadeonpathy can be appreciated on a lateral radiograph when it manifests with hilar lymphadenopathy. The combination of enlarged subcarinal and hilar lymph nodes creates a rounded mass that simulates a doughnut surrounding the mainstem bronchi. Posterior mediastinal lymph node enlargement causes localised displacement of the paraspinal and paraoesophageal lines. CT is an excellent method for detecting mediastinal lymph node enlargement. It is usually easy to distinguish between the normal vascular

structures and enlarged lymph nodes using contrast media-enhanced CT. The short-axis measurement provides the most representative guide to true size, because long-axis measurements vary to a significant degree according to the orientation of the lymph node within the CT section. In the assessment of lymph node enlargement, MRI provides essentially the same information as CT, although its use is limited to selected cases because of longer acquisition times and relatively limited spatial resolution (which may make measurement of individual nodes difficult). MRI is not very helpful for detecting calcification. Although high T2 signal may be seen in lymphadenopathy, this finding is rarely specific. Sarcoidosis.  Sarcoidosis is a common cause of intrathoracic lymph node enlargement. Mediastinal lymph node enlargement occurs at some stage in most patients, with the hilar nodes being enlarged in almost all cases. Additionally, tracheobronchial, aortopulmonary and subcarinal nodes are enlarged in over half the patients. Anterior mediastinal nodes occasionally increase in size, but posterior mediastinal and internal mammary node enlargement is rare. One important diagnostic feature of lymphadenopathy in sarcoidosis is its symmetry. Lymph node calcification may have a stippled or eggshell appearance. Tuberculosis and histoplasmosis.  Lymph node enlargement caused by tuberculous or fungal infection may affect any of the nodal groups in the hila or mediastinum. One or more lymph nodes may be visibly enlarged and an associated area of pulmonary consolidation may be present. Lymph node enlargement is usually seen ipsilateral to the side of lung disease, but involvement of the contralateral nodes may be present. Occasionally, widespread massive mediastinal and hilar node enlargement is seen. With healing, the nodes usually become smaller, often returning to normal size. Dense nodal calcification is frequent whether the nodes stay enlarged or shrink. The enlarged nodes, together with surrounding fibrosis, may compress the SVC or pulmonary veins and cause obstruction. Rim enhancement with a low-density centre may be seen with tuberculosis on contrast media-enhanced CT. Metastatic carcinoma. Mediastinal lymph node metastases can occur from primary bronchogenic carcinoma or from extrathoracic primary carcinomas. The extrathoracic tumours likely to metastasise to the mediastinum are head and neck cancer, breast cancer, genitourinary cancers and melanoma. In one large series, half the cases of mediastinal lymph node enlargement from extrathoracic primary carcinomas arose from tumours of the genitourinary tract, particularly the kidney and testis. Most metastatic tumours cause lymph node enlargement without distinguishing characteristics. Calcified lymph node metastases are typical of mucinous adenocarcinoma or thyroid carcinoma. Reactive hyperplasia. Reactive hyperplasia in nodes draining infection/inflammation may cause mild enlargement, recognisable on CT but not easily with CXR. Thoracic lymphadenopathy in AIDS.  Mediastinal lymphadenopathy is seen in 35%–40% of patients infected with human immunodeficiency virus (HIV) and may raise concern for infection or malignancy. Tuberculous and non-tuberculous mycobacterial disease and bacterial pneumonia are the primary infectious causes. Lymphoma and Kaposi sarcoma are the major causes of malignancy. Lymphadenopathy without parenchymal lung disease may occur in patients with tuberculosis, Mycobacterium avium-intracellulare or cryptococcal infection.

Foregut Duplication Cysts ‘Foregut duplication cyst’ is a term that covers various congenital cysts derived from the embryological foregut, including bronchogenic and enteric cysts. Although attempts have been made to distinguish bronchogenic from oesophageal duplication cysts, previous infection or haemorrhage may denude the epithelium. The net effect is that even the pathologist may not be more specific than the diagnosis of duplication cyst.

https://afkebooks.com

CHAPTER 4  The Mediastinum, Including the Pericardium

83

B A

C

Fig. 4.16  Castleman Disease. A 35-year-old woman with shortness of breath was further evaluated by contrast media-enhanced computed tomography (CT). Transaxial CT images (A and B) show an enhancing mediastinal mass displacing and compressing the trachea. (C) Coronal and (D) sagittal reformatted images display the craniocaudal extent of the lesion.

Bronchogenic Cysts Bronchogenic cysts are thought to result from abnormal budding of the developing tracheobronchial tree with separation of the buds from the normal airways. A good majority of bronchogenic cysts (65% to 90%) are mediastinal. Bronchogenic cysts are usually solitary asymptomatic mediastinal masses which may present at any age. Typically they have a thin fibrous capsule, are lined with respiratory epithelium and contain cartilage, smooth muscle and glandular tissue. The cyst contents usually consist of thick mucoid material. They may be filled with clear, serous fluid or thick, mucoid material. Most bronchogenic cysts are located adjacent to the trachea or main bronchi. The subcarinal location is most frequent. The cysts can grow very large without causing symptoms, but they may compress surrounding structures, particularly

D

the airways and give rise to symptoms. In rare cases they become infected or haemorrhage occurs into the cyst; these complications may be lifethreatening, particularly in infants and young children. On conventional CXRs, a bronchogenic cyst usually appears as a well-defined solitary spherical or oval mass with homogeneous opacity just inferior to the carina and often protruding slightly toward the right hilar shadow. Most are unilocular and do not have a lobulated outline. They usually contact the carina or main bronchi but may be seen anywhere along the course of the trachea and larger airways, and frequently project into the middle mediastinum. Calcification of the wall is rare. Occasionally, duplication cysts can contain milk of calcium which creates a cyst–liquid calcium level within the cyst. Foregut duplication cysts frequently push the carina anteriorly and the oesophagus posteriorly—displacements

https://t.me/afkebooks

84

SECTION A  The Chest and Cardiovascular System

that are almost never seen with other masses (the exceptions being thyroid masses and an aberrant left pulmonary artery). CT is an excellent method of demonstrating the size, shape and position of a bronchogenic cyst and defining its extent and relation to key structures. In some cases, it may demonstrate a thin-walled mass with contents of uniform CT attenuation close to that of water, thereby effectively making the diagnosis of a fluid-filled cyst (Fig. 4.17). In other cases, the CT attenuation is similar to soft tissue, and therefore to tumour, in which case the differential diagnosis becomes wider. Rarely, the cyst may show uniformly high density, probably caused by high protein content within the fluid. In these cases, unenhanced CT

images might be helpful. Calcification occurs occasionally in the wall or within the cyst contents. T1 weighted MRI show that the intrinsic signal intensity varies from low to high, depending on the cyst contents. T2 weighted images demonstrate high signal intensity. The possibility of malignancy should be considered when a solid component is seen in the cyst. One advantage of MRI is the ability to obtain fat-suppressed images before and after enhancement. The unenhanced images can be used as a mask, and subtraction images can be created. Subtraction images can be helpful in identifying subtle areas of enhancement and thus making duplication cyst unlikely.

A

B

C

D Fig. 4.17  Infected Bronchogenic Cyst. Contrast media-enhanced computed tomography (CT) of a 20-year-old man shows a middle mediastinal cystic lesion with no perceptible wall due to a bronchogenic cyst (A). A year later he presents with severe chest pain. (B) Coronal reformatted image of a contrast media-enhanced CT obtained at the time of symptomatic presentation demonstrates increased attenuation of the cyst content and development of an irregular thick wall. Further evaluation by contrast media-enhanced magnetic resonance imaging reveals intense enhancement of the wall of the infected bronchogenic cyst (C) and the surrounding mediastinal tissue due to development of acute mediastinitis (D).

CHAPTER 4  The Mediastinum, Including the Pericardium

SUMMARY BOX: Bronchogenic Cyst • Usually asymptomatic • Symptomatic if super-infected, has rapid growth due to internal haemorrhage, or compresses the airway • Most frequently located adjacent to the central airways in the subcarina • Solitary spherical or oval mass containing simple, low-attenuation fluid • Thin-walled unless super-infected • Can contain proteinaceous fluid, blood, or rarely, milk of calcium • CT and MRI demonstrate no internal enhancement • Imaging features similar to an oesophageal duplication cyst CT, Computed tomography; MRI, magnetic resonance imaging.

Oesophageal Duplication Cysts Oesophageal duplication cysts are uncommon. They usually present first in childhood but may not present until adulthood; initial presentation up to the age of 61 has been reported. They are distinguished from bronchogenic cysts pathologically by the presence of smooth muscle in the walls, absence of cartilage and presence of mucosa resembling that of the oesophagus, stomach or small intestine. Many are clinically silent and are first discovered as an asymptomatic mass on thoracic imaging, but they may cause dysphagia, pain or other symptoms caused by the compression of adjacent structures. A duplication cyst may become infected or ectopic gastric mucosa within the cyst may cause haemorrhage or perforation. The imaging features of oesophageal duplication cysts on CT and MRI are identical to those of bronchogenic cysts (Fig. 4.18) except that, in the former, the wall of the lesion may be thicker, the cyst may assume a more tubular shape and it may be in more intimate contact with the oesophagus.

Neurenteric Cysts Neurenteric cysts result from incomplete separation of the foregut from the notochord in early embryonic life and are far less common than oesophageal and bronchogenic cysts. The cyst wall contains both gastrointestinal and neural elements with an enteric epithelial lining. There is usually a fibrous connection to the spine or an intraspinal component. Communication with the subarachnoid space or the gastrointestinal tract may be present, but communication with the oesophageal lumen is rare. There are typically associated vertebral body anomalies such as butterfly vertebral deformity or hemivertebra. These cysts frequently produce pain and are often found early in life. Radiologically, a neurenteric cyst is a well-defined, round, oval or lobulated mass in the posterior mediastinum between the oesophagus (which is usually displaced) and the spine. Appearances on CT and MRI are similar to those of other foregut duplication cysts, with MRI being the investigation of choice for demonstrating the extent of intraspinal involvement.

Mediastinal Pancreatic Pseudocyst On rare occasions, pancreatic pseudocysts extend into the mediastinum. Most patients are adults and have a history of pancreatitis; in children, the usual cause of the pseudocyst is trauma. Most patients also have left-sided or bilateral pleural effusions. The mediastinal component of the pseudocyst is usually in the middle or posterior mediastinum adjacent to the oesophagus, having gained access to the chest via the oesophageal or aortic hiatus. CT is the optimal method of demonstrating these thin-walled cysts, which may show continuity with the pancreas and any peripancreatic fluid collections. MRI demonstrates the cystic nature of the mass. Isolated pancreatic mediastinal cysts are very rare. A history of pancreatitis will usually be present.

85

Neurogenic Tumours Neurogenic tumours represent 20% of all adult and 35% of all paediatric mediastinal neoplasms. Neurogenic tumours are the most common tumours to arise in the posterior mediastinum (paravertebral space), and most neurogenic tumours occur in this location. Most neurogenic tumours in adults are benign and are discovered as asymptomatic masses on chest radiography, though some, particularly the malignant lesions, cause chest pain. They can be classified as tumours arising from peripheral nerves, including neurofibromas (Fig. 4.19), schwannomas and malignant tumours of nerve sheath origin (neurogenic sarcomas), or as tumours arising from sympathetic ganglia such as neuroblastomas and ganglioneuroblastomas. MRI is the best technique for imaging these tumours. Neurofibromas or schwannomas are more common in adults, whereas neuroblastomas and ganglioneuroblastomas are more common in children. As a general guideline, peripheral nerve sheath tumours will have a craniocaudal dimension equal to their transverse dimension. Sympathetic ganglia tumours, however, tend to be longer in length than wide.

Peripheral Nerve Sheath Tumours Peripheral nerve tumours are the most common mediastinal neurogenic tumours. They typically originate in an intercostal nerve in the paravertebral region and most are benign. Radiologically, the benign tumours (neurofibromas and schwannomas) present as well-defined round or oval posterior mediastinal masses. Pressure deformity, causing a smooth, scalloped indentation on the adjacent ribs, vertebral bodies, pedicles or transverse processes, is common, particularly with larger lesions. The scalloped cortex is usually preserved and is often thickened. These bone changes are diagnostic of a neurogenic lesion, the only differential diagnosis being that of a lateral thoracic meningocele. The rib spaces and the intervertebral foramina may be widened by the tumour. A sudden increase in the size of a previously stable neurofibroma and the presence of neurological symptoms suggests malignant transformation. On CT the tumours may be homogeneous or heterogeneous, usually enhancing heterogeneously. Punctate foci of calcification may be seen. Care must be taken on CT, however, as these lesions are often homogeneous and low in attenuation (from the myelin content). The net effect is a lesion that can mimic a duplication cyst. As a general rule, a posterior mediastinal lesion should not be called a cyst unless there is a clear vertebral anomaly or communication with the spinal canal. On MRI, neurofibromas and schwannomas have low-to-intermediate signal intensity on T1 weighted images and may have characteristic high signal intensity peripherally, low signal intensity centrally (target sign) on T2 weighted images and enhance after gadolinium. Ten per cent of paravertebral neurofibromas extend into the spinal canal and appear as dumb-bell–shaped masses with widening of the affected neural foramen. Malignant tumours of nerve sheath origin (Fig. 4.20) are rare spindle cell sarcomas, typically occurring in the third to fifth decades, although they may occur earlier in patients with neurofibromatosis type 1. Radiologically, the masses are usually larger than 5 cm in diameter. Although MRI cannot reliably differentiate benign from malignant neurogenic tumours, sudden change in size of a pre-existing mass, the development of heterogeneous signal intensity (caused by haemorrhage and necrosis) or infiltration of adjacent mediastinum or chest wall is cause for concern. Hematogenous metastases to the lung have been reported, but lymph node metastasis is rare.

Sympathetic Ganglion Tumours Sympathetic ganglion tumours are rare neoplasms representing a biological continuum ranging from benign ganglioneuroma to malignant neuroblastoma, with ganglioneuroblastoma being an intermediate form.

86

SECTION A  The Chest and Cardiovascular System

A B

C

E

D

F Fig. 4.18  Oesophageal Duplication Cyst. A 59-year-old woman with hypertension and diabetes presented to the emergency department with acute chest pain. Computed tomography images of the chest obtained (A) before and (B) after administration of intravenous contrast show a low-attenuation mass adjacent to the lower oesophagus with no significant enhancement. Further evaluation by magnetic resonance imaging demonstrates intermediate signal with T1 weighted imaging (C), high signal intensity with T2 weighted imaging (D). Fat-saturated, T1 weighted imaging without (E) and with (F) intravenous contrast show enhancement of only the wall of the lesion, which is thin and smooth with no mural nodules. The remainder of the cystic lesion shows with no enhancement.

CHAPTER 4  The Mediastinum, Including the Pericardium

A

B Fig. 4.19  Neurofibromas. Contrast media-enhanced computed tomography images show low-attenuation lesions in the subcarina and subpleural chest wall, representing multiple neurofibromas involving the middle mediastinal and intercostal nerves (A and B) in a patient with neurofibromatosis.

B A

D

C Fig. 4.20  Plexiform Neurofibromas With Malignant Degeneration. A 17-year-old with neurofibromatosis type 1 presented with difficulty breathing. T1 weighted thoracic magnetic resonance images without (A) and with (B) fat saturation demonstrate a large mass with lobular contours occupying the mediastinum, with loss of signal upon application of a-saturation pulse due to the fatty content of the neurofibromas. (C) T2 weighted image shows high signal intensity and (D) contrast administration shows intense enhancement of the neurofibromas. Note compression of the airways. Biopsy of the neurofibroma revealed atypical cells suspicious of malignant degeneration on pathological assessment (not shown).

87

88

SECTION A  The Chest and Cardiovascular System

They originate from nerve cells rather than nerve sheaths and can occur in sympathetic ganglia and adrenal glands. Ganglioneuromas are benign neoplasms usually occurring in children and young adults. Ganglioneuroblastomas exhibit variable degrees of malignancy and usually occur in children. Neuroblastomas are highly malignant tumours that typically occur in children younger than 5 years of age. The posterior mediastinum is the most common extra-abdominal location of a neuroblastoma. Ganglioneuromas and ganglioneuroblastomas usually arise from the sympathetic ganglia in the posterior mediastinum and therefore usually present radiologically as well-defined elliptical masses, with a vertical orientation, extending over the anterolateral aspect of three to five vertebral bodies. Calcification occurs in approximately 25% and CT appearance is variable. On MRI, ganglioneuromas and ganglioneuroblastomas are usually of homogeneous intermediate signal intensity on T1 and T2 weighted images. Neuroblastomas are typically more heterogeneous, caused by areas of haemorrhage, necrosis, cystic degeneration and calcium. They may be locally invasive and have a tendency to cross the midline.

Mediastinal Paragangliomas Paraganglioma is a rare neuroendocrine tumour of chromaffin cell origin. Intrathoracic paragangliomas are of two types: chemodectomas or phaeochromocytomas (functioning paragangliomas), either of which may be benign or malignant. Almost all intrathoracic chemodectomas are in a location close to the aortic arch and are classified as aortic body tumours. Other mediastinal chemodectomas are very rare. They are usually single, but multicentric cases are reported. Fewer than 2% of phaeochromocytomas occur in the chest. Most intrathoracic phaeochromocytomas are found in the posterior mediastinum or closely related to the heart, particularly in the wall of the left atrium or the interatrial septum. Approximately one-third of mediastinal phaeochromocytomas are non-functioning and asymptomatic, the remainder presenting with the symptoms, signs and laboratory findings of overproduction of catecholamines. The various paragangliomas have similar appearances on chest radiography, CT and MRI. They form rounded, soft-tissue masses, which are usually very vascular and therefore enhance intensely on CT. On MRI, phaeochromocytomas usually show a signal intensity similar to that of muscle on T1 weighted images and very high signal intensity on T2 weighted images. MRI is particularly useful for demonstrating intracardiac phaeochromocytomas. Radio-iodine MIBG (131I-metaiodobenzylguanidine ) and somatostatin receptor scintigraphy both show increased activity in paragangliomas (Fig. 4.21) and are useful techniques for identifying extra-adrenal phaeochromocytomas.

A

B Fig. 4.21  Paraganglioma. Computed tomography of the chest demonstrates an enhancing mediastinal mass arising in the middle mediastinum adjacent to the right pulmonary artery (A). This lesion was resected and shown to be a paraganglioma. 131I- metaiodobenzylguanidine scintigraphy (posterior projection) of a different patient with a similar mediastinal lesion shows increased uptake, revealing that it is a paraganglioma (B).

Lateral Thoracic Meningocele Lateral thoracic meningoceles are rare posterior mediastinal cystic lesions characterised by redundant meninges (dura and arachnoid with small amounts of neural tissue within the wall) that protrude through the spinal foramen and are filled with cerebrospinal fluid (CSF). Like neurofibromas, they are commonly associated with neurofibromatosis. They present as an asymptomatic mass, often with pressure deformity of the adjacent bone, indistinguishable on plain radiographs from neurofibromas. CT and MRI can both indicate the correct diagnosis by showing the mass to be fluid filled rather than solid and demonstrating continuity between the CSF in the meningocele and that contained in the thecal sac. If necessary, the diagnosis can be established by CT with intrathecal contrast medium demonstrating flow into the lesion.

Extramedullary Haematopoiesis Extramedullary haematopoiesis can result in paravertebral masses caused by compensatory expansion of bone marrow in patients with severe

anaemia caused by inadequate production or excessive destruction of blood cells. Red blood cells are produced in the bone marrow of the long bones, pelvis, vertebrae, ribs, sternum and the skull. In conditions such as thalassaemia and sickle cell disease, there is a problem with erythropoiesis and red blood cells are produced in extramedullary sites such as the liver and spleen. The bone marrow of the vertebrae reacts to the increased demand for peripheral red blood cells by proliferating beyond the spongious bone of the vertebrae into the paraspinal regions of the posterior mediastinum. During the active phase these contain primarily red marrow tissue. Occasionally, when anaemia is treated, the haematopoietic tissue involutes and is replaced by fat and fibrosis. The mass itself almost never causes symptoms. Radiographically, lobulated paravertebral masses, usually multiple and bilateral and in the lower thoracic vertebra, are typically seen. They appear well marginated. The bones may be normal or may show an altered lace-like trabecular pattern caused by marrow expansion (Fig. 4.22). The masses are usually

CHAPTER 4  The Mediastinum, Including the Pericardium

89

B A

C

D Fig. 4.22  Extramedullary Haematopoiesis. A 43-year old man with haemolytic anaemia was found to have an abnormal chest radiograph (A) showing a large mediastinal mass which does not silhouette the cardiac borders, suggesting that the mass is located posterior to the heart. Contrast medium-enhanced computed tomography (B) shows a posterior mediastinal mass composed of areas of fat and soft attenuation. Coronal (C) and sagittal (D) reformatted images demonstrate that the mass is closely associated with the spine.

of homogeneous soft-tissue attenuation on CT, although, occasionally, when the anaemia resolves, a fatty component may be visible. Usually the masses are bilateral and reasonably symmetrical.

Mesenchymal Tumours and Tumour-Like Conditions Lymphangiomas (Cystic Hygromas)

Lymphangiomas are rare, benign congenital malformations consisting of focal proliferations of well-differentiated lymphatic tissue comprising complex lymph channels or cystic spaces containing clear or strawcoloured fluid. Most lymphangiomas are present at birth and are detected in the first 2 years of life. Lymphangiomas are most common in the neck and axilla. Ten per cent of lymphangiomas in the neck extend into the mediastinum. Lymphangiomas can occur in any part

of the mediastinum but are most common in the anterior or superior mediastinum. Mediastinal lymphangiomas may on occasion be wholly confined to the mediastinum but more frequently there will be an extension from a lymphangioma in the neck. Most cervicomediastinal lymphangiomas present in early life as a neck mass, whereas the purely mediastinal lymphangiomas usually present in older children and adults as an asymptomatic mediastinal mass. They are classified histologically as simple (capillary), cavernous or cystic (hygroma), depending on the size of the lymphatic channels they contain. Cystic lymphangiomas are most common. Lymphangiomas rarely produce symptoms caused by their soft consistency. However, compression of mediastinal structures can result in chest pain, cough and dyspnoea. Complications include airway compromise, infection, chylothorax and chylopericardium.

90

SECTION A  The Chest and Cardiovascular System compression. On CT, enhancement following contrast media administration can be dense, focal or diffuse and peripheral or central. Phleboliths or punctate calcifications are seen in 10% to 20% of the cases.

Fatty Lesions in the Mediastinum Fat is normally present in the mediastinum and its amount increases with age. Normal fat is equally distributed throughout the matrix of the mediastinum and is not encapsulated. Abnormalities of fat distribution in the mediastinum can be diffuse (mediastinal lipomatosis) or focal (fat-containing diaphragmatic hernia or mediastinal lipoma). Relatively large collections of fat are often present in the cardiophrenic angles, particularly in obese subjects. These cardiophrenic fat pads may resemble a mass.

Mediastinal Lipomatosis A

Mediastinal lipomatosis is a benign accumulation of excessive amount of redundant unencapsulated histologically normal fat in the mediastinum. Mediastinal lipomatosis is a phenomenon seen particularly in Cushing disease, in patients on steroid therapy and in obese subjects. When the fat deposits are extensive and symmetrical, the diagnosis is usually obvious. The excess fat deposition is most prominent in the upper mediastinum, resulting in a smooth symmetrical mediastinal widening on CXR. Chest CT shows exuberant, unconfined tissue of homogeneously low-attenuation fat that sharply outlines anatomical structures in all mediastinal compartments. The diffuse nature of the fatty mediastinal infiltration helps to differentiate this entity from a focal mediastinal tumour.

Fatty Tumours of the Mediastinum

B Fig. 4.23  Lymphangioma. Contrast medium-enhanced computed tomography (CT) (A) of a 32-year-old man shows a low-attenuation cystic lesion with no perceptible wall in the mediastinum. The lesion was surgically excised and pathological assessment revealed a lymphangioma. A follow-up CT (B) obtained several months later reveals a new small cystic lesion at the site of the resected lesion, due to recurrence of the lymphangioma.

On CT, a lobulated smooth mass envelopes the adjacent mediastinal structures rather than displaces them. This feature can be useful in distinguishing lymphangiomas from other mediastinal cysts. Usually they have a homogeneous fluid attenuation (Fig. 4.23) but can have a combination of fluid and soft tissue. Thin septations can sometimes be seen within the mass. On MRI the lesions may have heterogeneous T1 signal intensity but usually have high T2 signal intensity. Complete resection of lymphangiomas may be difficult because of their insinuating nature and follow-up may be needed to exclude recurrence.

Haemangiomas Haemangiomas are rare vascular tumours composed of interconnecting vascular channels with varying areas of thrombosis and fibrous stroma. Haemangiomas can be capillary, cavernous or venous, with cavernous haemangiomas accounting for approximately 75% of the cases. Haemangiomas occur in young patients, with half of the patients being asymptomatic. Symptoms, when they occur, are caused by

Fatty tumours of the mediastinum are rare. On chest radiography, regardless of whether they are benign or malignant, fatty tumours are seen as well-defined round or oval mediastinal masses. Mediastinal lipomas constitute 2% of all mediastinal tumours. They can occur in any part of the mediastinum but are most common in the prevascular space. Benign lipomas are soft and do not compress surrounding structures unless they are very large. On CT they show uniform fat attenuation. Their boundaries are smooth and sharply demarcated from adjacent mediastinal structures. Mediastinal liposarcomas are rare malignant fat-containing tumours. They may occur anywhere in the mediastinum. In contradistinction to benign lipomas, they usually contain large areas of soft-tissue density material. Histological differentiation between lipoma and liposarcoma depends on the presence of mitotic activity, cellular atypia, neovascularisation and tumour infiltration. CT findings include inhomogeneous attenuation with significant soft tissue within a mass with fat attenuation, poor definition of adjacent mediastinal structures and infiltration or invasion of mediastinal structures (Fig. 4.24). Lipoblastoma, a benign tumour of childhood, contains fat and soft tissue. Occasionally, the amount of fat attenuation is relatively small. CT findings are similar to liposarcomas. Angiomyolipoma and myelolipoma are both benign tumours which may show a combination of soft-tissue and fat attenuation on CT and therefore can be indistinguishable from liposarcoma on imaging. Angiomyolipomas and myelolipomas are rare in the mediastinum.

Fat-Containing Hernias Herniation of omental fat is a common cause of a localised fatty mass in the mediastinum. Omental fat can herniate through the foramen of Morgagni and give the appearance of a cardiophrenic angle mass on the right. Fat herniation through the foramen of Bochdalek occurs most frequently on the left side posteriorly. The fat may herniate through the

CHAPTER 4  The Mediastinum, Including the Pericardium

A

B Fig. 4.24  Liposarcoma. A 58-year-old man with gradually worsening dyspnoea was found to have a large mediastinal mass when evaluated by chest computed tomography (CT). This mass has a predominantly fat attenuation, with internal thick septations and mural nodules, and exerts significant mass effect upon the airway (A). Positron emission tomography/CT shows significant fluorodeoxyglucose uptake of the soft-tissue components of the mass (B).

oesophageal hiatus as well. Such herniations are usually readily diagnosed because of their characteristic locations. On CT or MRI, appearances consistent with fat eliminate confusion with other mediastinal masses.

OTHER MEDIASTINAL LESIONS Acute Mediastinitis Acute mediastinitis is a rare but a life-threatening condition with high mortality and morbidity. The most common causes of acute mediastinitis are postoperative complications and oesophageal perforation. Forceful vomiting may result in oesophageal perforation (Boerhaave syndrome) and a leak into the mediastinum can result in acute mediastinitis. Such tears are almost invariably just above the gastro-oesophageal junction. Other causes of acute mediastinal infection are leakage from the oesophagus into the mediastinum through a necrotic neoplasm, and extension of infection from the neck, retroperitoneum or adjacent intrathoracic or chest wall structures into the mediastinum. Clinically, the patients are often very ill with an abrupt onset of high fever, tachycardia and chest pain. Diffuse mediastinitis has a very poor prognosis. The mortality associated with acute mediastinitis from oesophageal perforation is 5%–30%, even with appropriate treatment.

91

The CXR may show widening and ill-defined mediastinal outline adjacent to the oesophagus. Streaks or collections of air may be seen within the mediastinum, and there may even be mediastinal air–fluid levels. Air may also be seen in the soft tissues of the neck. Pleural effusions are frequent and are usually on the left. The radiographic picture is often complicated by lower lobe pneumonia or atelectasis. Radiologically, detection of oesophageal perforation relies on the presence of indirect signs, including pneumomediastinum, left pleural effusion and pneumothorax. An oesophagram using water-soluble contrast medium may show the site of perforation, with extravasation of orally ingested contrast agent into the mediastinum. CT is the optimal technique in evaluating suspected mediastinitis and mediastinal abscess. Common CT findings of acute mediastinitis include increased attenuation of mediastinal fat, free gas bubbles in the mediastinum, localised fluid collections, enlarged lymph nodes, pleural effusions and empyema. Additional findings that depend on the cause of the condition are obliteration of the normal mediastinal fat planes, oesophageal thickening and extraluminal gas bubbles within the mediastinum. In advanced cases there may be walled-off discrete fluid or air–fluid collections indicating abscess formation (Fig. 4.25). There may be an associated pleural effusion, empyema, subphrenic or pericardial collection. When acute mediastinitis is suspected following sternotomy, CT shows the extent of inflammation and any drainable mediastinal or pericardial fluid collections. Distinguishing a retrosternal haematoma from reactive granulation tissue or cellulitis is difficult, as is distinguishing osteomyelitis from the direct effects of the surgical incision. It should be remembered that substernal fluid collections and tiny pockets of air are normal in the first 20 days following sternotomy. Therefore, before gas-forming infections can be diagnosed, the air collections must appear de novo or must progressively increase in the absence of any other explanation. In descending necrotising mediastinitis, CT shows solitary or multiple fluid collections, which may be contiguous with other fluid collections in the cervical region and diffuse obliteration of normal fat planes related to fasciitis.

Fibrosing Mediastinitis Fibrosing mediastinitis (sclerosing mediastinitis or mediastinal fibrosis) is a disorder that results in proliferation of fibrous tissue and collagen within the mediastinum. It is usually caused by previous infection from histoplasmosis or tuberculosis. Other causes include sarcoidosis, autoimmune diseases, retroperitoneal fibrosis, radiation and drugs such as methysergide maleate. The most common clinical consequences are obstruction to the SVC and, occasionally, obstruction to the central pulmonary arteries or veins. The CXR is non-specific and often underestimates the extent of mediastinitis. In fibrosing mediastinitis caused by previous tuberculous or fungal infection, the CXR may show calcification of mediastinal or hilar lymph nodes. CT typically shows an infiltrative, often extensively calcified, hilar or mediastinal process, which may be relatively focal when disease is caused by previous histoplasmosis or tuberculosis (Fig. 4.26), and more diffuse in the idiopathic form. Airway narrowing, vascular encasement and obstruction may also be seen. CT is excellent for the evaluation of the extent of mediastinal soft-tissue infiltration and identification of the degree of narrowing of the mediastinal structures. Two patterns of fibrosing mediastinitis have been described: a focal pattern and a diffuse pattern. The focal pattern caused by histoplasmosis, seen in 82% of cases, manifests as a mass of soft-tissue attenuation that is frequently calcified (63% of cases) and is usually located in the right paratracheal, subcarinal or hilar regions. The diffuse pattern, not related to histoplasmosis, often occurs in the setting of retroperitoneal fibrosis, seen in 18% of cases, and manifests as a diffusely infiltrating, non-calcified mass that affects multiple mediastinal compartments.

92

SECTION A  The Chest and Cardiovascular System

A B

C

Fibrosing mediastinitis typically demonstrates a heterogeneous, infiltrative mass of intermediate signal intensity on T1 weighted MRI. On T2 weighted MRI it is more variable, with areas of both increased and markedly decreased signal intensity seen in the same lesion. Areas of decreased signal intensity represent calcification or fibrous tissue, and areas of increased signal intensity may indicate more active inflammation. Extensive regions of decreased signal intensity within the lesion, when present, help differentiate fibrosing mediastinitis from other infiltrative lesions of the mediastinum, such as metastatic carcinoma and lymphoma, that typically have increased T2 signal intensity. Heterogeneous enhancement of the mass may be seen after administration of a gadolinium-based contrast medium. MRI lacks sensitivity for detection of calcification, which is an important feature for differentiating fibrosing mediastinitis from other infiltrative disorders of the mediastinum, such as lymphoma and metastatic carcinoma.

Mediastinal Haemorrhage Mediastinal haemorrhage is most commonly caused by trauma to the arteries and veins within the mediastinum, with other causes including

Fig. 4.25  Mediastinal Abscess. A 36-year-old man who has a history of polysubstance abuse presents with shoulder pain and erythema of the chest wall. Contrast media-enhanced computed tomography of the chest (A–C) shows an anterior mediastinal fluid collection containing bubbles of gas representing a mediastinal abscess. This abscess extends to the anterior left chest wall above the level of the manubrium, and there is thickening of the left pectoralis muscle, with bubbles of gas representing a chest wall myositis and developing abscess.

rupture of an aneurysm, aortic dissection and complications of central venous catheterisation. Radiologically, haemorrhage produces an increase in the mediastinal diameter, which is maximal at the point of bleeding. Blood may track through the mediastinum, frequently running over the apex of the left lung to produce a smooth and well-defined apical cap. When haemorrhage is severe, blood may rupture into the pleural cavity or dissect into lung along peribronchovascular sheaths, resulting in a radiographic pattern resembling interstitial oedema. On unenhanced CT, acute haemorrhage may appear of relative high attenuation. The appearance of mediastinal haematoma on MRI varies with the age of the haemorrhage.

Pneumomediastinum Pneumomediastinum is characterised by the presence of free air around the mediastinal structures. Common causes of pneumomediastinum include blunt or penetrating trauma, oesophageal perforation, recent interventions in the oesophageal or tracheobronchial tree, pulmonary infections, gas-forming infections in the mediastinum, cocaine inhalation, and extension of air from a pneumothorax.

CHAPTER 4  The Mediastinum, Including the Pericardium

A B

D

C

E

Fig. 4.26  Fibrosing Mediastinitis. Contrast media-enhanced computed tomography (A–D) shows a partially calcified mediastinal and right hilar mass consistent with fibrosing mediastinitis secondary to histoplasmosis in a 46-year old woman with chronic cough and facial swelling. The mass causes stenosis of the superior vena cava, with dilated right internal mammary and right superior intercostal and azygous veins due to collateral venous flow (A–C). Coronal reformatted image shows that the right upper lobe pulmonary artery is completely obliterated by the fibrosing mass (D). A scintigraphic sestamibi pulmonary perfusion scan (E) demonstrates no perfusion of the right upper lobe due to the arterial occlusion.

93

94

SECTION A  The Chest and Cardiovascular System

The presence of a pneumomediastinum is, in itself, of little significance (though it may be responsible for substernal chest pain), but the condition causing the air leak (particularly bronchial, oesophageal or pharyngeal perforation) may be of great significance to the patient. The radiographic signs of pneumomediastinum depend on the anatomical structures outlined by the air. Air around the pulmonary

artery (usually the right pulmonary artery) results in the ‘ring around the artery sign’ (Fig. 4.27). Elevation of the thymus causes the ‘sail sign’. Air anterior to the pericardium is best seen on the lateral radiograph. The ‘continuous diaphragm sign’ is seen because of the air trapped posterior to the pericardium, giving the appearance of a continuous collection of air on the AP projection (Fig. 4.28). The ‘tubular artery

A

B

C

D Fig. 4.27  ‘Double Bronchial Wall’ and ‘Ring Around the Artery’ Signs of Pneumomediastinum. A 68-year-old man admitted with shortness of breath and chest pain. (A) Frontal view of the chest shows parallel lucent lines outlining the left main stem bronchus, representing the ‘double bronchial wall sign’. (B) Lateral projection of the chest radiograph demonstrates the ‘ring around the artery sign’ manifesting as a lucent line encircling the right pulmonary artery, indicating that the mediastinal air is tracking into the right hilum. Axial and sagittal reformatted computed tomography images of the chest show the air surrounding the central bronchi in the mediastinum (C) and the right pulmonary artery in the hilum (D).

CHAPTER 4  The Mediastinum, Including the Pericardium

95

A

A

B

B Fig. 4.28  ‘Continuous Diaphragm Sign’ in Pneumomediastinum. Frontal chest radiograph shows an uninterrupted outline of the diaphragm, indicative of a pneumomediastinum (A). Computed tomography image of the lower chest demonstrates the air in the anterior mediastinum, adjacent to the diaphragm, which is responsible for this sign (B).

sign’ occurs when there is air adjacent to the major branches of the aorta, the mediastinal air outlines the medial side and the aerated lung outlines the lateral side of the vessel. The ‘double bronchial wall sign’ is seen when the air adjacent to the bronchus allows clear depiction of the bronchial wall. Air can also dissect through the perivascular tissues and may track up into the neck, supraclavicular areas and axillae, as well as down into the retroperitoneum. The differential diagnosis of a pneumomediastinum on CXR includes a medially located pneumothorax and a ‘Mach effect’ caused by the abrupt change in density between the lung and the adjacent heart and mediastinum. The Mach band effect is associated with convex surfaces, appearing as a region of lucency adjacent to structures with convex borders. The absence of an opaque line, which is typically seen in

Fig. 4.29  Comparison of Pneumomediastinum on Computed Tomo­ graphy. Computed tomography images of two different patients show pneumomediastinum in the first patient (A) and a combination of pneumopericardium and a left pneumothorax in the second patient (B). Note that in pneumomediastinum there are strands of soft tissue within the areas of mediastinal air, due to dissection of the tissue by the air. However, in pneumopericardium and in pneumothorax such strands of tissue are not seen because the air accumulates in a ‘true’ space between the two layers of pericardium (for pneumopericardium) and the two layers of pleura (for pneumothorax).

pneumomediastinum, can aid in differentiation. It is easy to appreciate why a medial pneumothorax can be mistaken for a pneumomediastinum, because in both instances there is a linear collection of air bounded on its lateral side by a thin line of pleura. Deciding whether the line is mediastinal parietal pleura or visceral pleura can be difficult; the distinction often depends on recognising the full extent of the air and looking carefully for a pneumothorax, or looking for evidence of air elsewhere in the mediastinum. Pneumomediastinum is easy to diagnose on CT as streaks or rounded collections of gas surrounding vessels and other structures (Fig. 4.29).

PERICARDIUM The pericardium is a compliant sac that consists of two layers, the parietal and visceral pericardium, separated by a small amount of fluid,

96

SECTION A  The Chest and Cardiovascular System

which is normally less than 50 mL. The pericardium envelops the cardiac chambers and the origins of the great vessels. The left atrium is partially covered by the pericardium. The thickness of the normal pericardium, as measured on CT and MRI, is less than 2 mm. Pericardial sinuses may be seen on CT and MRI, containing small amounts of fluid in normal healthy individuals. The oblique pericardial sinus behind the left atrium may be misinterpreted (e.g. bronchogenic cyst). The transverse pericardial sinus posterior to the ascending aorta may also be misinterpreted (aortic dissection, lymphadenopathy, etc.). The superior pericardial recess lies posterior to the ascending aorta.

IMAGING PERICARDIAL DISEASE Chest radiography is of limited use in the assessment of pericardial disease although pericardial effusions, calcification and secondary signs and complications of pericardial disease may be evident. Interval enlargement of the cardiac silhouette should raise the suspicion of pericardial effusion. Transthoracic echocardiography (TTE) is usually the initial investigation of suspected pericardial disease. It is inexpensive and widely available and has high accuracy for detecting pericardial effusions and signs of tamponade. TTE is also helpful for guiding diagnostic or therapeutic pericardiocentesis. Restricted acoustic windows limit evaluation of the entire pericardium; loculated collections, intrapericardial blood clot and pericardial thickening may be difficult to assess. TTE is not very accurate for depicting pericardial thickening, because echogenicity of the pericardium is similar to adjacent tissues. Transoesophageal echocardiography is limited by a narrow field ofview. CT and MRI have distinct advantages over echocardiography, including larger field of view, higher contrast media resolution, excellent anatomical delineation and multiplanar reformats. CT with multiplanar reformats, particularly if ECG gated, provides excellent motion-free assessment of the pericardium; advantages include speed and wide availability and accessibility. CT can also detect pericardial calcifications that may be indicative of constrictive pericarditis. Disadvantages of CT include ionising radiation and the need for IV iodinated contrast agent. MRI can provide a comprehensive assessment of the pericardium. When T1 and T2 weighted sequences (some with ECG-gated breath-hold techniques) are combined with cine-based functional cardiac imaging, both pericardial disease and its impact on cardiac function can be assessed. MRI has some advantages over US and CT in detecting and characterising pericardial collections and masses. Limitations of MRI include its inability to reliably depict calcification, and relatively long data acquisition times, especially with regard to breath-holding. Arrhythmias, which commonly occur in association with pericardial disease, may affect image acquisition and quality. Nevertheless, CT or MRI should be used when findings on echocardiography are difficult to interpret or non-diagnostic.

DEVELOPMENTAL ANOMALIES Congenital Absence of the Pericardium Compromise of the vascular supply to the pleuropericardial membrane during embryological development is associated with congenital defects in the pericardium. Pericardial defects are rare and are usually asymptomatic. The defects vary in size from small communications between the pleural and pericardial cavities to complete (bilateral) absence of the pericardium. The most common form is complete absence of the left pericardium, with preservation of the pericardium on the right. Bilateral and isolated right-sided lesions are very rare. Absence of the pericardium is rarely associated with congenital anomalies of the heart and lungs, including atrial septal defect, tetralogy of Fallot, patent ductus arteriosus, bronchogenic cysts and pulmonary sequestration. Pericardial

defects are frequently associated with large defects in the parietal pleura, through which the left lung can herniate and surround the intrapericardial vascular structures. Complete absence of the pericardium is usually asymptomatic, whereas partial or localised absence of the pericardium may be complicated by herniation and entrapment of a cardiac chamber, in particular the left atrial appendage in left-sided defects. CXR findings are frequently subtle and non-specific. In complete absence of the left pericardium they include displacement of the heart into the left chest and interposition of lung between the aorta and pulmonary artery (as well as between the left hemidiaphragm and cardiac silhouette). Both the medial and lateral borders of the main pulmonary artery may be visualised more clearly, caused by absence of the anterior pericardial reflection between the aorta and the pulmonary artery. Because of leftward displacement and rotation, the right cardiac border may not be seen. In partial pericardial defects, varying degrees of prominence of the pulmonary artery and/or left atrial appendage may be seen while the heart retains its normal position in the thorax. CT and MRI can depict herniation of cardiac structure through the defect. Discontinuation of the pericardial line can occasionally be detected in the partial form. The most reliable signs of complete absence of the left pericardium are interposition of lung between the aorta and main pulmonary artery (in the aortopulmonary window), and a rotation of the cardiac axis to the left side (rather like a right anterior oblique view).

Pericardial Cysts Pericardial cysts are formed when a portion of the pericardium is pinched off during early development and are thought to be the result of persistence of blind-ending ventral parietal pericardial recesses. Those cysts that communicate with the pericardial space are termed pericardial diverticula. With increase or decrease in pericardial fluid, diverticula can change in size. They almost invariably appear as a well-defined, oval or occasionally lobulated mass attached to the pericardium. More occur in the right cardiophrenic angle (~70%) than on the left (~20%); some are seen higher in the mediastinum. They contain clear fluid and can be recognised as fluid-filled cysts surrounded by normal pericardium on echocardiography, CT or MRI (Fig. 4.30). On MRI they have lowto-intermediate T1 signal intensity and homogeneous high T2 signal intensity. They do not enhance following IV gadolinium administration.

ACQUIRED PERICARDIAL DISEASE Pericardial Effusion Pericardial effusions are transudative or exudative accumulations of fluid in the pericardial space. Common causes of pericardial effusion include heart failure, renal insufficiency, infection (bacterial, viral or tuberculous), neoplasm (carcinoma of lung, breast or lymphoma) and injury (trauma and myocardial infarction). Transudative pericardial effusions may develop after cardiac surgery or in congestive heart failure, radiation, uraemia, post-pericardiectomy syndrome, myxoedema and collagen–vascular diseases. Haemopericardium may be caused by trauma, aortic dissection, aortic rupture or neoplasm. Interval enlargement of the cardiac silhouette on a radiograph over a short period of time should raise the suspicion of pericardial effusion. Filling in of the retrosternal space, effacement of the normal cardiac borders, development of a ‘flask’ or ‘water bottle’ cardiac configuration and bilateral hilar overlay are features of pericardial effusion. The epicardial fat pad sign may be seen on the lateral projection that demonstrates an anterior pericardial stripe (bordered by epicardial fat posteriorly and mediastinal fat anteriorly) thicker than 2 mm. This sign represents pericardial thickening or fluid (Fig. 4.31). TTE is highly sensitive and specific for evaluating

CHAPTER 4  The Mediastinum, Including the Pericardium

A

B

C

D Fig. 4.30  Pericardial Cyst. (A and B) Frontal and lateral chest radiographs of a 38-year-old woman show an abnormal mass-like contour of the left ventricle. (C) Axial non-contrast media-enhanced computed tomography shows a mass of fluid attenuation adjacent to the left ventricle apex and free wall. Magnetic resonance imaging (MRI) of the heart was performed for further evaluation. (D) Steady-state free-precession bright blood image shows the close association of this lesion with the pericardium and a uniform high signal of this cystic lesion with a thin internal septum. Continued

97

98

SECTION A  The Chest and Cardiovascular System

E

F Fig. 4.30, cont’d (E) T2 weighted fat-saturated sequence demonstrates intensely high signal similar to simple fluid. (F) Contrast media-enhanced T1 weighted MRI shows no enhancement of this lesion, indicating it is a simple pericardial cyst.

pericardial effusion is seen in patients with malignancy, the pericardium should be carefully evaluated for nodular (possibly metastatic) thickening of the pericardium. MRI is useful to differentiate small pericardial effusion from pericardial thickening. On spin-echo MRI, the signal characteristics of pericardial collections vary, depending on the composition of the fluid. In the absence of haemorrhage, effusions are typically of predominantly low T1 signal intensity, although intermediate signal intensity may be seen in inflammatory conditions such as uraemia, tuberculosis or trauma, possibly reflecting high protein content and, when more focal, the presence of adhesions limiting normal flow of pericardial fluid in the pericardial space. In haemorrhagic effusions, signal intensity varies, depending on the age of blood products.

Cardiac Tamponade

Fig. 4.31  Pericardial Effusion. A 30-year-old man with chronic kidney disease presented with dyspnoea. Lateral projection of a chest radiograph shows the sandwich sign of a pericardial effusion manifesting as a wide band of higher density in the retrosternal space, outlined by two bands of lower density. The higher density represents the fluid in the pericardium and the lower densities represent the epicardial fat located deep to the visceral pleura layer and the pericardial fat external to the visceral pleura layers.

pericardial disease although visualisation may be limited in some obese or emphysematous patients; loculated collections and intrapericardial clot in postoperative patients may be difficult to detect. CT and MRI are indicated when TTE is inconclusive or when loculated or haemorrhagic effusion or pericardial thickening is suspected. Increased attenuation in a pericardial effusion on CT suggests haemorrhage (Fig. 4.32). When

Gradual accumulation of pericardial fluid may fail to produce clinical signs or symptoms for an extended period of time. However, rapid accumulation of as little as 100–200 mL of fluid can cause a haemodynamically significant compression of the heart, which severely impedes diastolic filling, resulting in pericardial tamponade. Despite the fact that the left ventricular contractility is normal, the stroke volume is decreased because of the diminished end-diastolic volume, resulting in decreased cardiac output. Because acute tamponade may occur with small effusions, clinically important pericardial enlargement may be difficult to detect on CXR. Subtle changes in cardiac contour may only be detectable by comparison with previous studies. If there is decreased pulmonary vascularity despite the cardiac enlargement or if the SVC and azygos veins are dilated, tamponade may be suspected. Echocardiographic demonstration of pericardial effusion and the clinical findings are usually sufficient to make the diagnosis of tamponade. CT and MRI are frequently useful for determining the cause of the effusion, some of which include haemorrhage, neoplastic involvement, inflammation caused by tuberculosis, or other infectious processes.

Pericarditis Inflammation of the pericardium (pericarditis) may occur in response to a variety of insults. Viral infection is the most common cause of

CHAPTER 4  The Mediastinum, Including the Pericardium

99

myocardial infarction (acute or post-myocardial infarction referred to as Dressler syndrome), pericardiotomy, mediastinal irradiation, infection (viral or bacterial), connective tissue disease (rheumatoid arthritis, SLE), metabolic disorders (uraemia, hypothyroidism), neoplasia and AIDS. The most common imaging manifestation of acute pericarditis is a pericardial effusion, the nature of the fluid varying with the underlying cause. Thickened, inflamed pericardium can appear as moderate-to-high signal intensity on spin-echo MRI, and pericardial enhancement may be seen on either MRI or CT performed after IV contrast medium administration. Delayed images on contrast media-enhanced CT are useful for demonstrating pericardial enhancement (Fig. 4.33).

Constrictive Pericarditis

A

B Fig. 4.32  Pericardial Haemorrhage. A 76-year-old with poorly controlled hypertension presents with chest pain. Axial computed tomography images of the chest performed (A) without and (B) with intravenous contrast medium demonstrate a high-attenuation fluid collection in the pericardium, representing haemorrhage with no pericardial thickening or enhancement. This haemopericardium was caused by an acute intramural haematoma of the aorta (not shown).

pericarditis in the United States, with the most common agents being Coxsackie group B and echoviruses. Pericarditis typically results in cellular proliferation, or the production of fluid (pericardial effusion) or fibrin, either alone or in combination. Thickening of the pericardium occurs because of fibrinous exudates and oedema. Causes include

Constrictive pericarditis presents with symptoms of heart failure such as dyspnoea, orthopnoea and fatigue. The most common causes of constrictive pericarditis are cardiac surgery and radiation therapy. Other causes include infection (viral, tuberculous), connective tissue disease, uraemia, neoplasm or idiopathic. The aetiology is unknown in many cases, presumed to be secondary to an occult viral pericarditis and other causes of pericarditis. Outside the United States, the most common cause is probably infectious. Any insult to the pericardium can progress from an acute pericarditis with pericardial effusion to a subacute stage of resorption of the effusion with organisation, and then to a chronic phase of fibrous scarring, pericardial thickening and obliteration of the pericardial cavity. Constrictive pericarditis is the condition in which a thickened, fibrotic and often calcified pericardium restricts diastolic filling of the heart. Constriction caused by neoplastic infiltration of the pericardium is most commonly secondary to carcinoma of the lung or breast, lymphoproliferative malignancies and melanoma. Pericardial constriction after mediastinal irradiation, usually performed to treat breast carcinoma or Hodgkin disease, may occur months to years after treatment. Pericardial thickening is seen in up to 88% of confirmed cases of constrictive pericarditis. In most cases, constrictive pericarditis involves the entire pericardium, restricting filling of all cardiac chambers. Occasionally, particularly anterior to the right ventricle in postoperative patients, the pericardial thickening is more localised. Constrictive pericarditis and restrictive cardiomyopathy are both characterised by restriction in diastolic filling, which leads to increase of diastolic pressure in all four chambers and equalisation of pressures. The clinical manifestations and findings on cardiac catheterisation and echocardiography are similar in both conditions. It is important to differentiate between these two conditions because the management approach will differ. Patients with pericardial constriction may benefit from pericardial stripping, while restrictive cardiomyopathy is managed medically or by cardiac transplantation. Diagnosing constriction often proves challenging and usually requires more than one investigation before surgery. The hallmarks of pericardial constriction are pericardial thickening, calcification and abnormal diastolic ventricular function. Although echocardiography is routinely performed and provides an excellent assessment of haemodynamic function, it is not highly accurate at depicting pericardial thickening. CT and MRI are significantly more sensitive, with CT having the advantage over MRI of being able to demonstrate the presence of calcification, which is associated with pericardial constriction. Pericardial calcification can be seen in the atrioventricular groove in Fig. 4.34. Pericardial thickening of greater than 4 mm, when accompanied by clinical features of constriction, is highly suggestive of constrictive pericarditis. Both CT and MRI may show the secondary effects of constriction on the central cardiovascular structures. The right ventricle tends to have a conical configuration and reduced volume. A sigmoidshaped interventricular septum or prominent leftward convexity of the septum may be seen. Hepatomegaly and ascites may also be seen. The right atrium, the SVC and, in particular, the inferior vena cava, and

100

SECTION A  The Chest and Cardiovascular System

A B

C

hepatic veins may be dilated. Cardiac MRI can also be used to provide a more detailed assessment of cardiac function. Diastolic septal bounce can be seen on cardiac MRI. A free-breathing sequence on cardiac MRI in which a patient performs a ‘sniff ’ while the images are acquired, which demonstrates an exaggerated septal bounce (often referred to as ventricular interdependence), is helpful in leading to the diagnosis.

Pericardial Neoplasms Pericardial metastases are as much as 20 to 40 times more common than primary pericardial neoplasms. They are identified at autopsy in approximately 10% of all patients with malignancy. The most common malignancies encountered are lung (Fig. 4.35), lymphoma, breast, melanoma and colon. Primary pericardial neoplasms are rare, with approximately equal incidence of benign versus malignant pericardial neoplasms. Benign tumours include teratomas, fibromas, neurofibromas, lipomas, haemangiomas

Fig. 4.33  Pericarditis. A 31-year-old-man with new onset of chest pain which developed a few days after an upper respiratory tract infection. (A) Axial, (B) coronal- and (C) sagittal-reformatted images of a contrast media-enhanced computed tomography show a large pericardial effusion outlined by a smoothly thickened, enhancing pericardium. The cranial extent of the pericardium is displayed on the sagittal reformatted image (C).

SUMMARY BOX: Constrictive Pericarditis • Presents with dyspnoea, orthopnoea and fatigue • Most common cause in the USA is iatrogenic (radiation and cardiac surgery) • Most common cause in the world is infectious • Pericardial thickening >4 mm • Pericardial thickening can be localised adjacent to right ventricular free wall or the right atrioventricular groove • Pathophysiology based on restricted diastolic filling of the cardiac chambers • CT more sensitive in depicting pericardial calcification • MRI can show ventricular interdependence on a free breathing sequence • Cardiac cirrhosis may ensue and can be reversible with pericardectomy CT, computed tomography; MRI, magnetic resonance imaging; RV, right ventricle.

CHAPTER 4  The Mediastinum, Including the Pericardium

101

B

A

D

C Fig. 4.34  Constrictive Pericarditis. A 58-year-old man presented with several months of lower-extremity oedema, shortness of breath and ascites. A chest radiograph shows a curvilinear calcific density over the heart (A). Axial reformatted images of contrast media-enhanced computed tomography show diffuse pericardial thickening with areas of pericardial calcification (B) and elongation of the ventricles with inward bowing of the right ventricle free wall (C). Secondary findings of diastolic heart failure due to the severe constrictive pericarditis are seen, manifesting as bilateral pleural effusions, ascites, inferior vena cava and hepatic venous distension, and a ‘nutmeg’ liver parenchyma due to liver congestion and development of hepatic cirrhosis (D).

and lymphangiomas. Although these patients are usually symptom free, pericardial effusion or constriction, particularly in the case of childhood teratomas, may occur. Malignant mesothelioma is the most common primary pericardial malignancy (Fig. 4.36) and is almost certainly related to asbestos exposure. Mesothelioma may present as a well-defined single mass, multiple nodules or diffuse plaques involving the visceral and parietal pericardium and wrapping around the cardiac chambers and great vessels. Clinically, it presents with haemorrhagic effusion and tamponade, congestive heart

failure, arrhythmia and occasionally pericardial constriction. Other malignant primary tumours include lymphoma, sarcoma, paraganglioma and liposarcoma. Teratomas of the pericardium may also be malignant and are most commonly seen in children. A pericardial effusion is the most common finding in pericardial malignancy, whether primary pericardial or metastatic. Intrapericardial neoplasms tend to compress and deform normal intrapericardial structures, whereas extrapericardial masses tend to displace the intrapericardial structures without compression or distortion.

102

SECTION A  The Chest and Cardiovascular System

A

B

Fig. 4.36  Mesothelioma. A 38-year-old woman with progressive shortness of breath. Contrast media-enhanced computed tomography images obtained several months later demonstrate marked nodular pericardial thickening with a small pericardial effusion. The nodules enhance and compress the right atrium and right ventricle. Surgical resection revealed primary mesothelioma of the pericardium.

CXRs are often abnormal but are non-specific. Alteration of fat-pad contours, cardiac enlargement, mediastinal widening, hilar adenopathy or a hilar mass may be seen. Echocardiography is usually the initial technique for evaluating a suspected pericardial neoplasm, with MRI and CT being useful for further evaluation. Both MRI and CT are excellent at providing information regarding the size, location and extent of pericardial neoplasms, but are not tissue specific. Fatty tumours (lipomas, fat-containing teratomas) are the exception, because of their typically low attenuation on CT and increased signal intensity on spin-echo T1 weighted MRI. Fatty tumours must be differentiated from the focal deposits of subepicardial fat and non-neoplastic lesions that can simulate fatty tumours, such as mesenteric fat in a hiatal hernia. Metastatic melanoma may have high signal intensity on T1 and T2 weighted images, a feature that may be useful in differentiating it from other metastatic neoplasms, which are frequently of low signal intensity on T1 weighted images and high signal intensity on T2 weighted images. In addition to discrete masses and effusions, metastatic involvement of the pericardium may cause focal or diffuse pericardial thickening, which may be irregular and usually enhances. Primary lipoma, liposarcoma and lymphoma of the pericardium typically appear as large heterogeneous masses frequently associated with a serosanguineous pericardial effusion.

ACKNOWLEDGEMENT C Fig. 4.35  Pericardial Metastasis. A 63-year-old man with stage IV non-small-cell lung cancer. Axial contrast media-enhanced computed tomography (CT) images (A and B) demonstrate nodular thickening of the pericardium, representing metastases, with a small pericardial effusion. Note the nodular pericardial metastasis extends into the adjacent pericardial fat in the anterior mediastinum. CT image optimised for the lung (C) demonstrates a spiculated nodule in the right upper lobe consistent with known non-small-cell lung cancer.

The authors acknowledge the contribution of Shahrzad Aziziddini, MD to this chapter in the current edition of this book.

FURTHER READING Ackman, J.B., 2015. MR imaging of mediastinal masses. Magn. Reson. Imaging Clin. N. Am. 23 (2), 141–164. Bonekamp, D., Horton, K.M., Hruban, R.H., et al., 2011. Castleman disease: the great mimic. Radiographics 31 (6), 1793–1807. Carter, B.W., Okumura, M., Detterbeck, F.C., et al., 2014. Approaching the patient with an anterior mediastinal mass: a guide for radiologists. J. Thorac. Oncol. 9 (9 Suppl. 2), S110–S118.

CHAPTER 4  The Mediastinum, Including the Pericardium Carter, B.W., Tomiyama, N., Bhora, F.Y., et al., 2014. A modern definition of mediastinal compartments. J. Thorac. Oncol. 9 (9 Suppl. 2), S97–S101. Carter, B.W., Benveniste, M.F., Truong, M.T., et al., 2015. State of the art: MR imaging of thymoma. Magn. Reson. Imaging Clin. N. Am. 23 (2), 165–177. Fujimoto, K., Hara, M., Tomiyama, N., et al., 2014. Proposal for a new mediastinal compartment classification of transverse plane images according to the Japanese Association for Research on the Thymus (JART) General Rules for the Study of Mediastinal Tumors. Oncol. Rep. 31 (2), 565–572. Jerushalmi, J., Frenkel, A., Bar-Shalom, R., et al., 2009. Physiologic thymic uptake of 18F-FDG in children and young adults: a PET/CT evaluation of incidence, patterns, and relationship to treatment. J. Nucl. Med. 50 (6), 849–853. Juanpere, S., Cañete, N., Ortuño, P., et al., 2013. A diagnostic approach to the mediastinal masses. Insights Imaging 4 (1), 29–52.

103

Lewis, R.B., Mehrotra, A.K., Rodriguez, P., et al., 2013. Esophageal neoplasms: radiologic-pathologic correlation. Radiographics 33 (4), 1083–1108. Molinari, F., Bankier, A.A., Eisenberg, R.L., 2011. Fat-containing lesions in adult thoracic imaging. AJR Am. J. Roentgenol. 197 (5), W795–W813. Nasseri, F., Eftekhari, F., 2010. Clinical and radiologic review of the normal and abnormal thymus: pearls and pitfalls. Radiographics 30 (2), 413–428. Takahashi, K., Al-Janabi, N.J., 2010. Computed tomography and magnetic resonance imaging of mediastinal tumors. J. Magn. Reson. Imaging 32 (6), 1325–1339. Tatci, E., Ozmen, O., Dadali, Y., et al., 2015. The role of FDG PET/CT in evaluation of mediastinal masses and neurogenic tumors of chest wall. Int. J. Clin. Exp. Med. 8 (7), 11146–11152. Tomiyama, N., Müller, N.L., Ellis, S.J., et al., 2001. Invasive and noninvasive thymoma: distinctive CT features. J. Comput. Assist. Tomogr. 25 (3), 388–393.

5  Pulmonary Infection in Adults Tomás Franquet, Gustavo Meirelles

CHAPTER OUTLINE Types of Pneumonias, 104 Clinical Utility and Limitations of Chest Radiography and Computed Tomography, 104 Patterns of Pulmonary Infection, 105 Complications of Pneumonia, 106

Respiratory infections are the most common illnesses occurring in humans and pneumonia is the leading cause of death due to infectious disease and the sixth most common cause of death in the United States. Pneumonia is an acute infection of the pulmonary parenchyma that is associated with at least some symptoms of acute infection, accompanied by the presence of an acute infiltrate on a chest radiograph.

TYPES OF PNEUMONIAS Currently accepted classifications of pneumonia include communityacquired pneumonia (CAP), hospital-acquired pneumonia (HAP), ventilator-associated pneumonia (VAP) and health care–associated pneumonia (HCAP).

Community-Acquired Pneumonia (CAP) (Box 5.1) The diagnosis of CAP is based on the presence of select clinical features (e.g. cough, fever, sputum production and pleuritic chest pain) and is supported by imaging of the lung, usually by chest radiography (Box 5.1). The spectrum of causative organisms of CAP includes gram-positive bacteria such as Streptococcus pneumoniae (pneumococcus), Haemophilus influenzae and Staphylococcus aureus, as well as atypical organisms such as Mycoplasma pneumoniae, Chlamydia pneumoniae, or Legionella pneumophila, and viral agents such as influenza A virus and respiratory syncytial viruses. Pulmonary opacities are usually evident on the radiograph within 12 hours of the onset of symptoms. Although the radiographic findings do not allow a specific aetiological diagnosis, they may be helpful in narrowing down the differential diagnosis.

Integrating Clinical and Imaging Findings, 107 Changing Spectrum of Human Immunodeficiency Virus Infections: 40 Years Later, 115 Parasitic Infections, 120

Ventilator-Associated Pneumonia (VAP) Microorganisms responsible for VAP may differ according to the population of patients in the ICU, the duration of hospital and ICU stays, and the specific diagnostic method(s) used. The spectrum of causative pathogens of VAP in humans is S. aureus, Pseudomonas aeruginosa and Enterobacteriaceae.

Health Care–Associated Pneumonia (HCAP) When pneumonia is associated with health care risk factors such as prior hospitalisation, dialysis, residing in a nursing home, and immunocompromised state, it is now classified as health care–associated pneumonia (HCAP). The number of individuals receiving health care outside the hospital setting, including home wound care or infusion therapy, dialysis, nursing homes and similar settings is constantly increasing.

Aspiration Pneumonia Infectious pneumonia needs to be differentiated from aspiration pneumonia, which also presents with patchy consolidations typically in the dependent portions of the lung (superior segments of the lower lobes and posterior segments of the upper lobes). The pattern is very variable, dependent on the quantity and quality of aspirated material, and ranges from tree-in-bud to patchy consolidations, usually multilobar and bilateral in distribution, though more frequently and more extensively to the right side due to the vertical position of the right-sided central airways.

Hospital-Acquired Pneumonia (HAP)

CLINICAL UTILITY AND LIMITATIONS OF CHEST RADIOGRAPHY AND COMPUTED TOMOGRAPHY

Hospital-acquired pneumonia (HAP) may be defined as one occurring after admission to hospital and was neither present nor in a period of incubation at the time of admission. Hospital-acquired pneumonia (nosocomial) is the leading cause of death from hospital-acquired infections and a serious public health problem. It occurs most commonly among intensive care unit (ICU) patients, predominantly in individuals requiring mechanical ventilation.

A clinical diagnosis of pneumonia can usually be readily established on the basis of signs, symptoms and chest radiographs. However, distinguishing pneumonia from conditions such as left heart failure, pulmonary embolism and aspiration pneumonia may sometimes be difficult, especially as patients with pre-existing lung disease (severe emphysema, interstitial lung disease etc.) may develop very atypical patterns of pneumonia.

104

CHAPTER 5  Pulmonary Infection in Adults

105

BOX 5.1  Differential Diagnosis of

Community-Acquired Pneumonia Radiographic Findings

Most Common Organisms

Lobar consolidation

Streptococcus pneumoniae, Klebsiella pneumoniae S. pneumoniae Staphylococcus aureus, gram-negative bacilli, anaerobes, S. pneumoniae Virus, Mycoplasma pneumoniae Mycobacterium tuberculosis, S. aureus, gram-negative bacilli

Round pneumonia Bronchopneumonia Interstitial pneumonia Cavity formation

Although different patterns of pneumonia are associated with certain underlying microorganisms, it has to be clearly stated that there is no specific radiological pattern of pneumonia caused by one particular microbe. Overlap of imaging findings also with respect to course over time makes the differentiation of aetiologies based solely on the radiograph unreliable. The spectrum of causative pathogens of pneumonia in humans includes gram-positive bacteria (S. pneumoniae and S. aureus), gramnegative bacteria (H. influenzae, Escherichia coli and Klebsiella pneumoniae), atypical bacteria (Mycoplasma pneumoniae, C. pneumoniae and L. pneumophila), oral anaerobes and viral agents, fungi, protozoa and parasites. Differentiation of aetiologies based solely on the radiograph is not reliable, yet the pattern of abnormalities can be very useful in formulating a differential diagnosis of the nature of the disease. New emerging pathogens have been recognised such as community-acquired methicillinresistant S. aureus, human metapneumovirus (hMPV), avian influenza A viruses (H5N1), coronavirus associated with severe acute respiratory syndrome (SARS), swine flu (H1N1) and Middle East respiratory syndrome coronavirus (MERS-CoV). Chest radiography remains an important component of evaluating a patient with a suspicion of pneumonia, and is usually the first examination to be obtained. Although chest radiographs are of limited value in predicting the causative pathogen, they are of good use to determine the extent of pneumonia and to detect complications (e.g. cavitation, abscess formation, pneumothorax and pleural effusion).

A

Computed Tomography High-resolution computed tomography (HRCT) with thin < 2 mm thick slices, has been shown to be more sensitive than the radiograph in the detection of subtle abnormalities and may show findings suggestive of pneumonia up to 5 days earlier than chest radiographs High-resolution CT is recommended in patients with clinical suspicion of infection and normal or non-specific radiographic findings and in patients with increased risk of pulmonary infections (e.g. neutropenia) (Fig. 5.1). CT is also indicated in patients with pneumonia and persistent or recurrent pulmonary opacities to diagnose or rule out underlying or alternative disease processes. According to the American Thoracic Society (ATS), the presence of radiographically visible opacification is part of the definition of pneumonia, although there may be a time delay of several hours between onset of clinical symptoms and radiographic changes. Specific conditions may further the delay or cause a negative chest radiograph. Regression of pneumonia over time varies with the underlying organism, patient comorbidity and patient age and can take between 1 to 2 weeks or up to 2 months.

B Fig. 5.1  Cellular Bronchiolitis. A 71-year-old man with fever of 48 hours duration. (A) Posteroanterior chest radiograph is normal. (B) Complementary CT shows centrilobular branching nodular and linear opacities resulting in a ‘tree-in-bud’ appearance (arrows). Mycoplasma bronchiolitis was diagnosed.

PATTERNS OF PULMONARY INFECTION Pneumonia is usually divided according to the chest imaging appearance into lobar pneumonia, bronchopneumonia, and interstitial pneumonia. Common associated findings include hilar and mediastinal

106

SECTION A  The Chest and Cardiovascular System

A Fig. 5.3  Bronchopneumonia Caused by Haemophilus influenzae. A 48-year-old man with productive cough and fever. Coronal reformatted CT shows a focal area of consolidation in the right lower lobe with visible air bronchogram and poorly defined margins (arrows). Also evident are small nodular opacities and a few ‘tree-in-bud’ opacities (arrowhead).

B Fig. 5.2  Round Pneumonia. A previously healthy 64-year-old man with fever and productive cough. (A) Chest radiograph shows a mass-like area of consolidation in the left upper lobe (arrow). (B) CT shows a discrete fairly well-marginated opacity in the left upper lobe containing small areas of low attenuation. The abnormality resolved following appropriate antibiotic therapy and Gram stains of sputum demonstrated Streptoccocus pneumoniae.

lymphadenopathy, pleural effusion, cavitation, and chest wall invasion. These findings are not specific and may be seen in other conditions. In up to 10% of patients with proven Pneumocystis pneumonia (PCP), the chest radiograph is normal. In lobar pneumonia, the inflammatory exudate begins in the distal air spaces adjacent to the visceral pleura, and then spreads via collateral air drift routes (pores of Kohn) to produce uniform homogeneous opacification of partial or complete segments of lung and, occasionally, an entire lobe. Occasionally, infection is manifested as a spherical focus of consolidation (Fig. 5.2). An air bronchogram is frequently seen. S. pneumoniae is by far the most common cause of complete lobar consolidation. Other causative agents that produce complete lobar consolidation include K. pneumoniae and other gram-negative bacilli, L. pneumophila, H. influenzae, and, occasionally, M. pneumoniae. Characteristic manifestations on CT are lobar or sublobar consolidations, sharply demarcated by the interlobar fissure.

Bronchopneumonia (lobular pneumonia) is characterised histologically by predominantly peribronchiolar inflammation. Although initially patchy, progression of disease results in lobular and segmental consolidation (Fig. 5.3). The initial peribronchiolar inflammation manifests radiologically as patchy airspace nodules with poorly defined margins. An air bronchogram is usually absent. The most common causative organisms of bronchopneumonia are S. aureus, H. influenzae, P. aeruginosa and anaerobic bacteria. Characteristic manifestations of bronchopneumonia on HRCT include centrilobular ill-defined nodules and branching linear opacities, airspace nodules, and multifocal lobular areas of consolidation. The term atypical pneumonia (interstitial pneumonia) was initially applied to the clinical and radiographic appearance of lung infection not behaving or looking like that caused by S. pneumoniae. In the literature, the term ‘atypical pneumonia’ (as opposed to ‘bacterial pneumonia’) is still in wide usage, although technically incorrect. Many causative organisms are identified as bacteria, albeit unusual types (Mycoplasma is a type of bacteria without a cell wall and Chlamydia are intracellular parasites). It is therefore important to realise that the term ‘atypical pneumonia’—if more correctly based on the underlying type of causative pathogen—does not only refer to an interstitial pattern (Pneumocystis jiroveci pneumonia, certain viral infections but also infections presenting with dense consolidation such as Legionella, Mycoplasma, Chlamydia, etc.). The usual causes of interstitial pneumonia are viral and mycoplasmal infections that radiographically present with focal or diffuse small heterogeneous opacities.

COMPLICATIONS OF PNEUMONIA See Box 5.2. Lung abscess is defined as a localised necrotic cavity containing pus and the most common cause is aspiration. A lung abscess occurs most commonly in the posterior segment of an upper lobe or the superior segment of a lower lobe (Fig. 5.4). Common causes of lung abscess include anaerobic bacteria (most commonly Fusobacterium nucleatum and Bacteroides sp.), S. aureus, P. aeruginosa and K. pneumoniae and

CHAPTER 5  Pulmonary Infection in Adults

107

BOX 5.2  Differential Diagnosis of

Cavitating/Necrotising Community-Acquired Pneumonia in Immunocompetent Patients • Staphylococcus aureus, including methicillin-resistant S. aureus • Anaerobic aspiration syndrome • Klebsiella spp. • Streptococcus milleri • Right-sided endocarditis • Tuberculosis • Nontuberculous mycobacteria

Fig. 5.4  Lung Abscess. A 35-year-old man with high fever and large purulent sputum production with positive culture for Pseudomonas aeruginosa. Coronal reformatted CT shows a large cavity in the left upper lobe. Note intracavitary thick septa.

radiologically manifest with single or multiple masses that are often cavitated. Pulmonary gangrene is a rare complication of pneumonia characterised by the development of fragments of necrotic lung within an abscess cavity (pulmonary sequestrum). Radiological manifestations consist initially of small lucencies within an area of consolidated lung, usually developing within lobar consolidation associated with enlargement of the lobe and outward bulging of the fissure (bulging fissure sign). Pneumatocele is a thin-walled, gas-filled space that usually develops in association with infection. It presumably results from drainage of a focus of necrotic lung parenchyma followed by check-valve obstruction of the airway subtending it, enabling air to enter the parenchymal space during inspiration but preventing its egress during expiration. The complication is caused most often by S. aureus in infants and children and P. jiroveci in patients who have acquired immune deficiency syndrome (AIDS) (Fig. 5.5). Septic emboli to the lungs originate in a variety of sites, including cardiac valves (endocarditis), peripheral veins (thrombophlebitis), and venous catheters or pacemaker wires. On cross-sectional CT images the nodules often appear to have a vessel leading into them (‘feeding vessel’ sign) (Fig. 5.6). Dependent on the underlying organism, nodules cavitate typically at different time points, resulting in the simultaneous appearance of solid nodules and nodules with varying sizes of cavitations.

Fig. 5.5  Pneumocystis Pneumonia and Cysts. A 47-year-old man with AIDS. Close-up view from a CT at the level of the right upper lobe shows a focal consolidation with multiple cysts (pneumatoceles) (arrows). A Pneumocystis jiroveci pneumonia was diagnosed.

Empyema occurs in less than 5% of pulmonary infections. The pathogens traditionally associated with empyema are S. pneumoniae, Streptococcus pyogenes and S. aureus. Radiographically, early signs include obliteration of the costophrenic angle. Complete opacification of a hemithorax and contralateral mediastinal displacement may occur in large effusions. Typically, an infected pleural effusion is encapsulated. Other CT features include (1) pleural enhancement and thickening of the parietal pleura (split pleura sign), (2) increased density of extrathoracic fat and (3) thickening and increased density of the extrapleural subcostal fat. Bronchopleural fistula is a sinus tract between the bronchus and the pleural space that may result from necrotising pneumonias, lung surgery, lung neoplasms and trauma. Imaging features consist of (1) increase in intrapleural air space, (2) appearance of a new air-fluid level, (3) changes in an already present air-fluid level, (4) development of tension pneumothorax and (5) demonstration of actual fistulous communication by CT.

INTEGRATING CLINICAL AND IMAGING FINDINGS The clinician evaluating the patient with a known or suspected diagnosis of pulmonary infection faces a diagnostic challenge. This is because most processes present with similar signs and symptoms, and the radiographic findings of an individual pneumonia do not provide a specific aetiological diagnosis. Furthermore, radiographic manifestations of a given infectious process may be variable, depending on the immunological status of the patient as well as the pre- or coexisting lung disease.

108

SECTION A  The Chest and Cardiovascular System

Fig. 5.6  Septic Embolism. A 40-year-old male intravenous drug user with fever. CT shows multiple cavitated nodules in the left upper lobe. Different vessels (arrows) course into the nodules. Blood cultures were positive for Staphyloccocus aureus.

The most useful imaging techniques available for the evaluation of the patient with known or suspected pulmonary infection are chest radiography and CT.

Lobar Pneumonia

Most Common Organisms Streptococcus pneumoniae.  S. pneumoniae is responsible for approximately one-third of all cases of CAP. Pneumoccocal infections occur predominantly in the winter and early spring and are often associated with prior viral infection. Risk factors for the development of pneumococcal pneumonia include the extremes of age, chronic heart or lung disease, immunosuppression, alcoholism, institutionalisation and prior splenectomy. The characteristic clinical presentation is abrupt in onset, with fever, chills, cough and pleuritic chest pain. In the elderly, these classic features of disease may be absent and pneumonia may be confused with or confounded by other common medical problems, such as congestive heart failure, pulmonary thromboembolism or malignancy. The typical radiographic appearance of acute pneumococcal pneumonia consist of an homogeneous consolidation that crosses segmental boundaries (nonsegmental) but involves only one lobe (lobar pneumonia) (Fig. 5.7). Occasionally, infection is manifested as a spherical focus of consolidation that simulates a mass (round pneumonia). Complications such as cavitation and pneumatocele formation are rare. Pleural effusion is common and is seen in up to half of patients. The CT ‘angiogram sign’, initially described in the lobar form of lepidic adenocarcinomas as the enhancement of branching pulmonary vessels in a homogeneous low-attenuation consolidation of lung parenchyma, may also occur in lobar pneumonia.

Fig. 5.7  Lobar Pneumoccocal Pneumonia. Close-up view from a CT at the level of the right inferior lobe shows a dense consolidation with visible air bronchogram (white arrows) and normal vasculature (CT angiogram sign) (black arrow).

Klebsiella.  K. pneumoniae is among the most common gram-negative bacteria, accounting for 0.5%–5.0% of all cases of pneumonia. The radiographic features include bulging fissures due to volume increase of the infected lobe, sharp margins of the advancing border of the pneumonic infiltrate and early abscess formation (Fig. 5.8). CT findings consist of ground-glass attenuation, consolidation and abscess formation. Legionella sp.  Legionella is one of the most common causes of severe CAP in immunocompetent hosts. Human infection may occur when Legionella contaminates water systems, such as air conditioners and condensers. Risk factors for the development of L. pneumophila pneumonia include immunosuppression, post-transplantation, cigarette smoking, renal disease and exposure to contaminated drinking water. Patients with Legionella pneumonia usually present with fever, cough (initially dry and later productive), malaise, myalgia, confusion, headaches and diarrhoea. Imaging findings include peripheral airspace consolidation similar to that seen in acute S. pneumoniae pneumonia. In many cases, the area of consolidation rapidly progresses to occupy all or a large portion of a lobe (lobar pneumonia) to involve contiguous lobes or to become bilateral (Fig. 5.9). Occasionally, Legionella pneumonia may result in a round area of consolidation simulating a mass (round pneumonia). Pleural effusion may occur in 35%–63% of cases. Chlamydia.  C. pneumoniae (strain TWAR) is the most commonly occurring gram-negative intracellular bacterial pathogen. It is frequently involved in respiratory tract infections and has also been implicated in the pathogenesis of asthma in both adults and children. On CT, C. pneumoniae pneumonia demonstrates a wide spectrum of imaging findings that are similar to those of S. pneumoniae pneumonia and M. pneumoniae pneumonia, consisting of areas of consolidation, bronchovascular bundle thickening, nodules, small pleural effusion, lymphadenopathy, reticular or linear opacities and airway dilatation (Fig. 5.10). Moraxella catarrhalis.  Moraxella catarrhalis (formerly known as Branhamella catarrhalis) is an intracellular gram-negative coccus now

CHAPTER 5  Pulmonary Infection in Adults

Fig. 5.8  Klebsiella Pneumonia. A 50-year-old man with fever and a severe right pneumonia. Posteroanterior chest radiograph shows dense consolidation of the right upper lobe with visible areas of abscessification (arrowhead). Note an inferior convexity of the major fissure (‘bulging fissure’ sign) (arrows) characteristic of lobar expansion.

109

Fig. 5.10  Chlamydia pneumoniae Pneumonia. A 67-year-old woman with chest pain, fever and non-productive cough. Coronal reformatted CT shows multiple ill-defined, rounded areas of consolidation in the left upper lobe with visible air bronchogram and poorly defined margins (arrows).

of lung disease by other causes. Chest radiographs show bronchopneumonia or lobar pneumonia that usually involves a single lobe. Additional CT findings include ground-glass opacities, bronchial wall thickening and centrilobular nodules. Small effusions occur in one-third of patients.

Immunocompromised Host

Fig. 5.9  Legionella Pneumonia. A 51-year-old man with cough and fever. Posteroanterior chest radiograph at emergency area shows dense heterogeneous consolidation in the left lung. A community-acquired Legionella pneumophila pneumonia was diagnosed.

recognised as one of the common respiratory pathogens. M. catarrhalis causes otitis media and sinusitis in children and relatively mild pneumonia and acute exacerbation in older patients with chronic obstructive pulmonary disease (COPD). It is currently considered the third most common cause of community-acquired bacterial pneumonia (after S. pneumoniae and H. influenzae). M. catarrhalis seldom results in pneumonia in previously healthy individuals. Most patients with this type of pneumonia (80% to 90%) have underlying chronic pulmonary disease and their clinical illness may be difficult to distinguish from exacerbations

Nocardia sp.  Nocardia is a genus of filamentous gram-positive, weakly acid-fast, aerobic bacteria that affects both immunosuppressed and immunocompetent patients. Nocardiosis usually begins with a focus of pulmonary infection and may disseminate through haematogenous spread to other organs, most commonly to the central nervous system (CNS). Imaging findings are variable and consist of unifocal or multifocal consolidation and single or multiple pulmonary nodules. Cavitation is common and lymphadenopathy or chest wall involvement may occur. Nocardia asteroides infection may complicate alveolar proteinosis (Fig. 5.11). Actinomyces sp.  Thoracic actinomycosis is a chronic suppurative pulmonary or endobronchial infection caused by Actinomyces sp., most frequently Actinomyces israelii, which is considered to be a gram-positive branching filamentous bacterium. Actinomycosis has the ability to spread across fascial planes to contiguous tissues without regard for normal anatomic barriers. On CT, parenchymal actinomycosis is characterised by airspace consolidation with cavitation, or central areas of low attenuation and adjacent pleural thickening (Fig. 5.12). Endobronchial actinomycosis can be associated with a foreign body (direct aspiration of a foreign body contaminated with Actinomyces organisms) or a broncholith (secondary colonisation of a pre-existing endobronchial broncholith by aspirated Actinomyces organisms).

Endemic in Certain Geographic Areas Coxiella burnetii (Rickettsial Pneumonia). The most common rickettsial lung infection is sporadic or epidemic Q-fever pneumonia caused by Coxiella burnetii, an intracellular, gram-negative bacterium.

110

SECTION A  The Chest and Cardiovascular System

Fig. 5.11  Alveolar Proteinosis and Nocardia Pneumonia. A 42-year-old man with alveolar proteinosis who presented with fever. CT at the level of the lower lobes shows bilateral areas of extensive ground-glass opacities with superimposed smooth septal lines and intralobular lines, resulting in a pattern known as ‘crazy-paving’. Note a localised area of consolidation (arrows) and a right pleural effusion.

Fig. 5.13  Staphylococcus aureus Pneumonia. A 13-year-old boy with fever. CT at the level of the aortic arch shows a necrotising pneumonia in the upper left lobe. A community-acquired methicillin-resistant S. aureus (MRSA) was diagnosed.

Infection is mainly acquired by inhalation from farm livestock or their products, and occasionally from domestic animals. Imaging findings consist of multilobar airspace consolidation, solitary or multiple nodules surrounded by a halo of ‘ground-glass’ opacity and vessel connection, and necrotising pneumonia. Other rickettsial infections such as Rocky Mountain spotted fever are usually tick-borne and occasionally demonstrate diffuse heterogeneous or homogeneous opacities on chest radiographs, perhaps representing vasculitis or cardiogenic pulmonary oedema. Francisella tularensis.  Tularaemia is an acute, febrile, bacterial zoonosis caused by the aerobic gram-negative bacillus Francisella tularensis. It is endemic in parts of Europe, Asia and North America. Primary pneumonic tularaemia occurs in rural settings. Humans become infected after introduction of the bacillus by inhalation, intradermal injection or oral ingestion. Chest radiographic findings are scattered multifocal consolidations, hilar adenopathy and pleural effusion.

Bronchopneumonia

Most Common Organisms

Fig. 5.12  Pleuropulmonary Actinomycosis. A 52-year-old alcoholic man with fever, cough and left chest pain. Contrast-enhanced CT shows consolidation in the left lower lobe containing multiple areas of decreased attenuation with small air bubbles (arrow). Also evident are pleural thickening (arrowhead) and a small pleural effusion.

Affected patients are invariably debilitated by a chronic medical or pulmonary disease. These bacteria are generally aspirated from a colonised upper respiratory tract or may be inhaled or spread haematogenously. The lower lobes predominantly tend to be affected and the radiographic pattern is similar to that seen with S. aureus infections in adults.

Staphylococcus aureus.  Pneumonia caused by S. aureus usually follows aspiration of organisms from the upper respiratory tract. Risk factors for the development of staphylococcal pneumonia include underlying pulmonary disease (e.g. COPD, carcinoma), chronic illnesses (e.g. diabetes mellitus, renal failure) or viral infection. A severe CAP caused by associated methicillin-resistant S. aureus (MRSA) carrying genes for Panton–Valentine leukocidin has been described in immunocompetent young adults. This bronchopneumonia (lobular pneumonia) is bilateral in approximately 40% of patients (Fig. 5.13). Other features are cavitation, pneumatoceles, pleural effusions and spontaneous pneumothorax. Pleural effusions occur in 30% to 50% of patients and abscesses develop in 15% to 30% of patients. The CT manifestations of S. aureus pneumonia include centrilobular nodules and branching opacities (tree-in-bud pattern) and lobular, subsegmental or segmental areas of consolidation with or without abscess formation. Escherichia coli.  E. coli accounts for approximately 4% of cases of CAP and 5%–20% of cases of HAP or HCAP. It occurs most commonly in debilitated patients. The typical history is one of abrupt onset of fever, chills, dyspnoea pleuritic pain and productive cough in a patient with pre-existing chronic disease.

CHAPTER 5  Pulmonary Infection in Adults

111

Fig. 5.14  Escherichia coli Pneumonia. A 54-year-old man with fever. Minimum intensity projection CT demonstrates multiple bilateral peripheral areas of consolidation (arrows).

The radiographic manifestations are usually those of bronchopneumonia; rarely, a pattern of lobar pneumonia may be seen. The pneumonia tends to be severe. Involvement is usually multilobar and predominantly in the lower lobes (Fig. 5.14). Pseudomonas aeruginosa.  P. aeruginosa is a gram-negative bacillus that is the most common cause of nosocomial pulmonary infection. It causes confluent bronchopneumonia that is often extensive and frequently cavitates. The radiological manifestations are non-specific and consist most commonly of patchy areas of consolidation and widespread poorly defined nodular opacities. CT findings consist of multifocal, predominantly upper lobe, airspace consolidation, random large nodules, tree-in-bud opacities, ground-glass opacity, necrosis and pleural effusion (Fig. 5.15). Haemophilus influenzae.  H. influenzae is a pleomorphic, gramnegative coccobacillus that accounts for 5% to 20% of CAP in patients in whom an organism can be identified successfully. Factors that predispose to Haemophilus pneumonia include COPD, malignancy, human immunodeficiency virus (HIV) infection and alcoholism. It is often associated with a previous history of upper respiratory tract infection followed by onset of high fever, cough, dyspnoea, purulent sputum and pleuritic chest pain. The typical radiographic appearance of H. influenzae pneumonia consists of multilobar involvement with lobar or segmental consolidation and pleural effusion (Fig. 5.16). In 30% to 50% of patients, the pattern is that of lobar consolidation similar to that of S. pneumoniae.

Atypical Pneumonia Mycoplasma pneumoniae.  M. pneumoniae is one of the most common causes of CAP. It accounts for up to 37% of CAP in persons treated as outpatients and 10% of pneumonia in persons requiring hospitalisation. It occurs most commonly in younger persons and infection is particularly common among military recruits. Patients with COPD appear to be more severely affected with M. pneumoniae than normal hosts. The radiographic findings in M. pneumoniae are variable and, in some cases, closely resemble those seen in viral infections of the lower respiratory tract. A focal reticulonodular opacification confined to a single lobe is a radiographic pattern that seems to be closely associated with Mycoplasma infection.

Fig. 5.15  Pseudomonas aeruginosa Pneumonia With Abscess Formation. A 45-year-old woman with chest pain and fever. Contrast-enhanced CT shows extensive dense consolidation in the right upper lobe with associated areas of necrosis and abscessification (arrows). Note vascular structures visible within the consolidated lung (arrowheads).

Although lymphadenopathy is uncommon in Mycoplasma pneumonia, unilateral hilar lymph node enlargement has been described. These findings may be indistinguishable from those seen in children with primary tuberculosis. CT findings consist of patchy segmental and lobular areas of ground-glass opacity or airspace consolidation, centrilobular nodules and thickening of the bronchovascular bundles (Fig. 5.17).

Viral The clinical signs and symptoms of viral pneumonia are often nonspecific and the clinical course of infection will be highly dependent on the overall immune status of the host. Acute bronchiolitis is a term most often used to describe an illness in infants and children characterised by acute wheezing with concomitant signs of respiratory viral infection. Viral infections can result in several forms of lower respiratory tract disease including tracheobronchitis, bronchiolitis and, usually, bilateral pneumonia. Viral infections predispose to secondary bacterial pneumonia. Organising pneumonia, a non-specific reparative reaction, may result from a variety of causes or underlying pathological processes including viral infections. Influenza A.  Influenza type A is the most important of the respiratory viruses with respect to the morbidity and mortality in the general population. They are more common during infancy and may often lead to severe lower respiratory tract disease. In adults, infections are usually mild and restricted to the upper respiratory tract. Influenza A virus is transmitted from person to person by aerosolised or respiratory droplets.

112

SECTION A  The Chest and Cardiovascular System

Fig. 5.16  Haemophilus influenzae Pneumonia. A 49-year-old man with fever. Posteroanterior chest radiograph shows bilateral areas of consolidation with ill-defined margins (arrows). A community-acquired H. influenzae pneumonia was diagnosed.

Fig. 5.17  Mycoplasma Pneumonia. A 35-year-old man presents with non-productive cough and fever. CT shows ill-defined rounded groundglass opacities (arrows) and lobular consolidations (arrowheads).

In recent years, both influenza and parainfluenza viruses have been recognised as a significant cause of respiratory illness in immunocompromised patients, including solid organ transplant recipients. The predominant HRCT findings are ground-glass opacities, consolidation, centrilobular nodules and branching linear opacities. Adenovirus.  Adenovirus accounts for 5%–10% of acute respiratory infections in infants and children but for less than 1% of respiratory illnesses in adults. Swyer–James–MacLeod syndrome is considered to be a post-infectious bronchiolitis obliterans (BO) secondary to adenovirus infection in childhood.

Fig. 5.18  Adenovirus Pneumonia. A 46-year-old man with fever. Coronal reformatted CT shows bilateral multiple small branching centrilobular opacities, representing dilated peripheral bronchioles (arrows) associated with bilateral focal areas of consolidation (arrowheads).

CT findings in post-infectious BO consist of sharply marginated focal areas of increased and decreased lung opacity with reduced vessel size in lucent lung regions, bronchial wall thickening and bronchiectasis. Air-trapping is commonly visible on expiratory CT as lucent areas that represent regions of lung that are poorly ventilated and perfused. In children, adenovirus may result in lobar collapse, especially of the right upper lobe. Adenovirus infections in immunocompromised individuals, such as stem-cell and solid-organ transplant recipients, are increasingly recognised as significant causes of morbidity and mortality. In the stem-cell transplantation population, the incidence of disease ranges from 3% to 47%. Adenovirus pneumonia has only been sporadically reported in lung transplantation recipients. A rapidly fatal adenovirus necrotising pneumonia, early in the post-transplantation course, may occur in the paediatric population. The CT findings consist of patchy bilateral areas of consolidation in a lobular or segmental distribution and/or bilateral ground-glass opacities with a random distribution (Fig. 5.18). Respiratory syncytial virus (RSV).  Respiratory syncytial virus (RSV) is the most frequent viral cause of lower respiratory tract infection in infants. The major risk factors for severe RSV disease in children are prematurity (90% of patients show radiographic abnormalities, mainly the classical findings of diffuse bilateral interstitial infiltrates in a perihilar distribution, although normal radiographs do not exclude the diagnosis. In these patients, CT may be helpful in confirming the diagnosis of PCP when clinical suspicion is high, typically showing images with perihilar ground-glass opacities, occasionally combined with focal consolidation, with a patchy or geographical distribution (Fig. 5.34). In the acute phase there may be subpleural sparing and in the subacute phase, cysts may develop during treatment.

Mucormycosis Mucormycosis is an opportunistic fungal infection of the order Mucorales, characterised by broad, nonseptated hyphae that randomly branch at right angles. The most common radiographic findings consist of lobar or multilobar areas of consolidation and solitary or multiple pulmonary nodules and masses; associated cavitation is found in 26% to 40% of cases. An air-crescent sign, highly suggestive of an invasive fungal infection, can be identified in 5% to 12.5% of cases and cannot be differentiated from an angioinvasive Aspergillus infection. CT features are non-specific and consist of solitary or multiple areas of consolidation and solitary or multiples nodules surrounded by a halo of ground-glass attenuation (‘halo sign’) and cavitation. A pattern of multifocal pneumonia in patients with severe clinical disease is associated with a high mortality rate.

Cryptococcosis Cryptococcosis is caused by inhaling spores of Cryptococcus neoformans, a fungus of worldwide distribution found in soil and in bird droppings. Many patients have no symptoms and the pulmonary lesions heal spontaneously. Cryptococcal pneumonia is a common pulmonary infection in AIDS patients with CD4 counts below 100 cells/mm3. The most typical radiographic manifestation consists of pulmonary masses, homogeneous segmental or lobar opacifications, and miliary, reticular or reticulonodular interstitial patterns (Fig. 5.35). The masses, ranging from 5 mm to very large size, are usually ill-defined and may show a halo similar to an invasive Aspergillus lesion, which may eventually cavitate.

Histoplasmosis Histoplasma capsulatum is a fungus found in moist soil and in bird or bat excreta in many parts of the world, but human infection is endemic

Fig. 5.36  Histoplasmoma. A 62-year-old asymptomatic woman living in an area endemic for histoplasmosis. An incidental nodule was found on routine chest radiography. CT shows a rounded opacity in the left upper lobe (arrow). A 3-mm nodule (arrowhead) is also seen in the superior segment of the left lower lobe.

in areas such as Ohio, Mississippi and the St. Lawrence River valleys (North America). Occasional epidemics of symptomatic infection (similar to the flu) are reported in areas where construction is occurring or following exposure from entering bat caves or cleaning out chicken pens; however, most cases are asymptomatic. Though nodules greater than 3 cm may be seen (Fig. 5.36), the most common radiographic findings consist of diffuse nodular opacities of 3 mm or less in diameter, nodules greater than 3 mm in diameter, small linear opacities and focal or patchy areas of consolidation. As these findings are not seen in chest radiography in approximately 40% of patients with pulmonary histoplasmosis, CT is requested as a more sensitive imaging investigation. Hilar and mediastinal lymph nodes are frequently enlarged. Chronic pulmonary histoplasmosis radiologically resembles post-primary tuberculosis, with upper lobe contraction,

120

SECTION A  The Chest and Cardiovascular System

calcification and cavitation. In some cases, fibrosing mediastinitis may develop and can lead to constriction of mediastinal structures, including the airways, superior vena cava, pulmonary arteries and pulmonary veins.

cause miliary nodules, particularly in immunocompromised patients. A chronic fibrocavitary form of the disease is also seen.

Coccidioidomycosis

Parasitic infections of the lung occur worldwide among both immunocompetent and immunocompromised patients. Parasitic diseases account for an increasing presence in industrialised countries because of returning travellers, immigration and mass movements as a result of political or socioeconomic reasons.

Coccidioidomycosis is caused by Coccidioides immitis, a fungus found in soil in arid regions of the southwestern United States, Northern Mexico and in the semi-arid northeastern region of Brazil. Infection is acquired by inhaling dust containing the fungus. Almost half the patients develop a febrile illness; the rest are asymptomatic. In primary coccidioidomycosis, unifocal or multifocal homogeneous opacities resembling community-acquired bacterial pneumonia may be seen. Cavitation and hilar/mediastinal adenopathy may be seen with approximately 20% of these lesions. Primary disease almost invariably resolves spontaneously or reveals only small residual linear or nodular scars. A characteristic CT appearance consisting of a central area of soft-tissue attenuation with a surrounding halo of ground-glass attenuation may be seen around these nodules. Disseminated coccidioidomycosis may cause miliary nodules. Chronic fibronodular cavitary disease may resemble post-primary tuberculosis.

Paracoccidioidomycosis (South American Blastomycosis) Paracoccidioidomycosis (PCM), an endemic disease caused by the dimorphic fungus Paracoccidioides brasiliensis, is the most frequent systemic mycosis in Latin America, especially in Brazil. The predominant HRCT findings consist of areas of ground-glass opacities, nodules, the ‘halo’ and ‘reversed halo’ signs, interlobular septal thickening, airspace consolidation, cavitation and fibrosis (Fig. 5.37).

North American Blastomycosis North American blastomycosis is due to Blastomyces dermatiditis. Pulmonary infection may be accompanied by infection of the skin, bones and genitourinary tract. The chest radiograph reveals homogeneous unifocal or multifocal segmental or lobar opacification indistinguishable from acute pneumonia. Cavitation occurs in approximately 15% of cases. Sometimes, the pneumonia is spherical in shape, closely resembling bronchial carcinoma. Pleural thickening or pleural effusion may accompany the pneumonia in 10%–15% of cases. Blastomycosis may

PARASITIC INFECTIONS

Protozoa

Amoebiasis Pleuropulmonary amoebiasis caused by Entamoeba histolytica is usually secondary to liver involvement. The lung is the second most common extraintestinal site of amoebic involvement after the liver. Pleuropulmonary amoebiasis is a significant complication of amoebic liver abscess. Right-sided abnormalities are found in 86% of cases and consists of hemidiaphragmatic elevation, pleural effusion or empyema and/or thickening and plate-like atelectasis. Liver abscess can extend directly into the lung. Causing pulmonary consolidation. If communication with a major bronchus occurs, haemoptysis can develop, containing the ‘anchovy paste’ pus coming from the amoebic abscess (Fig. 5.38).

Malaria Malaria is transmitted by the bite of Anopheles mosquito. The microorganisms Plasmodium vivax, P. falciparum, P. malariae and P. ovale are responsible for the disease, and P. falciparum is the deadliest type of infection. Adult respiratory distress syndrome is the most common lung-finding manifestation. Septal thickenings, pleural effusions and airspace consolidations are seen on HRCT and are consistent with noncardiogenic pulmonary oedema. Cryptogenic organising pneumonia has also been reported (Fig. 5.39).

Nematodes

Dirofilariasis Although not common in humans, dirofilariasis is occurring with increasing frequency as the canine population grows. Dirofilaria immitis, or the dog heartworm, is a rare cause of pulmonary nodules in humans. The majority of patients with dirofilariasis are asymptomatic. Ascaris immitis is transmitted by mosquitoes from dogs to humans. An immature adult worm unable to mature in the accidental human host can reach a peripheral vein and travel in the bloodstream until it lodges in a pulmonary vein. The disease has been reported predominantly in the temperate climates of the east and south coasts of the United States, but sporadic cases have been found worldwide (Fig. 5.40).

Ascariasis Ascariasis is one of the most common parasitic infections, affecting 1.3 billion people worldwide. The disease is caused by ingestion of food or fluids contaminated with faeces with A. lumbricoides. The main signs and symptoms are those of Loeffler syndrome, characterised by cough, fever, expectoration and eosinophilia. Chest radiography and CT may show patchy acinar opacities, usually bilateral and migratory.

Strongyloidiasis Fig. 5.37  Paracoccidioidomycosis. A 61-year-old patient with cough and fever. CT at the level of the lung bases shows a combination of subsegmental areas of consolidation and randomly distributed bilateral, ill-defined nodules of variable size. (Courtesy Dr. Edson Marchiori, Rio de Janeiro, Brazil.)

Strongyloidiasis is a chronic parasitic infection caused by Strongyloides stercoralis. S. stercoralis filariform larvae invade the lungs and small intestine through the skin from the soil. Continuous autoinfection can lead to a massive parasitic infestation (hyperinfection syndrome), especially in immunosuppressed patients. The main imaging findings include ill-defined, patchy, migratory airspace consolidation (Fig. 5.41).

CHAPTER 5  Pulmonary Infection in Adults

A

121

Fig. 5.39  Malaria. A 38-year-old man with fever, headaches and a sore throat. Chest CT shows diffuse septal thickening (arrows), sparse consolidation (arrowhead) and right pleural effusion (asterisk). (Courtesy Dr. Edson Marchiori, Petrópolis, Brazil.)

and ventricular aneurysms. Late gastrointestinal compromise is caused by damage to neurones of the myenteric plexus, with achalasia, mega­ oesophagus and megacolon. Oesophageal manifestations are similar to those of idiopathic achalasia. Radiographically, oesophageal dilatation shows a shadow projecting to the right of the mediastinum, with or without the presence of an air-fluid level (Fig. 5.42).

Cysticercosis Cysticercosis is a common parasitic disease in Latin America caused by infection with the larval stage of the pork tapeworm Taenia solium. Disseminated cysticercosis mainly involves the CNS and, occasionally, heart, lung, striated muscles and subcutaneous tissue (Fig. 5.43). Subcutaneous cysticercosis presents as small, moveable, painless nodules, usually in the arms or chest. CT may depict cystic lesions, commonly with a hyperdense central nodule, which represents the parasite head, called the scolex. Pulmonary cysticercosis mimics many other diseases presenting with nodules, cavitary lesions and pleural effusion. If association of chest wall and cardiac muscles lesions is seen, cysticercosis should be the first diagnosis to be considered.

Toxocariasis

B Fig. 5.38  Pulmonary Amoebiasis. A 47-year-old man with cough and fever. (A) Posteroanterior chest radiography shows a right upper smoothbordered paramediastinal lesion. (B) Close-up view from an IV-enhanced CT shows pulmonary consolidation (arrows) containing multiple small air bubbles (arrowhead) within small abscesses.

Trypanosomiasis Caused by Trypanosoma cruzi, which is acquired through the bite of a triatomine insect, trypanosomiasis is also known as Chagas disease, being endemic in South America. The acute phase is usually asymptomatic but can present with febrile illness with facial or palpebral oedema and acute myocarditis. A nodular lesion or furuncle, usually called chagoma, can appear at the site of inoculation. Chronic manifestations include cardiomyopathy, bundle branch blocks, complete atrioventricular block

Toxocariasis is distributed worldwide. Humans can be accidental hosts or may develop the disease, which is caused by the larvae of Toxocara canis or T. cati. In humans, the larvae do not develop into adult worms but migrate through host tissues. Therefore, the disease is also called visceral larva migrans. The most common clinical features of the disease in humans are peripheral eosinophilia, abdominal pain, hepatosplenomegaly, fever and hypergammaglobulinaemia. CT findings consist of ground-glass opacities, solid nodules, areas of consolidation and linear opacities (Fig. 5.44).

Cestodes

Echinococcosis (Hydatid Disease) Hydatid disease (echinococcosis) is caused by the larval forms of Echinococcus granulosus, E. multilocularis and E. vogeli. E. granulosus (unilocular cystic echinococcosis) is the most common form affecting man and is seen in the Mediterranean area, Eastern Europe, Africa, South America, the Middle East, Australia and New Zealand. Humans are accidental hosts and acquire infection by ingesting ova from fomites or contaminated water and by direct contact with dogs.

122

SECTION A  The Chest and Cardiovascular System

Fig. 5.41  Strongyloidiasis. A 42-year-old male patient with AIDS and a massive infestation (hyperinfection syndrome) of Strongyloides stercoralis. Chest CT at the level of the carina shows multiple patchy foci of consolidation in the right lung (arrows) and ground-glass opacities in the left upper lung (arrowheads). (Courtesy Dr. Edson Marchiori, Petrópolis, Brazil.)

A

B Fig. 5.40  Dirofilariasis. A 20-year-old man with recent close interactions with a dog, presenting with fever, chest pain and cough. (A) Close-up view from PA chest radiograph shows ill-defined opacities in the RLL. (B) Chest CT shows multiple scattered bilateral pulmonary nodules (arrowheads), some of them with a ground-glass halo (arrowheads).

Fig. 5.42  Trypanosomiasis. A 36-year-old man with chronic Chagas disease presenting with chest pain and regurgitation. Close-up view from an enhanced CT shows a dilated oesophagus (arrows) with intraluminal content with air bubbles (small arrows) and a localised liquid/contrast level (arrowhead). Also note a right pleural effusion (asterisk).

Hydatid cyst has been reported in almost all human tissues and organs. Cysts in the mediastinum, heart and pulmonary arteries are rare. The clinical manifestations of pulmonary hydatidosis are non-specific. Complications occur because of cyst rupture. Hydatid cysts are usually solitary but may be multiple and/or bilateral in 10% of cases. They may be ruptured (two-thirds) or unruptured (one-third) at the time of presentation. Aggressive invasion of vascular structures such as bronchial and pulmonary arteries may result in massive haemoptysis and haemorrhage.

The radiological findings in patients with unruptured pulmonary cysts are one or more homogeneous, roughly spherical or oval, sharply demarcated lesions with mass effect. Cyst rupture is usually associated with secondary infection and may spread into the airways or pleural space. The radiographic appearance resembles the air crescent of a mycetoma. Should there be disruption of the inner layers, a complex cavitary lesion results with one or more of the following radiographic features: an air-fluid level, a floating membrane (water lily sign, camalote sign)

CHAPTER 5  Pulmonary Infection in Adults

123

Fig. 5.43  Disseminated Cysticercosis. A 49-year-old asymptomatic male rural worker. Close-up view of a posteroanterior chest radiograph shows large numbers of calcified cysticerci in chest wall muscles (arrows).

Fig. 5.45  Ruptured Hydatid Cyst. A 65-year-old male shepherd with abrupt onset of expectoration and pruritus. Close-up view of the right upper lung shows a cystic lesion surrounded by a parenchymal consolidation due to a massive aspiration of intracystic content. Note a rounded opacity immediately above the fluid level (‘water lily’ sign) (arrows).

(Fig. 5.45), a double wall, an essentially dry cyst with crumpled membranes lying at its bottom (rising sun sign, serpent sign) and a cyst with all its contents expectorated (empty cyst sign). Secondary infection of a hydatid cyst may produce a lung abscess with or without surrounding lung opacity. Rupture into the pleural space causes an effusion or, if there is airway communication, a hydropneumothorax.

Trematodes

Paragonimiasis A

B Fig. 5.44  Toxocariasis. A 14-year-old male patient with cough, fever, tachypnoea and weight loss in the last 10 days. (A) Posteroanterior chest radiograph and (B) chest CT show a diffuse micronodular pattern. (Courtesy Dr. Dante Escuissato, Curitiba, Brazil.)

Pleuropulmonary paragonimiasis is a disease caused by a fluke (Paragonimus westermani) characterised by migration of a juvenile worm in the early stage and by formation of cysts around the worm later on. Water snails and crustaceans are intermediate hosts and infestations are acquired from eating raw or incompletely cooked fresh water crabs and crayfish. The disease mainly occurs in the Far East, southeast Asia and Africa. Radiological changes tend to be bilateral, including a mixture of consolidation, nodules and band, tubular and ring opacities. In the lower lobes, parenchymal changes mimic bronchiectasis, and in the upper lobes, tuberculosis. The constellation of focal pleural thickening and subpleural linear opacities leading to a necrotic peripheral pulmonary nodule is another frequent CT finding of paragonimiasis.

Schistosomiasis Schistosomiasis is an acute and chronic parasitic disease caused by a trematode worm of the genus Schistosoma. Chronic granulomatous inflammation of pulmonary arterioles can result in arteriolitis obliterans, pulmonary hypertension and cor pulmonale Formation of a pulmonary artery aneurysm is a common complication related to this disease

124

SECTION A  The Chest and Cardiovascular System

B

A

C Fig. 5.46  Schistosomiasis. A 41-year-old man with schistosomiasis and pulmonary hypertension. (A) Anteroposterior chest radiograph shows a significant enlargement of the main pulmonary artery (arrows). Descending aorta is also seen (arrowhead). (B) Close-up view of a contrast-enhanced CT shows a huge dilatation of the main pulmonary artery (arrows). Note an eccentric filling defect along the posterior margin of the descending aorta with a peripheral calcification (arrowhead). (C) A 3D external volume rendering is also useful to show the pulmonary artery dilatation (arrows).

(Fig. 5.46). Schistosomiasis should be suspected in native populations and travellers with pulmonary arterial hypertension coming from endemic regions.

FURTHER READING Althoff Souza, C., Muller, N.L., Marchiori, E., et al., 2006. Pulmonary invasive aspergillosis and candidiasis in immunocompromised patients: a

comparative study of the high-resolution CT findings. J. Thorac. Imaging 21 (3), 184–189. Balkhy, H.H., Alenazi, T.H., Alshamrani, M.M., et al., 2016. Description of a hospital outbreak of Middle East respiratory syndrome in a large tertiary care hospital in Saudi Arabia. Infect. Control Hosp. Epidemiol. 37 (10), 1147–1155. Chou, S.H., Prabhu, S.J., Crothers, K., et al., 2014. Thoracic diseases associated with HIV infection in the era of antiretroviral therapy: clinical and imaging findings. Radiographics 34 (4), 895–911.

CHAPTER 5  Pulmonary Infection in Adults Das, K.M., Lee, E.Y., Langer, R.D., et al., 2016. Middle East respiratory syndrome coronavirus: what does a radiologist need to know? AJR Am. J. Roentgenol. 206 (6), 1193–1201. Franquet, T., 2001. Imaging of pneumonia: trends and algorithms. Eur. Respir. J. 18 (1), 196–208. Franquet, T., 2011. Imaging of pulmonary viral pneumonia. Radiology 260 (1), 18–39. Fraser, R.S., Colman, N., Müller, N.L., et al., 2005. Synopsis of Diseases of the Chest. Elsevier Saunders, Philadelphia. Gaur, P., Dunne, R., Colson, Y.L., et al., 2014. Bronchopleural fistula and the role of contemporary imaging. J. Thorac. Cardiovasc. Surg. 148 (1), 341–347. Gavilanes, F., Piloto, B., Fernandes, C.J.C., 2018. Giant pulmonary artery aneurysm in a patient with schistosomiasis-associated pulmonary arterial hypertension. J. Bras. Pneumol. 44 (2), 167. Hoette, S., Figueiredo, C., Dias, B., et al., 2015. Pulmonary artery enlargement in schistosomiasis associated pulmonary arterial hypertension. BMC Pulm. Med. 15, 118.

125

Ketai, L., Jordan, K., Busby, K.H., 2015. Imaging infection. Clin. Chest Med. 36 (2), 197–217, viii. Khanna, N., Steffen, I., Studt, J.D., et al., 2009. Outcome of influenza infections in outpatients after allogeneic hematopoietic stem cell transplantation. Transpl. Infect. Dis. 11 (2), 100–105. Müller, N.L., Fraser, R.S., Colman, N., et al., 2001. Radiologic Diagnosis of Diseases of the Chest. W.B. Saunders, Philadelphia. Musher, D.M., Thorner, A.R., 2015. Community-acquired pneumonia. N. Engl. J. Med. 372 (3), 294. Nambu, A., Ozawa, K., Kobayashi, N., et al., 2014. Imaging of community-acquired pneumonia: roles of imaging examinations, imaging diagnosis of specific pathogens and discrimination from noninfectious diseases. World J. Radiol. 6 (10), 779–793. Nelson, A.M., Manabe, Y.C., Lucas, S.B., 2017. Immune Reconstitution Inflammatory Syndrome (IRIS): What pathologists should know. Semin. Diagn. Pathol. 34 (4), 340–351. www.who.int/csr/don/2009_09_04/en/index. WHOW, html. Global alert and response: pandemic (H1N1) 2009: update 64. 2009.

6  Large Airway Disease and Chronic Airflow Obstruction Philippe A. Grenier, Catherine Beigelman-Aubry, Pierre-Yves Brillet, Catalin I. Fetita

CHAPTER OUTLINE Introduction, 126 Tracheal Disorders, 126 Bronchiectasis, 136 Broncholithiasis, 144

INTRODUCTION The purpose of this chapter is to review lesions involving the trachea and proximal bronchi, to describe the radiological signs of bronchiectasis and to discuss the role of imaging in obstructive lung disease, a group of diffuse lung diseases associated with chronic airflow obstruction that includes chronic obstructive pulmonary disease (COPD), asthma, and obliterative bronchiolitis. In obstructive lung disease, decreased expiratory flow may be related to loss of lung recoil or small airway obstruction or a combination of both. The most common abnormality correlated with loss of recoil is emphysema. The process that causes the small airway obstruction is inflammatory in nature and is characterised by thickening of all the layers of the bronchiolar walls as well as an accumulation of mucus in the airway lumen (COPD and asthma), and/or an irreversible fibrosis (COPD and obliterative bronchiolitis).

TRACHEAL DISORDERS The trachea may be affected by a variety of extrinsic or intrinsic processes. Extrinsic processes, particularly masses, displace and distort the trachea, while intrinsic ones cause narrowing, widening, or a mass effect. Tracheal narrowing may affect a short or a long segment and may extend to the mainstem bronchi. Tracheal disease, though commonly missed on the chest radiography, is usually evident on careful evaluation of the frontal and lateral radiograph. Computed tomography (CT) allows precise delineation of the intratracheal and extratracheal extent of the abnormality. Multidetector CT, by combining helical volumetric CT acquisition and thin collimation during a single breath hold, provides an accurate assessment of proximal airways, allowing multiplanar reformations and 3D rendering of very high quality. Complementary CT acquisition at suspended or continuous expiration demonstrates tracheal collapsibility. This expiratory acquisition may be performed at reduced dose or even better at ultra-low dose, by using thin slices reconstructed with iterative reconstruction or a soft kernel when using filtered back projection mode (Box 6.1).

Post-Traumatic Strictures Strictures of the trachea are usually secondary to damage from a cuffed endotracheal or tracheostomy tube or to external neck trauma. The

126

Obliterative (Constrictive) Bronchiolitis, 144 Chronic Obstructive Pulmonary Disease, 147 Asthma, 159

lesions consist of granulation tissue followed by the development of dense mucosal and submucosal fibrosis associated with the distortion of cartilage plates. The two principal sites of stenosis following intubation or tracheostomy tube are at the stoma or at the level of the endotracheal or tracheostomy tube balloon. On radiographs, the stenosis may be seen as a focus of circumferential or eccentric narrowing associated with a segment of increased soft tissue. The size of the narrowing is usually easily seen at CT. The narrowing is often concentric. Post-intubation stenosis can extend for several centimetres and typically involves trachea above the level of the thoracic inlet. Post-tracheostomy stenosis typically begins 1 to 1.5 cm distal to the inferior margin of the tracheostomy stoma and involves 1.5 to 2.5 cm of tracheal wall. Multiplanar reformations are particularly helpful in defining accurately the site, the length and the degree of the stenosis (Fig. 6.1). In selected cases, the degree of stenosis may also be shown by virtual bronchoscopy.

Infectious Tracheobronchitis A number of infections, both acute and more often chronic, may affect the trachea and proximal bronchi, resulting in both focal and diffuse airway disease. Subsequent fibrosis may result in localised airway narrowing. The most common causes of infectious tracheobronchitis are bacterial tracheitis in immunocompromised patients (Fig. 6.2), tuberculosis, rhinoscleroma (Klebsiella rhinoscleromatis), and necrotising invasive aspergillosis. On CT, the extent of irregular and sometimes circumferential tracheobronchial narrowing is clearly demonstrated, and in some patients an accompanying mediastinitis (opacification of the mediastinal fat) is evident. In active disease, the narrowed trachea and frequently one or more main bronchus have an irregularly thickened wall. In the fibrotic or healed phase, the trachea is narrowed but has a wall that is smooth and of normal thickness.

Primary Malignant Neoplasms These are uncommon tumours, accounting for less than 1% of all thoracic malignancies. The vast majority are squamous cell carcinoma and adenoid cystic carcinoma (Fig. 6.3). Other neoplasms, such as mucoepidermoid carcinoma, carcinoid tumour (Fig. 6.4), lymphoma, plasmocytoma and adenocarcinoma are rare. On CT, they appear as a soft-tissue mass, usually in the posterior and lateral wall (see Fig. 6.3A). Often sessile and

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

A

B

C

Fig. 6.1  Post Intubation Tracheal Stenosis in a Severe Chronic Obstructive Pulmonary Disease Patient. (A) Axial computed tomography (lung window). (B) Coronal oblique multiplanar reformation (MPR) image (mediastinal window) along the long axis of the trachea. (C) Coronal oblique MPR image (lung window). Continued

127

E

D

A

Fig. 6.1, cont’d (D) Coronal oblique average image (21-mm-thick slab). Note the visibility of the ring cartilages of the trachea. (E) Endoscopic view. There is a circumferential luminal narrowing of the trachea extending along 2 cm associated with soft-tissue thickening which produces the characteristic ‘hourglass’ configuration, well assessed on coronal views (C and D). Note the roughly triangular shape on axial views (A and E) and the slightly irregular and nodular aspect on 3D image (E).

B

Fig. 6.2  Infectious Tracheobronchitis. Bacterial tracheitis in a severely immunocompromised patient suffering from a rheumatoid arthritis with vasculitis. She presented with dyspnoea and cough as she was in agranulocytosis secondary to cyclophosphamide treatment. A severe stenosis of the distal trachea (orange arrows) and proximal main bronchi predominant on the left side associated with a fistulous tract (blue arrow) connecting with a paratracheal submucosal abscess was shown during bronchoscopy. This was related to Pseudomonas aeruginosa, Escherichia coli and Streptococcus infection. (A) Axial computed tomography (CT; mediastinal window) at the level of the distal part of the trachea showing the irregular thickening with a lucency on the left side (blue arrow) related to the fistulous tract. (B) Axial CT (lung window) at the same level.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

129

D

C

Fig. 6.2, cont’d (C) 3D reconstruction of the tracheobronchial tree perfectly demonstrating the whole stenosis and the fistula. (D and E) Axial CT images at the level of the mainstem bronchi showing a significant decrease of the bronchial thickening after 2 weeks of antibiotic treatment (D: before; E: after treatment)

BOX 6.1  Multiplanar Reformations to

Assess Tracheal Disorders

In tracheal disorders, multiplanar reformations after MDCT acquisition at full inspiration help assess: • the site, length and degree of luminal narrowing, • the site, degree and extent of wall thickening, • the regular, irregular or nodular appearance of the inner surface of tracheal lumen.

eccentric, resulting in asymmetric luminal narrowing, they may appear rarely circumferential. They can be polypoid and mostly intraluminal with mediastinal extension representing 30–40%. The surface of tumour is often irregular in squamous cell carcinoma, whereas it is smooth in adenoid cystic carcinoma. Multiplanar reformation and volumetric

E

rendering images are recommended for a precise pre-therapeutic assessment of tumour extent (see Figs 6.3B, 6.4B and D). These tumours are best treated surgically especially with primary resection and reanastomosis followed by radiation.

Secondary Malignant Neoplasms The large airways may be involved secondarily by malignant neoplasms as a result of either haematogenous metastasis or direct invasion from the oesophagus, thyroid, mediastinum or lung. Neoplasms that have a propensity to metastasise to the trachea and major bronchi include renal cell carcinoma and melanoma. On CT, the abnormalities are usually focal and include intraluminal soft-tissue nodules and wall thickening (Fig. 6.5). Distal bronchoceles may also be seen.

Benign Neoplasms The most common are hamartoma, leiomyoma, neurogenic tumour and lipoma. They are usually well-demarcated, round and less than

130

SECTION A  The Chest and Cardiovascular System

A

Fig. 6.3  Adenoid Cystic Carcinoma of the Trachea. (A) Axial compute tomography (CT) at the level of the supra-aortic part of the mediastinum. Soft-tissue mass arising from the posterior wall of the trachea and bulging into the lumen of the trachea. (B) Sagittal reformation showing the smooth appearance of the surface of the tumour, and the posterior extent of the extraluminal tumour growth.

A

B

B Fig. 6.4  Atypical Carcinoid Tumour of the Intermediate Trunk. Atypical carcinoid tumour revealed by recent recurrent haemoptysis. (A) Axial computed tomography (CT; lung window) showing the upper portion of the endobronchial lesion with a rounded shape. (B) Axial CT (mediastinal window) showing strong enhancement after contrast administration.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

C

131

D Fig. 6.4, cont’d (C) Sagittal oblique reformation (mediastinal window) demonstrating the filled bronchiectasis distal to the tumour. (D) Coronal oblique reformation (lung window) showing the upper limit of the tumour obstructing the intermediate trunk with peripheral atelectasis.

2 cm in diameter. The radiologic appearance typically consists of a smoothly marginated intraluminal polyp. Hamartomas and lipomas may demonstrate fat attenuation on CT. Tracheobronchial papillomatosis is a particular entity caused by human papillomavirus infection usually acquired at birth from an infected mother. The larynx is most commonly affected, with occasional extension into the trachea and proximal. Exceptionally, the infection spreads into the lung parenchyma. The typical radiological findings consist of multiple small nodules projecting into the airway lumen or diffuse nodular thickening of the airway wall. Although benign, papilloma may undergo transformation to squamous cell carcinoma.

Wegener Granulomatosis Involvement of the large airways is a common manifestation of Wegener granulomatosis. Inflammatory lesions may be present with or without subglottic or bronchial stenosis, ulcerations and pseudotumours. Radiological manifestations include thickening of the subglottic region and proximal trachea with a smooth symmetric or asymmetric narrowing over variable length. Stenosis may also be seen in any main lobar or segmental bronchus. Nodular or polypoid lesions may also be seen on the inner contour of the airway lumen.

Relapsing Polychondritis Relapsing polychondritis is a rare systemic disease of autoimmune pathogenesis that affects cartilage at various sites, including the ears, nose, joints, and tracheobronchial tree. Histologically, the acute inflammatory infiltrate present in the cartilages and perichondrial tissue induces progressive dissolution and fragmentation of the cartilage followed by fibrosis. Symmetric subglottic stenosis is the most frequent manifestation in the chest. As the disease progresses, the distal trachea and bronchi may be involved. CT shows smooth thickening of the airway wall associated with more or less diffuse narrowing (Fig. 6.6). In the early stage, the posterior wall of the trachea is spared but in advanced disease circumferential wall thickening occurs (Fig. 6.7). The trachea may become flaccid with considerable collapse at expiration. Gross destruction of the cartilaginous rings with fibrosis may cause stenosis.

Tracheobronchial Amyloidosis Deposition of amyloid in the trachea and bronchi may be seen in association with systemic amyloidosis or as an isolated manifestation. As a result, the amyloid forms are either multifocal or diffuse submucosal plaques or masses. The overlying mucosa is usually intact. Dystrophic

132

SECTION A  The Chest and Cardiovascular System

A

B Fig. 6.5  Endobronchial Metastasis. Patient suffering from lung and liver metastasis of a colon carcinoma. (A) Axial CT (lung window) showing a peribronchial metastasis. (B) 5 months later: Oblique reformation along the axis of the upper segmental bronchus of the left lower lobe. The enlarged and filled bronchus reflects the growth of the metastasis.

A

B

Fig. 6.6  Relapsing polychondritis (A and B). Axial computed tomography images at the levels of the distal part of the trachea and mainstem bronchi. Abnormal thickening of the anterior and lateral walls of the trachea and mainstem bronchi and right upper lobar bronchus associated with calcium deposits. The posterior membranous wall of the trachea is unaffected.

calcification or ossification is frequently present. CT shows focal or, more commonly, diffuse thickening of the airway wall and narrowing of the lumen. Calcification may be seen. Narrowing of the proximal bronchi can lead to distal atelectasis, bronchiectasis, or both, or to obstructive pneumonia.

Sarcoidosis Involvement of the trachea is rare and when it occurs, it is associated with laryngeal involvement. The proximal and distal parts of the trachea may be affected, and the appearance of the stenosis may be smooth, irregular and nodular, or even mass-like. Bronchial involvement is much more

common as a manifestation of sarcoidosis. The most common signs at CT are regular or nodular bronchial wall thickening, reflecting the presence of granulomas and fibrous tissue in the peribronchial interstitium. This bronchial wall thickening may result in smooth or irregular bronchial narrowing, which correlates with the presence of mucosal thickening at bronchoscopy and presumably reflects prominent inflammation in this location. Obstruction of lobar or segmental bronchi may occur as a result of granulomata, airway wall fibrosis, peribronchial lymph node compression and conglomerate fibrosis or some combination of these phenomena. Bronchial stenosis may clear spontaneously or with steroid treatment.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

133

of the trachea and proximal airways. Men are more frequently involved than women and most patients are more than 50 years of age. Histologically, the nodules contain heterotopic bone, cartilage and calcified acellular protein matrix. The overlying bronchial mucosa is normal and because it contains no cartilage, the posterior wall of the trachea is spared. The chest radiograph may be normal or may demonstrate lobar collapse or infective consolidation. If the tracheal air column is clearly seen, multiple sessile nodules that project into the tracheal lumen extending over a long segment of the trachea can be appreciated. On CT, tracheal cartilages are thickened and show irregular calcifications. The nodules may protrude from the anterior and lateral walls into the lumen; they usually show foci of calcification (Fig. 6.9).

Sabre-Sheath Trachea A

Characterised by a diffuse narrowing involving the intrathoracic trachea, this entity is almost always associated with COPD (Fig. 6.10). The pathogenesis of the lesion is obscure, but probably it is an acquired deformity related to the abnormal pattern and magnitude of intrathoracic pressure changes in COPD. On radiographs and CT, the condition is easily recognised by noting that the internal side-to-side diameter of the trachea is decreased to half or less than the corresponding sagittal diameter. On the postero-anterior radiograph and CT multiplanar reformations, the narrowing usually affects the whole intrathoracic trachea, with an abrupt return to normal calibre at the thoracic inlet (see Fig. 6.10). The trachea usually shows a smooth inner margin but occasionally has a nodular contour. Calcification of the tracheal cartilage is frequently evident.

Tracheobronchomegaly (Mounier-Kuhn Disease)

B Fig. 6.7  Late Stage Relapsing Polychondritis. (A) Axial computed tomography at the level of aortic arch. Thickening of the anterior and lateral walls associated with narrowing of the tracheal lumen which presents a circular shape. (B) Coronal oblique reformation with minimum intensity projection: thickening of the tracheolateral walls with tracheal luminal narrowing extending from the cervical part of the trachea to the carina.

This refers to patients who have marked dilatation of the trachea and mainstem bronchi. It is often associated with tracheal diverticulosis, recurrent lower respiratory tract infection and bronchiectasis. Atrophy affects the elastic and muscular elements of both the cartilaginous and membranous parts of the trachea. The diagnosis is based on radiological findings. The immediately subglottic trachea has a normal diameter, but it expands as it passes to the carina and this dilatation often continues into the major bronchi. Atrophic mucosa prolapses between cartilage rings and gives the trachea a characteristically corrugated outline on a plain radiograph. Corrugations may become exaggerated to form sacculations or diverticula. On CT, a tracheal diameter of greater than 3 cm (measured 2 cm above the aortic arch) and diameter of 2.4 and 2.3 cm for the right and left bronchi respectively are diagnosing criteria (Fig. 6.11). Additional findings include tracheal scalloping or diverticula (especially along the posterior membranous tracheal wall).

Tracheobronchomalacia Inflammatory Bowel Disease Ulcerative tracheitis and tracheobronchitis are rare complications and occur more often in association with ulcerative colitis than Crohn disease. In most but not all cases, the diagnosis of inflammatory bowel disease precedes the presence of airway disease. Histologically, tracheobronchitis is characterised by chronic inflammation and more or less concentric mucosal and submucosal fibrosis. Ulceration and luminal narrowing may be evident. Cartilaginous plates are not destroyed. On CT, the tracheobronchial walls are thickened and produce irregular luminal narrowing (Fig. 6.8). Bronchial wall thickening and bronchiectasis may also be present with or without mucoïd impaction.

Tracheobronchopathia Osteochondroplastica This rare disorder is characterised by the presence of multiple cartilaginous nodules and bony submucosal nodules on the inner surface

Resulting from weakened tracheal cartilages, this abnormality may be seen in association with a number of disorders including tracheobronchomegaly, COPD, diffuse tracheal inflammation such as relapsing polychondritis, as well as following trauma. On chest radiographs, a reduction by almost 300% of the sagittal diameter at expiration is an excellent indicator of the diagnosis. At CT, the diagnosis is based on a narrowing of the lumen of diameter by more than 70% on expiration compared with that on inspiration. The increase in compliance is due to the loss of integrity of the wall’s structural components and is particularly associated with damaged or destroyed cartilage. The coronal diameter of the trachea becomes significantly larger than the sagittal one, producing a lunate configuration to the trachea. The flaccidity of the trachea or bronchi is usually most apparent during coughing or forced expiration. In patients with COPD with high downstream resistance, particularly high dynamic pressure gradients can be generated across the tracheal wall and it is likely that calibre changes of more

134

SECTION A  The Chest and Cardiovascular System

Fig. 6.8  Tracheal Involvement in Crohn Disease. Axial computed tomography images at the levels of subglottic and upper thoracic parts of the trachea. Circumferential thickening of the trachea walls associated with irregularities of the inner surface of the posterolateral trachea wall, and slight deformity of the tracheal lumen. Note the right aberrant retro-oesophageal subclavian artery.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

135

Fig. 6.9  Tracheopathy Osteochondroplastica. Axial computed tomography at the level of the upper part of the intrathoracic trachea. Calcified or partly calcified nodules arising from the inner surface of the trachea which protrude into the lumen.

A

B

C

Fig. 6.10  Saber-Sheath Trachea in a Chronic Obstructive Pulmonary Disease Patient. (A) Axial computed tomography at the level of the upper lobes shows a significant reduction of the coronal diameter of the trachea. Bilateral centrilobular and paraseptal emphysematous areas are also present in the upper lobes. (B) Coronal oblique reformation along the long axis of the trachea. Reduction of the coronal diameter of the trachea lumen (arrows). Note the upper part of the trachea above the thoracic inlet has a normal appearance. (C) Endoscopic view.

than 50% can occur at expiration with normal tracheal compliance. As a result, only a decrease in cross-section area of the tracheal lumen greater than 70% at expiration indicates tracheomalacia. Dynamic expiratory multislice CT may offer a feasible alternative to bronchoscopy in patients with suspected tracheobronchomalacia. Dynamic expiratory CT may show complete collapse or collapse of greater than 80% of airway lumen (Fig. 6.12). Involvement of the central tracheobronchial tree may be diffuse or focal. The reduction of the airway may have an

oval or crescentic shape. The crescentic form is due to the bowing of the posterior membranous trachea (Box 6.2).

Tracheobronchial Fistula and Dehiscence Multidetector computed tomography (MDCT) with thin collimation is the most accurate technique to identify peripheral bronchopleural fistula that are most commonly caused by necrotising pneumonia (Fig. 6.13) or secondary to traumatic lesions. Minimum intensity projection

136

SECTION A  The Chest and Cardiovascular System

TABLE 6.1  Mechanisms and Causes of

Bronchiectasis

Bronchial Obstruction • Carcinoma • Fibrous stricture (e.g. tuberculosis) • Broncholithiasis • Extrinsic compression (lymphadenopathy, neoplasm) Parenchymal Fibrosis (Traction Bronchiectasis) • Tuberculosis • Sarcoidosis • Idiopathic pulmonary fibrosis

A

Bronchial Wall Injury • Cystic fibrosis • Childhood viral and bacterial infection • Immunodeficiency disorders • Dyskinetic cilia syndrome • Allergic bronchopulmonary aspergillosis • Graft vs. host disease in lung and bone marrow transplantation • Panbronchiolitis • Systemic disorders (rheumatoid arthritis, Sjögren syndrome, inflammatory bowel disease, yellow nail syndrome) • α-1-antitrypsin syndrome Congenital • Williams Campbell syndrome

(minIP) reconstructions provide good delineation of the fistulous tract. Nodobronchial and nodobronchoesophageal fistulas caused by mycobacterium tuberculosis infection, are depicted by the presence of gas in cavitated hilar or mediastinal lymphadenopathy adjacent to the airways. Tracheobronchoesophageal fistula may also be diagnosed even in adults (Fig. 6.14). Malignant neoplasia, particularly oesophageal, is the most common cause of tracheoeosophageal fistula in adults. Occasionally congenital fistulas are first manifested in adults. Infection and trauma are the most frequent non-malignant causes. MDCT has a high degree of sensitivity and specificity for depicting bronchial dehiscence occurring after lung transplantation. Bronchial dehiscence is seen as a bronchial wall defect associated with extraluminal air collections. SUMMARY BOX: Tracheal Disorders

B Fig. 6.11  Tracheobronchomegaly. (A) Axial computed tomograms at the upper part of the chest. Dilatation of the trachea lumen. (B) Coronal oblique reformatted slab with application of minimum intensity projection. The dilatation of the tracheal lumen is extended to the mainstem bronchi lumen.

BOX 6.2  Tracheobronchomalacia and

Dynamic Expiratory Computed Tomography In patients with suspected tracheobronchomalacia, dynamic expiratory MDCT may help depict expiratory dynamic airway collapse greater than 80% of airway lumen.

Tracheal diseases that may extent along the mainstem bronchi may be posttraumatic, inflammatory or infectious, or neoplastic. Although often evident on frontal and lateral chest radiographies, both intraluminal and extraluminal extent of the disease are particularly well delineated and characterized at computed tomography (CT) using volumetric thin collimation acquisition and multiplanar reformations. CT abnormalities include tracheal lumen narrowing, endoluminal mass, wall thickening, calcifications and fistula. Additional CT acquisition at full expiration may depict excessive collapsibility of the airway.

BRONCHIECTASIS Bronchiectasis is a chronic condition characterised by local, irreversible dilatation of bronchi, usually associated with inflammation. Despite its decreased prevalence in developed countries, bronchiectasis remains an important cause of haemoptysis and chronic sputum production. Although the causes of bronchiectasis are numerous, there are three

Fig. 6.12  Tracheobronchomalacia. Axial computed tomography and sagittal reformation acquired during dynamic expiratory manoeuvre. Almost complete collapse of the trachea, left mainstem and right intermediate bronchi lumen. The airway lumen is crescentic in shape because of the anterior bowing of the posterior membranous trachea.

Fig. 6.13  Peripheral bronchopleural fistula secondary to a necrotising pneumonia in the right lower lobe. Axial computed tomography targeted on the right lung (left) and sagittal reformation of the right hemithorax (right). The arrow shows the small fistula leading to a large pleural gas collection. Presence of lung consolidation in the right lower lobe.

138

SECTION A  The Chest and Cardiovascular System

A

A

B Fig. 6.14  Congenital Trachea-Oesophageal Fistula Revealed in an Adult. (A) Axial computed tomography showing a large communication (arrow) between the posterior wall of the trachea and the oesophagus. (B) Oblique longitudinal reformation (left) and 3D reconstruction (right) showing the fistula.

mechanisms by which the dilatation can develop: bronchial obstruction, bronchial wall damage, and parenchymal fibrosis (Table 6.1). In the first two mechanisms, the common factor is the combination of mucus plugging and bacterial colonisation. Cytokines and enzymes released by inflammatory cells plus toxins from the bacteria result in a vicious cycle of increasing airway wall damage, mucus retention and bacterial proliferation. In cases of parenchymal fibrosis, the dilatation of bronchi is caused by maturation and retraction of fibrous tissue located in the parenchyma adjacent to an airway (traction bronchiectasis). Pathologically, bronchiectasis has been classified into three subtypes, reflecting increasing severity of disease: cylindrical, characterised by relatively uniform airway dilatation, varicose, characterised by nonuniform and somewhat serpiginous dilatation, and cystic. As the extent and degree of airway dilatation increase, the lung parenchyma distal to the affected airway shows increasing collapse of fibrosis.

Radiographic Findings Chest radiography reveals abnormalities in most cases. Thickened bronchial walls are visible either as single thin lines or as parallel line

B Fig. 6.15  Bronchiectasis and Obliterative Bronchiolitis. (A) Chest radiography shows chest radiograph shows oligaemia in the lung bases with pulmonary blood flow redistribution in the upper parts of the lungs, and slight overinflation of the lungs predominant on the right side. (B) Targeted image on the right lung base in the same patient shows tramlines and ring opacities reflecting the presence of dilated and wall-thickened bronchi.

opacities (tramline) (Fig. 6.15). When seen end-on, bronchiectatic airway appears as a poorly defined ring or curvilinear opacities. Dilated bronchi filled with mucus or pus result in tubular or ovoid opacities of variable size. Cystic bronchiectasis manifests as multiple thin-walled ring shadows often containing air-fluid levels (Fig. 6.16). Pulmonary vessels may

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

139

appear increased in size and may be indistinct because of adjacent peribronchial inflammation fibrosis. In generalised bronchiectasis, such as that associated with cystic fibrosis, overinflation is often present. Localised forms are frequently accompanied by atelectasis which may be mild and detected only because of vascular crowding, fissure displacement, or obscuration of part of the diaphragm.

Computed Tomographic Findings

Fig. 6.16  Cystic Fibrosis. Pulmonary artery shows a slight overinflation and the presence of multiple thin wall ring shadows in the right lung and the left upper lung, reflecting cystic bronchiectasis. Some ring shadows contain air fluid levels.

The major sign of bronchiectasis on thin collimation high resolution computed tomography (HRCT) scan is dilatation of the bronchi (with or without bronchial wall thickening). The CT findings of bronchial dilatations include lack of tapering of bronchial lumina (the cardinal sign of bronchiectasis), internal diameter bronchi greater than that of the adjacent pulmonary artery (PA) (signet ring sign) (Fig. 6.17), visualisation of bronchi within 1 cm of the costal pleura or abutting the mediastinal pleura, and mucus-filled dilated bronchi. In varicose bronchiectasis, the bronchial lumen assumes a beaded configuration. Cystic bronchiectasis is seen as a string of cysts caused by sectioning irregular dilated bronchi along their lengths, or a cluster of cysts, caused by multiple dilated bronchi lying adjacent to each other. Clusters of cysts are most commonly seen in the atelectatic lobe. Air fluid levels, caused by retained secretions, may be present in the dependent portion of the dilated bronchi. Secretion accumulation within bronchiectatic airways is generally easily recognisable as lobulated glove-finger, -V or -Y shaped densities (Fig. 6.18). When oriented perpendicular to the avail acquisition plane, the filled dilated bronchi are visualised as nodular opacities and recognised by the observation of the homologous pulmonary arteries,

Fig. 6.17  Post-Infectious Bronchiectasis. Axial computed tomography (left) and coronal multiplanar reformation (right). Bilateral cylindrical bronchiectasis involving the right upper and the lower lobes. Note the presence of bronchial wall thickening, mucoid impactions with slight volume loss of the right lower lobe. Note also a lung cyst in the posterior part of the right upper lobe.

140

SECTION A  The Chest and Cardiovascular System

A

B

C

Fig. 6.18  Bronchiectasis in a Patient With Cystic Fibrosis Suffering From Chronic Infectious Bronchiolitis. Bilateral cylindrical, varicose and cystic bronchiectasis with thickened walls predominating at the level of the upper lobes. (A) Axial computed tomography (CT) at the level of the upper lobes. Note a moderate volume loss of these lobes with some degree of alveolar consolidation on the right side. (B) Coronal oblique reformation targeted on the left side demonstrates the beaded configuration of varicose bronchiectasis at the level of the lingula (blue arrows). Note also the mucoid impaction appearing as lobulated glove-finger (orange arrow). (C) Axial CT targeted on the left lower lobe–Centrilobular nodules predominating at the level of the lateral segment. (D) Axial maximum intensity projection image (5-mm-thick slab) clearly demonstrating the tree-in-bud appearance related to infectious bronchiolitis.

D

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction whose diameters are smaller than those of the dilated filled bronchi. CT may show a completely collapsed lobe containing bronchiectatic airways. Subtle degrees of volume loss may be seen in lobes in relatively early disease. This is most evident in the lower lobes on the basis of crowding of the mildly dilated bronchi and posterior displacement of the oblique fissure (see Fig. 6.17). Multiplanar reformations along the long axis of airways may help in assessing the dilated bronchi, recognising clusters of cystic bronchiectasis and differentiating them from lung abscess. Associated CT findings of bronchiolitis are seen in about 70% of patients with bronchiectasis. Small centrilobular nodular and linear branching opacities (tree-in-bud sign) express inflammatory and infectious bronchiolitis (see Fig. 6.18). Areas of decreased attenuation and vascularity, mosaic perfusion pattern, and expiratory air trapping reflect the extent of obliterative bronchiolitis (Fig. 6.19). These abnormalities are very common in patients with severe bronchiectasis and can even precede the development of bronchiectasis. The obstructive defect found at pulmonary tests in patients with bronchiectasis seems not to be related to the degree of collapse of large airways on expiratory CT or to the extent of mucus plugging of the airway but is the consequence of an obstructive involvement of the peripheral airways (obliterative bronchiolitis). The extent of CT evidence on small airway disease (decreased lung attenuation, expiratory air trapping) and bronchial wall thickening commonly present in patients with bronchiectasis have proven to be the major determinants of airflow obstruction. The bronchial wall thickening assessed by CT has been demonstrated to be the primary determinant of subsequent major functional decline. In one study, pulmonary arterial enlargement as shown by CT was the best predictor of mortality and was associated with outcomes independently of CT signs of bronchiectasis. Pulmonary hypertension (PH) may be reflected by pulmonary arterial enlargement on CT scans and is a highly significant prognostic indicator in the evaluation of patients with bronchiectasis.

Accuracy of Computed Tomography By combining helical volumetric CT acquisition and thin collimation, CT has gained greater advantages by circumventing the limitations of HRCT, particularly the risk of missing bronchiectasis strictly localised within the intervals between slices. At the present time, multidetector CT with thin collimation is a highly recommended technique for assessing the presence and extent of bronchiectasis. A CT Dose Index (CTDI) of less than 1 may be proposed with iterative reconstruction or filtered back projection with soft kernel, especially for young people. Multiplanar reformations increase the detection rate and the reader’s confidence, with regard to the distribution of bronchiectasis, and improve agreement between observers, with regard to the diagnosis of bronchiectasis. In addition, maximum intensity projections (MIPs) improve the detection and display of both mucoïd impactions and small centrilobular and linear branching opacities (tree in bud sign) characteristic of infectious bronchiolitis. minIP performed on a slab reconstructed with a soft-tissue kernel or iterative reconstruction help detect hypoattenuated lung areas reflecting obstruction lesion in the small airways (obliterative bronchiolitis) associated with bronchiectasis (see Fig. 6.19). The reliability of CT for distinguishing between the causes of bronchiectasis is somewhat controversial. An underlying cause for bronchiectasis is found in fewer than half of patients and CT features alone do not usually allow a confident distinction between idiopathic bronchiectasis versus other causes of bronchiectasis. However, bilateral upper lobe distribution is most commonly seen in patients with cystic fibrosis and allergic bronchopulmonary aspergillosis (ABPA), while unilateral upper lobe distribution is most common in patients with tuberculosis, and a lower lobe distribution is most often seen in patients after childhood viral infections.

141

Cystic Fibrosis Cystic fibrosis results from an autosomal recessive genetic defect in the structure of the cystic fibrosis transmembrane regulation protein, which leads to abnormal chloride transport across epithelial membranes. Although the mechanisms by which this defect leads to lung disease are not entirely understood, an abnormally low water content of airway mucus is at least partially responsible for decreased mucus clearance, mucus plugging of airways, and an increased incidence of bacterial airway infection. Bronchial wall inflammation progressing to secondary bronchiectasis is always present in patients with longstanding disease. In patients with early or mild disease, chest radiographs may be quite subtle. Hyperinflation reflects the presence of obstruction of the small airways. Thickening of the wall of the upper lobar bronchi can also be seen on the lateral film. In more advanced disease, the radiographs can be diagnostic, showing increased lung volume, accentuated linear opacities in the upper lung areas, resulting from bronchial wall thickening or bronchiectasis, proximal bronchiectasis and mucoïd impaction. Additional findings include cystic regions of the upper lobes, representing cystic bronchiectasis, healed abscess cavities, bullae, atelectasis, findings of PH or cor pulmonale, pneumothorax or pleural effusion (see Fig. 6.16). Chest radiography may be sufficient for clinical management, but there is often little visible radiographic change during clinical exacerbation. These studies consistently document a close correlation between HRCT findings and both clinical and pulmonary functional evaluation of these patients. On CT, peripheral and/or central bronchiectasis is present in all patients with advanced cystic fibrosis (see Figs 6.18 and 6.19). All lobes are typically involved, although early in the disease abnormalities are often predominantly distributed in the upper lobes, and sometimes with right upper lobe predominance. Bronchial wall and/or peribronchial interstitial thickening is also commonly present. It is generally more evident than bronchial dilatation in patients with early disease. Mucus plugging is present in 25–50% of patients and may be seen in all lobes. Collapse or consolidation is visible in up to 80% of patients. Lobar volume loss is often present in patients with advanced disease. Bullae may be difficult to distinguish from cystic bronchiectasis, particularly in fibrotic upper lobes. Abscesses may be difficult to distinguish from cystic bronchiectasis, particularly as both may contain air fluid levels. Pleural thickening, often apparent on chest radiography, is better demonstrated by CT. Small centrilobular nodular and branching linear opacities (tree in bud sign) can be an early sign of disease. They reflect presence of mucous impactions in dilated bronchioles associated with peribronchiolar inflammation. Focal areas of decreased lung attenuation are frequently present, representing air trapping and mosaic perfusion due to obstruction of the small airways (obliterative bronchiolitis). At an early stage of disease, CT can demonstrate abnormalities of the airways in patients who are asymptomatic and have normal pulmonary functions and normal chest radiographs. In patients with more advanced disease, CT is superior to chest radiograph in detecting bronchiectasis and mucus plugging. Magnetic resonance imaging has been recommended as a substitute for CT in the assessment of patients with cystic fibrosis in order to avoid the use of ionising radiation. Although the spatial resolution of pulmonary MR is lower than that of CT, it has the advantage of being able to distinguish different aspects of tissue on the basis of different contrast on T1 weighted and T2 weighted images as well as enhancement after contrast media administration. Perfusion defects assessed by contrast enhanced perfusion MR imaging show good correlation with the degree of tissue destruction in patients with cystic fibrosis. If applied to therapeutic monitoring, it might be possible to differentiate between lung areas with reversible and irreversible disease. The recent development

142

SECTION A  The Chest and Cardiovascular System

A

C

B

D Fig. 6.19  Cystic Bronchiectasis and Obliterative Bronchiolitis. Cystic fibrosis in a young female patient chronically infected with Pseudomonas aeruginosa Mycobacterium abscessus and Aspergillus fumigatus—Low dose computed tomography (CT) performed on inspiration and expiration with a CT dose index of respectively 0.66 and 0.33 mGy, resulting in a dose-length product (DLP) of respectively 24 and 11 mGy/cm. (A) Axial slice at the level of the upper lobes showing alveolar consolidation with cystic lesions predominating on the right side. (B) Coronal oblique maximum intensity projection (MIP) image (3-mm-thick slab) perfectly assess the varicose and cystic bronchiectatic nature of the cystic lesions. (C) Sagittal MIP image (3-mm-thick slab) targeted on the right lung on inspiration. (D) Sagittal MIP image (3-mm-thick slab) at the equivalent level on expiration. Note the multifocal air trapping on (D) perfectly matched with areas of low attenuation that reflect hypoperfusion due to hypoventilation secondary to obliterative bronchiolitis. (mosaic perfusion) on (C) well assessed by adapting the window width and window level.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

bronchi. Beyond the proximal bronchi, more distal airways remain normal and patent, though small airway abnormalities are present on CT. These abnormalities include a tree-in-bud appearance reflecting mucoïd impaction in dilated bronchioles and focal areas of decreased lung attenuation and air trapping reflecting obstruction of the small airways. Compared with other bronchiectatic diseases, bronchiectasis in ABPA is more commonly widespread and central, and more likely to contain cystic or varicose components. Mucus plugs within the ectatic airways are frequently seen. High attenuation within the plugs is also relatively frequently seen, reflecting the presence of calcium salts and

of ultrashort TE sequences provides the differentiation between bronchial wall thickening and mucus and permits the visual assessment of bronchiolar abnormalities. In addition, the ultrashort TE (time of echo) sequences can detect air trapping areas (making gadolinium administration unnecessary).

Allergic Bronchopulmonary Aspergillosis A hypersensitivity reaction to aspergillus species, ABPA is characterised by asthma, blood eosinophilia, radiographic pulmonary opacities and evidence of allergy to antigens of aspergillus species. It may also occur in patients with cystic fibrosis. Recurrent acute episodes cause progressive lung damage that can be controlled by steroids. The radiological features can be classified as acute and transient, or chronic and permanent. The most common acute changes are transient consolidation, mucoïd impaction, and atelectasis. Consolidation ranges from massive and homogeneous to lobar or segmental in configuration, or, to subsegmental or smaller. When consolidation clears, it often leaves residual bronchiectasis that creates a favourable condition for fungal recolonisation, a finding that accounts for the fact that consolidation recurs often in the same area. Mucoid impaction obstructs the airway lumen which becomes distended by retained secretions. At the same time, lung parenchyma remains aerated by collateral drift, permitting the visualisation of the impacted airway. Bronchoceles appear as opacities of a variety of shapes (linear, branched or unbranched, bandlike opacities that point to the hilum, tooth-paste opacities, V- and Y-shape opacities, glove finger opacities) (Fig. 6.20). These opacities disappear once their airway contents have been coughed up, leaving ring or parallel linear opacities. Atelectasis can be subsegmental, segmental, lobar or even affecting a whole lung, with a tendency to recur in the same area. Permanent changes indicate irreversible lung damage and provide the clue that an asthmatic has ABPA when he/she is in remission. Bronchiectasis is responsible for most of the permanent radiological changes. It affects lobar bronchi and the first- and second-order segmental

A

143

Fig. 6.20  Allergic Bronchopulmonary Aspergillosis. Axial computed tomography in the upper lobes. Presence of mucoïd impactions within segmental and subsegmental dilated bronchi of the upper lobes. Small centrilobular linear branching opacities are seen in the periphery of the right upper lobe.

B

Fig. 6.21  Allergic Bronchopulmonary Aspergillosis. Axial computed tomography targeted on the right lung at the level of the right upper lobar bronchus in lung windowing (A) and mediastinal windowing (B). The oval mass located in the posterior segment of the right upper lobe presents a hyperattenuated component reflecting the presence of calcium and other metallic salts in a large mucoïd impaction within a dilated bronchus.

144

SECTION A  The Chest and Cardiovascular System

BRONCHOLITHIASIS

Fig. 6.22  Dyskinetic Cilia Syndrome. Axial computed tomography at the level of the lower part of the chest. Bilateral bronchiectasis in the right middle lobe and the left lower lobe with some mucoïd impactions. Note the presence of bronchial wall thickening and multiple foci of ‘tree-in-bud’ sign reflecting infectious or inflammatory bronchiolitis. This patient also has situs inversus (Kartagener syndrome).

metals (the ions of iron and manganese) (Fig. 6.21). Hyperattenuated mucus plugs may be depicted within the areas of consolidation. Parenchymal scarring represents the fibrotic stage of the disease. It commonly follows bronchiectasis and manifests by linear opacities and lobar shrinkage. Mirroring the distribution of bronchiectasis, these features have a strong upper zone predilection. Despite this upper lobar shrinkage, the lung volume is frequently increased, reflecting overinflation in the lower lobes due to obstruction of the small airways and the presence of bullae and cavitation in the upper lobes.

Diskinetic Cilia Syndrome Resulting from a genetic abnormality having autosomal recessive transmission, dyskinetic cilia syndrome is characterised by abnormal ciliary structure and function, leading to a reduced mucociliary clearance and chronic airway infection. Bronchiectasis and sinusitis are common manifestations. About half of patients also have situs inversus. The combination of bronchiectasis, sinusitis and situs inversus is termed Kartagener syndrome (Fig. 6.22). Men and women are equally affected, but in men the syndrome may be associated with immotile spermatozoa and infertility. Respiratory symptoms can generally be traced back to childhood. Bronchiectasis develops in childhood and adolescence and is associated with recurrent pneumonia. Both radiography and CT typically show bilateral bronchiectasis with a basal (lower or middle lobe) predominance, similar to that seen in patients with other causes of postinfectious bronchiolitis. Cylindrical bronchiectasis is most common, and a diffuse bronchiolitis may be present.

SUMMARY BOX: Bronchiectasis Although chest radiography may reveal abnormalities in most cases of bronchiectasis, computed tomography (CT) using multidetector technique and thin collimation is the imaging modality of reference to assess the presence, distribution and severity of bronchiectasis. Maximum intensity projection technique improves the detection of mucoid impactions and associated inflammatory and infectious bronchiolitis (tree-in -bud sign). Minimum intensity projection technique and additional acquisition at full expiration help assess the extent of associated obliterative bronchiolitis (mosaic perfusion pattern and expiration gas trapping).

Broncholithiasis is a condition in which peribronchial calcified nodal disease erodes into or distorts an adjacent bronchus. The underlying abnormality is usually granulomatous lymphadenitis caused by mycobacterium tuberculosis or fungi such as histoplasma capsulatum. A few cases have been reported with silicosis. Calcified material in a bronchial lumen or bronchial distortion by peribronchial disease results in airway obstruction. This leads to collapse, obstructive pneumonitis, mucoïd impaction, or bronchiectasis. Symptoms include cough, haemoptysis, recurrent episodes of fever and purulent sputum. Broncholithiasis is more common on the right, and obstructive changes particularly affect the right middle lobe. On chest radiographs, three major types of changes may be seen: • disappearance of a previously identified calcified nidus • change in position of a calcified nidus • evidence of airway obstruction, including segmental or lobar atelectasis, mucoïd impaction, obstructive pneumonitis, obstructive oligaemia with air trapping. Calcified hilar or mediastinal nodes are a key feature. CT and fibreoptic bronchoscopy complement each other in this condition. Broncholithiasis is recognised at CT by the presence of a calcified endobronchial or peribronchial lymph node, associated with bronchopulmonary complication due to obstruction (including atelectasis, pneumonia, bronchiectasis, and air trapping), in the absence of an associated soft-tissue mass.

OBLITERATIVE (CONSTRICTIVE) BRONCHIOLITIS Inflammation of the bronchioles (bronchiolitis) is very common, although it is rarely extensive enough to cause clinical symptoms. Pathological studies have repeatedly emphasised the frequent involvement of the bronchioles in diverse diffuse disease. Inflammation of the bronchioles may be reversible under specific or anti-inflammatory treatment or lead to subsequent scarring and obliteration. Obliterative bronchiolitis is a condition characterised by bronchiolar and peribronchiolar inflammation and fibrosis that ultimately leads to luminal obliteration affecting membranous and respiratory bronchiolitis. Obliterative bronchiolitis is the result of a variety of causes and only rarely idiopathic (Table 6.2). When a large proportion of the airways are affected, patients usually present with progressive shortness of breath and functional evidence of airflow obstruction.

Pathological Features The pattern of obliterative bronchiolitis is characterised by the development of an irreversible circumferential submucosal fibrosis, resulting in bronchiolar narrowing or obliteration of bronchioles in the absence of intraluminal granulation tissue polyps or surrounding parenchymal inflammation. Proliferation of fibrosis extends predominantly between the epithelium and the muscular mucosa and along the long axis of the airway, impairing collateral ventilation, and leading to airflow obstruction. The epithelium overlying the abnormal fibrosis tissue may be flattened or metaplastic and is usually intact without any ulceration. In some instances, the accompanying artery is also obliterated by the same fibrotic process.

Radiological Findings The chest radiograph is often normal. In a small number of patients, mild hyperinflation, subtle peripheral attenuation of the vascular makings, widespread and conspicuous abnormalities in lung attenuation, and central bronchiectasis may be seen.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

145

TABLE 6.2  Causes of and Association

With Obliterative (Constrictive) Bronchiolitis Post-infection Childhood viral infection (adenovirus, respiratory syncytial virus, influenza, parainfluenza) Adulthood and childhood (Mycoplasma pneumoniae, Pneumocystis jiroveci) in AIDS patients, endobronchial spread of tuberculosis, bacterial bronchiolar infection) Post-inhalation (toxic fume and gases) Nitrogen dioxide (silo-filler’s disease), sulphur dioxide, ammonia, chlorine, phosgene Hot gases Gastric aspiration Diffuse aspiration bronchiolitis (chronic occult aspiration in the elderly, patients with dysphagia) Connective tissue disorders Rheumatoid arthritis Sjögren syndrome Allograft recipients Bone-marrow transplant Heart-lung or lung transplant Drugs Penicillamine Lomustine Ulcerative colitis Other conditions Bronchiectasis Chronic bronchitis Cystic fibrosis Hypersensitivity pneumonitis Sarcoidosis Microcarcinoid tumorlets (neuroendocrine cell hyperplasia) Sauropus androgynus ingestion Idiopathic

Thin section CT is superior to radiography in demonstrating the presence and extent of abnormalities. The main CT findings usually consist of areas of decreased lung attenuation associated with vessels of decreased calibre during inspiration and air trapping on expiratory CT. Because the lesions of bronchiolar narrowing or obstruction are heterogeneously distributed throughout the lungs, redistribution of blood flow to areas of normal lung or less diseased areas results in a pattern of mosaic perfusion (Fig. 6.23). Bronchial wall thickening and bronchiectasis, both central and peripheral, are also commonly present (see Fig. 6.19). Although the vessels within areas of decreased attenuation on thin-section CT may be of markedly reduced calibre, they are not distorted as in emphysema. The lung areas of decreased attenuation related to decreased perfusion can be patchy or widespread. They are poorly defined or sharply demarcated, giving a geographical outline, representing a collection of affected secondary pulmonary lobules. Redistribution of blood flow to the normally ventilated areas causes increased attenuation of lung parenchyma in these areas. The patchwork of abnormal areas of low attenuation and normal lung or less diseased areas, appearing normal in attenuation or hyper-attenuated, gives the appearance of mosaic attenuation (Fig. 6.24). The vessels in the abnormal hypo-attenuated areas are reduced in calibre, whereas the vessels in normal areas are increased in size, and the resulting pattern is called ‘mosaic perfusion’. The difference in vessel size between low- and

Fig. 6.23  Post-Infectious Obliterative Bronchiolitis. Axial computed tomography at full inspiration (top) and full expiration (bottom). Mosaic perfusion pattern with multiple areas of hypoattenuation containing pulmonary vessels reduced in number and calibre. The contrast between the hypoattenuated abnormal areas and normal lung parenchyma is accentuated at expiration (normal areas increase in attenuation at expiration as normally expected while the abnormal areas do not).

high-attenuation areas allows one to distinguish the mosaic perfusion pattern from mosaic attenuation due to an infiltrative lung disease with patchy distribution, in which the vessels have the same calibre in both high-attenuation and normal-attenuation areas. The areas of decreased lung attenuation and perfusion may be confined to or predominant in one lung, particularly in Swyer-James or MacLeod syndrome, that is a variant form of post-infectious obliterative bronchiolitis in which the obliterative bronchiolar lesions affect predominantly one lung. Usually the regional heterogeneity of the lung density seen at end-inspiration on thin-section CT is accentuated on sections obtained at end, or during, expiration because the high-attenuation areas increase in density and the low-attenuation areas remain unchanged. In the case of more global involvement of the small airways, the lack of regional homogeneity of the lung attenuation is difficult to perceive on inspiratory CT images, and as a result, mosaic perfusion becomes visible only on expiratory scans. In patients with particularly severe and widespread involvement of the small airways, the patchy distribution of hypoattenuation and mosaic pattern is lost. Inspiratory CT shows an apparent uniformity of decreased attenuation in the lungs, and CT at end-expiration may appear unremarkable (Fig. 6.25). In these patients, the most striking

146

SECTION A  The Chest and Cardiovascular System

Fig. 6.24  Obliterative Bronchiolitis. Coronal CT reformation at inspiration (left) and expiration (right). Minimum intensity projection was applied on the expiratory reformation. Large hypo-attenuated areas expressing gas trapping are distributed in the lower parts of the lung.

A

B Fig. 6.25  Post-Bone Marrow Transplantation Obliterative Bronchiolitis. (A) Axial computed tomography (CT) at the level of the lower part of the chest. Diffuse hypoattenuation of lung parenchyma. Lung vessels are reduced in number and in calibre. Note the slight dilatation of the bronchi lumens and the presence of bronchial wall thickening. (B) Low dose axial CT performed at short suspended end expiration at the same level as A. The absence of increase in lung attenuation and significant reduction in lung cross section area reflect the presence of diffuse air trapping. The complete collapse of the bronchial lumens in the lower lobes testifies that CT was acquired at the end of a forced expiratory manoeuvre.

features are paucity of pulmonary vessels and lack of change of the cross-sectional areas of the lung at comparable levels on inspiratory and expiratory CT. In such a situation, there is a risk of misdiagnosis between obliterative bronchiolitis and panlobular emphysema. Both conditions are characterised by bronchial wall thickening and generalised decreased attenuation of the lung parenchyma and bronchial dilatation. However, patients with panlobular emphysema demonstrate parenchymal destruction with higher frequency and to a greater extent than those with obliterative bronchiolitis. Long lines reflecting limited thickened

interlobular septa were significantly more frequent in patients with panlobular emphysema.

Computed Tomography Assessment of Air Trapping The most commonly used technique for the assessment of air trapping is postexpiratory thin section CT obtained during suspended respiration following a forced exhalation. Each of the postexpiratory images should be compared with the inspiratory image that most closely duplicates

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

BOX 6.3  Minimal Intensity Projection to

Assess Obliterative Bronchiolitis

• In patients with obliterative bronchiolitis, multiplanar volume rendering slab associated with minimal intensity projection helps the depiction of mosaic perfusion pattern. • Its application on expiratory computed tomography facilitates assessment of the presence and extent of air trapping.

its level to detect air trapping. Dynamic expiratory manoeuvre performed during helical CT acquisition has been described. It permits a small increase in the degree of expiration, which leads to a better detection of air trapping. This technique is recommended when patients have difficulty performing the suspended end-expiration manoeuvre adequately. Using MDCT with thin collimation over the lungs and low dose or ultra-low dose with iterative reconstruction or soft-tissue kernel with filtered back projection has become routine in many institutions to improve the conspicuity and the apparent extent of air trapping (Box 6.3). Multiplanar volume rendering slab associated with the technique of minIP increases the contrast between areas of normal lung attenuation and areas of lung hypoattenuation. This helps the depiction of mosaic perfusion pattern. Its applications on expiratory CT can also facilitate assessment of the presence and extent of air trapping. The extent of air trapping present on expiratory images can be measured using a semiquantitative scoring system that estimates the percent of lung that appears abnormal on each section. In the scoring system proposed by Stern et al., estimates of air trapping were made at each level and for each lung on a four-point scale: 0: no air trapping; 1: 1% to 25%; 2: 26% to 50%; 3: 51% to 75%; and 4: 76% to 100% of cross-sectional areas of lung affected. The air trapping score is the summation of these numbers for the different levels studied. This scoring system provides good interobserver and intraobserver agreement. The extent of expiratory air trapping at CT has proved to be correlated with the degree of airflow obstruction at pulmonary function tests in patients with obliterative bronchiolitis. Objective measurement of air trapping can be done using CT densitometry. In the density mask technique, all the pixels included in areas of air trapping are segmented by thresholding at −910 HU and are highlighted and automatically counted. Density changes between full inspiration and full expiration can be compared, and expiratory/ inspiratory ratios can be calculated. The density mask has the advantage that it combines density measurement with the visual assessment of pathology. Using multislice CT with thin collimation over the lungs performed at full expiration, an exhaustive assessment of the volume of air trapping may be provided as well as a 3D visualisation of distribution of air trapping.

CHRONIC OBSTRUCTIVE PULMONARY DISEASE Characterised by functional abnormalities, COPD is a slowly progressive airway obstructive disorder resulting from an exaggerated inflammatory response to cigarette smoke or other inhaled pollutants that ultimately destroy lung parenchyma (emphysema) and induce irreversible reduction of the calibre of the small airways (obstructive bronchiolitis). Both lesions may be associated in the same patient. On the other hand, narrowing and loss of terminal bronchioles clearly precede the appearance of microscopic emphysematous destruction. That explains why the use of CT in evaluation of patients with COPD has made it clear that individuals with identical severity of airflow obstruction may exhibit different morphological appearances. Some have extensive emphysema

147

while others have minimal emphysema suggesting more significant small airway disease. These differences in morphological appearances may be related to differences in pathophysiology and genomic profile. As a result, CT imaging may be employed to objectively classify individuals as having either emphysema or airway predominant disease. This better phenotyping of COPD patients may help select stratify patients in clinical trials and in given individuals help optimise treatment.

Pathological Findings Inflammatory changes in the airways in COPD patients involve both small airways and large airways. Small airway disease in COPD is characterised at the beginning by inflammatory change in the walls and around the respiratory bronchioles (respiratory bronchiolitis characterised by pigmented macrophages). In more advanced disease, inflammatory changes in respiratory bronchioles are associated or replaced by obstruction of the lumen of the small airways by plugs of inflammatory exudates and mucus. In still more advanced involvement, the lumen of the terminal bronchioles are narrowed by peribronchiolar fibrosis (obstructive bronchiolitis). Large airway disease in COPD includes inflammation and remodelling of the trachea and bronchi. Histologically this is characterised by wall inflammation, squamous metaplasia of bronchial epithelium, slight increase of basal membrane and smooth muscle, submucosal cells, glands hyperplasia, and deficit in cartilage. Emphysema is defined as a condition of the lung characterised by permanent abnormal enlargement of airspaces distal to the terminal bronchiole, accompanied by the destruction of their walls without obvious fibrosis. The most important aetiologic factor by far is cigarette smoking. There is also a causal relationship between HIV infection and the development of early emphysema. Various genetic disorders may be associated with emphysema including α1-antitrypsin deficiency, heritable diseases of connective tissue such as cutix laxa, Marfan syndrome and familial emphysema. Emphysema is thought to result from the destruction of elastic fibres caused by an imbalance between proteases and protease inhibitors in the lung and from the mechanical stresses of ventilation and coughing. Proteases are normally released in low concentration by phagocytes in the lung. Protease inhibitors, mainly α1-protease inhibitor (α1antitrypsin), prevent them from causing structural damage to the lung. Imbalance in the protease-antiprotease activity may result from antiprotease deficiency (α1-antitrypsin deficiency), from excess release of protease stimulated by environmental agents, or from the defective repair of protease-induced damage. Tobacco smoke increases the number of pulmonary macrophages and neutrophils, reduces antiprotease activity, and may impair the synthesis of elastin. As emphysema develops, lung destruction progresses, airspaces enlarge and elastic recoil declines, reducing radial traction on bronchial walls and on blood vessels and allowing airways and vessels to collapse. Emphysema is traditionally classified on the microscopic localisation of disease within the secondary pulmonary lobule. The principal types are centrilobular, panlobular, paraseptal and irregular emphysema. Centrilobular (centriacinar) emphysema affects mainly the proximal respiratory bronchioles and alveoli in the central part of the acinus. The process tends to be most developed in upper parts of the lungs. It is strongly associated with cigarette smoking. Paraseptal emphysema selectively involves the alveoli adjacent to connective tissues in septa and bronchovascular bundles, particularly at the margins of the acinus and lobule but also subpleurally and adjacent to the bronchovascular bundles. Airspaces in paraseptal emphysema may become confluent and develop into bullae, which may be large. Panlobular (panacinar emphysema) is characterised by a dilatation of the airspaces of the entire acinus and lobule. With progressive destruction, all that eventually remains are thin strands of deranged tissue surrounding blood vessels.

148

SECTION A  The Chest and Cardiovascular System

It is the most widespread and severe type of emphysema with changes distributed throughout the lungs although often basely predominant. Panlobular emphysema is the type occurring in α1-antitrypsin deficiency and in familial cases.

Radiographic Findings Chest radiography may be normal in COPD. When radiographic abnormalities are present, they can include hyperinflation, oligaemia, bronchial wall thickening and accentuation of linear lung markings. Thickening of the bronchial walls leads to tubular and ring shadows (Fig. 6.26). Increased lung markings cause the appearance of ‘dirty chest’, a term widely used for describing a loss in clarity of the lung vessels (see Fig. 6.26). Sabre-sheath trachea may be present. Cor pulmonale is a recognised complication which is seen most exclusively in hypoxic patients. With the onset of heart failure, the heart and hila and intermediate lung vessels become enlarged. Enlargement of vessels is present in all zones and affects particularly segmental vessels and a few divisions beyond. Signs of overinflation are the best predictors of the presence and severity of emphysema. Signs of overinflation include the height of the right lung being greater than 29.9 cm, location of the right hemidiaphragm at or below the anterior aspect of the seventh rib, flattening of the hemidiaphragm, widening of the sternodiaphragmatic angle, narrowing of the transverse cardiac diameter and enlargement of the retrosternal space on the lateral view. Alterations in lung vessels include arterial depletion, whereas vessels of normal, or occasionally increased, calibre are present in unaffected areas of the lung, absence or displacement of vessels caused by bullae, widened branching angles with loss of side branches and vascular redistribution. With the development of cor pulmonale, or left heart failure, the radiographic appearances will alter and may become less obviously abnormal. The heart may then appear to be normal in size, or sometimes enlarged, and the diaphragm becomes less flat while the pulmonary vessels less attenuated. Bullae may be as small as 1 cm in diameter or may occupy the whole hemithorax causing marked relaxation collapse of the adjacent lung (Fig. 6.27). Bullae caused

by paraseptal emphysema are much more common in the upper zones, but when they are associated with widespread panlobular emphysema, the distribution is much more even. Occasionally the wall is completely absent and in such a case bullae can be difficult to detect. The presence of emphysema associated with large bullae is referred to as bullous emphysema (Fig. 6.28). An entity mainly seen in young men, characterised by the presence of large progressive upper lobe bullae which occupy a significant volume of a hemithorax and are often asymmetrical, is referred as giant bullous emphysema, vanishing lung syndrome or primary bullous disease of the lung. Large bullae may be seen as avascular transradiant areas, usually separated from the remaining lung parenchyma by a thin curvilinear wall. They can cause marked relaxation collapse of the adjacent lung and can even extend across into the opposite hemithorax, particularly by way of the anterior junctional area. Spontaneous pneumothorax commonly occurs in association with localised areas of emphysema or bullae affecting the lung apices. Bullae may enlarge progressively over months or years; a period of stability may be followed by a sudden expansion. Bullae may also disappear, either spontaneously or following infection or haemorrhage. The main complications of bullae include pneumothorax, infection and haemorrhage. In case of infection or haemorrhage, bullae contain fluid and may show an air-fluid level. When a bulla becomes infected the hairline wall becomes thickened and may mimic a lung abscess. Carcinoma arising in or adjacent to bullae should be suspected when there is a mural nodule, mural thickening, a change in diameter of the bulla, pneumothorax, and the accumulation of fluid within the bulla.

Computed Tomographic Findings The Fleischner Society published a statement in 2015 on CT-definable subtypes of COPD. The purpose of this statement was to describe and define the phenotypic abnormalities that can be identified on the visually and quantitative evaluation of CT images in subjects with COPD, with the goal of contributing to a personalised approach to the treatment of patients with COPD. The classification system provides a structured approach to visual and quantitative assessment of COPD. Emphysema is classified as centrilobular, panlobular and paraseptal. Additional important visual features include airway wall thickening, inflammatory small airway disease, expiratory gas trapping, and associate features (Box 6.4).

Visually Defined Computed Tomography Pattern of Emphysema It is possible to recognise the subtypes of emphysema on the basis of their CT appearances. Centrilobular emphysema is characterised by poorly defined local lucencies without visible walls and surrounded by normal lung and having a predominant distribution in the upper lobes. According to the Fleischner Society subclassification, centrilobular emphysema is regarded as mild when scattered centrilobular lucencies, usually separated by large regions of normal lung, involving an estimated 0.5–5% of an upper lung zone (Fig. 6.29). It is regarded as moderate centrilobular emphysema when many well-defined centrilobular or

BOX 6.4  Computed Tomography to Assess

Bronchiectasis

Fig. 6.26  Chronic Bronchitis and Obstructive Lung Disease. Posteroanterior chest radiograph shows mild overinflation. A ring shadow is visible above the left hilum (arrow) reflecting bronchial wall thickening. There is also an accentuation of linear markings in the right lung basis.

• In patients with bronchiectasis, the extent of computed tomography (CT) evidence of decreased lung attenuation and expiratory air trapping and the presence of bronchial wall thickening are the major determinants of airflow obstruction. • Bronchial wall thickening at CT is the primary determinant of subsequent functional decline.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

A

149

B Fig. 6.27  Severe Diffuse Emphysema. Posteroanterior (A) and lateral (B) chest radiographs. The diaphragm is displaced downwards and appears flattened. On the pulmonary artery radiograph (A), the transverse cardiac diameter is reduced. The diaphragm appears irregular in contours due to an abnormal visibility of diaphragmatic insertions on the ribs. Note the depression of vessels in the periphery of the lungs. On the lateral view (B), there is a widening of the sternodiaphragm angle and an increase of dimensions of the retrosternal transradiant area.

Fig. 6.28  Giant Bullous Emphysema. The pulmonary artery chest radiograph shows large avascular transradiant areas in the upper and lower parts of the right lung. The bullae are marginated with thin curvilinear opacities.

Fig. 6.29  Mild Centrilobular Emphysema. HRCT targeted on the right lung. Multiple small, round areas of low attenuation that are distributed through the lungs, mainly around the centrilobular arteries.

150

SECTION A  The Chest and Cardiovascular System

BOX 6.5  Assessment of Emphysema at

Computed Tomography

In chronic obstructive pulmonary disease patients, visual assessment of the presence and extent of emphysema subtypes including centrilobular (traces, mild, moderate or confluent), paraseptal (mild or significant) and advanced destructive, can be reliably estimated at a lobar level.

Fig. 6.30  Moderate Centrilobular Emphysema. Axial computed tomography images of the upper and lower parts of the lungs. Many well-defined centrilobular lucencies without visible walls occupy more than 5% of any lung zone. Note the predominance of the number of lesions in the upper lobes.

lobular lucencies are occupying more than 5% of any lung zone (Fig. 6.30). Small vessels, often seen traversing the hypoattenuated focal areas, are centrilobular pulmonary arteries or arterioles marking the centre of each lobule. Centrobronchovascular bundles are preserved. As emphysema becomes more severe, the areas of hypoattenuation appear confluent and multiple regions of lucencies span several secondary pulmonary lobules but do not involve extensive hyperexpansion of the secondary pulmonary lobules or gross distortion of pulmonary architecture (Fig. 6.31). Advanced destructive emphysema represents very severe centrilobular emphysema, characterised by panlobular lucencies with hyperexpansion of the secondary pulmonary lobules and distortion of pulmonary architecture most often with upper lung predominance (Fig. 6.32A). On maximal intensity projection CT images, the pulmonary vessels located in the emphysematous parenchyma are straightened and splayed with decreasing branching (see Fig. 6.32B). In case of lung destruction involving the lung bases, long lines reflecting the thickening of the interlobular septa by slight fibrosis may be seen crossing the hypoattenuated destroyed lung parenchyma. Panlobular emphysema is a term that is associated with α-1-antitrypsin deficiency characterised by a lower lobe predominant pattern involving generalised destruction of all acini more or less equally. On CT, the visual pattern is characterised by a generalised decrease of lung attenuation involving the lower lobes with an increase of lower lobes volume with anterior displacement of the major fissures (Fig. 6.33).

Paraseptal emphysema is characterised by well-demarcated rounded juxtapleural lucencies, aligned in a row along a pleural margin, sometimes including an interlobular fissure, and along the peribronchovascular bundles. Paraseptal emphysema is characterised as mild when the rounded lucencies are small (1 cm) cystic-like lucencies or bullae are present (Fig. 6.34). Bullae may be large enough to compress the adjacent lung parenchyma. Although such compression is usually relatively mild, it may occasionally result in compressive atelectasis appearing as a parenchymal band or mass-like opacity. Bullous emphysema is a pattern characterised by multiple large avascular lucencies partly bordered by a thin wall (Fig. 6.35). Most patients with bullous lung disease have concomitant centrilobular or paraseptal emphysema. The term ‘giant bullous emphysema’ has been used to describe the presence of bullae occupying at least one third of a hemithorax. Giant bullae may be compressive not only on the lung parenchyma but also on the diaphragm and the right atrium with a risk of tamponade. Visual assessment of emphysema may be done at the lobar level. Several different subtypes may be present in different lobes due to heterogeneous severity of disease. Kim et al. showed that visual subtypes can be reliably estimated at a lower lobar level with good agreement between readers (Box 6.5).

Visually Defined Computed Tomography Patterns of Airway Diseases These include bronchial wall thickening and features of small airway diseases. Bronchial wall thickening is a feature often observed on CT of smokers with or without COPD. Bronchial wall thickening is recognised as a relative increase in the thickness of the bronchial wall by comparison with the bronchial lumen and the diameter of the homologous PA (Figs 6.36 and 6.37). Unfortunately, this feature suffers from interobserver variability and is better assessed by comparison with visual standards obtained from subjects with normal and abnormal findings. At the earlier stage of the disease, inflammatory changes in the small airways are seen on CT as multiple areas of ground glass attenuation and small centrilobular ill-defined nodular opacities (see Fig. 6.37). These abnormalities, predominant in the upper lobes or sometimes more diffuse in distribution, have been reported to be present in about 22 to 25% of asymptomatic smokers. At a later stage of the disease (obstructive bronchiolitis), the CT findings include mosaic perfusion pattern (low attenuation and low perfusion areas where the terminal bronchioles are obstructed) and expiratory air trapping in the same areas (Fig. 6.38). Obstruction of the small airways may also be suggested in patients presenting airflow limitation at spirometry in the absence of emphysema at CT.

Associated Features They include bronchiectasis, large airway disease, interstitial lung abnormalities and PH. Evidence of moderate tubular bronchiectasis, mostly in the lower lobes, may be present and is usually associated with most severe airflow obstruction and with hospitalizations for exacerbation. Varicose and cystic bronchiectasis are often associated

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

151

Fig. 6.31  Confluent Centrilobular Emphysema. Axial computed tomograms (left) on the upper and lower parts of the lungs, and maximum intensity projection (MIP; (top right) and minimum intensity projection (bottom right) on axial 7 mm thick slabs of the lung bases. minIP, Minimum intensity projection. Note multiple regions of lucency that span several secondary pulmonary lobules, but that do not involve hyper-expansion of the secondary pulmonary lobules and without gross distortion of pulmonary architecture.

with panlobular emphysema in patients with α-1-antitrypsin deficiency. In smokers, bronchial diverticula or outpouchings may be seen as small airway collections in the wall of the main and lobar bronchi particularly well displayed on coronal reformations with minimal intensity projection (Fig. 6.39). They express the fusion of multidepressions and dilatations of the bronchial gland ducts forming a diverticulum which herniates between and through the smooth muscle cellular bundles. In one study, 12% of smokers presented more than three bronchial diverticula in one or more bronchi. An increased number of diverticula is associated with cigarette smoking and with symptoms of cough. An expiratory central airway collapse, defined by a 50% reduction of cross-sectional area of tracheal lumen at expiration in a large study of current and former smokers, was reported in 5% of subjects. This feature was more prevalent in those with COPD and its presence was associated with worse quality of life as well as greater frequency of acute respiratory events. Tracheobronchomalacia, defined as a reduction in the tracheal luminal cross-sectional area by more than 80% at dynamic expiratory imaging, was found in about 20% of COPD patients but was not correlated with physiological impairment. The presence of non-dependent interstitial lung abnormalities, including ground glass and reticular abnormalities that affected more than 5% of any lung zones, were reported in 8% of smokers with or without COPD (Fig. 6.40). The prevalence of these interstitial lung abnormalities rises with age, tobacco exposure, and current smoking. They are associated with the lower total lung capacity and a lesser degree of emphysema. They were independently associated with reduced exercise capacity and higher mortality.

Pulmonary hypertension frequently complicates COPD and its presence is often seen in early disease. Pulmonary arterial enlargement may be measured as a surrogate for PH. A ratio of the trunk of pulmonary artery to aorta diameter (PA/AA) greater than 1.0 correlates with PH and independently predicts future severe exacerbations requiring hospitalisation.

Quantitative Computed Tomography Image Analysis Quantitative CT (QCT) is useful for identifying and sequentially evaluating the extent of emphysematous lung destruction, changes in the airway walls, and expiratory air trapping. The goals of QCT in COPD are to quantify the presence and percentage of emphysema-like lung (low-attenuation areas), the lobar and zonal distribution of the low-attenuated regions, changes in airway walls and luminal calibre, and the severity of gas trapping at expiratory CT. Volumetric thin section CT is generally recommended for characterisation of COPD. Submillimetric z-axis resolution with overlapping section reconstruction is recommended for optimal airway analysis. A high spatial resolution algorithm is better for visual assessment of the lung whereas a smooth reconstruction algorithm facilitates computerised analysis by reducing image noise. The CT radiation dose delivery is driven by the balance between radiation dose and image quality. Excessive image noise with a reduced CT dose can simulate emphysema and may impair segmentation of the airways and quantitative evaluation of the airway dimensions. Expiratory CT is required for determining the severity of airway obstruction. It is performed as suspended full expiration. It may be performed at a lower CT radiation dose. Text continued on p. 156

Fig. 6.32  Advanced Destructive Emphysema. (A) Axial computed tomography in the upper lobes. (B) Coronal reformations without (left) and with (right) maximum intensity projection image. Pan lobular lucencies in the upper parts of the lungs with hyperexpansion of the secondary pulmonary lobules and distortion of the pulmonary architecture with upper predominance. The pulmonary vessels in the upper parts of the lungs are thin, straightened with pruning and decreased branching.

A

B

Fig. 6.33  Panlobular Emphysema in a Patient With α-1-Antitrypsin Deficiency. Coronal and left sagittal reformations. Lower lobe predominant pattern of a generalised decrease of lung attenuation with slight increase of lower lobe volume as illustrated by anterior convexity of the left major fissure (yellow arrows).

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

Fig. 6.34  Substantial Paraseptal Emphysema. Axial computed tomography in the upper and middle parts of the lungs. Many large (>1 cm diameter) juxtapleural cyst-like lucencies or bullae aligned in a row along pleural margins, including interlobar fissures and adjacent to peribronchovascular bundles. Note the presence of a few small centrilobular emphysematous spaces.

Fig. 6.35  Bullous Emphysema. Axial computed tomography in the middle and lower parts of the lung. Bilateral voluminous bullae in the subpleural areas associated with moderate centrilobular emphysema, and paraseptal emphysema along the mediastinal pleural surfaces.

153

154

SECTION A  The Chest and Cardiovascular System

A

Fig. 6.37  Inflammatory Bronchiolitis in a Heavy Smoker With Chronic Bronchitis. Multiple ill-defined small centrilobular nodular opacities distributed mainly in the upper lobes. Bronchial wall thickening and paraseptal emphysema are also present.

B Fig. 6.36  Chronic Obstructive Pulmonary Disease Patient With Airway Disease Predominant Phenotype. Axial computed tomography sections at the levels of the upper and lower parts of the chest. There are a few small centrilobular and paraseptal emphysematous spaces in the upper lobes. Bronchial wall thickening, slight bronchial dilatation, and lung parenchyma hypoattenuation reflecting obstructive bronchiolitis in the lower lobes.

Fig. 6.38  Expiratory Gas Trapping in Chronic Obstructive Pulmonary Disease Patients With Airway Predominant Disease. Inspiratory (left) and expiratory (right) axial computed tomography (CT) images. The heterogeneity of lung attenuation present on inspiratory images is accentuated at expiration. There is a lack of increase in attenuation at expiration, and an absence of antero-posterior gradient of attenuation on expiratory CT as normally expected.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction

Fig. 6.39  Bronchial Diverticula in a Chronic Obstructive Pulmonary Disease Patient With Chronic Bronchitis. Coronal (left) and oblique (right) reformations with minimum intensity projection on 7-mm thick slab. Multiple small diverticula are visible along the wall of the left main and upper lobar bronchi. Note the presence of paraseptal and centrilobular emphysema.

A

B Fig. 6.40  Interstitial Lung Abnormalities in a Chronic Obstructive Pulmonary Disease Patient. (A) Axial computed tomography images in upper and lower parts of the lungs. (B) Coronal and right sagittal reformations. Paraseptal and confluent centrilobular emphysema are seen in the upper lobes. Posterior subpleural paraseptal emphysema is seen in the lower lobes. Presence of reticulation in the peripheral parts of the lower lobes surrounding enlarged airspaces of emphysema in subpleural parenchyma mimicking honeycombing of usual interstitial pneumonitis (UIP).

155

156

SECTION A  The Chest and Cardiovascular System

Quantitative Analysis of Emphysema Extent So far the most commonly used technique for QCT analysis of emphysema remains lung densitometry. It consists of automatically segmenting voxels having an attenuation value below a given threshold cut off of the CT density histogram to generate CT density mask. Although the highest correlations between QCT metrics and histology were found when using thresholds of −960 and −970 HU on MDCT and thin collimation CT, the threshold of −950 HU is commonly used in the interest of balancing between sensitivity and specificity (Fig. 6.41). An alternative is to use the percentile density, that is, a density value below which a given percentage of the lung pixels falls. Although correlations with histology have shown the 1st percentile to be the optimal percentile for this determination, most studies used the 15th percentile because of concern regarding the presence of noise and artefact at the 1st percentile. The 15th percentile has proven to be the most sensitive index

to detect disease progression in longitudinal studies. As emphysema is a regionally distributed disease, it makes sense to provide 2D and 3D colour display in longitudinal reformats with automatic extraction of proximal airway lumens and segmentation of the lungs with automatic calculation of lung volume and percentage of emphysema (see Fig. 6.41). Specific software may divide the lung into upper, middle and lower parts of equal height or volume. Other software may segment the lobar boundaries automatically and calculate the percentage of lobar emphysema (Fig. 6.42). Sources of variations in density-based measures include inspired lung volume, CT system used (make and model), increased body-mass index and increased lung density in individuals who are currently smoking. Precise calibration of the CT system, ideally with a standardised phantom, is important for ensuring the accuracy of the CT number. When different CT models are used, the results should be adjusted for tracheal air attenuation. Adjustments for demographic factors are needed. Other sources of variation deserve

Fig. 6.41  Lung densitometry using mask density technique (threshold: −950 HU) in a patient with moderate centrilobular emphysema. The voxels of emphysema are highlighted in red. Axial, coronal and left sagittal reformations and 3D reconstruction of tracheobronchial lumen associated with projection of voxels of emphysema in red (bottom right). Gas included in the lumen of the proximal airways is excluded from the fraction of emphysema.

CHAPTER 6  Large Airway Disease and Chronic Airflow Obstruction specific attention. Differences in emphysema measurement at varying inspiration levels are not clinically relevant above 90% of vital capacity. Spirometric control of lung volume is not necessary but careful coaching of the patient by the radiographers is important. Lung densitometry is highly reproducible when it is corrected for differences in lung volume

Fig. 6.42  Results of automatic segmentation of lung parenchyma with segmentation of the lobes reconstructed in 3D. The segmentation of the lobes allows for the calculation of the volume and fraction of emphysema per lobe.

157

(percentage of emphysema on the follow-up CT studies corrected using the same achieved lung volume on the baseline study). Patients referred for QCT of emphysema should be allowed to rest before the examination to increase the reliability of emphysema volume or index values. Variation in technical parameters should be minimised by using standardised acquisition and reconstruction parameters and using the same equipment and the same software. Differences in reconstruction algorithms have a large effect (9.5%) in CT measurement of a low attenuation area. Excessive image noise with a reduced CT dose can simulate emphysema at QCT. It is recommended to not use less than 60 mAs. Iterative reconstruction algorithm improves consistency between low dose CT and standardised dose CT for emphysema quantification. The extent of emphysema is more consistent across different tube currents when using iterative reconstruction algorithm. The extent of emphysema appears to increase quite rapidly after smoking cessation, reflecting a fall in lung attenuation due to the decrease of inflammatory changes. Moreover, emphysema measurements increase with increasing body mass. As a result, longitudinal analysis of emphysema must adjust for both smoking status and body-mass index both of which may change over time. QCT analysis of emphysema extent correlates significantly with both FEV1/FVC as well as FEV1 and predicts COPD exacerbations that require hospitalisations. Quantitative emphysema is associated with respiratory quality of life and with increased all-cause mortality in patients with COPD. When the main sources of variations are minimised or under control, QCT has ability to detect progression of emphysema. According to Gietama et al, a change of 1.1% in the emphysema extent may be detected with 95% probability by using MDCT (Fig. 6.43).

Quantitative Computed Tomography Analysis of Gas Trapping Gas trapping on expiratory CT has been used as a surrogate for small airway disease. The expiration to inspiration ratio of mean lung density

Fig. 6.43  Progression of emphysema extent by subsequent computed tomography (CT) examination with quantitative CT densitometry after control of sources of variation. Emphysema extent progress between 2009 and 2015 from 4% to 10% of the right lung volume and from 2% to 6% of the left lung volume.

158

SECTION A  The Chest and Cardiovascular System

(E/I-ratioMLD) has been shown to be the most suitable. As opposed to other gas trapping severity metrics, E/I-ratioMLD is not influenced by iterative reconstruction techniques. The percentage of low attenuation area at −856 or −850 HU at end-expiration CT has proved to provide remarkable high correlations with predicted forced expiratory volume (FEV)1% and FEV1/forced vital capacity (FVC) ratio. However, both these metrics are influenced by underlying emphysema. Patients with similar values of emphysema extent (percentage of low attenuation area 5 to ≤7 cm in greatest dimension or tumour of any size that involves chest wall, pericardium, phrenic nerve or satellite nodules in the same lobe >7 cm in greatest dimension or any tumour with invasion of mediastinum, diaphragm, heart, great vessels, recurrent, laryngeal nerve, carina, trachea, oesophagus, spine or separate tumour in different lobe of ipsilateral lung Ipsilateral peribronchial and/or hilar nodes and intrapulmonary nodes Ipsilateral mediastinal and/or subcarinal nodes Contralateral mediastinal or hilar; ipsilateral/contralateral scalene/supraclavicular Distant metastasis Tumour in contralateral lung or pleural/pericardial nodule/ malignant effusion Single extrathoracic metastasis, including single non-regional lymph node Multiple extrathoracic metastases in one or more organs

TNM 8th edition.

despite no change in size, it would instead be designated a T4 tumour. For the radiologist, the distinction between a T3 tumour and a T4 tumour requires careful thought in certain locations. For example, any invasion of the mediastinum is a T4 descriptor. Phrenic nerve involvement is a T3 descriptor, yet many radiologists would consider that the phrenic nerve lies within the mediastinum. However, the TNM 8th edition clearly states that the fibrous pericardium and phrenic nerve involvement indicates T3 stage. Assessment of pleural and chest wall invasion has long been recognised as a difficult task for CT analysis. Where there is obvious soft-tissue extension into the intercostal muscles or bone destruction, the issue is easily resolved. Subtle chest wall parietal pleural invasion is more difficult to define and magnetic resonance imaging (MRI), high-resolution targeted ultrasound and even diagnostic artificial pneumothorax may be helpful. Pancoast (superior sulcus) tumours are individually staged according to the involved tissues. For example, an apical tumour with parietal pleural involvement is defined as a T3 lesion, but a similar tumour extending into a vertebral body or involving subclavian vessels becomes

190

SECTION A  The Chest and Cardiovascular System

a T4 lesion. Because of the importance of this differentiation, the utility of multiplanar reformatted images on CT and targeted MRI examination has been highlighted.

Additional Pulmonary Nodules in the Presence of Lung Cancer In the 7th edition, nodules in the same lobe as the primary tumour conferred T3 status (unchanged in the 8th edition). A nodule in the ipsilateral lung but in a different lobe, whilst predicting a poor 5-year survival, confers a slightly better outlook than M1 disease of other types. Therefore, the combination of a primary lesion with a further nodule within an ipsilateral different lobe remains described as T4 disease in the 8th edition.

N Descriptors Nodes are described as either N0 (no involvement), N1 (nodes up to and including hilar stations), N2 (ipsilateral mediastinal nodes) or N3 (contralateral mediastinal or hilar nodes and supraclavicular or scalene nodes). Nodes beyond these stations are designated M1 nodes. The nodal mapping system used by the American Thoracic Society (ATS) is given in Fig. 8.8. Note the following: • Supraclavicular and sternal notch nodes are designated station 1. • There is a shift of the midline for designation of right and left level 2 and level 4 nodes such that the ‘midline’ lies at the left lateral border of the trachea. Therefore a lymph node lying directly anterior to the trachea using the 8th edition nodal map would be designated as a right paratracheal lymph node. • Station 7 is defined by the undersurface of the carina superiorly. The lower border of station 7 is the upper border of the lower lobe bronchus on the left and the lower border of the bronchus intermedius on the right. • Hilar lymph nodes are those adjacent to the mainstem bronchus and hilar vessels, including the proximal portions of the pulmonary veins and main pulmonary artery, with station 10R on the right and station 10L on the left. The upper border of station 10R is the lower rim of the azygos vein, and the upper border of station 10L is the upper rim of the pulmonary artery on the left. The lower borders of stations 10R and 10L are the interlobar regions bilaterally. For more detail, readers are referred to El-Sherief et al. 2013 (see further reading).

M Descriptors M disease is now divided into: • M1a, indicating additional tumour nodule(s) in the contralateral lung or pleural or pericardial nodules or malignant effusions. • M1b, indicating a single extrathoracic metastasis, including in a non-regional lymph node, and • M1c, indicating multiple extrathoracic metastases in one or more organs. These staging changes are related to more comprehensive examination of survival patterns, a direct result of the larger databases from which the 7th edition, and now the 8th edition, have been derived. The stage groups from the 8th TNM classification are shown in Table 8.4.

Small Cell Lung Cancer SCLC presents a different phenotype to NSCLC. It is relatively common (15%–20% of all lung cancers). The disease is characterised by rapid growth rate, early metastatic spread and an association with smoking. Characteristically, these tumours are initially responsive to radiation and chemotherapeutic treatment but are also associated with early recurrence. In the 7th edition of the lung cancer TNM staging, SLCLC was divided simply into limited disease and extensive disease groups.

Fig. 8.8  American Thoracic Society nodal map for use with the 8th tumour node metastasis staging classification for lung cancer staging. Supraclavicular nodes. 1. Low cervical, supraclavicular and sternal notch nodes. From the lower margin of the cricoid to the clavicles and the upper border of the manubrium. The midline of the trachea serves as border between 1R and 1L. Superior Mediastinal Nodes 2-4: 2R. Upper Paratracheal. 2R nodes extend to the left lateral border of the trachea. From upper border of manubrium to the intersection of caudal margin of innominate (left brachiocephalic) vein with the trachea. 2L. Upper Paratracheal. From the upper border of manubrium to the superior border of aortic arch. 2L nodes are located to the left of the left lateral border of the trachea. 3A. Pre-vascular. These nodes are not adjacent to the trachea like the nodes in station 2, but they are anterior to the vessels. 3P. Pre-vertebral. Nodes not adjacent to the trachea like the nodes in station 2, but behind the esophagus, which is prevertebral. 4R. Lower Paratracheal. From the intersection of the caudal margin of innominate (left brachiocephalic) vein with the trachea to the lower border of the azygos vein. 4R nodes extend from the right to the left lateral border of the trachea. 4L. Lower Paratracheal. From the upper margin of the aortic arch to the upper rim of the left main pulmonary artery. Aortic Nodes 5-6: 5. Subaortic. These nodes are located in the AP window lateral to the ligamentum arteriosum. These nodes are not located between the aorta and the pulmonary trunk but lateral to these vessels. 6. Para-aortic. These are ascending aorta or phrenic nodes lying anterior and lateral to the ascending aorta and the aortic arch. Inferior Mediastinal Nodes 7-9. 7. Subcarinal. 8. Paraesophageal. Nodes below carina. 9. Pulmonary Ligament. Nodes lying within the pulmonary ligaments. Hilar, Lobar and (sub)segmental Nodes 10-14: These are all N1-nodes. 10. Hilar nodes. These include nodes adjacent to the main stem bronchus and hilar vessels. On the right they extend from the lower rim of the azygos vein to the interlobar region. On the left from the upper rim of the pulmonary artery to the interlobar region. (Source: Redrawn from Radiology Assistant. https://radiologyassistant. nl/chest/lung-cancer-tnm-8th-edition)

CHAPTER 8  Pulmonary Neoplasms

TABLE 8.4  Stage Groupings From the 8th

Edition of the TNM Staging Scheme T1 T2a T2b T3 T4 M1a M1b M1c

N0 IA IB IIA IIB IIIA IVA IVA IVB

N1 IIB IIB IIB IIIA IIIA IVA IVA IVB

N2 IIIA IIIA IIIA IIIB IIIB IVA IVA IVB

N3 IIIB IIIB IIIB IIIC IIIC IVA IVA IVB

TNM 8th edition.

Limited disease indicates disease confined to one hemithorax but includes contralateral mediastinal and supraclavicular nodes and malignant pleural effusions. Patients with disease beyond these parameters are described as having extensive disease. Patients with limited disease typically receive chemotherapy and possibly radiotherapy. Patients with extensive disease will have chemotherapy alone. It is possible to stage SCLC using the same TNM system utilised for NSCLC, but this is not likely to be important for decision-making purposes. Surgery is sometimes an option for SCLC in localised disease.

Bronchopulmonary Carcinoid Tumour Bronchopulmonary carcinoid tumours are staged in the same way as NSCLC and is also classified under the 8th edition of the American Joint Commission on Cancer (AJCC) TNM staging system. Carcinoid tumours are potentially malignant neuroendocrine tumours. The spectrum of disease ranges from low-grade typical carcinoids, through atypical carcinoids to higher-grade large cell and small cell carcinomas. The distinction between these neuroendocrine tumours is based on pathological analysis. The field is also slightly complicated by the relatively small numbers available for analysis and the phenomenon of preinvasive lesions seen in diffuse idiopathic pulmonary neuroendocrine-cell hyperplasia (DIPNECH), to be distinguished from genuine metastatic disease. The presence of multiple small nodules of less than 5 mm in size in association with mosaic attenuation pattern within the lungs in the setting of a known carcinoid tumour should raise the possibility of DIPNECH. Bronchial carcinoids are uncommon, constituting less than 5% of pulmonary tumours. The peak age at diagnosis is in the fifth decade, but the age range is wide and includes children. Two forms of bronchial carcinoid are described: typical (85% to 90%) and atypical (10% to 15%). Typical carcinoids most commonly arise in central airways. Atypical carcinoids usually arise in the lung periphery. Bronchial carcinoids can invade locally and may metastasise to hilar and mediastinal lymph nodes as well as to the brain, liver and bone. The atypical carcinoids have histological and clinical features intermediate between typical bronchial carcinoid and small cell carcinoma of the lung and have a poorer prognosis. Bronchial carcinoid may present with wheeze, pneumonia or haemoptysis. Even when small, tumours may secrete adrenocorticotrophic hormone (ACTH) in sufficient quantities to cause Cushing syndrome. Carcinoid syndrome is very rare if the tumour is still confined to the lung. Radiographic appearances vary with location of the tumour. There is no lobar predilection and on rare occasions carcinoids may arise in the trachea. Bronchial carcinoids, particularly those located centrally, may calcify and occasionally ossify. Calcification is seen on CT in up to one-third of cases, but is only occasionally visible on chest radiography. Marked contrast enhancement may be seen on CT.

191

Carcinoids arising in central bronchi (80%–90% of cases) often show a larger mass external to the bronchus than within the lumen (so-called ‘iceberg’ lesions), and the extrabronchial component may be visible as a hilar mass (Fig. 8.9). Central lesions usually produce partial or complete bronchial obstruction, resulting in atelectasis with or without pneumonia. Central bronchial obstruction may be complicated by development of distal bronchiectasis or lung abscess. Occasionally, a bronchial carcinoid in a segmental or subsegmental bronchus may obstruct bronchial secretions, thereby causing a mucocele. Peripheral lesions (10%–20% of carcinoids) present as solitary spherical or lobular nodules, 2–4 cm in diameter, with a well-defined smooth edge (see Fig. 8.9). Non-calcified peripheral bronchial carcinoid tumours closely resemble bronchial carcinomas, both radiologically and cytologically, and are therefore frequently removed surgically in the belief that they are carcinomas.

Summary The use of the 8th edition of the AJCC TNM staging system in NSCLC, SCLC and bronchopulmonary carcinoid informs the stage grouping system that guides treatment choices. Since the adoption of the 8th edition TNM descriptors, the group staging has become more complex. These stage groupings are of relatively little importance for the thoracic radiologist in day-to-day practice but will have a significant effect in some patients in determining treatment options and trial eligibility.

IMAGING PROTOCOLS FOR LUNG CANCER STAGING Clinical features vary with cell type and extent of disease. Approximately 25% of patients are asymptomatic at the time of diagnosis, following the discovery of an abnormality on a chest radiograph or CT. Pneumonia is another common presentation. Cough, wheeze, haemoptysis, symptoms of pneumonia and paraneoplastic syndromes, such as the inappropriate secretion of antidiuretic hormone or a peripheral neuropathy, are the cardinal symptoms at a stage when lobectomy or pneumonectomy may be curative. Hoarseness, chest pain, brachial plexus neuropathy and Horner syndrome (Pancoast tumour), superior vena caval obstruction, dysphagia and the problems of pericardial tamponade indicate invasion of the mediastinum or chest wall and a much poorer prognosis. The chest radiograph will remain the initial investigation in all patients suspected of lung cancer. As the lung cancer screening studies have shown, the ability of a chest radiograph to detect all lung cancers is distinctly limited. Therefore the mainstay of staging investigation for a patient with suspected lung cancer (in distinction to patients undergoing lung cancer screening) is contrast-enhanced CT supplemented by PET/ CT and MRI when required. In certain circumstances ultrasound assessment of peripheral tumours and supraclavicular lymph nodes adds further useful information.

The Current Standards of Computer Tomography Technology The current standard of CT technology is a multislice (usually at least 64 detector rows) CT system, able to acquire submillimetre collimation images through the thorax in a short breath-hold. The decision to include abdominal, pelvic and intracranial assessment varies from centre to centre. However, a comprehensive brain, chest, abdomen and pelvis acquisition can be undertaken in very short order and the main rate-limiting steps are now patient identification, preparation and documentation rather than the acquisition of the data sets. Most institutions will routinely use intravenous contrast agents, although this is not mandatory. Not infrequently, difficulty with venous access, asthma, allergy or previous contrast reaction and impaired renal function will prevent contrast

192

SECTION A  The Chest and Cardiovascular System

A

B

C

D Fig. 8.9  Carcinoid Tumour. (A) A tumour is partially occluding the left main bronchus. (B) A well-defined perihilar carcinoid tumour (arrows) is demonstrated anterior to the artery to the right lower lobe. (C) On lung windows there is only a small band of atelectasis in the middle lobe. (D) A small peripheral carcinoid tumour indistinguishable from a number of other causes of a solitary pulmonary nodule.

administration. A standard thoracic CT will be undertaken with the patient in the supine position and the arms elevated. Imaging planning and dose reduction optimisation require an anteroposterior (AP), and sometimes a lateral scout projection. If contrast medium is administered, then image acquisition is timed to optimise opacification of central pulmonary vasculature (usually 20–30 seconds). If an abdominal and pelvic study is also undertaken, this component is best acquired during the portal venous phase of contrast enhancement (65 seconds). This will require two short breath-holds and two pre-planned acquisition ranges with some overlap at the lung bases. Alternative injection protocols combined with fast CT scanners now allow a single volumetric acquisition with both arterial and portal venous phase enhancement simultaneously. The ability to produce isotropic voxels allows multiplanar reformatting to be undertaken as a routine, either by the radiographic staff or at the time of reporting by the radiologist.

The PET/CT technique is similar for the CT acquisition of the study, which is usually obtained from skull base to upper thigh. If the patient is also undergoing a conventional CT, the CT acquisition for PET coregistration can be low dose and unenhanced. As the PET component of the acquisition takes considerably longer than the CT acquisition, in this situation the unenhanced low-dose CT will usually be undertaken during gentle respiration, to allow optimum co-registration with the PET data. Usually the PET acquisition takes between 5 and 7 bed couch positions, with each position taking up to 5 minutes, and therefore the whole study may take up to 35 minutes to acquire. More modern systems achieve the entire study in considerably less time. MRI of the thorax is usually undertaken to answer a particular question as a problem-solving tool. In the context of lung cancer staging, the MRI study is often to assess superior sulcus tumours, chest wall or thoracic invasion or to assess the integrity of the diaphragm. Usually

CHAPTER 8  Pulmonary Neoplasms

A

193

B

Fig. 8.10  (A, B) Bronchial carcinoma in the left lower lobe showing typical rounded, slightly lobular configuration. The mass shows a notch posteriorly.

triplanar examinations are untaken with respiratory gating and T1 and T2 weighting. A variety of more refined techniques, including dynamic contrast-enhanced MRI sequences for the evaluation of lung nodules as well as use of diffusion-weighted imaging, are commonplace. Diffusionweighted imaging can differentiate tumour from surrounding lung and may be helpful in assessing response to treatment before clear dimensional changes are evident. Dynamic cine acquisitions can be utilised to assess fixation of a peripheral tumour to chest wall or mediastinal structures.

IMAGING FEATURES OF BRONCHOGENIC CARCINOMA The thoracic imaging features of bronchial carcinoma are discussed under three headings: (1) peripheral tumours; (2) central tumours (arising in a large bronchus at or close to the hilum); and (3) staging intrathoracic spread of bronchial carcinoma.

Peripheral Tumours Approximately 40% of bronchial carcinomas arise beyond the segmental bronchi, and in 30% a peripheral mass is the sole radiographic finding (Fig. 8.10).

Tumour Shape and Margins Tumours at the lung apex (Pancoast tumours, superior sulcus tumours) may resemble apical pleural thickening; however, most peripheral lung cancers are approximately spherical or oval in shape. Lobulation, a sign that indicates uneven growth rates in different parts of the tumour, is common. Occasionally, a dumb-bell shape is encountered or two nodules are seen next to one another. The term ‘corona radiata’ is used to describe numerous fine strands radiating into the lung from a central mass, sometimes with transradiant lung parenchyma between these strands. While not specific, this sign is highly suggestive of bronchial carcinoma (Fig. 8.11). Absolutely spherical, sharply defined, smooth-edged nodules due to carcinoma of the lung are rare. A peripheral line shadow or ‘tail’ may be seen between a peripherally located mass lesion and the pleura, a phenomenon that occurs in both benign and malignant lesions. When associated with carcinoma of the lung, the ‘tail’ probably represents either plate-like atelectasis secondary to bronchial obstruction beyond the mass or septal oedema due to lymphatic obstruction.

Fig. 8.11  Computed tomography demonstrating a second primary bronchogenic carcinoma in the right lung. The patient had undergone a previous left pneumonectomy 7 years earlier. The new tumour has spiculated edges, infiltrating into the adjacent lung (corona radiata).

Although the edges of a tumour are frequently well defined, some peripheral cancers, notably some types of adenocarcinoma, have illdefined edges similar to pneumonia (Fig. 8.12).

Cavitation Cavitation may be identified in tumours of any size (Fig. 8.13) and is best demonstrated by CT (Fig. 8.14). Squamous cell carcinoma is the most likely cell type to show cavitation. The walls of the cavity are of irregular thickness and may contain tumour nodules, but sometimes the wall has smooth inner and outer margins. The cavity wall is usually 8-mm thick or greater. Fluid levels are common.

Calcification Calcification within bronchogenic carcinomas is rarely seen on chest radiography but is identified on CT in 6% to 10% of cases. Some foci of calcification represent pre-existing calcified granulomatous disease engulfed by tumour (Fig. 8.15). However, amorphous or cloud-like calcification consistent with dystrophic tumour calcification is still seen in a small proportion of cases (5 cm) often show irregular low-attenuation

Fig. 9.28  Progressive Massive Fibrosis in Coal Worker’s Pneumoconiosis. Mass-like opacities are seen bilaterally in the upper lobes in association with multiple small nodules that cluster into pseudo-plaques (arrow).

regions on CT indicative of necrosis. Frank cavitation is a less frequent finding; it should always be put in differential with tuberculosis (conventional or atypical). Unilateral or asymmetric PMF may be distinguished from lung cancer by the presence of lobar volume loss and peripheral emphysema.

Classical HRCT Findings • Well-defined dense centrilobular and subpleural nodules, upper and posterior lung predominance • Hilar and mediastinal lymph node enlargement, with or without calcification • Upper lobe irregular masses characterise PMF

Asbestos-Related Disease Asbestos is the generic term for a group of fibrous silicates that share the property of heat resistance. They are classified into two groups: the serpentines and the amphiboles. The only serpentine asbestos used commercially is chrysolite, which accounts for more than 90% of the asbestos in the USA. The pathological hallmark of asbestos exposure

228

SECTION A  The Chest and Cardiovascular System

is the asbestos body consisting of a fibre (2–5 µm in width) that is inhaled and accumulated in the lung. These bodies can be identified in tissue sections in interstitial fibrous tissue and intra-alveolar macrophages in bronchoalveolar lavage (BAL). The effects of asbestos on the lung are diverse. Clinical onset of disease is usually delayed by some 20 years or more after initial exposure. The exception is asbestos-related pleural effusions, which may be present as early as 5 years after first exposure.

Benign Pleural Effusions The exact prevalence of benign pleural effusions is unknown, as many are subclinical. The effusions are typically haemorrhagic exudates of mixed cellularity and usually do not contain asbestos bodies. Their diagnosis is mostly made on exclusion of other causes. The development of effusions is thought to be exposure-dependent. The effusions are often small, may be persistent or recurrent, and may be simultaneously or sequentially bilateral. Diffuse pleural thickening, with or without areas of subjacent folded lung, is the usual consequence.

A

Pleural Plaques The most common manifestation of asbestos exposure is the presence of pleural plaques that, macroscopically, are discrete foci of pearly white fibrous tissue (2–5 mm thick). They involve the parietal pleura almost exclusively and are classically distributed under the anterior ends of the upper ribs, the paravertebral gutters and the diaphragmatic surface. Calcification is reported in 10%–15% of cases. CT is undoubtedly more sensitive for the detection of pleural plaques. Only 50%–80% of cases of documented pleural thickening are detected by chest radiography; on chest radiography, pleural plaques were most commonly missed in the paravertebral and posterior regions of the costal pleura. Studies have suggested that pleural plaques are not associated with significantly impaired lung function.

Diffuse Pleural Thickening The frequency of diffuse pleural thickening increases with time from first exposure and is thought to be dose-related. It results from thickening and fibrosis of the visceral pleura, which leads to fusion with the parietal pleura and may be caused by extension of interstitial fibrosis to the visceral pleura, consistent with the pleural migration of asbestos fibres. Diffuse pleural thickening superimposed on circumscribed plaques has been observed, often after a pleural effusion. CT is more sensitive and specific than chest radiography in the detection of diffuse pleural thickening and can make the distinction between mild pleural disease and extra-pleural fat.

Round Atelectasis Round atelectasis, also known as folded lung, is a form of parenchymal collapse that occurs in association with pleural thickening, most commonly in the peripheral lung in the dorsal regions of the lower lobes. Pathological examination shows pleural fibrosis overlying the abnormal parenchyma as well as invaginations of fibrotic pleura into the region of collapse. Because of the pathogenetic association with fibrosis, the areas of atelectasis are always seen adjacent to the visceral pleura. A characteristic finding is the presence of crowding of bronchi and blood vessels that extend from the border of the mass to the hilum (‘comet tail’ sign). In most cases, the collapsed lung has a rounded or oval shape; however, wedge- and irregularly-shaped masses can also occur (Fig. 9.29). Volume loss of the affected lobe is a key sign. Serial examinations show a relatively stable appearance, and the differentiation from a lung neoplasm is usually straightforward on CT by characteristic findings such as the subpleural location, comet tail sign and the strong and homogeneous

B Fig. 9.29  Round Atelectasis. High-resolution computed tomography in axial (A) and sagittal reformation (B) show the subpleural consolidation and the curve bronchovascular bundle (‘comet tail’) towards the atelectasis.

enhancement after intravenous contrast medium, the latter indicative of atelectasis rather than tumour.

Asbestosis Asbestosis is defined as pulmonary parenchymal fibrosis secondary to inhalation of asbestos fibres. The time lag between exposure and onset of symptoms is usually 20 years or longer. Histologically, fibrosis is first seen in the interstitium of respiratory bronchioles, particularly in the lower lobes adjacent to the visceral pleura. With advancing disease, the fibrous tissue extends into the adjacent alveolar septa, eventually involving the entire lobule. In the most severe cases there is diffuse interstitial fibrosis associated with parenchymal remodelling and honeycombing. Asbestos bodies are almost always identifiable microscopically in the fibrous tissue or macrophages in residual air spaces. Early CT changes

CHAPTER 9  High-Resolution Computed Tomography of Interstitial and Occupational Lung Disease

229

history and exclusion of other plausible conditions associated with pulmonary fibrosis. It is noteworthy that subjects exposed to asbestos likely have risk factors for IPF (male gender, smoking habit and age). Distinguishing asbestosis from IPF is also desirable, as asbestosis is associated with a much slower rate of progression and hence a better prognosis. Although it may be difficult to differentiate between asbestosis and IPF on HRCT, the presence of pleural disease may provide a pointer: Akira et al. reported pleural disease in 83% (66/80) of patients with asbestosis but only in 4% (3/80) of patients with IPF. Copley et al. found no statistically significant differences in the coarseness of fibrosis between individuals with asbestosis and a cohort of individuals with biopsy-proven UIP, although the CT findings of asbestosis were strikingly different from NSIP; the quality of fibrosis was coarser, there was a lower proportion of ground-glass opacification, and a higher likelihood of a basal and subpleural distribution.

Classical HRCT Findings A

• Asbestos-related benign pleural disease consists of either parietal pleural plaques in characteristic locations or diffuse visceral pleural thickening • The presence of such pleural disease in individuals with HRCT findings compatible with UIP/NSIP pattern suggests the diagnosis of asbestosis in patients with an appropriate exposure history.

FURTHER READING

B Fig. 9.30  Asbestosis. (A) High-resolution computed tomography (HRCT) features of early asbestosis include subpleural lines (arrowheads) and fine reticulation (arrows). These subtle abnormalities persisted on prone sections. (B) In more advanced disease, a coarse reticular pattern with honeycombing, often indistinguishable from usual interstitial pneumonia on HRCT, is seen in the left lower lobe. Note the calcified pleural plaques in both examples.

indicative of asbestosis are the presence of subpleural curvilinear lines and dots, pleural-based nodular irregularities, parenchymal bands and septal lines. The fine reticulation eventually progresses to a coarse linear pattern with honeycombing (Fig. 9.30). These abnormalities are usually most severe in the subpleural regions of the lower lobes. HRCT– pathological correlation studies have shown that subpleural dots and branching structures correspond to peribronchiolar fibrosis. The sensitivity of HRCT over the chest radiograph for the identification of early fibrosis in asbestos-exposed individuals is well established; however, sensitivity is not 100% and a histopathological diagnosis of asbestosis can be present in patients with normal or near-normal HRCTs. The diagnosis of asbestosis is prognostically relevant, including work ability and the possibility of receiving legal compensation. Although both the chest radiograph and HRCT can confirm previous exposure, the diagnosis of asbestosis is largely inferential from critical investigation of exposure

Akira, M., 2008. Imaging of occupational and environmental lung diseases. Clin. Chest Med. 29 (1), 117–131, vi. Akira, M., Yamamoto, S., Inoue, Yoshikazu, et al., 2003. AJR Am. J. Roentgenol. 181 (1), 163–169. doi:10.2214/ajr.181.1.1810163. Brownell, R., Moua, T., Henry, T.S., et al., 2017. The use of pretest probability increases the value of high-resolution CT in diagnosing usual interstitial pneumonia. Thorax 72 (5), 424–429. Castaner, E., Alguersuari, A., Gallardo, X., et al., 2010. When to suspect pulmonary vasculitis: radiologic and clinical clues. Radiographics 30 (1), 33–53. Chung, J.H., Chawla, A., Peljto, A.L., et al., 2015. CT scan findings of probable usual interstitial pneumonitis have a high predictive value for histologic usual interstitial pneumonitis. Chest 147 (2), 450–459. Egashira, R., Jacob, J., Kokosi, M.A., et al., 2017. Diffuse pulmonary ossification in fibrosing interstitial lung diseases: prevalence and associations. Radiology 284 (1), 255–263. Fell, C.D., Martinez, F.J., Liu, L.X., et al., 2010. Clinical predictors of a diagnosis of idiopathic pulmonary fibrosis. Am. J. Respir. Crit. Care Med. 181 (8), 832–837. Fischer, A., Antoniou, K.M., Brown, K.K., et al., 2015. An official European Respiratory Society/American Thoracic Society research statement: interstitial pneumonia with autoimmune features. Eur. Respir. J. 46 (4), 976–987. Hansell, D.M., 2010. In: Hansell, David M., et al. (Eds.), Imaging of Diseases of the Chest, fifth ed. Mosby, Edinburgh. Hansell, D.M., Bankier, A.A., MacMahon, H., et al., 2008. Fleischner Society: glossary of terms for thoracic imaging. Radiology 246 (3), 697–722. Jennette, J.C., Falk, R.J., Bacon, P.A., et al., 2013. 2012 revised International Chapel Hill Consensus Conference Nomenclature of Vasculitides. Arthritis Rheum. 65, 1–11. Lynch, D.A., Sverzellati, N., Travis, W.D., et al., 2017. Diagnostic criteria for idiopathic pulmonary fibrosis: a Fleischner Society White Paper. Lancet Respir. Med. Nishino, M., Hatabu, H., Sholl, L.M., et al., 2017. Thoracic complications of precision cancer therapies: a practical guide for radiologists in the new era of cancer care. Radiographics 37 (5), 1371–1387. Raghu, G., Lynch, D., Godwin, J.D., et al., 2014. Diagnosis of idiopathic pulmonary fibrosis with high-resolution CT in patients with little or no

230

SECTION A  The Chest and Cardiovascular System

radiological evidence of honeycombing: secondary analysis of a randomised, controlled trial. Lancet Respir. Med. 2 (4), 277–284. Raghu, G., Remy-Jardin, M., Myers, J.L., et al., 2018. Diagnosis of idiopathic pulmonary fibrosis. An official ATS/ERS/JRS/ALAT clinical practice guideline. Am. J. Respir. Crit. Care Med. 198 (5), e44–e68. Raghu, G., Wells, A.U., Nicholson, A.G., et al., 2017. Effect of nintedanib in subgroups of idiopathic pulmonary fibrosis by diagnostic criteria. Am. J. Respir. Crit. Care Med. 195 (1), 78–85. Reddy, T.L., Tominaga, M., Hansell, D.M., et al., 2012. Pleuroparenchymal fibroelastosis: a spectrum of histopathological and imaging phenotypes. Eur. Respir. J. 40 (2), 377–385. Schwarz, M.I., King, T.E., Jr, eds. 2011. Interstitial Lung Disease, fifth ed. People’s Medical Publishing House, Shelton, CT, McGraw-Hill [distributor].

Silva, M., Nunes, H., Valeyre, D., et al., 2015. Imaging of sarcoidosis. Clin. Rev. Allergy Immunol. 49 (1), 45–53. Sverzellati, N., Lynch, D.A., Hansell, D.M., et al., 2015. American Thoracic Society-European Respiratory Society Classification of the Idiopathic Interstitial Pneumonias: Advances in Knowledge since 2002. Radiographics 35 (7), 1849–1871. Travis, W.D., Costabel, U., Hansell, D.M., et al., 2013. An official American Thoracic Society/European Respiratory Society statement: update of the international multidisciplinary classification of the idiopathic interstitial pneumonias. Am. J. Respir. Crit. Care Med. 188 (6), 733–748. Webb, W.R., Müller, N.L., Naidich, D.P., 2008. High-Resolution CT of the Lung. Philadelphia.

11  Airspace Diseases Gurinder S. Nandra, Sujal R. Desai

CHAPTER OUTLINE Introduction, 250 Suggested Approach to the Radiological Diagnosis of Airspace Diseases, 250 Pulmonary Oedema, 251

INTRODUCTION Diseases of the air spaces are remarkably common, yet, the radiological approach to diagnosis is often considered challenging. In part, this is because a pattern of airspace opacification is non-specific (Table 11.1). However, at its simplest, this radiological pattern simply indicates that air has been displaced, to a greater or lesser degree, from the lung. In clinical practice, airspace opacification is most commonly a manifestation of pulmonary oedema or infection. This chapter considers not only some of the common but also a few of the more unusual causes of airspace opacification in clinical practice. Airspace diseases caused by infection and cancer are considered in detail elsewhere.

SUGGESTED APPROACH TO THE RADIOLOGICAL DIAGNOSIS OF AIRSPACE DISEASES The plain chest radiograph (CXR) is usually the first imaging test requested by clinicians. In patients with an abnormal CXR, radiologists should aim to formulate a sensible diagnosis or, at most, a short list of differential diagnoses. To this end, the radiologist must pay heed to the following: the clinical background, the distribution of radiographic abnormalities and serial changes (i.e. progression/resolution, time course etc.) on repeat studies where available. Clinical context is, of course, important when reporting imaging studies. For instance, the most likely cause of lobar consolidation in a patient with pyrexia and a productive cough is infection, whereas the same radiological pattern (only bilateral) in a critically ill patient is most likely to indicate noncardiopulmonary oedema/acute respiratory distress syndrome (ARDS) (Fig. 11.1). The distribution of airspace opacities on imaging studies can also provide important clues: in cryptogenic organising pneumonia (COP), for instance, areas of consolidation tend to be most obvious in the periphery and lower zones. By contrast, upper zone infiltrates parallel to the chest are typical in chronic eosinophilic pneumonia (Fig. 11.2). A review of serial radiographs should be considered de rigeur in the radiologist’s routine. Rapid clearing—occurring over a period of hours or, at most, a few days—suggests oedema fluid (Fig. 11.3)

250

Diffuse Pulmonary Haemorrhage, 256 Organising Pneumonia, 258 Pulmonary Alveolar Proteinosis, 263 Alveolar Microlithiasis, 263

or pulmonary haemorrhage as the likely cause as opposed to, say, pneumonia. Opacities that are transient and migratory in a patient with constitutional symptoms should make the radiologist consider an eosinophilic pneumonia in the differential diagnosis. Computed tomography (CT) is frequently requested in patients with airspace disease and, occasionally, the CT features will be helpful; the so-called ‘crazy-paving’ pattern is an example that immediately comes to mind, which, at least in its classical form, should be considered pathognomonic of pulmonary alveolar proteinosis (PAP). In other instances, the radiologist may only be able to limit the list of diagnostic possibilities despite the additional information from CT (e.g. cavitation that may not have been evident on plain radiographs). Therefore, except in certain circumstances, the advantages of CT over plain radiography in the diagnosis of airspace diseases are not clearly defined.

Anatomical Considerations The air spaces are defined as the air-containing part of the lung, which includes the respiratory bronchioles but excludes the terminal bronchioles; the latter are the last purely conducting airways of the bronchial tree and the region of lung subtended by a terminal bronchiole is the acinus. Important pathways of collateral ventilation (the pores of Kohn) link different alveolar units and maintain lung inflation in the presence of proximal airway obstruction. These normal collateral pathways also facilitate the spread of certain diseases (most notably infections) into adjacent alveolar units. An important unit of lung structure is the pulmonary lobule, defined as the smallest unit of lung bounded by connective tissue septa. Individual lobules are irregular polyhedrons, best seen in the subpleural lung and measuring between 5 and 30 mm in diameter, incorporating between 3 and 24 acini. The lobular bronchiole and accompanying artery form the core structures. Normal centrilobular arteries (with a maximum diameter of 0.2 mm) can be resolved on high-resolution computed tomography (HRCT), but the wall of the accompanying bronchiole is too thin to be seen (Fig. 11.4). The implication is that when bronchioles are visible within 2 cm of the subpleural space (either because of wall thickening, dilatation and/or mucous plugging of the lumen) there is disease. Infiltration of the interlobular septa by oedema fluid or malignant

CHAPTER 11  Airspace Diseases

251

TABLE 11.1  Causes of Airspace

Opacification

• Cardiogenic oedema (cardiogenic, non-cardiogenic) • Pneumonia • Inflammatory • Organising pneumonia (cryptogenic or other aetiologies) • Diffuse alveolar damage • Vasculitides (e.g. Wegener’s granulomatosis) • Haemorrhage • Idiopathic pulmonary haemosiderosis • Antibasement membrane antibody disease • Systemic lupus erythematosus • Neoplasm • Adenocarcinoma with lepidic growth pattern (former bronchoalveolar cell carcinoma) • Lymphoma (MALToma) • Miscellaneous causes • Eosinophilic pneumonia • Alveolar proteinosis • Alveolar microlithiasis • Sarcoidosis Fig. 11.2  Chronic Eosinophilic Pneumonia. Coronal CT reconstruction showing the typical peripheral and upper zone distribution of infiltration.

Fig. 11.1  Chest Radiograph in a Patient With Acute Respiratory Distress Syndrome. There is bilateral airspace opacification with striking air bronchograms.

cells, or thickening caused by fibrosis, will also render individual pulmonary lobules visible on HRCT (Fig. 11.5).

Radiological Signs of Airspace Disease One of the principal limitations of imaging airspace diseases is that a multitude of pathological processes manifest as a limited number of patterns; thus, for most airspace diseases, a nodular pattern, ground-glass opacification and consolidation represent the range of radiological abnormalities. 1. A nodular pattern as a sole manifestation of airspace disease is relatively uncommon. The term ‘acinar nodules’ or ‘acinar rosettes’ has been used in the past to describe the appearance of poorly defined infiltrates on a CXR and HRCT. However, the diagnostic value of localising disease to the acinus is questionable; in pathological studies, the acinar pattern on plain radiographs, as described in radiology reports,

does not generally correspond to the filling of acini as per strict anatomical definitions. This notwithstanding, the so-called acinar pattern is most frequently encountered in the context of bacterial infection or pulmonary haemorrhage (Fig. 11.6). 2. Ground-glass opacification is a relatively common sign that can reflect airspace disease. On plain radiography, ground-glass opacification is seen as hazy, increased lung opacity in which the margins of pulmonary vessels are obscured. Because of the greater contrast resolution, ground-glass opacification on CT appears as a hazy increase in lung attenuation but without obscuration of bronchial and vascular markings (Fig. 11.7). It is important to remember that ground-glass opacification can be a manifestation of airspace and/ or interstitial disease (Fig. 11.8). Sometimes, particularly when there is diffuse disease, ground-glass opacification on CT may be subtle and barely perceptible. In such cases, a noticeable difference between the density of air in the lumen of an airway and that in the adjacent lung (the ‘black bronchus’ sign) (Fig. 11.9) might be the clue needed to confirm the suspicion of lung infiltration: in the normal lung, the two densities will be roughly equal. 3. Consolidation refers to the increase in lung density on a CXR or CT in which the margins of vessels and airways are obscured (Fig. 11.10). An air bronchogram may or may not be seen. This radiological pattern indicates that air in the air spaces has been replaced (e.g. by inflammatory cells, blood or tumour). In some patients, the distribution of consolidation in relation to the pulmonary lobule is an important diagnostic pointer: a perilobular distribution in which there is dense opacification apparently ‘smeared’ around the lobule is a characteristic finding in organising pneumonia (Fig. 11.11).

PULMONARY OEDEMA Pulmonary oedema—defined as an excess of extravascular lung water—is caused either by an increase in hydrostatic pressure (sometimes termed ‘cardiogenic’ oedema) or increased vascular permeability (or ‘noncardiogenic’ oedema) (Table 11.2). However, despite the attraction of simplicity, the clinical utility of this dichotomous classification of

252

SECTION A  The Chest and Cardiovascular System

A

C

Fig. 11.4  Magnified View of Multiple Adjacent Pulmonary Lobules on Thin-Section CT. The targeted image clearly demonstrates the interlobular septa (comprising pulmonary venules, lymphatics and interstitium) and the centrilobular arteriole (arrow) but because the wall of the homologous airway is too thin, it is not resolved on CT.

B

Fig. 11.3  (A to C) Rapid Changes in Radiographic Appearances in Pulmonary Oedema. Serial chest radiographs over approximately a 48-hour period show rapid fluctuation in the extent/severity of airspace opacification, reflecting relatively rapid shifts of fluid between the intravascular compartment and the air spaces/ interstitium.

Fig. 11.5  Thickened Interlobular Septa in a Patient With Diffuse Pulmonary Lymphangiomatosis. There are smoothly thickened interlobular septa, delineating pulmonary lobules, in the upper lobes.

CHAPTER 11  Airspace Diseases

253

A

Fig. 11.7  Diffuse Ground-Glass Opacification on CT in Pneumocystis jiroveci Infection. There is widespread ground-glass opacification in both lungs. There are two simple subpleural air cysts (on in each lung) and note is made of the pneumothorax on the right.

TABLE 11.2  Computed Tomography Signs

of Pulmonary Oedema

Common Findings • Ground-glass opacification (patchy or diffuse) ± consolidation • Smooth interlobular septal thickening • Peribronchovascular thickening • Vascular dilatation B Fig. 11.6  Nodular Airspace Opacities on CXR and CT. (A) Widespread acinar nodules in a patient with disseminated tuberculosis. (B) Multiple ‘soft’ ground-glass opacities on CT also reflect filling of acini with inflammatory/infective exudate in a patient with cough and pyrexia.

pulmonary oedema is debatable. That said, hydrostatic oedema occurs when there is a shift of fluid out of the vascular compartment caused by an increase in venous/capillary pressure. Perhaps the commonest cause of increased hydrostatic pressure is left heart failure. A reduction in plasma osmotic pressure (as in hypoalbuminaemic patients) will have the same effect. Non-cardiogenic pulmonary oedema occurs in conditions where the permeability of the alveolar-capillary barrier is increased. The archetypal example of increased permeability oedema is ARDS.

Chest Radiography in Pulmonary Oedema Plain CXR is undoubtedly more sensitive than clinical examination for the early detection of pulmonary oedema. At a pathological level, there is a roughly predictable sequence, with fluid first moving into the interstitium and then the alveoli. Accordingly, on CXRs, the signs of interstitial oedema generally precede frank airspace opacification. In the following sections, the radiographic features of pulmonary oedema

Ancillary Findings • Pleural effusions • Enlargement of mediastinal lymph nodes/‘hazy’ opacification of mediastinal fat (in heart failure)—reversible

are considered; for clarity, the vascular, interstitial and intra-alveolar changes are discussed separately.

Vascular Alterations The signs of raised pulmonary venous pressure on a CXR are well documented, although the mechanisms causing blood flow ‘redistribution’ are not entirely clear. Signs of vascular redistribution (from bases to apex), namely balanced flow or inverted flow, often suggest elevation of the pulmonary venous pressure (Fig. 11.12). Both vascular dilatation and redistribution are more appreciable in chronic or, at least, subacute left heart dysfunction. The ratio of the diameter of adjacent pulmonary arteries and bronchi seen end-on, particularly at the level of the upper lobes, is useful when judging whether vessels are abnormally enlarged.

Interstitial Oedema Thickening of the interlobular septa is a classical chest x-ray sign of interstitial oedema; prominent septal (Kerley B) lines, indicating fluid

254

SECTION A  The Chest and Cardiovascular System

A

Fig. 11.9  ‘Black Bronchus’ Sign on CT. Air in the segmental/segmental airways is of much lower attenuation than the air in the surrounding parenchyma.

B

Fig. 11.10  Consolidation on CT. Unenhanced CT in a patient with acute respiratory distress syndrome shows bilateral-dependent dense consolidation. Airways are rendered conspicuous because of surrounding consolidation but note that bronchial walls and vessels are not visible.

C Fig. 11.8  Variable Causes of Ground-Glass Opacification on CT Caused by Interstitial, Airspace and ‘Mixed’ Disease Processes. (A) Fibrotic non-specific interstitial pneumonia in a patient with an established connective tissue disease. Note the non-tapering subsegmental airways (‘traction bronchiectasis’) indicating fine fibrosis below the limits of CT resolution. (B) Acute intra-alveolar haemorrhage with widespread ‘bland’ ground-glass opacification and (C) pulmonary oedema causing subtle ground-glass opacification with smooth thickening of intralobular septa.

in the interlobular septa (typically 1–2 mm wide and 30–60 mm long), are only really seen in the subpleural lung, perpendicular to the pleural surface (Fig. 11.13). By contrast, Kerley A lines are longer (up to 80–100 mm), are occasionally angulated and cross the inner two-thirds of the lung in varying directions, but tend to point medially towards the hilum. In practice, Kerley A lines are more difficult to see. In left heart failure, septal lines become visible as they distend with extravascular fluid. Naturally, thickened oedematous septal lines will not be seen if neighbouring alveoli are also opacified. It should also be remembered that the demonstration of thickened interlobular septa is not diagnostic of pulmonary oedema; fibrosis and malignant infiltration (as in lymphangitis carcinomatosa) will also render interlobular septa visible. Another useful sign of interstitial oedema on frontal chest radiographs is peribronchial cuffing, in which the normally thin and well-defined wall of the airway becomes thickened and indistinct. A loss of conspicuity

CHAPTER 11  Airspace Diseases

255

Fig. 11.13  Thickening of Interlobular Septa/Kerley B Lines. CXR in a patient with left ventricular dilatation: there are multiple thin linear opacities, roughly perpendicular to the pleural surface, in the right mid/ lower zone. Fig. 11.11  Perilobular Consolidation in Organising Pneumonia. Targeted image of the left lower lobe with consolidation that outlines a number of adjacent pulmonary lobules. This distribution of perilobular consolidation is virtually pathognomonic for a pattern of organising pneumonia.

Fig. 11.14  Lamellar Effusion. There is exudation of fluid between the visceral pleural surface and the lung (red arrows).

Fig. 11.12  Upper Lobe Blood Diversion. Vessels in the upper zones in contrast with vessels in the lower lobes, which are barely perceptible. Note also that the heart is enlarged.

of the central pulmonary vessels (termed a perihilar haze) also occurs and, as with peribronchial cuffing, is believed to be caused by oedema of the perivascular interstitium. Oedema fluid can also collect in the potential space between the visceral pleura and lung; on a CXR this may be seen as thickening of the interlobar fissures or as a lamellar

‘effusion’ in the costophrenic recesses (Fig. 11.14). The latter (admittedly something of a misnomer) indicates fluid between the lung and visceral pleura.

Alveolar Oedema Airspace opacification becomes apparent on the CXRs as oedema fluid passes from the interstitium into the alveoli. The distribution of oedema is variable and bilateral opacification is the norm. However, an asymmetric distribution or oedema restricted to one lung on CXRs also occurs. Not infrequently, oedema fluid spares the apices and extreme lung bases. Sparing of the lung peripheries with involvement of the central lungs produces the so-called ‘bat’s wing’ distribution (Fig. 11.15).

256

SECTION A  The Chest and Cardiovascular System

Fig. 11.15  ‘Bat’s Wing’ Distribution of Pulmonary Oedema. There is bilateral airspace opacification, which spares the peripheries and the lung bases/apices.

As oedema progresses, opacities may coalesce to produce a general ‘white-out’ and an air bronchogram or alveologram may be seen. The density of airspace opacification caused by pulmonary oedema can change relatively quickly; indeed, the speed of change (i.e. sometimes over hours as opposed to days or weeks) is a useful pointer to the diagnosis of oedema fluid rather than another airspace disease. In specific settings, the radiographic signs of intra-alveolar oedema may be modified. For instance, as hinted above, the distribution of pulmonary oedema can vary with posture so that in patients lying on one side for a prolonged period, the dependent lung becomes more oedematous and there is unilateral airspace opacification. Coexisting diseases (e.g. fibrosis or emphysema) will also influence the distribution and appearance of oedema fluid.

Radiographic Differentiation of Cardiogenic and Non-Cardiogenic Oedema Making the distinction between hydrostatic and permeability oedema is of clinical value and, for this purpose, a CXR is certainly more reliable than physical examination. However, whether a CXR can consistently differentiate between cardiogenic and non-cardiogenic oedema is doubtful. Moreover, in clinical practice, both patterns of oedema not infrequently coexist. This notwithstanding, the distribution of blood flow (i.e. upper vs lower zone) and oedema (i.e. peripheral vs central) together with the width of the vascular pedicle may be discriminatory: in patients with hydrostatic oedema, upper lobe blood diversion is said to be more common. By comparison, in patients with non-cardiogenic oedema caused by ARDS, a minority show this inverted pattern. A peripheral distribution of oedema is uncommon in hydrostatic oedema, being seen more frequently in ARDS. The discriminatory value of signs of interstitial fluid accumulation and pleural effusions, when distinguishing cardiogenic from non-cardiogenic causes, is questionable. In summary, analysis of the radiographic pattern will sometimes allow a distinction to be made, but the inconsistency of radiographic signs suggests that radiographic distinction between the various forms of pulmonary oedema is unreliable.

Computed Tomography in Pulmonary Oedema Not surprisingly, CT is more sensitive to small changes in lung water than CXR: CT may demonstrate clinically ‘silent’ oedema or help to

differentiate oedema from other disease processes. The latter is of particular value in critically ill patients with multiple co-morbidities. The typical CT signs of pulmonary oedema are summarised in Table 11.2. However, the appearances of pulmonary oedema on CT can vary. Because of excellent contrast resolution, CT may detect abnormalities before the transudation of fluid into the interstitium and air spaces: in an animal model of fluid overload there was an increase in background lung attenuation, attributed to an expansion of intra-capillary volume. Vascular dilatation, particularly in the perihilar regions, may be observed in association with other CT abnormalities before the development of frank pulmonary oedema. The CT equivalent of radiographic upper lobe blood diversion may be seen with preferential dilatation of vessels in the anterior (non-dependent) lung. It is likely that the earliest detectable findings associated with enlarged vessels are scant, thickened interlobular septa and ground-glass opacification. Smooth septal lines are often limited to the lung apices reflecting engorged septal veins. With more florid transudation of oedema fluid, peribronchovascular cuffing, prominent interlobular septa, ground-glass opacification and consolidation become more obvious (Fig. 11.16). The absence of any lung parenchymal distortion and the more linear, smooth septal thickening should differentiate cardiogenic interstitial oedema from other causes including lymphangitis carcinomatosis and sarcoidosis. As with a CXR, the changes on CT are usually bilateral but, occasionally, confined to one lung, and the appearances may be modified by coexistent disease such as emphysema. A perihilar, ‘bat’s wing’ appearance is seen in some patients but it is important to stress that this is not specific and also occurs in other diseases including PAP or pulmonary haemorrhage. An important ancillary finding in congestive heart failure is the enlargement of mediastinal lymph nodes and a hazy increase in the attenuation of mediastinal fat. Overall, ~50% of patients with heart failure have nodal enlargement on CT and this rises as the ejection fraction falls. Enlarged lymph nodes may have blurred margins. With medical therapy, a significant reduction in the volume of nodes occurs, often within days.

DIFFUSE PULMONARY HAEMORRHAGE Bleeding into the lungs is a relatively common—albeit sometimes subclinical—event. It is known, for instance, that patients with pneumonia and lung cancer frequently aspirate blood into the air spaces. However, because the bleeding tends to be localised and there is often an established underlying cause, the diagnosis is generally straightforward. In addition to these more common clinical settings, there are numerous pulmonary haemorrhage syndromes characterised by diffuse intra-alveolar bleeding. The severity of haemorrhage is variable, ranging from small subclinical episodes to catastrophic, life-threatening haemorrhage. One scheme for classifying diffuse pulmonary haemorrhage (DPH) categorises the various syndromes according to the presence/absence of immunocompromise (Table 11.3). In immunocompetent subjects, DPH may be immunologically mediated (e.g. antiglomerular basement membrane disease), have a presumed immunological basis (e.g. systemic lupus erythematosus, Wegener’s granulomatosis), or be unrelated to immunological mechanisms (e.g. idiopathic pulmonary haemosiderosis (IPH), drug reactions). In immunocompromised patients, infection, tumours and blood dyscrasias account for most cases. The clinical presentation of the DPH syndromes also varies but many patients give a history of recurrent haemoptysis, dyspnoea and chronic cough. Non-specific clinical features include intermittent fever, headache, lethargy, basal crackles on auscultation and clubbing. On histopathological examination, alveolar blood and haemosiderin-laden macrophages are the cardinal findings. With repeated episodes there is

CHAPTER 11  Airspace Diseases

A

D

B

257

C

Fig. 11.16  Pulmonary Oedema on CT. (A) There is subtle ground-glass opacification, smooth thickening of multiple interlobular septa and peribronchovascular cuffing. Bilateral pleural effusions are also seen. (B) More conspicuous ground-glass opacification and consolidation as the severity of oedema increases. (C and D) Another patient with pulmonary oedema on (C) lung windows showing ground-glass opacifcation, thickened interlobular septa and bilateral pleural effusions and (D) soft-tissue windows with enlarged ‘reactive’ mediastinal lymph node enlargement.

thickening of alveolar septa, indicating fibrosis, which, in occasional patients, may be florid. The chest x-ray and CT appearances of DPH are fairly stereotypical and differentiation between different causes of DPH, on the basis of radiological findings alone, is not possible. Following acute bleeding the CXR is usually, but not invariably, abnormal. On chest x-ray there may be small acinar nodules or patchy consolidation and ground-glass opacification more pronounced in the perihilar region of the mid/lower zones. A useful diagnostic clue is that compared with other causes of widespread airspace opacification (but with the notable exception of pulmonary oedema), the changes of diffuse intra-alveolar haemorrhage will generally clear quickly (i.e. over a few days) as blood is removed by lung macrophages (Fig. 11.17). With repeated episodes, ill-defined nodular or reticulonodular opacities are seen and there may be enlargement of hilar lymph nodes. On CT there may be poorly defined centrilobular acinar nodules and patchy ground-glass opacities. Abnormal thickening of interlobular septa may be present and, in some patients, a combination of ground-glass opacities with thickening of inter- and intralobular septa (the crazy-paving pattern) is seen. Of the many DPH syndromes, IPH and antiglomerular basement membrane antibody disease (Goodpasture’s syndrome) usually receive the greatest attention and are considered briefly below.

Idiopathic Pulmonary Haemosiderosis Idiopathic pulmonary haemosiderosis (IPH) is a rare disorder of unknown aetiology. The majority of patients are children (typically in the first decade), although sporadic cases in older subjects have been recorded. The clinical picture is that of episodic intra-alveolar haemorrhage, haemoptyses, iron-deficiency anaemia and airspace opacification on CXRs. Repeated bouts of bleeding may lead to lung fibrosis. The outlook for patients with IPH varies with survival, ranging from a few days (following massive haemorrhage) to years. The pathogenesis of IPH is not clear, although a number of hypotheses have been proposed. The imaging findings are non-specific and, as for other haemorrhage syndromes, the clue to a radiological diagnosis may only come after a thorough review of clinical features and the exclusion of other causes of widespread pulmonary haemorrhage.

Antibasement Membrane Antibody Disease (Goodpasture’s Syndrome) A link between renal disease and diffuse intra-alveolar bleeding has long been known. Although the eponymous title Goodpasture’s syndrome was, until relatively recently, in common use, the pathogenetically accurate term antibasement membrane (anti-BM) antibody disease is preferred.

258

SECTION A  The Chest and Cardiovascular System

TABLE 11.3  Proposed Classification of

Causes of Diffuse Pulmonary Haemorrhage Immunocompetent Patients • Immunologically mediated • Antibasement membrane antibody disease (Goodpasture’s syndrome) • Disorders with a presumed immune aetiology, with or without nephropathy • Connective tissue diseases • Systemic lupus erythematosus • Rheumatoid arthritis • Mixed connective tissue disease • Systemic sclerosis • Anti-neutrophil cytoplasmic antibodies–associated vasculitis • Wegener’s granulomatosis • Churg–Strauss syndrome • Microscopic polyangiitis • Anti-phospholipid syndrome • Pauci-immune glomerulonephritis • Henoch–Schönlein purpura • Cryoglobulinaemia • Diseases with no known immune aetiology • Idiopathic pulmonary haemosiderosis • Rapidly progressive glomerulonephritis without immune complexes • Drug-induced (anticoagulants, trimellitic anhydride, cocaine) • Valvular heart disease • Disseminated intravascular coagulation • Diffuse alveolar damage • Tumours Immunocompromised Patients • Blood dyscrasias • Infection • Tumours Adapted from Lara and Schwarz, Miller and Lee and D’Cruz.

Anti-BM antibody disease typically affects young men, with a male-tofemale ratio of around 3 : 1. On histopathological examination there is glomerulonephritis with circulating serum antibodies directed against components of basement membrane in the lungs and kidneys. The pulmonary manifestations of anti-BM antibody disease often dominate the clinical presentation, though renal disease is present in the majority.

Granulomatosis With Polyangiitis (GPA; Formerly Wegener’s Granulomatosis) Granulomatosis with polyangiitis (GPA), together with microscopic polyangiitis, eosinophilic GPA (Churg–Strauss syndrome) and isolated pauci-immune pulmonary capillaritis, is best classified as one of the primary (idiopathic) small vessel vasculitides. GPA is multisystem disorder with necrotising granulomatous inflammation of small vessels in the upper and lower respiratory tracts. There is an equal gender predilection and a wide age range of presentation (from childhood to >70 years). The lungs are affected in approximately 90% of patients with GPA. Most patients present with symptoms referable to the nose, paranasal sinuses or chest; in some patients, the disease manifests solely in the respiratory tract and is termed ‘limited’ GPA. Chest symptoms include cough, dyspnoea, pleuritic chest pain and haemoptysis. The aetiology of GPA is unclear but there is a strong link with a cytoplasmic-staining

pattern of anti-neutrophil cytoplasmic antibodies (c-ANCA) directed against proteinase-3 (PR3-ANCA). The spectrum of morphological abnormalities in GPA is potentially legion. Bilateral nodules or masses are the most prevalent finding, seen in 70%–90% of patients. Nodules range in size from a few millimetres up to 10 cm in diameter, are frequently multiple and can increase in size and number as the disease progresses (Fig. 11.18). Nodules have no specific zonal predilection and generally cavitate when more than 2 cm in diameter. Cavitation is generally regarded as the classical radiological finding in pulmonary GPA (Fig. 11.19). However, cavitation is by no means an invariable feature and the absence of this sign does not preclude a diagnosis of GPA. On CT, a halo of ground-glass opacification (believed to reflect surrounding haemorrhage) may be seen around nodules. In some patients there may be a ‘feeding’ vessel leading to a nodule; linear bands, spiculation and pleural tags may also be seen. With treatment, nodules generally regress, but it should be noted that CXRs may not return to normal for up to a month after the start of treatment. Moreover, there may be residual parenchymal scarring on CT despite resolution of nodules and consolidation. Interestingly, the converse is apparently true in children, in whom nodules are seen less frequently. Consolidation and ground-glass opacities are recognised features on CT but are less common than nodules. The distribution of consolidation is variable and might include peripheral wedge-shaped foci abutting the pleura (mimicking pulmonary infarcts), a peribronchovascular predilection (Fig. 11.20), a ‘reverse halo’ pattern, or multifocal areas of consolidation with or without cavitation. Airway disease is also reported in GPA. Stenoses of large airways leading to subglottic, tracheal or bronchial narrowing are well documented. Bronchiectasis is an additional feature and seen in ~40% of cases. CT has also highlighted some of the less common features of GPA, including areas of lobar or segmental atelectasis, pleural effusions or thickening and, rarely, hilar and mediastinal lymph node enlargement.

ORGANISING PNEUMONIA The historical aspects of cryptogenic organising pneumonia (COP) are worth discussing briefly. The pattern of COP was first reported as a clinico-radiological-pathological entity by Davison and colleagues in 1983. In their paper, the authors described the case studies of eight patients presenting with an illness of insidious onset characterised by cough, night sweats, generalised malaise and weight loss. On chest x-ray, there were bilateral patchy areas of consolidation and, on biopsy, buds of fibrous connective tissue (‘bourgeons conjunctifs’) were seen in the alveoli and alveolar ducts and, crucially, only infrequently in the airways. Despite thorough investigation—specifically for infections—the investigators found no cause. Another key aspect was that there was a striking response to corticosteroid treatment. The authors noted that organising pneumonia per se is simply a histological response to ‘injury’ and that the label ‘cryptogenic’ should be used only when other potential causes of an organising pneumonia pattern have been excluded, since organising pneumonia occurs in a variety of clinical contexts (Table 11.4). Unfortunately, the more confusing term ‘idiopathic brochiolitis obliterans organising pneumonia (BOOP)’, now, thankfully, confined to the historical waste-bin, was also used to describe the same clinicoradiological-pathological entity described by Davison, based on the findings of a paper published two years later. However, because the airway changes are secondary to the dominant process in the air spaces and there is no evidence of a ‘bronchiolitis obliterans’, the term organising pneumonia is now preferred. The typical chest x-ray and CT signs of COP may be predicted from knowledge of the histopathological changes. Bilateral patchy areas of consolidation, which tend to be peripheral, are the characteristic findings

CHAPTER 11  Airspace Diseases

A

B

D

C Fig. 11.17  (A to C) Chest Radiographs and (D) High-Resolution Computed Tomography (HRCT) in a Patient with Idiopathic Pulmonary Haemosiderosis. (A) There is diffuse ground-glass opacification with no zonal predilection during an acute hospital admission with haemoptysis. (B) Radiograph taken 4 days later shows striking (but incomplete) resolution of airspace opacities. There is residual opacification around the right hilum. (C) Chest radiograph obtained 1 month following admission demonstrates ground-glass opacification in the right mid zone. (D) HRCT through the lower zones (concurrent with the radiograph in (A)) shows patchy ground-glass opacification with a somewhat unusual geographical distribution; there are thickened interlobular septa in areas of ground-glass opacification.

259

260

SECTION A  The Chest and Cardiovascular System

A

A

B Fig. 11.19  Granulomatosis With Polyangiitis. CT through (A) the upper and (B) mid zones demonstrating multiple irregular ‘mass-like’ foci of consolidation and cavitation.

B Fig. 11.18  Granulomatosis With Polyangiitis. Chest radiographs in two patients demonstrating (A) multiple and typically large (>2 cm) cavitating nodules with irregular wall thickness and (B) smaller, but again, multiple cavitating lesions.

on chest x-ray and CT (Fig. 11.21). Although earlier series suggested a predilection for the mid and lower zones, it is clear that all lung zones may be affected. The changes of COP may be confined to one lung but this is uncommon. Consolidation in COP has a propensity for the subpleural and/or peribronchovascular regions in approximately twothirds of patients. Cavitation is not a feature of COP but multifocal areas of ground-glass opacification with a surrounding rim of consolidation (termed the ‘reverse halo’ or ‘Atoll’ sign), has been described (Fig. 11.22). Needless to say, the radiological distinction from pure infection (e.g. bacterial pneumonia) may be difficult, particularly when abnormalities are unilateral. Nodules, sometimes measuring up to 1 cm in diameter and representing focal areas of organising pneumonia, are seen in some patients and, occasionally, these may be the sole radiographic manifestation of COP;

Fig. 11.20  Bronchocentric Disease in Granulomatosis With Polyangiitis. Many of the segmental and subsegmental bronchi are thick walled (arrows). (Courtesy Dr. Kate Pointon, Nottingham City Hospital, United Kingdom.)

as with areas of consolidation, there is no definite zonal predilection. In rare instances, a large solitary nodule or mass with irregular margins may be seen, usually prompting investigations for lung cancer. Linear opacities may be the dominant CT pattern in some patients with COP. Two types of opacity (termed types I and II) are recognised

CHAPTER 11  Airspace Diseases

261

B

A

C D Fig. 11.21  Variable CXR and CT Patterns of Organising Pneumonia. (A) CXR in a patient with organising pneumonia showing a typical mid/lower zone consolidation. (B) CT in another patient with classical mid/lower zone peripheral consolidation. (C) Striking bronchocentric consolidation in organising pneumonia as a manifestation of nitrofurantoin toxicity and (D) radial (type I) and subpleural curvilinear (type II) bands of consolidation in organising pneumonia.

TABLE 11.4  Recognised Causes of

Organising Pneumonia and Associations With Other Conditions • Unknown (cryptogenic organising pneumonia) • Infections (bacterial, viral) • Toxic fume exposure • Drugs (antibiotics, chemotherapeutic, anti-inflammatory, etc.) • Connective tissue diseases (particularly polymyositis dermatomyositis, rheumatoid arthritis, Sjögren’s syndrome, etc.) • Vasculitis (particularly Churg–Strauss disease, Wegener’s granulomatosis, etc.) • Immunological disorders (common variable immunodeficiency syndrome, essential mixed cryoglobulinaemia) • Organ transplantation (bone marrow, lung, renal) • Radiation pneumonitis • Neoplasms (particularly lymphoma)

Fig. 11.22  Atoll or Reverse ‘Halo’ Sign in Organising Pneumonia. CT shows bilateral foci of ground-glass opacities surrounded by a ‘halo’ of consolidation.

262

SECTION A  The Chest and Cardiovascular System

(see Fig. 11.21). Type I opacities are intimately related to bronchi, extend radially towards the pleura, measure 2–4 cm in length and are 1–2 mm in thickness. Areas of consolidation may coexist with type I linear opacities. Type II linear opacities are subpleural and, unlike the type I pattern, are not related to airways. Type II linear opacities also tend to parallel the pleural surface and are frequently associated with multifocal airspace consolidation. A perilobular distribution of consolidation, giving rise to a distinctive CT appearance, is also recognised in COP and, unlike a multitude of other radiological signs, may be regarded as pathognomonic of an organising pneumonia pattern (see Fig. 11.11).

Eosinophilic Lung Disease The eosinophilic lung diseases are a diverse group of disorders associated with peripheral or tissue eosinophilia. Infiltration of the lung by eosinophils is surprisingly common, and reported in conditions as diverse as asthma, opportunistic infections and certain cancers. However, by common convention, such processes are not usually considered to be eosinophilic lung diseases per se. A simplified classification of the pulmonary eosinophilias is given in Table 11.5 and a few of the disorders under the broad category of eosinophilic pneumonia are discussed below.

Simple Pulmonary Eosinophilia (Löffler’s Syndrome) The term Löffler’s syndrome (first described in 1932 and synonymous with simple pulmonary eosinophilia) describes patients with transient

TABLE 11.5  Modified Classification of

Eosinophilic Lung Disease

Idiopathic • Simple pulmonary eosinophilia (Löffler’s syndrome) • Acute eosinophilic pneumonia • Chronic eosinophilic pneumonia • Hypereosinophilic syndrome Drug-Induced • Aminosalicylic acid • Para-aminosalicylic acid • Non-steroidal anti-inflammatory drugs • Captopril • Cocaine • Minocycline • Nitrofurantoin • Phenytoin Infection • Parasitic (ascariasis, paragonimiasis, tropical eosinophilia) • Fungal (Aspergillus) • Bacterial (tuberculosis, atypical mycobacterial infection, Brucella) • Viral (respiratory syncytial virus) Immunological Diseases • Wegener’s granulomatosis • Churg–Strauss syndrome • Rheumatoid disease • Sarcoidosis Neoplasms • Brochogenic carcinoma • Bronchial carcinoid • Lymphoma (Hodgkin’s, non-Hodgkin’s)

radiographic infiltrates, minimal constitutional upset and an elevated eosinophil count in peripheral blood. The airspace opacification in Löffler’s syndrome is classically fleeting and may be either uni- or bilateral. Resolution of opacities within a period of days and, by definition, within a month is the rule. In many cases, no underlying cause is found but there is an association with parasitic infection, in particular infestation with Ascaris lumbricoides. CT findings include ground-glass opacities or consolidation principally in the periphery of the middle/upper lung zones, as well as single or multiple acinar nodules.

Acute Eosinophilic Pneumonia (AEP) In rare patients with pulmonary eosinophilia there is a more fulminant clinical illness beginning with a short febrile episode (25% eosinophils in lavage fluid (± variable increase in lymphocytes and neutrophils) or evidence of an eosinophilic pneumonia on biopsy; and (4) the exclusion of other specific eosinophilic diseases (e.g. eosinophilic GPA (formerly, Churg-Strauss syndrome) and allergic bronchopulmonary aspergillosis). AEP may be idiopathic but an aetiological link with cigarette smoke seems very likely. On this note, an important finding is the observation that AEP has been reported in subjects who have just started to smoke or in those with an established habit but which may have changed recently. Exposure to other inhaled ‘toxins’ (e.g. cocaine, marijuana, firework dust and tear gas), certain drugs (e.g. antimicrobials, antidepressants, non-steroidal anti-inflammatory drugs (NSAIDs) and infections (parasitic, fungal and viral) have also be implicated. Spontaneous regression of AEP is reported but most patients respond relatively quickly to corticosteroid treatment. Treatment of the underlying cause (in particular, those cases linked to infection) is important. Disease relapses tend to be rare, except in smokers who continue to smoke. On CXR, there is bilateral airspace opacification and/or reticular infiltrates. Pleural effusions are common. Areas of ground-glass opacification and consolidation are seen on CT and there may be smooth thickening of interlobular septa. The CT signs are similar to those seen in patients with pulmonary oedema or diffuse alveolar damage and the initial serum eosinophil count may be normal. Therefore, establishing a confident diagnosis of AEP can be difficult.

Chronic Eosinophilic Pneumonia The clinical and radiological features of chronic eosinophilic pneumonia are strikingly different from the entities described above. As the term suggests, the clinical course of chronic eosinophilic pneumonia is generally more protracted and symptoms often more marked than in patients with simple pulmonary eosinophilia. There is frequently mildto-moderate eosinophilia and increased serum IgE levels in the peripheral blood. The prognosis is good and most patients respond to steroid therapy. The plain radiographic abnormalities in chronic eosinophilic pneumonia can be characteristic with patchy, non-segmental areas of consolidation typically in the mid and upper zones. A distinctive feature is that the opacities are peripheral and seem to parallel the chest wall, a finding that was considered to be the ‘photographic negative of pulmonary oedema’ by Gaensler and Carrington (Fig. 11.23). Not surprisingly, the peripheral location of the consolidation and ground-glass opacity is more readily appreciated on CT. Other disorders may mimic chronic eosinophilic pneumonia at CT, including organising pneumonia. In one study, the most helpful distinguishing feature was the presence of

CHAPTER 11  Airspace Diseases

263

A

Fig. 11.23  Chronic Eosinophilic Pneumonia. Coronal CT reconstruction in a patient with a dry cough constitutional symptoms peripheral eosinophilia. There is predominant upper zone airspace opacification which is peripheral, paralleling the chest wall (the so-called ‘photographic negative’ of pulmonary oedema) and not respecting normal segmental/ fissural boundaries.

nodules, seen in 32% of patients with COP but only 5% of patients with chronic eosinophilic pneumonia.

PULMONARY ALVEOLAR PROTEINOSIS PAP (also called alveolar lipoproteinosis and alveolar phospholipoproteinosis) is a rare disease characterised by the accumulation of a periodic acid–Schiff-positive lipoproteinaceous material in the alveoli. Abnormal surfactant clearance from the lungs caused by a fault in macrophage function is at the heart of the pathogenesis of adult forms of PAP. The most common cause of such impairment is a fault in granulocytemacrophage colony stimulating factor (GM-CSF) signalling. Historically, most cases were regarded as ‘idiopathic’ but those patients are now believed to have an autoimmune form of PAP in which anti-GM-CSF antibodies lead to inadequate macrophage maturation/function. Secondary PAP occurs in association with other disorders, most notably haematological diseases (e.g. myelodysplasia, lymphoma, myeloid leukaemia), dust inhalation (e.g. silica) and infections. PAP usually affects adults aged 20–50 years, with a male preponderance, but has been described in children (in whom the outlook tends to be worse). A definitive diagnosis usually requires bronchoscopic lavage and/or biopsy, supported by imaging findings. Although spontaneous resolution is reported, most patients require therapeutic (whole lung) bronchoalveolar lavage, a technique that has improved the outlook of patients with PAP (Fig. 11.24). The chest radiographic changes of alveolar proteinosis are nonspecific. In general, both lungs are affected and airspace opacification is most pronounced in the central lung, sometimes producing a ‘bat’s wing’ appearance. The CT features are much more suggestive of alveolar proteinosis: a ‘crazy-paving’ pattern (comprising geographical areas of ground-glass opacification with thickened inter- and intralobular septa) is the characteristic feature, which, in its classical form, is virtually

B Fig. 11.24  ‘Crazy-Paving’ Pattern in Alveolar Proteinosis. CT at the level of the carina (A) before and (B) 6 months after therapeutic bronchoalveolar lavage. There is a classical geographical ‘crazy-paving’ pattern with thickened inter- and intralobular septa in regions of ground-glass opacification. The follow-up study shows considerable reduction in the extent of disease.

diagnostic of PAP. However, it is worth remembering that from time to time, a similar CT appearance is seen in some patients with adenocarcinoma, exogenous lipoid pneumonia and pulmonary oedema.

ALVEOLAR MICROLITHIASIS Pulmonary alveolar microlithiasis is a rare cause of airspace disease characterised by the deposition of tiny stones or calcipherites (measuring 250–750 µm in diameter and composed mainly of calcium phosphate) in alveoli. There is a high familial incidence and a genetic abnormality has been identified: mutations in the SLC34A2 gene (which codes for a sodium-dependent phosphate transporter in alveolar type II cells) leads to phosphate accumulation. A wide age range (the peak incidence is between 30 and 50 years) has been reported but it is believed that alveolar microlithiasis begins in early life. Most patients are asymptomatic at the time of diagnosis and disease progression is variable. However, there is a tendency for pulmonary fibrosis and the development of cor pulmonale. The fibrosis of alveolar microlithiasis is associated with the formation of bullae, particularly at the apices. The classical finding on chest radiography is the widespread discrete high-density opacities (resembling grains of sand) seen in both lungs; when the infiltration is profuse there may be a ‘white-out’, with

264

SECTION A  The Chest and Cardiovascular System and interlobular septa may be present and, in rare patients, a crazy-paving pattern has been reported.

FURTHER READING

Fig. 11.25  Alveolar Microlithiasis. (A and B) CT through the mid zone showing diffuse micronodular infiltration but with a characteristic subpleural low density.

obscuration of the heart borders and diaphragm and the tiny stones may then only be seen on an overexposed radiograph. A tell-tale line of black subpleural lung (caused by 5–10 mm diameter small cysts or paraseptal emphysema) may be seen and there may be thickening or beading of fissures. Less commonly, apical blebs and thickened septal (Kerley B) lines may be seen. On CT, the characteristic finding is of widespread ground-glass opacification and/or a micronodular pattern (Fig. 11.25). The changes are more pronounced in the lower zones and posteriorly. In about half of cases, calcification is uniform, producing a pattern of dense parenchymal opacification. Thickening of fissures

Chen, X.Y., et al., 2017. Idiopathic pulmonary haemosiderosis in adults: review of cases reported in the last 15 years. Clin. Respir. J. 11, 677–681. Chung, M.P., et al., 2018. Serial chest CT in cryptogenic organizing pneumonia: evolutional changes and prognostic determinants. Respirology 23, 325–330. Cottin, V., 2016. Eosinophilic lung diseases. Clin. Chest Med. 37, 535–556. Hansell, D.M., Lynch, D.A., McAdams, H.P., et al. (Eds), 2010. Basic HRCT patterns of lung disease. Drug- and radiation-induced lung disease In: Imaging of Diseases of the Chest, 5th ed. Elsevier Mosby, Philadelphia. Hunter Guevara, L.R., et al., 2018. Whole-lung lavage in a patient with pulmonary alveolar proteinosis. Ann. Card. Anaesth. 21, 215–217. Kumar, A., et al., 2018. Pulmonary alveolar proteinosis in adults: pathophysiology and clinical approach. Lancet Respir. Med. 6, 554–565. Martínez-Jiménez, S., Rosado-de-Christenson, M.L. (Eds.), 2017. Section 2: Pathological patterns of injury In: HRCT of the Lung, 2nd ed. Elsevier, Philadelphia. Pearce, F.A., et al., 2018. Novel insights into the aetiology of granulomatosis with polyangiitis: a case-control study using the Clinical Practice Research Datalink. Rheumatology 57, 1873. doi:10.1093/rheumatology/key249. Vergani, G., et al., 2017. A morphological and quantitative analysis of lung CT scan in patients with acute respiratory distress syndrome and in cardiogenic pulmonary edema. J. Intensive Care Med. doi:10.1177/0885066617743477. 885066617743477. Von Ranke, F.M., et al., 2013. Infectious diseases causing diffuse alveolar hemorrhage in immunocompetent patients: a state-of-the-art review. Lung 191, 9–18. Webb, W.R., Müller, N.L., Naidich, D.P., 2015. Approach to HRCT diagnosis and findings of lung disease. In: High-Resolution CT of the Lung, 5th ed. Wolters Kluwer, Philadelphia. Zare Mehrjardi, M., et al., 2017. Radio-pathological correlation of organizing pneumonia (OP): a pictorial review. Br. J. Radiol. 90, 20160723.

12  Cardiac Anatomy and Imaging Techniques Hans-Marc J. Siebelink, Jos J.M. Westenberg, Lucia J.M. Kroft, Albert de Roos

CHAPTER OUTLINE Normal Chest Radiography, 265 Cardiac Magnetic Resonance, 265

Knowledge of the cardiac anatomy is essential for identifying and understanding cardiovascular disease in patients and is therefore important in clinical practice. To date, various imaging techniques such as conventional chest radiography, cardiac magnetic resonance (CMR), computed tomography (CT) and echocardiography are all used to assess aspects of cardiac and vascular anatomy. This chapter contains an overview of the techniques used and provides examples of the anatomy identified with these techniques.

NORMAL CHEST RADIOGRAPHY Postero-anterior and lateral chest radiographs are commonly obtained in patients with cardiovascular disease. The chest radiograph provides an impression of the size of cardiovascular structures and the lung parenchyma. Specific cardiac chambers and large vessel anatomy can be appreciated on chest radiography. Evaluation of the lung parenchyma and lung vasculature is helpful for assessing and grading heart failure. Valvular calcifications may be recognised as a clue for specific valvular disease. Fig. 12.1 illustrates the normal chest radiograph, noting normal cardiovascular structures. SUMMARY BOX: Chest Radiography • Cardiovascular structures and lung parenchyma Cardiac Magnetic Resonance • Various imaging planes (body axis- and cardiac planes) • Anatomy of thoracic vascular structures • Cardiac anatomy and function

CARDIAC MAGNETIC RESONANCE An advantage of choosing magnetic resonance for cardiac imaging is the free choice in obtaining imaging planes of cardiovascular anatomy in any arbitrary view, since this technique is not hampered by the limited availability of acoustic windows, as with ultrasound. This benefit is especially advantageous when imaging the morphology of the right ventricle (RV), which is excellently delineated by CMR, whereas in echocardiography the assessment of RV geometry and function is challenging because of the particular crescentic shape of the RV as it wraps around the left ventricle (LV). Furthermore, the unrestricted

Computed Tomography Imaging Techniques, 269 Echocardiography, 281

field of view of CMR allows superior visualisation of extracardiac and large vessel anatomy. Single-plane two-dimensional (2-D) or multiple-plane 2-D or three-dimensional (3-D) imaging is possible with CMR. Dynamic functional information can be obtained by synchronising image acquisition to the interval of the R-waves on the electrocardiogram, using either prospective triggering or retrospective gating. With prospective triggering, the operator needs to set the expected heart rate before the acquisition and triggering will then be performed according to this defined heart rate. With retrospective gating, imaging is performed continuously and the ECG signal is stored additionally. In retrospect, image reconstruction is synchronised to the stored ECG, providing time-resolved imaging in multiple phases of the cardiac cycle, which can then be presented in cine mode. Imaging planes in CMR are usually obtained in the orientation to the axes of the heart, or oriented to the major axes of the body. Therefore, the standard CMR planes of the heart are comparable to the standard cardiac views, well-known and established in other non-invasive imaging techniques such as echocardiography, cardiac CT, x-ray LV angiography and nuclear medicine. The choice for a specific CMR protocol is mainly determined by the clinical questions that need to be answered. Standardised nomenclature for cross-sectional anatomy has been described, facilitating comparison between different techniques and proper communication between imaging specialists. Another important issue in clinical CMR imaging is the ability of the patient to collaborate during the examination and to perform breath-holding repeatedly and consistently. If a patient is capable of performing breath-holding, successive imaging planes are obtained with accelerated imaging, with the patient usually performing breath-holding in end expiration, as the anatomical level may be more reproducible than planes that are examined in inspiration. CMR techniques for anatomical evaluation include bright-blood and black-blood imaging, which essentially determines the contrast in signal intensity between myocardium and the intracardiac blood pool. For the assessment of left and right ventricular function, fast-gradient echo sequences are usually performed in combination with steady-state free-precession (SSFP) technique (balanced-Turbo Field Echo [TFE], True-Fast Imaging Steady state Precession [FISP], Fast Imaging Employing Steady-state Acquisition [FIESTA]) for optimal contrast. On these

265

266

SECTION A  The Chest and Cardiovascular System

Fig. 12.1  Normal Postero-Anterior (Left) and Lateral (Right) Chest Radiographs. Note normal cardiovascular structures. 1, Contour of superior vena cava; 2, contour of right atrium; 3, aortic knob; 4, left pulmonary artery at hilar level; 5, contour of left ventricle; 6, anterior contour of right ventricle and pulmonary outflow tract; 7, aortic arch; 8, upper posterior contour of left atrium; 9, lower posterior contour of left ventricle. Note relative bulging of left ventricular contour in relationship with inferior vena cava.

bright-blood images, the blood pool is presented with bright signal, whereas the myocardium is represented dark with low signal. This results in an excellent definition of the left ventricular endocardial and epicardial borders, which is required for accurate image segmentation during cardiac volume and function quantification. Typically, SSFP images should be acquired with slice thickness of 6–8 mm and temporal resolution less than 45 ms to obtain optimal accuracy in ventricular function assessment. Additionally, cardiac morphology can be evaluated by doubleinversion, black-blood, spin-echo sequences with fat suppression, providing gated, static images of the heart with high spatial resolution (optimally, in-plane-acquired resolution of less than 2 × 2 mm2 and slice thickness of 5–8 mm) in the orientation of the heart or the patient’s body axes. Multiple other magnetic resonance imaging techniques are available for tissue characterisation (e.g. T2 weighted sequences, delayed gadolinium-based contrast enhancement), extending the capabilities of CMR beyond anatomical and functional evaluation of the cardiovascular system.

Cardiac Axis Imaging Planes To acquire imaging planes in the direction of the cardiac axes, SSFP scout views are used for planning. If available, free-breathing images obtained during real-time imaging can be used instead. Perpendicular to an anatomical transverse image, which displays the heart’s four chambers, an acquisition plane is chosen through the middle of the atrioventricular junction at the level of the mitral valve and running through the apex (Figs 12.2A and B). This plane is the so-called vertical long-axis (VLA) plane (see Fig. 12.2C). On this VLA view, a plane is defined intersecting the apex and the middle of the mitral valve, resulting in the horizontal long-axis (HLA) view (see Fig. 12.2D). This HLA view

is almost comparable to the four-chamber view; however, often only a part of the left ventricular outflow tract is visualised in this HLA view. On the acquired HLA plane, the short-axis (SA) views (see Figs 12.2E and F) covering the entire LV are planned parallel to the ring of the mitral valve and perpendicular to the line intersecting the apex. For reproducibility and comparison purposes, the true two- and four-chamber views can still be obtained (see Fig. 12.2G and H). The two-chamber view is planned perpendicular to the anterior and inferior wall of the LV through the centre of the left ventricular cavity on a mid-ventricular SA image intersecting the apex. On the two-chamber view the apex, anterior and inferior wall of the LV, the mitral valve and left atrium can be evaluated. The four-chamber view is also planned on a mid-ventricular SA image by a plane through the centre of the left ventricular cavity and the acute margin of the RV, also intersecting the apex (see Fig. 12.2E). The four-chamber view depicts the interventricular septum, the lateral wall of the LV, the free wall of the RV, the left and right atrium as well as the interatrial septum and both the mitral and tricuspid valves. Routinely, the three-chamber or so-called left ventricular outflow tract (see Fig. 12.2I) view is planned on a basal SA plane (see Fig. 12.2F), and also intersects the apex. The left ventricular outflow tract view depicts the apex, the anteroseptal interventricular wall, the left ventricular outflow tract, the inferolateral wall, and the aortic and mitral valve. The standard SSFP cine CMR protocol for assessing left ventricular function should include the two-, three and four-chamber views in combination with SA images covering the entire LV, resulting in images covering all described 17 left ventricular segments in two directions. Additionally, the right ventricular outflow tract (RVOT) can be obtained (Fig. 12.3). This view can be planned on a coronal image, depicting the outflow tract of the RV. Alternatively, an optimised view

A

B

C

D

E

G

F

H

I

Fig. 12.2  Planning Acquisition of Standard Cardiac Views. On two transverse slices (A) and (B), the left ventricular vertical long axis (VLA) (C) is planned by a plane transecting the mitral valve and the apex. The horizontal long axis (HLA) (D) is obtained by acquiring a plane transecting the VLA through the mitral valve and apex. A short-axis image can be obtained perpendicular to HLA, at mid-ventricular (E) and basal level (F). The four-chamber (G) of the left ventricle (LV) is obtained as indicated from a plane transecting both LV and the right ventricle. The two-chamber (H) of the LV is acquired as a plane perpendicular to the four-chamber. The three-chamber LV (I) is obtained from a plane transecting the LV through the LV outflow tract. In Fig E the planes of the four-chamber and two-chamber are indicated as the white lines, perpendicular to each other.

268

SECTION A  The Chest and Cardiovascular System

Normal Anatomy on Cardiac Magnetic Resonance Images

Fig. 12.3  Bright-Blood Magnetic Resonance Acquisition of the Right Ventricle. Ao, Aorta; PA, main pulmonary artery; PV, pulmonary valve; RA, right atrium; RV, right ventricle; RVOT, right ventricular outflow tract, TV, tricuspid valve.

of the RVOT view can be obtained from a plane outlining the tricuspid valve plane and the outflow tract. On this plane the outflow tract, pulmonary valve, tricuspid valve and the basal (diaphragmatic) part of the right ventricular wall are all visualised.

Body Axes Imaging Planes For the evaluation of cardiac morphology, the pericardium, the great thoracic vessels and (para-) cardiac masses, imaging planes oriented to the main body axes are obtained. Also, the transverse (or axial), coronal (frontal) and sagittal planes are well known to clinicians, as these anatomical orientations are similar to clinical (cardiac) CT. Black- and brightblood sequence approaches (Fig. 12.4) can be used in optimally adjusted planes to answer specific clinical questions. Black-blood images provide only static information in a single phase and are not suitable for quantification of left or right ventricular dimensions. For this analysis, SSFP multiphase images with appropriate temporal resolution are necessary. Transversely orientated planes (Fig. 12.5) are especially useful for the evaluation of thoracic vascular structures including the ascending and descending thoracic aorta, the superior and inferior vena cava, the pulmonary trunk and right and left pulmonary artery. The right and left pulmonary veins entering the left atrium are also well depicted. Images in transverse orientation through the heart allow the evaluation of morphology of the ventricles and atria. Also, the right ventricular free wall, the RVOT, the pericardium and mediastinum are well depicted. It has been suggested that right ventricular volume and function quantification by planimetry can be performed more accurately on transversely oriented images instead of SA images. Coronal or frontal anatomical views can be instructive for analysing the connection between the heart and the great vessels. An advantage of the frontal view is the similarity to the well-known anatomy from chest radiography. On sagittal images, the RVOT in relation to the pulmonary valve is well outlined and the connection of the right atrium with the superior and inferior vena cava can be studied.

CMR images present distinct anatomical features of both atria and ventricles. For evaluating anatomy, either cardiac axes (Fig. 12.6) or body axes imaging planes (see Figs 12.4 and 12.5) can be chosen. The pericardial sac encloses the heart and the roots of the great vessels. The pericardial cavity is outlined by the parietal and visceral layer of the inner pericardium. Normal pericardium has a longer T1 than fat tissue and, therefore, yields low-signal intensity on T1 weighted MR images and can be well visualised due to the surrounding epicardial and pericardial fat. Normally, the thickness of the pericardium measures less than 4 mm on CMR images. In normal cardiac anatomy, the right atrium can be recognised by identifying the corresponding broad-based triangular appendage. At the base, the tricuspid valve, positioned between the right atrium and the RV, is located closer to the apex compared with the mitral valve. The right atrium receives venous blood from the superior and inferior vena cava and the coronary sinus. The coronary sinus enters the right atrium in the posterior atrioventricular groove. The appendage of the morphological left atrium has a narrow attachment to the atrium and is more tubular shaped. Characteristically, the left atrium receives four pulmonary veins in total—two on either side—although several variations occur. To date, imaging the venous anatomy of the heart is becoming more relevant. For example, during pre-ablation work-up for supraventricular arrhythmias, the clinician needs to be informed about the exact anatomy of the left atrial morphology and number of the pulmonary ostia, as the left atrium and pulmonary veins are used to guide the interventional procedure. The interatrial septum separates the two atria. As part of the interatrial septum, the fossa ovalis is very thin and can hardly be depicted on CMR images due to the limited spatial acquisition resolution. The RV is normally triangular in shape and anteriorly located relative to the LV, directly behind the sternum. Morphologically, the RV has typical features that can be depicted on CMR images. The RV shows a muscular moderator band (Fig. 12.7) carrying branches of the conducting system. Furthermore, the RV contains a muscular outflow tract (infundibulum or conus arteriosus) and typically, the RV wall is more trabeculated than the left. In normal anatomy, the LV is positioned posteriorly and to the left. The septum is smooth with no trabeculae and the left ventricular outflow tract lacks a muscular part. The interventricular septum consists of a muscular and a membranous part. In particular, the membranous part is very thin and is sometimes not depicted on CMR images. It is important to recognise these normal anatomical features of atrial and ventricular morphology because the position of the atria and ventricles may be inversed in complex congenital heart disease. At the outlet of each of the heart’s four chambers, one-way valves are positioned to ensure that blood flows in the proper direction. The blood flow through the atria into the ventricles is regulated by the atrioventricular valves (the tricuspid valve is related to the morphological RV, the mitral valve to the morphological LV). The pulmonary valve connects the outflow tract of the RV to the pulmonary trunk and the aortic valve connects the left ventricular outflow tract to the thoracic aorta. The normal tricuspid valve consists of three cusps, whereas the mitral valve consists of two cusps. Both the normal pulmonary and the aortic valve (Fig. 12.8) normally consist of three cusps. Opening of the atrioventricular valves is predominantly determined by pressure differences between the atria and ventricles, which are the result of the isovolumetric relaxation of the ventricles during diastole. Furthermore, the motion of the valves is regulated by papillary muscles, which originate from the inferolateral and anterolateral left ventricular myocardial wall and are connected to the valve leaflets by chordae tendineae. During

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

A

B

C

D

269

Fig. 12.4  Normal Cardiac Anatomy on Black-Blood and Bright-Blood MR Acquisitions, in Sagittal (A and B) and Coronal (C and D) Views. Ao-Arch, Aortic arch; Ao-Asc, ascending aorta; Ao-Desc, descending aorta; LA, left atrium; LV, left ventricle; PA, pulmonary artery; RA, right atrium; RPA, right pulmonary artery; RV, right ventricle.

contraction of the ventricle, the papillary muscles also contract, pulling on the chordae tendineae, closing the valves and preventing blood flow from the ventricles into the atria (i.e. regurgitation). Normally, in the RV three small papillary muscles can be depicted: the anterior, posterior and septal papillary muscle. The LV reveals two larger papillary muscles, the anterior and posterior papillary muscle; however, there is quite a wide variation in this standard morphology. Cine SSFP long-axis and SA images, as well as transverse images, are all well suited for depicting morphology and function of the valvular apparatus. The valve leaflets can be depicted if spatial resolution is adequate. Dedicated acquisitions of specific valvular planes are used to

image the valve area, which is especially useful when studying aortic valve stenosis or incompetence. Both SSFP and fast gradient-echo sequences are used for valvular imaging. Papillary muscles are well visualised on both cine bright- and black-blood sequences. Chordae tendineae, on the other hand, are difficult to visualise on CMR due to the limited spatial resolution (see Fig. 12.6).

COMPUTED TOMOGRAPHY IMAGING TECHNIQUES With the introduction of ≥64 slice CT imaging devices, CT has been accepted as a diagnostic imaging tool for the evaluation of patients

270

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E

F

G

H

I

Fig. 12.5  Normal Cardiac Anatomy on Transverse Black-Blood MR Acquisitions. Ao-Arch, Aortic arch; Ao-Asc, ascending aorta; Ao-Desc, descending aorta; AV, aortic valve; C, carina; cs, coronary sinus; IVC, inferior vena cava; LA, left atrium; LAA, left atrial appendage; LAD, left anterior descending coronary artery; LPA, left pulmonary artery; LV, left ventricle; MV, mitral valve; P, papillary muscle; PA, main pulmonary artery; pc, pericardium; RA, right atrium; RCA, right coronary artery; RPA, right pulmonary artery; RV, right ventricle; RVOT, right ventricular outflow tract; SVC, superior vena cava; T, trachea; TV, tricuspid valve.

A

B

C

Fig. 12.6  Normal Cardiac Anatomy on Bright-Blood MR. Two-chamber (A), four-chamber (B) and threechamber (C) views. Ao, Aorta; AV, aortic valve; ch, chordae tendineae; LA, left atrium; LV, left ventricle; M, moderator band; MV, mitral valve; P, papillary muscle; RA, right atrium; RV, right ventricle; TV, tricuspid valve.

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

271

B

A

Fig. 12.7  Transverse Black-Blood (A) and Bright-Blood (B) MR Acquisition Illustrating the Moderator Band in the Right Ventricle. Ao-Desc, Descending aorta; LA, left atrium; LV, left ventricle; M, moderator band; RA, right atrium; RV, right ventricle; TV, tricuspid valve.

with suspected coronary artery disease and for several other cardiac indications. This includes evaluation of cardiac structures and function in adult congenital heart disease, evaluation of ventricular morphology, left and right ventricular systolic function, the pericardium and evaluation of intra- and extracardiac structures such as valves, masses and pulmonary veins. Stress CT myocardial perfusion may be used for evaluating myocardial perfusion for the detection of functionally significant coronary artery disease. Cardiac CT requires intravenous injection of iodinated contrast agent, with the exception of CT calcium scoring, which requires unenhanced cardiac CT. Various cardiac CT imaging techniques are applied in clinical practice, depending on the CT device used, on the clinical question that must be answered and on patient-related parameters. Imaging can be performed with helical acquisition, with step-and-shoot ‘sequential’ acquisition or with wide-volume acquisition. With helical and with step-and-shoot acquisition, multiple heart beats are needed for cardiac imaging, requiring a breath-hold of approximately 10 seconds. Images are reconstructed with a slice thickness down to 0.5 mm, with a temporal resolution down to 83 ms. Also, fast dual-source helical techniques and wide-volume detectors that cover the whole heart allow for imaging of the heart within a single heart beat at a radiation dose below 5 mSv. In some protocols, the dose can be as low as 1 mSv. ECG recording is central in all cardiac CT techniques to avoid motion artefacts. For optimal coronary artery imaging, after assessment of blood pressure, patients are generally prepared with β-blockers to slow the heart rate to below 65 beats/min and with nitroglycerine to dilate the coronary arteries. In patients with a regular heart rhythm, the acquisition is preferably performed with prospective ECG triggering. This allows imaging during a predefined cardiac rest phase with least motion, which is at mid-diastole at approximately 75% of the cardiac -R cycle. In patients with high heart rates that are susceptible to motion artefacts, image acquisition may be performed, with a wider acquisition interval during the cardiac cycle, allowing for selection and reconstruction of multiple cardiac phases for optimal motion-free imaging of each coronary artery. Functional acquisition may be needed such as in patients with valvular disease (e.g. before trans-catheter aortic valve implantation), or in patients with congenital heart disease and with contraindications for magnetic resonance imaging. For functional imaging, including

calculation of the end-diastolic and end-systolic volumes and ejection fraction, data acquisition throughout the cardiac cycle is required. The current prospective triggering acquisition techniques have effectively reduced patient exposure to radiation. With the older retrospective gating techniques, CT data were continuously acquired throughout several consecutive cardiac cycles with simultaneous recording of the ECG. This allows any cardiac phase to be reconstructed but at the expense of a relatively high radiation dose in the range of 12–21 mSv. The technique may be used for combined imaging of the coronary arteries and ventricular function (ejection fraction), or in patients with irregular heart rhythm or high heart rate. Radiation dose for functional analysis may be reduced by ECG-dose modulation techniques if coronary artery information is also needed, or by low-dose CT data acquisition throughout the cardiac cycle if only functional information is required. Cardiac CT imaging is a three-dimensional volume technique, which implies that any imaging plane can be reconstructed. Therefore, the use of ‘standard views’ is less critical than with projection invasive coronary angiography or echocardiography, or than with image-stack CMR. CT investigations are evaluated by reviewing the appropriate (multiplanar) image reconstructions that depend on the topic of interest and the clinical question that must be answered.

Computed Tomography Imaging of Ventricles and Myocardial Tissue The myocardial tissue can be visualised in any plane. The RV can be best evaluated by scrolling through the transverse images (Fig. 12.9). Two-, three- and four-chamber views, and especially SA views, are helpful for visualising the LV and left ventricular myocardium (Fig. 12.10). The SA views can be used for 17-segment evaluation of the left ventricular myocardium, and enable good correlation with echocardiography, CMR and nuclear medicine techniques.

Coronary Arteries by Computed Tomography Because of the good spatial and temporal resolution, the main coronary arteries and large side branches can be well visualised by coronary CT angiography. CT can establish the dominance of coronary circulation. The location of origins and courses of the coronary arteries can be visualised, along with coronary anatomical anomalies. CT allows

272

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E Fig. 12.8  A Segmented Gradient-Echo MR Acquisition of the Aortic Valve. In (A and B), the planning of the acquisition plane (dashed line) is presented in black-blood coronal view of the aorta during end-diastole (A) and bright-blood at peak systole (B). In (C), a closed normal valve at end-diastole and in (D), an opened normal valve with three cusps at peak systole is presented (L, Left coronary cusp; N, non-coronary cusp; R, right coronary cusp). In (E), a bicuspid aortic valve is presented, with a fused non-coronary and right coronary cusp.

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

Fig. 12.9  Right Ventricle. Transverse reconstruction showing the right ventricle (RV). Ao-Desc, Descending aorta; FO, fossa ovalis; LA, left atrium; LV, left ventricle; M, moderator band; MV, mitral valve; P, papillary muscle; RA, right atrium; RCA, right coronary artery; RV, right ventricle.

273

evaluation of coronary lumen and vessel wall for assessing coronary artery stenoses. Dominance of the coronary arteries refers to the artery that gives rise to the posterior descending artery. The right coronary artery is dominant in approximately 80% of the population and the left circumflex artery in approximately 10%. The circulation is balanced (co-dominant) in the remaining population. Evaluating the dominancy of the coronary arteries prevents confusion with branch occlusion, as in right dominant circulation a relatively small circumflex artery can be expected and in left dominant circulation a small right coronary artery is expected (Fig. 12.11). Fig. 12.12 shows normal coronary anatomy in transverse orientation with proximal, middle and distal segments of coronary arteries. Volume-rendered images show coronary anatomy in Fig. 12.13. The right coronary artery arises from the right sinus of Valsalva and courses in the right atrioventricular groove (between the right atrium and RV). Its side branches are usually visualised by scrolling through the transverse image stack: the conus branch coursing along the RVOT, the atrioventricular branch coursing posteriorly to the sinus node (at the junction of the superior vena cava and right atrium), the acute marginal/right ventricular branches along the right ventricular free wall (see Fig. 12.11A), the posterior descending branch in the posterior interventricular groove and a posterolateral branch that continues in the left atrioventricular groove (see Figs 12.11–12.13). The left main coronary artery arises from the left sinus of Valsalva (see Figs 12.11–12.13). It divides into the left anterior descending artery that runs in the anterior interventricular groove (between the LV and

A

B

C

D

E

F

Fig. 12.10  Left Ventricular Orientation. Longitudinal two-chamber (A), three-chamber (B) and four-chamber (C) reconstructions. Left ventricular short-axis reconstructions at the base (D), mid-ventricular (E) and apical level (F). Ao, Aorta; LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

274

SECTION A  The Chest and Cardiovascular System

A

B Fig. 12.11  Coronary Dominance. Three-dimensional volume-rendered computed tomography images with frontal view and view from below showing right dominant coronary artery circulation (A) and left dominant coronary artery circulation (B). In right dominant coronary artery circulation, the posterior descending artery (PD) arises from the right coronary artery (RCA). In left dominant coronary artery circulation, the PD arises from the circumflex artery (Cx) (B). Note the short RCA with empty right atrioventricular groove (RAVG), which is normal in left dominant coronary artery circulation, and should not be confused with RCA occlusion. Side branches visualised: AM, acute marginal branch; D, diagonal branch; LAD, left anterior descending coronary artery; LV, left ventricle; OM, obtuse marginal branch; PL, posterolateral branch.

RV) and the circumflex artery that runs in the left atrioventricular groove (see Fig. 12.12). In about one-third of the population, an intermediate artery arises from the left main artery which courses along the left ventricular wall between the left anterior descending and circumflex artery. The left main coronary artery may be absent in 1% of the population. In these cases, the left anterior descending and circumflex arteries arise from a common origin or separately from the left sinus

of Valsalva (Fig. 12.14). The left anterior descending artery gives rise to septal branches that run straight down into the interventricular septum, and to diagonal branches (usually one to three) that course along the anterolateral left ventricular wall (see Figs 12.11 and 12.12). The circumflex artery gives rise to obtuse marginal branches (usually one to three) supplying the lateral free wall of the LV (see Figs 12.11 and 12.12). In approximately one-third of the population, the sinusnode

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

A

B

C

D

275

Fig. 12.12  Coronary Anatomy: Segments. Transverse computed tomography reconstructions showing (A) the left main coronary artery (LM); (B) proximal right coronary artery (proximal-RCA), proximal left anterior descending artery (proximal-LAD) with diagonal side branch (D), and proximal circumflex artery (proximal-Cx) with obtuse marginal branch (OM); (C) mid-LAD, mid-RCA and mid-Cx/OM; (D) distal-RCA and posterior descending branch (PD), distal-LAD. Pericardium is visualised as a thin line (arrows in D). Visualised pulmonary veins: D, diagonal branch; LA, left atrium; LAA, left atrium appendage; LIPV, left inferior pulmonary vein; LV, left ventricle; RMPV, right middle pulmonary vein; RSPV, right superior pulmonary vein.

branch arises from the left circumflex artery instead of the right coronary artery. Furthermore, a left atrial circumflex branch that supplies part of the left atrium may be observed. The coronary arteries and their major side branches can be classified by location, by segment numbers or segment names for locating coronary artery stenosis. As several numeric classification systems are used in clinical and research practice, using numbers may be

confusing. Therefore, it may be more practical to use the segment names (see Fig. 12.12). Evaluation of the coronary arteries is done on original standard transverse views or orthogonal reconstructions because these images contain all information without risk of reconstruction artefacts. (Curved) multiplanar and surface rendering reconstruction can be additionally helpful for overview and fast reading by projecting larger parts of the

276

SECTION A  The Chest and Cardiovascular System

A

B

C

D Fig. 12.13  Coronary Arteries and Cardiac Veins. Volume-rendered computed tomography reconstructions for coronary artery and cardiac venous anatomy. Coronary arteries and side branches: AM, acute marginal branch of right coronary artery (RCA) (A); Cx, circumflex artery (B and C); D1 and D2, first and second diagonal branch (B); LAD, left anterior descending artery (B); LM, left main coronary artery (B); OM, obtuse marginal branch (B and C); PD, posterior descending branch (D); PL, posterolateral branch from RCA (D); RCA, right coronary artery (A and D); RVB, right ventricle branch (running to distal part of posterior interventricular groove, A and D). Cardiac veins: AIV, anterior interventricular vein (B); CS, coronary sinus (D); GCV, great cardiac vein (C and D); LMV, left marginal vein (C); PIV, posterior interventricular vein (D); PLVV, posterior left ventricular vein (C and D); RAV, right atrial vein draining directly into right atrium (A).

coronary arteries within single images (Fig. 12.15). For comparison with CT, Figs 12.16 and 12.17 show views of the coronary anatomy obtained with invasive angiography. Invasive angiography is still considered the gold standard for evaluation of the coronary anatomy. It has a high temporal and spatial resolution and provides lumen evaluation that is not limited by the presence of calcium. The main disadvantage of

invasive angiography is its invasive fashion, although at low risk of complications.

Valves Echocardiography and CMR are the preferred imaging techniques for evaluating the cardiac valves, as these techniques provide advanced

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

Fig. 12.14  Absent Left Main. Double oblique orientation parallel to the aortic root showing separate coronary ostia of left anterior descending artery (LAD) and circumflex artery (Cx). The left main artery is absent. Right coronary artery (RCA) with conus branch (CB).

A

B

C

D

E

F

Fig. 12.15  Full-Length Display of Coronary Arteries. Three-dimensional volume-rendered computed tomography reconstructions in right anterior oblique (A) and left anterior oblique (D) view and curved multiplanar reconstructions (B, C, E, and F) showing each coronary artery in two longitudinal perpendicular directions: the right coronary artery (RCA, in B), left anterior descending coronary artery (LAD, in C), circumflex artery (Cx, in E) and obtuse marginal branch (OM, in F). Note that the OM is much larger than the Cx (E) itself, which is usually the case. LM, Left main.

277

278

SECTION A  The Chest and Cardiovascular System

Fig. 12.16  Invasive Coronary Angiography of the Right Coronary Artery in Two Different Directions. Left panel, left anterior oblique 45 degrees; right panel, right anterior oblique 35 degrees; conus, Conus branch; RDP, right posterior descending branch; RPL, right posterolateral branch; RV, right ventricular branch.

Fig. 12.17  Invasive Coronary Angiography of the Left Coronary Artery in Two Different Directions. Left panel, right anterior oblique 30 degrees, 25 degrees caudal angulation; right panel, left anterior oblique 50 degrees, 25 degrees cranial angulation. Cx, Circumflex coronary artery; D-branches, diagonal branches; LAD, left anterior descending coronary artery; LM, left main coronary artery; LPL, left posterolateral branch, OM, obtuse marginal branch; S-branches, septal branches.

functional imaging with superior temporal resolution without the use of ionising radiation. However, CT allows detailed information on aortic and mitral valve morphology and can provide functional information as well. The pulmonary and tricuspid valve can also be visualised but have thinner cusps and are, therefore, more difficult to appreciate on CT. For each valve, multiphase cine views may be helpful for evaluation.

The normal aortic valve is composed of a fibrous annulus, three cusps (right coronary cusp, left coronary cusp, posterior or non-coronary cusp) and three commissures that separate the cusps. The dilatations in the aorta at each cusp are the sinuses of Valsalva. The aortic cusps open at systole and close at diastole with a small area of overlap. The normal aortic valve area at opening is 3.0–4.0 cm2. The aortic valve is

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

A

B

279

C

Fig. 12.18  Aortic Valve. Multiplanar reconstructions at mid-diastole showing the closed aortic valve (AV) in coronal view (A), three-chamber view (B) and short-axis parallel to the aortic valve (C). Arrows in (C) point at the commissures. L, Left coronary cusp; LA, left atrium; LM, left main coronary artery; LV, left ventricle; MV, mitral valve; N, non-coronary cusp; R, right coronary cusp; S, sinus of Valsalva.

A

B Fig. 12.19  Bicuspid Aortic Valve. ‘Short-axis’ double oblique transverse images parallel to the aortic valve, showing bicuspid aortic valve at diastole (A, closed) and at systole (B, slit-like opening due to fusion of left and right coronary cusp). Note the difference in noise level between the images, caused by ECG-dose modulation with full dose at diastole (during the cardiac rest phase for sharp imaging of the coronary arteries), and lower radiation dose at systole to save radiation dose.

well visualised on coronal view and with multiplanar reconstruction (MPR) on coronal three-chamber view and parallel at the valvular plane itself (Fig. 12.18). A bicuspid aortic valve is present in 1–2% of the population (Fig. 12.19) and often results in complications such as aortic valve stenosis and/or regurgitation. The mitral valve is composed of a saddle-shaped fibrous annulus, two leaflets (a semicircular anterior leaflet and a rectangular posterior leaflet), two commissures, two papillary muscles (anterolateral and posteromedial) and chordae tendineae, which are fibrous tendons that arise from the papillary muscles and insert on the free edges of the leaflets. The normal mitral valve area during opening in diastole is 4.0–5.0 cm2. The mitral valve leaflets close with some overlap to prevent regurgitation. The mitral valve can be visualised on two-, three- and four-chamber views and on SA MPR (Fig. 12.20).

The pulmonary valve contains a fibrous annulus, three cusps and three commissures separating the cusps. Optimum imaging views are the sagittal plane showing the RVOT and pulmonary valve, and with MPR at the valvular plane itself (Fig. 12.21). If the pulmonary cusps are easily visible, they are likely to be thickened. The tricuspid valve has three cusps (anterior, superior, inferior). The tricuspid valve is usually moderately visualised, due to the thin cusps; its optimal views are transverse or four-chamber long- and short-axis views (Fig. 12.22).

Pulmonary Veins CT provides excellent view on the left atrium and pulmonary veins that can be used as a road map for guiding radiofrequency catheter ablation therapy. Volume rendering (Fig. 12.23) and transverse views (see Fig.

280

SECTION A  The Chest and Cardiovascular System

A

B

C

D Fig. 12.20  Mitral Valve. Multiplanar reconstructions at mid-diastole showing the mitral valve (MV) in longitudinal two-chamber view (A), three-chamber view (B), four-chamber view (C) and short-axis view (D). Anterior mitral leaflet (arrow), posterior mitral leaflet (arrowhead). AV, Aortic valve; Ch, chordae tendineae; LV, left ventricle; P, papillary muscles.

12.12) show the pulmonary vein and left atrium anatomy well. Orthogonal sagittal views may be used for pulmonary vein ostium evaluation.

Other Structures CT also visualises other (cardiovascular) structures in the same acquisition, including cardiac veins (see Fig. 12.13) and the normal pericardium being visualised as a thin line (see Fig. 12.12). Pulmonary CT angiography is the imaging test of first choice in patients with suspected pulmonary embolism and aortic CT angiography is the preferred imaging technique for evaluating patients with suspected acute aortic syndrome. Other

thoracic structures beyond the heart and great vessels—the hilum, lungs, chest wall and bones—can be optimally visualised by viewing the images at the appropriate window-width and window-level settings. SUMMARY BOX: Computed Tomography • Various imaging planes (body axis- and cardiac planes) • Anatomy of thoracic vascular structures • Cardiac anatomy • Anatomy coronary arteries and veins

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

A

B

281

C

Fig. 12.21  Pulmonary Valve. Coronal (A), sagittal (B) and multiplanar reconstruction parallel to the pulmonary valve (C) showing the pulmonary valve (arrows). AV, Aortic valve; LA, left atrium; LV, left ventricle; RV, right ventricle.

A

B Fig. 12.22  Tricuspid Valve. Transverse reconstruction (A) and short-axis reconstruction parallel to the tricuspid valve (B), showing the tricuspid valve (arrows). Note that the tricuspid valve is difficult to recognise compared with the mitral valve (B). LV, Left ventricle; MV, mitral valve; PV, pulmonary valve; RV, right ventricle.

ECHOCARDIOGRAPHY Cardiac ultrasound is currently one of the most important imaging techniques in clinical cardiology because it is quickly performed, readily available (even at bedside) and bears low costs. Echocardiography has evolved from M-mode ultrasound to 2-D and, recently, to 3-D image orientation. Echocardiography uses high-frequency ultrasound (2.0–7.5 MHz), and the nature of the ultrasound waves are such that the use of echocardiography is harmless and no x-rays are involved. Ultrasound waves do not traverse interfaces with air (lung) or bone (rib/ sternum) and are attenuated by body fat. Therefore, in patients with chronic obstructive pulmonary disease and in obese patients, suboptimal windows with limited quality are almost inevitable. Cardiac and vascular anatomy can be assessed with 2-D echocardiography and cardiac and vascular function can be assessed with ECG

traced cine images. The temporal resolution of the cine images depends on the various settings and ranges typically from 30 to 100 frames per second (temporal resolution: 10–33 ms). With the use of the Doppler technique, blood-flow velocities can be measured to estimate pressure gradients in a non-invasive fashion and colour Doppler allows for non-invasive assessment of blood flow. The Doppler techniques are particularly important for the assessment of valvular function. Since the focus of this chapter is mainly on cardiac anatomy, functional and Doppler aspects are not addressed. Trans-thoracic echocardiography consists of several standard views and usually starts with the patient in the left lateral decubitus position at the left parasternal position at the third to fourth intercostal space. Then, apical views are performed from the apex of the heart. In the supine position, subcostal views and suprasternal views can be obtained. Additional customoriented views may be obtained depending on the clinical questions.

282

SECTION A  The Chest and Cardiovascular System

B

A

Fig. 12.23  Pulmonary Veins. Volume-rendered (A) and maximum intensity projection (B) reconstructions, dorsal view (A) and frontal view (B). LA, Left atrium; LIPV, left inferior pulmonary vein; LSPV, left superior pulmonary vein; RIPV, right inferior pulmonary vein; RSPV, right superior pulmonary vein.

Fig. 12.24  Parasternal Long-Axis View. Ao, Aorta; AV, aortic valve; LA, left atrium; LV, left ventricle; MV, Mitral valve; RVOT, right ventricular outflow tract.

Fig. 12.25  Parasternal Short-Axis Aorta View. AV, Aortic valve; IAS, interatrial septum; LA, left atrium; LPV, left pulmonary veins; PV, pulmonary valve; RA, right atrium; RVOT, right ventricular outflow tract; TV, tricuspid valve.

Fig. 12.26  Parasternal Short-Axis Pulmonary Artery View. AV, Aortic valve; LPA, left pulmonary artery; PA, main pulmonary artery; PV, pulmonary valve; RPA, right pulmonary artery; RVOT, right ventricular outflow tract.

The parasternal position enables LA and SA views. The parasternal long-axis view (Fig. 12.24) shows the anteroseptal and the inferolateral/ posterior wall of the LV. This view is used to assess end-diastolic and end-systolic LV dimensions and normal values are derived from this orientation. The dimensions of the left atrium, the aorta and the RVOT and left ventricular outflow tract can also be measured from this view. Furthermore, identification of the aortic valve and mitral valve is possible. By rotating the transducer 90 degrees, the parasternal SA view at the level of AV (Fig. 12.25) allows assessment of the right atrium, left atrium, tricuspid valve, RVOT (in transverse orientation from the parasternal long-axis view), pulmonary veins, pulmonary valve and the interventricular septum. Of note, the view of the AV is a transverse orientation from the parasternal long axis and provides clear identification of the aortic leaflets and also the pulmonary artery branches may be identified (Fig. 12.26). At the level of the basis of the LV (Fig. 12.27), all six basal segments of the myocardium referring to the 17-segment model can be identified. Furthermore, a part of the RV, mitral leaflets and, occasionally, the moderator band are shown. At the mid-level of the LV (Fig. 12.28), the six mid-segments of the myocardium from the 17-segment model are identified, as well as a mid-portion of the RV and the papillary

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

283

A Fig. 12.27  Parasternal Short-Axis Basal Left Ventricle View. M, Moderator band; MV, mitral valve; MV aml, anterior leaflet mitral valve; MV pml, posterior leaflet mitral valve; RV, right ventricle.

B Fig. 12.30  (A) Apical Four-Chamber View and (B) Detail of the Interatrial Septum. FO/SS, Fossa ovalis/secondary septum; IAS, inter-atrial septum; L, lateral myocardium; LA, left atrium; LPV, left pulmonary veins; LV, left ventricle; M, moderator band; MV, mitral valve; PS, primary septum; PV, pulmonary vein; RA, right atrium; RV, right ventricle; S, septal myocardium; TV, tricuspid valve. Fig. 12.28  Parasternal Short-Axis Mid-Left Ventricle View. AP, Anterior papillary muscle; LV, Left ventricle; PP, posterior papillary muscle; RV, right ventricle.

Fig. 12.29  Parasternal Short-Axis Apex Left Ventricle. A, Apex left ventricle; RV, right ventricle.

muscles. The last parasternal SA view is the apical view and shows the four apical segments (Fig. 12.29). The apical views show the apex of the heart usually on top and the atria on the bottom of the image. The first apical view shows the LV, RV, left atrium and right atrium and is, therefore, named the ‘fourchamber’ view, although the left atrium and right atrium are not real chambers (Fig. 12.30A). Furthermore, the interatrial septum (consisting of the primary and secondary parts; Fig. 12.30B), the pulmonary veins, mitral valve, tricuspid valve and the septal and lateral myocardium are identified. With a 90-degree rotation of the transducer from the four-chamber view, the two-chamber view is obtained (Fig. 12.31), showing the LV with the anterior and inferior myocardium, and the left atrium as the second ‘chamber’. From the four-chamber view, the five-chamber view (Fig. 12.32) is then obtained, showing the four chambers and the aorta as the fifth chamber. This view identifies the anteroseptal- and inferolateral-posterior myocardium, the aortic valve and the left ventricular outflow tract. When the regular four-chamber view is angulated towards the RV, the apical RV view (Fig. 12.33) is obtained, showing a greater part of the RV, and allows assessment of right ventricular function. The subcostal views are performed via a sub-xiphoid approach with the transducer in the direction of the heart and show again a fourchamber view (Fig. 12.34), which has the advantage of a detailed view

284

SECTION A  The Chest and Cardiovascular System

Fig. 12.31  Apical Two-Chamber View. A, Anterior myocardium; I, inferior myocardium; LA, left atrium; LV, left ventricle; MV, mitral valve.

Fig. 12.34  Subcostal Four-Chamber View. LA, Left atrium; LV, left ventricle; MV, mitral valve; pc, pericardium; RA, right atrium; RV, right ventricle; TV, tricuspid valve.

Fig. 12.32  Apical Five-Chamber View. Ao, Aorta-ascendens; AS, anteroseptal myocardium; AV, aortic valve; IL-P, inferolateral-posterior myocardium; LA, left atrium; LV, left ventricle; LVOT, left ventricular outflow tract; MV, mitral valve; RA, right atrium; RV, right ventricle.

Fig. 12.35  Subcostal Inferior Vena Cava View. IVC, Inferior vena cava; L, liver; RA, right atrium.

Fig. 12.33  Apical Right Ventricle View. RA, Right atrium; RV, right ventricle; RV-L, right ventricle lateral myocardium; S, septal myocardium; TV, tricuspid valve.

Fig. 12.36  Suprasternal View. AC, Communal carotid artery; Ao-Arch, aortic arch; Ao-Asc, ascending aorta; Ao-Desc, descending aorta; LSA, left subclavian artery; RPA, right pulmonary artery.

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

285

Fig. 12.37  Aortic Valve Short-Axis View. Normal aortic valve with right coronary cusp (R), left coronary cusp (L) and non-coronary cups (N) in closed (upper panel) and open position (lower panel).

Fig. 12.38  Bicuspid Aortic Valve. Functionally bicuspid aortic valve with non-coronary cusp (N), and fusion (Raphe) of the right coronary cusp (R) and left coronary cusp (L). Upper panel shows the aortic valve in closed and lower panel in open position.

on the pericardium and the pericardial space to assess pericardial effusion. This view is often used after thoracic operations in which the apical and parasternal windows are usually limited. Also, the inferior vena cava is seen and the inferior vena cava dimension, indicative of right atrial pressure, can be measured (Fig. 12.35). The last standard view is the suprasternal view (Fig. 12.36) obtained from the suprasternal notch and this view enables assessment of the proximal ascending aorta, the aortic arch and its upper side branches and the descending aorta.

is situated posteriorly to the heart with almost no interposed tissue. The main disadvantage is discomfort for the patient, sometimes even requiring light anaesthesia. Fig. 12.39 provides detailed anatomy of the aortic valve and other atrial structures such as left atrial appendage and pulmonary veins. Fig. 12.40 shows anatomy of the mitral valve with identification of all the different parts of the anterior and posterior mitral valve leaflets. These detailed views provide better insight into the mechanism causing mitral valve regurgitation or stenosis, compared with trans-thoracic echocardiography. For the mitral valve, an ‘en face’ view like in the aortic valve (see Fig. 12.37), is usually not possible since the mitral valve is a saddle-shaped structure and, therefore, the valve cannot be viewed in a single plane with a 2-D technique. Three-dimensional echocardiography was introduced for identifying valvular anatomy, and examples are shown in Figs 12.41 and 12.42. Three-dimensional images provide a better orientation of the valve anatomy in relation to the surrounding structures. However, acquisition and spatial and temporal resolution of 3-D images are still to be improved.

Valves Apart from the above-mentioned structures, echocardiography is particularly suitable for assessing valvular anatomy and function with Doppler techniques. All four cardiac valves can be assessed with regular 2-D echocardiography but in clinical practice, most lesions affect the aortic valve and mitral valve. Echocardiographic example of a normal aortic valve (a detail from the parasternal SA aorta view; see Fig. 12.25) allows identification of the three leaflets (Fig. 12.37). In case of a bicuspid aortic valve, echocardiography can distinguish the anatomy of a true bicuspid valve or a bicuspid valve with fused leaflets with a raphe (Fig. 12.38). Also, the mitral valve anatomy can be looked at with 2-D echocardiography in the parasternal long-axis view, the apical fourchamber view and the two-chamber view (see Figs 12.24, 12.30 and 12.31). For detailed assessment of the cardiac valves, trans-oesophageal echocardiography can also be performed. Compared with trans-thoracic echocardiography, trans-oesophageal echocardiography provides closer views on the valves and atria with less attenuation since the oesophagus

SUMMARY BOX: Echocardiography Echocardiography • Transthoracic and transoesophageal modalities • Cardiac anatomy and function • Detailed anatomy of cardiac valves

286

SECTION A  The Chest and Cardiovascular System

Fig. 12.39  Trans-Oesophageal Echocardiographic View of the Aortic Valve in Transverse/Short-Axis View (Left Panel) and Longitudinal View (Right Panel). Ao, Ascending aorta; AV, aortic valve; L, left coronary cusp; LA, left atrium; LAA, left atrial appendage; LV, left ventricle; N, non-coronary cusp; R, right coronary cusp; RVOT, right ventricular outflow tract.

A

B

C

D Fig. 12.40  Trans-Oesophageal Echocardiographic Views of the Mitral Valve. TOE views of the mitral valve; four-chamber view (A), bi-commissural view (B), two-chamber view (C), three-chamber view (D). With multiplanar views the different parts of the anterior mitral valve leaflet (A1, A2, A3) and the posterior leaflet (P1, P2, P3) can be visualised. LA, Left atrium.

CHAPTER 12  Cardiac Anatomy and Imaging Techniques

Fig. 12.41  Three-Dimensional View of the Aortic Valve in Closed (Upper Panel) and in Open Position (Lower Panel). The image is obtained with trans-oesophageal echocardiography. L, left coronary cusp; N, non-coronary cusp; R, right coronary cusp.

ACKNOWLEDGEMENT G. Kracht, F. van der Kley and E. R. Holman are acknowledged for their assistance with the figures of this chapter.

FURTHER READING Achenbach, S., Marwan, M., Ropers, D., et al., 2010. Coronary computed tomography angiography with a consistent dose below 1 mSv using prospectively electrocardiogram-triggered high-pitch spiral acquisition. Eur. Heart J. 31, 340–346. Alfakih, K., Plein, S., Bloomer, T., et al., 2003. Comparison of right ventricular volume measurements between axial and short axis orientation using steady-state free precession magnetic resonance imaging. J. Magn. Reson. Imaging 18, 25–32. Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography), 2003. ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography: summary article. A report of the American College of Cardiology/American Heart Association. Circulation 108, 1146–1162. Cerqueira, M.D., Weismann, N.J., Dilsizian, V., et al., 2002. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart association. Circulation 105, 539–542.

287

Fig. 12.42  Three-Dimensional View on the Mitral Valve With Anterior Leaflet and Posterior Leaflet in Closed (Upper Panel) and in Open Position (Lower Panel). The image is obtained with trans-oesophageal echocardiography and the valve is seen from the left atrium. The different parts of the anterior mitral valve leaflet (A1, A2, A3) and the posterior leaflet (P1, P2, P3) can be determined. Chen, J.J., Manning, M.A., Frazier, A.A., et al., 2009. CT angiography of the cardiac valves: normal, diseased, and postoperative appearances. Radiographics 29, 1393–1412. Chhedda, S.V., Srichai, M.B., Donnino, R., et al., 2010. Evaluation of the mitral and aortic valves with cardiac CT angiography. J. Thorac. Imaging 25, 76–85. Earls, J.P., Berman, E.L., Urban, B.A., et al., 2008. Prospectively gated transverse coronary CT angiography versus retrospectively gated helical technique: improved image quality and reduced radiation dose. Radiology 246, 742–753. Einstein, A.J., Elliston, C.D., Arai, A.E., et al., 2010. Radiation dose from single-heartbeat coronary CT angiography performed with a 320-detector row volume scanner. Radiology 254, 698–706. Ho, S.Y., Nihoyannopoulos, P., 2006. Anatomy, echocardiography, and normal right ventricular dimensions. Heart 92, i2–i13. Jongbloed, M.R., Dirksen, M.S., Bax, J.J., et al., 2005. Atrial fibrillation: multi-detector row CT of pulmonary vein anatomy prior to radiofrequency catheter ablation—initial experience. Radiology 234, 702–709. Jongbloed, M.R., Lamb, H.J., Bax, J.J., et al., 2005. Noninvasive visualization of the cardiac venous system using multislice computed tomography. J. Am. Coll. Cardiol. 45, 749–753. Kramer, C.M., Barkhausen, J., Flamm, S.D., et al., 2008. Standardized cardiovascular magnetic resonance imaging (CMR) protocols, Society for Cardiovascular Magnetic Resonance: Board of Trustees Task Force on Standardized Protocols. J. Cardiovasc. Magn. Reson. 10, 10–35. Lang, R.M., Bierig, M., Devereux, R.B., et al. Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards

288

SECTION A  The Chest and Cardiovascular System

Committee; European Association of Echocardiography., 2005. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J. Am. Soc. Echocardiogr. 18, 1440–1463. Manghat, N.E., Rachapalli, V., van Lingen, R., et al., 2008. Imaging the heart valves using ECG-gated 64-detector row cardiac CT. Br. J. Radiol. 81, 275–290. Mansour, M., Holmvang, G., Sosnovik, D., et al., 2004. Assessment of pulmonary vein anatomic variability by magnetic resonance imaging: implications for catheter ablation techniques for atrial fibrillation. J. Cardiovasc. Electrophysiol. 15 (4), 387–393. Meijer, A.B., O, Y.L., Geleijns, J., et al., 2008. Meta-analysis of 40- and 64-MDCT angiography for assessing coronary artery stenosis. AJR. Am. J. Roentgenol. 191, 1667–1675.

Miller, S., Simonetti, O.P., Carr, J., et al., 2002. MR imaging of the heart with cine true fast imaging with steady-state precession: influence of spatial and temporal resolutions on left ventricular functional parameters. Radiology 223, 263–269. Taylor, J.A., Cerqueira, M., Hodgson, J.M., et al., 2010. ACCF/SCCT/ACR/ AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010 appropriate use criteria for cardiac computed tomography: a report of the American College of Cardiology Foundation Appropriate Use Criteria Task Force, the Society of Cardiovascular Computed Tomography, the American College of Radiology, the American Heart Association, the American Society of Echocardiography, the American Society of Nuclear Cardiology, the Society for Cardiovascular Angiography and Interventions, and the Society for Cardiovascular Magnetic Resonance. J. Am. Coll. Cardiol. 56, 1864–1894. Uribe, S., Muthurangu, V., Boubertakh, R., et al., 2007. Whole-heart cine MRI using real-time respiratory self-gating. Magn. Reson. Med. 57, 606–613.

13  Congenital Heart Disease: General Principles and Imaging Michael A. Quail, Andrew M. Taylor

CHAPTER OUTLINE Introduction, 289 Clinical Presentation, 289 Morphological Description and Sequential Segmental Analysis, 290 Physiological and Functional Assessment, 292 Non-Invasive Imaging Techniques, 292

INTRODUCTION Although rare, with an incidence of 8 per 1000 births, congenital heart disease (CHD) has increased in prevalence due to the success of surgical and medical management in childhood. A significant proportion of patients with repaired CHD surviving to adulthood fall under the care of cardiologists outside tertiary centres for congenital cardiac care. Specialist cardiovascular and general radiologists require an understanding of the underlying morphological abnormalities and their physiology, methods of repair and how potential complications may be detected and assessed in their practice, using appropriate imaging techniques, such as echocardiography, magnetic resonance imaging (MRI), cardiac magnetic resonance (CMR) and computed tomography (CT). CHD is any developmental malformation of the heart. The spectrum of disease falling into this classification ranges from simple lesions—for example bicuspid aortic valve—through to more complex diseases involving single ventricle lesions, such as hypoplastic left-heart syndrome. The underlying causes of CHD remain relatively poorly understood, although the epidemiology suggests a genetic basis contributing to the majority of CHD. Aneuploidies—for example, trisomy 21 (Down syndrome, septal defects) and monosomy X (Turner syndrome, bicuspid aortic valve and coarctation)—are the earliest identified causes and account for 10%–20% of CHD. Copy number variations (small to large deletions or duplications) lead to altered dosage of genes and may represent another important mechanism; an example is Del22q11, which causes DiGeorge syndrome (interrupted aortic arch, tetralogy of Fallot (TOF), truncus arteriosus). Unfortunately, the cause of CHD in most patients remains unknown.

CLINICAL PRESENTATION The clinical presentation of CHD in infancy may be dominated by a number of physiological states. 1. Left-to-Right Shunts: Redirection of blood from the systemic (left) to the pulmonary circulation (right) may occur at atrial, ventricular or great vessel level. A proportion of already oxygenated blood is

Specific Lesions, 296 Acyanotic Lesions, 296 Cyanotic Congenital Heart Disease, 304 Single Ventricles, 311 Conclusion, 313

recirculated to the lungs with each heartbeat, resulting in inefficiency. The volume of the shunt and its location accounts for the observed signs. Chambers and vessels receiving the excessive volume enlarge and high pulmonary blood flow results in pulmonary plethora. Typical examples include atrial septal defects (ASDs), ventricular septal defects (VSDs) and patent ductus arteriosus (PDA). Patients are pink but increasingly breathless with larger shunts. 2. Compromised Systemic Perfusion: This may result from low stroke volume of a systemic ventricle (hypoplastic left-heart syndrome), outflow tract obstruction (critical aortic stenosis (AS)) or aortic obstruction (interrupted aortic arch or coarctation). The clinical picture is one of poor peripheral perfusion, with low-pulse volume; patients may be pink or blue (cyanotic). The ductus arteriosus may provide an effective temporary bypass for the obstruction, facilitating systemic perfusion with deoxygenated or mixed blood; however, as the duct closes (some days after birth), life-threatening systemic or lower body hypoperfusion ensues and, often, pulmonary venous hypertension. Therapy is directed at maintaining the patency of the arterial duct using intravenous prostaglandins, intensive care for critically ill patients and planning for surgical relief of the obstruction. 3. Pulmonary Venous Congestion: Obstruction to pulmonary venous return results in increased pulmonary venous pressure (elevated pulmonary capillary wedge pressure); at progressively higher transvascular gradients, oncotic pressure is exceeded and extravasation of fluid into the interstitial and alveolar space occurs. Obstruction may occur in the pulmonary venous pathway (total anomalous pulmonary venous connection, TAPVD), in the atrium (cor triatriatum) or at the level of the left ventricle (LV) inflow (supravalvular, valvular or subvalvular mitral stenosis or mitral regurgitation). Pulmonary venous congestion may also occur as a function of elevated left atrial pressure secondary to LV diastolic dysfunction; increased LV end diastolic pressure (valve disease, aortic coarctation, myocardial disease). The degree of pulmonary venous hypertension determines the clinical presentation. Patients with severe obstruction may present with hypoxia, cyanosis and dyspnoea caused by pulmonary oedema,

289

290

SECTION A  The Chest and Cardiovascular System

whilst patients with less severe obstruction may remain pink but may present later with failure to thrive. The following three physiological states predominantly account for patients with cyanosis. 4. Low Pulmonary Blood Flow: Reduction in pulmonary blood flow is commonly caused by obstruction to outflow from the right ventricle (RV), e.g. TOF or severe pulmonary stenosis (PS). The increased resistance to RV outflow results in a redirection of systemic venous return to the left heart (right to left shunt) via an interatrial communication (patent foramen ovale or ASD) or a VSD. Elevated pulmonary vascular resistance (PVR) without anatomical obstruction may also result in shunt redirection. As lung function is normal, any pulmonary venous blood returns to the left atrium fully saturated and mixes with the shunted systemic venous blood; however, because pulmonary flow is so low, there is insufficient oxygenated blood in the mix, resulting in cyanosis. 5. Parallel Circulations: This occurs in transposition of the great arteries (TGA), where the aorta arises from the morphological right ventricle and the pulmonary artery from the left ventricle. In this condition, deoxygenated systemic venous return recirculates into the aorta and the oxygenated pulmonary venous return recirculates to the pulmonary artery, a situation clearly incompatible with life. Patients can only survive if there is sufficient mixing of the streams (shunt); this can occur best at atrial level through a large interatrial communication, less well at ventricular level (via a VSD) and even less well at great vessel level (via a PDA). Critical cyanosis may be managed medically by maintaining patency of the PDA by prostaglandins, but may require the creation of an artificial interatrial communication using cardiac catheterisation until definitive treatment by surgically switching the great vessels. 6. Intracardiac Mixing: Complete intracardiac mixing of blood may occur at atrial level (common atrium), ventricular level (all univentricular hearts) or great artery level (common arterial trunk). Patients are expected to be mildly cyanosed, depending on the relative amount of deoxygenated blood in the mix, and breathless, according to the amount of pulmonary blood flow.

Later Clinical Presentation The majority of adult patients with CHD are survivors from childhood. This group may present with an interesting array of problems related to residual lesions or deteriorations of their initial repair or palliation (heart failure, valve regurgitation, conduit stenosis, baffle leaks). They require life-long surveillance for anticipated problems arising from the ‘unnatural history’ of their underlying disorder. New presentations of CHD continue beyond infancy into adulthood, usually because the underlying disorder has not yet produced symptoms. Common lesions include left-right shunts, such as ASDs, VSDs or partial anomalous pulmonary venous drainage, which only begin to be symptomatic in older patients; milder forms of LV or aortic obstruction such as coarctation of the aorta or valvular AS, which did not compromise systemic perfusion, may progress and become symptomatic in later childhood or adulthood; and milder forms of RV obstruction such as PS. A very rare late presentation of CHD is congenitally corrected TGA (atrioventricular (AV) and ventriculo-arterial discordance). Here, the right atrium connects to the left ventricle, then the pulmonary trunk, and the left atrium to the right ventricle, then the aorta. Whilst blood flow is ‘normal’, the right ventricle is the systemic ventricle and may fail late in life or even found to be undiagnosed until a post-mortem study. An important presentation for unrepaired CHD is that of pulmonary arterial hypertension (PAH). PAH is a common complication of adult congenital heart disease (ACHD), affecting up to 10% of patients. PAH related to CHD is characterised by a rise in PVR with normal left atrial

pressure. It is typically the result of pulmonary vascular disease caused by chronically elevated pulmonary arterial pressures in patients with large post-tricuspid defects such as VSDs, PDAs or aortopulmonary windows. Chronic pressure and volume load cause proliferative lesions to develop in the small muscular pulmonary arteries, which result in elevated PVR. Clinical features include signs of elevated pulmonary artery (PA) pressure and subpulmonary (usually RV) maladaptation (Fig. 13.1). Clearly, life-limiting conditions such as parallel circulation, significant shunts, severe intracardiac mixing or compromised systemic perfusion would not normally be expected beyond childhood. SUMMARY BOX: Major Modes of Clinical Presentation of Congenital Heart Disease  • Left-to-right shunt (e.g. ASDs, VSDs, PDA). • Compromised systemic perfusion (e.g. critical aortic stenosis). • Pulmonary venous congestion (e,g, obstructed TAPVD). • Low pulmonary blood flow (e,g, tetralogy of Fallot). • Parallel circulation (e,g, transposition of the great arteries). • Intracardiac mixing (e,g, truncus arteriosus). ASDs, Atrial septal defects; PDA, patent ductus arteriosus; TAPVD, total anomalous pulmonary venous connection; VSDs, ventricular septal defects.

MORPHOLOGICAL DESCRIPTION AND SEQUENTIAL SEGMENTAL ANALYSIS A significant diversity of morphological abnormalities may be responsible for the physiological phenomena described above and although initial management of an infant simply requires correct classification of the initial physiological pattern, subsequent surgical correction and medical management requires a precise anatomical diagnosis. The potential intrinsic complexity of CHD necessitates a systematic scheme of nomenclature that captures precisely the unique anatomy of each patient, called sequential segmental analysis. Using this approach, the clinician describes how the components of the heart and blood vessels are connected. This entails describing atrial situs (location of the atrial chambers and whether they are of left or right morphology), atrioventricular (AV) connections, ventriculo-arterial (VA) connections and other associated lesions in turn. Any cross-sectional imaging technique may be used for this purpose but transthoracic echocardiography is most commonly used for routine inpatient and outpatient assessment. In more complex lesions or when echocardiography provides an inadequate assessment (e.g. poor acoustic windows), CMR represents a powerful non-invasive technique giving morphological and haemodynamic information that echocardiography alone cannot provide.

Sequential Segmental Analysis Step 1—Atrial Situs

Atrial situs is determined by an assessment of the morphology of each atrial appendage. Correct identification of the atrium allows the subsequent determination of the AV connection. The atrial appendages are the most consistent feature of the atrial mass; indeed, the venous attachments to each atrial chamber can form a variety of combinations. The right atrial appendage is a triangular shape, with a broad base and prominent pectinate muscles that extend around the right AV valve, whilst the left atrial appendage is a more elongated, tubular structure and has less extensive pectinate muscles that are confined within the appendage. The most common lesions involve inversion of situs, or isomerism of the left or right atrial appendages. The non-cardiac thoracic and abdominal organs usually (but not always) demonstrate a similar ‘sidedness’ to that of the atrial chambers.

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

A

B

C

D

291

Fig. 13.1  A 7-Year-Old Child With Pulmonary Arterial Hypertension and an Atrial Septal Defect. (A) Tricuspid regurgitation velocity 4 m/s, indicating an estimated right ventricular (RV) systolic pressure of 64 mm Hg + right atrial pressure. (B) Dilated and hypertrophied RV. (C) Abnormal interventricular septum curvature in late RV systole consistent with elevated RV systolic pressure. (D) Phase-contrast flow curve in right pulmonary artery. Arrow indicates mid-systolic ‘notch’ caused by pathological wave reflection arising in the abnormal pulmonary vasculature.

In the normal heart the morphological right atrium is located to the right of the morphological left atrium (situs solitus). The right lung is trilobed, with a shorter, early-branching bronchus and the left lung is bilobed. In addition, the inferior vena cava (IVC) is to the right of the abdominal aorta, with a right-sided liver and left-sided spleen. In situs inversus the mirror image of the normal anatomy is present. Isomerism of the left atrial appendages is usually associated with bilateral bilobed lungs, polysplenia and IVC interruption. Isomerism of the right atrial appendages is usually associated with bilateral triilobed lungs, asplenia and a midline liver. In isomeric lesions there is often a common AV junction (instead of two separate and offset left and right junctions) with varying degrees of AV septal defect (AVSD). Gut malrotation is associated with both right and left-sided isomerism.

Step 2—Ventricular Morphology Determination of ventricular morphology allows analysis of AV and ventriculo-arterial connections. An AV connection is described as ‘concordant’ when the atria are connected to the expected ventricle (i.e. left atrium with left ventricle and right atrium with right ventricle); ‘discordant’ if the left atrium is connected to the right ventricle and right atrium to the left ventricle; ‘ambiguous’ if there is isomerism of the atrial appendages (e.g. two morphologically right atria connected

to a left and right ventricle, respectively (one connection is concordant, the other discordant)); and, finally, ‘univentricular’ if both atria predominately connect to a single ventricle. Irrespective of AV concordance, the AV valve is always concordant with the ventricle—that is the tricuspid valve connects to the morphological right ventricle and the mitral valve connects to the morphological left ventricle. The most distinguishing feature of the tricuspid valve is the direct attachments to the septum of cords from the septal leaflet. Unlike the tricuspid valve, the mitral valve has no direct septal attachments. The septal insertion of the tricuspid valve is more apical (apically ‘offset’) than that of the mitral valve and these features aid determination of the ventricular morphology. The muscular structure of the ventricles also differs, with the RV being more trabeculated than the LV, with a muscular infundibulum and mid-ventricular ‘moderator band’. Although they are different in normal subjects, the size, shape and degree of trabeculation of the ventricles are not good indicators of ventricular origin, as all are dependent on load effects.

Step 3—Ventriculo-Arterial Connection Description of ventriculo-arterial connections represents the final element of sequential segmental analysis. This entails describing how each great vessel (aorta, pulmonary artery [PA], or common trunk) is

292

SECTION A  The Chest and Cardiovascular System

connected to its respective ventricle. A ventriculo-arterial connection may be concordant (RV-PA, LV-aorta), discordant (RV-aorta, LV-PA), double outlet (e.g. RV-PA and aorta) or single outlet (e.g. LV and RV to common arterial trunk). The aorta and pulmonary arteries are defined by their typical branching patterns. Three-dimensional balanced steady-state free-precession (b-SSFP) and contrast-enhanced magnetic resonance angiography (MRA) techniques are particularly useful in determining the arrangement of the great vessels and the connections with their respective ventricles.

Flow: Using phase-contrast MRI, blood flow and its direction (mL/ beat) across valves and vascular structures can be quantified. In valvar regurgitation, backward flow can be measured and expressed as a regurgitant fraction (backward flow/forward flow). Combined with cine imaging, flow data can be used to calculate and localise intra-/extracardiac shunts.

Step 4—Identification of Other Abnormalities

Imaging is fundamental to the diagnosis of CHD and is required at all stages of patient care. An ideal non-invasive technique for imaging of CHD should be able to accurately and reproducibly delineate all aspects of the anatomy, including intracardiac abnormalities and abnormalities of extracardiac vessels; evaluate physiological consequences of CHD such as measurement of blood flow and pressure gradients across stenotic valves or blood vessels; be cost-effective and portable; provide data from fetal life to adulthood; not cause excessive discomfort and morbidity; and not expose patients to harmful effects of ionising radiation. No single technique has fulfilled these entire requirements and in the delivery of a CHD service, the imaging techniques discussed below play an important complementary role.

Other abnormalities to be considered include abnormal systemic and pulmonary venous connections, intracardiac shunts, valvar abnormalities and vascular abnormalities (PDA, right/left aortic arch, coarctation/ interruption or pulmonary arterial abnormalities). In general, most congenital cardiac lesions are single abnormalities that are easily described; however, almost any combination of abnormalities and connections can occur, and using the sequential segmental analysis method, the description of all conceivable combinations and diagnoses is possible. For more advanced reading, the reader is referred to the textbook Paediatric Cardiology by Anderson and colleagues (see Further Reading).

NON-INVASIVE IMAGING TECHNIQUES

Echocardiography SUMMARY BOX: Summary of Sequential Segmental Analysis 1. Describe the atrial arrangement. 2. Describe the type of atrioventricular connection. 3. Describe the ventriculo-arterial connection. 4. Describe the position of the heart (particularly if abnormal or unexpected). 5. Describe associated abnormalities (e.g., venous abnormalities, septal defects, valvar lesions, great vessel and coronary abnormalities). 6. Describe acquired or iatrogenic abnormalities

PHYSIOLOGICAL AND FUNCTIONAL ASSESSMENT Whilst the correct morphological analysis is a critical first step, it must be incorporated into a complete physiological assessment to understand the clinical problem. It is helpful to briefly consider a few parameters relating to normal cardiac function, which are commonly calculated by techniques such as CMR. Stroke Volume: This is the volume of blood (mL) pumped (displaced) by a ventricle with each heartbeat. The displaced volume is calculated by subtracting the volume of the ventricular cavity at end diastole from the volume at end systole. In the normal heart the stroke volume for each ventricle is the same and is also the same as the forward flow in the associated great artery. It may not be the same in the presence of a shunt or a regurgitant valve. Here, discrepancies in interventricular volumes or great artery flows help locate and quantify the severity of shunts and valve regurgitation. Cardiac Output: This is how much blood each ventricular chamber pumps in 1 minute (L/min). It is calculated by multiplying the stroke volume (or great artery flow (mL)) by the heart rate (stroke (beat)/min). Cardiac output is increased by physiological stress (e.g., exercise that increases both heart rate and stroke volume) and depressed in conditions that reduce either heart rate (bradyarrhythmias) or stroke volume (dilated cardiomyopathy, heart failure). Ejection Fraction: A useful assessment of gross systolic cardiac function is the percentage of blood ejected from the heart during each beat. This is calculated by dividing the stroke volume by the end-diastolic volume. Ejection fraction may be decreased if the systolic performance of the ventricle is impaired (cardiomyopathy).

Echocardiography is the initial imaging technique used in the evaluation of patients with suspected CHD and should always be performed before other techniques are used. In most patients, echocardiography alone provides sufficient information to complete the diagnostic evaluation using a sequential segmental and functional analysis. In UK clinical practice, paediatric cardiologists have traditionally performed echocardiography; however, more recently, neonatologists and radiologists have begun to use echocardiography in patients with suspected CHD where paediatric cardiology services are not immediately available. Cardiac anaesthetists also increasingly perform perioperative assessment using transoesophageal echocardiography. For a more comprehensive discussion of echocardiography in CHD, the reader is referred to Lai et al (see Further Reading).

Magnetic Resonance Imaging As previously alluded, CMR probably provides the most comprehensive assessment available from a single non-invasive imaging technique but its immobility, cost and limited availability constrain its general applicability. In our clinical practice it is used to define the morphology and physiology of the most complex CHD cases as well as providing routine surveillance for patients with repaired CHD such as TOF and TGA. Extracardiac anatomy, including the great arteries and systemic veins, can be delineated with high spatial resolution. Vascular and valvular flow can be assessed, shunts can be quantified and myocardial function can be measured accurately with high reproducibility, regardless of ventricular morphology. Finally, CMR provides high-resolution, isotropic, three-dimensional data sets. This allows for reconstruction of data in any imaging plane, facilitating visualisation of complex cardiac anomalies without the use of ionising radiation. The majority of CMR images are acquired using cardiac (vectorcardiograph) gating during a single breath-hold to reduce the artefacts associated with cardiac and respiratory motion. For a complex case, CMR is performed over approximately 1 hour, although this time can be considerably reduced if a focused question is being addressed or by the incorporation of newer real-time sequences. Imaging sequences can be broadly divided into: • ‘Black blood’ spin-echo images, where signal from blood is nulled and thus not seen—for accurate anatomical imaging.

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging • ‘White blood’ gradient-echo or SSFP images, where a positive signal from blood is returned—for anatomical, cine imaging and quantification of ventricular volumes, mass and function. • Phase-contrast imaging, where velocity information is encoded for quantification of vascular flow, including newer 4D (or 7D) phasecontrast imaging. • Contrast-enhanced MRA, where non-echocardiogram (ECG)-gated 3D data are acquired after gadolinium contrast medium has been administered for thoracic vasculature imaging. • Tissue characterisation imaging, where innate contrast between normal myocardium and disease can be imaged: T1 mapping and extracellular volume imaging, T2 oedema imaging, T2* iron deposition imaging, late gadolinium enhancement fibrosis and necrosis imaging. All these sequences can be acquired in a single breath-hold, reducing the overall time in the CMR machine and enabling the acquisition of accurate data in the majority of patients. Importantly, ‘white blood’ cine images can be acquired in a continuous short-axis stack along the heart, enabling accurate quantification of RV and LV function. Imaging should be performed in the presence of a cardiovascular MRI clinician in conjunction with an MRI technician to ensure that the appropriate clinical questions are answered. A comprehensive treatment of cardiovascular MRI is provided in the textbook by Bogaert and colleagues (see Further Reading).

Computed Tomography Cardiac CT is now well established for the assessment of the thoracic vasculature and large and small airways. Recent advances in multidetector CT (MDCT) with high-pitch spiral and volumetric scan modes have resulted in significant advances in spatial and temporal resolution and a decrease in radiation dose, often less than 1 mSv. Most studies are fully diagnostic without ECG gating. ECG-triggered and ECG-gated acquisitions should therefore only be utilised when cardiac motion may produce non-diagnostic images. Newer-generation scanners permit rapid imaging with minimal motion artefact and thus remove the need for gated scans and general anaesthesia. General anaesthesia may still be needed to control breathing when detailed coronary artery imaging is required. The route and contrast administration protocol are critical to successful imaging. The right upper limb in most cases provides a good site of administration of contrast (avoiding streak artefact from contrast in the innominate vein, which can obscure head and neck vessels). Generally, a biphasic injection protocol using a power injector will be appropriate, deploying a neat contrast bolus (1–3 mL/kg) followed by saline chaser. In neonates and low-bodyweight children requiring small absolute contrast doses, short contrast transit time increases the risk of suboptimal opacification of essential structures. Using the full available contrast dose, reducing the injection rate and mixing the contrast bolus with saline increases the transit time. Empirically, dilution with saline to 70%–80% contrast concentration gives good opacification. Visual bolus triggering from a low-resolution monitoring scan is our favoured approach for timing of acquisition because it more reliably ensures opacification of the appropriate cardiac structures. This does incur a small, added radiation dose, which can be minimised by delaying monitoring toward the end of the injection and reducing the frequency of monitoring scans. A pre-scan timing bolus is often avoided because it utilises part of the contrast bolus available and increases the radiation dose. We currently use MDCT for the following indications in patients with CHD: • Thoracic aorta disease, including vascular rings where airway information is critical. Aim for a narrow bolus via right arm injection to avoid streak artefact from innominate vein over head and neck vessels.

293

• Pulmonary arterial disease. Aim for a prolonged contrast transit time to ensure all sources of pulmonary blood supply are evaluated. The infradiaphragmatic area should also be imaged because collateral vessels may arise here. If thromboembolic disease must be excluded, a rapid undiluted bolus must reach the pulmonary arteries—this can be extremely challenging in cavo-pulmonary connections. • Coronary artery abnormalities. ECG-triggered or ECG-gated protocols may be necessary. Pharmacological heart rate reduction may be required in cases of very high heart rate. • Pulmonary venous abnormalities. Aim for a prolonged contrast transit time to ensure there has been ample time for delayed filling of venous collaterals or slow flow segments. Care should be taken to avoid very dense contrast that may cause streak artefact and obscure anomalous entry points into the systemic venous system. The scan should include the hepatic inferior vena cava in suspected infradiaphragmatic total anomalous pulmonary anomalous connections or partial anomalous pulmonary venous drainage to the hepatic inferior vena cava. • Patients with implants or devices that cannot be imaged by CMR. Imaging should take account of the material to be imaged. Higher kVp, edge-enhancing reconstruction kernels and iterative reconstruction improve the diagnostic accuracy. The reader is referred to the expert consensus document on CT imaging in patients with CHD of the Society for Cardiovascular Computed Tomography (SCCT) (see Han et al., Further Reading).

Conventional Radiology Although CHD may be suspected on the basis of the chest x-ray (CXR), the technique precludes the detailed morphological assessment necessary for diagnosis and determination of specific underlying pathology. The diagnostic accuracy of the CXR in the assessment of infants with asymptomatic murmurs is poor. Despite its poor performance as a screening tool, CHD may be suspected on the basis of a CXR because of higher specificity, but false-positive rates are significant. The CXR, however, is not dispensable and remains important in the subsequent management of patients with CHD, particularly in three situations: 1. Postoperatively, for identification of the position of intravascular catheters, chest drains and endotracheal tubes (Fig. 13.2A). 2. Identification of postoperative complications: consolidation, collapse, pleural effusion, pneumothorax, pneumomediastinum or pericardial collections (see Fig. 13.2B). 3. Perioperative, physiological assessment of the lungs and cardiomediastinal contour (see below).

Diagnostic Features The ubiquity of the CXR in clinical practice warrants discussion of the diagnostic features that should prompt suspicion of CHD. It is suggested that when reporting images, the reader avoid such terms as ‘boot-shaped’ or ‘snowman’ typically associated with specific lesions because they can be misleading and often erroneous. More appropriate is a descriptive consideration of the cardiomediastinal contour and lungs, attempting to evaluate the predominant physiological profile discussed above. The reader may find it helpful to read this section in conjunction with the section on clinical presentation.

The Pulmonary Vasculature Radiologically normal pulmonary vascularity is present in CHD if the patient is not in heart failure, if no large shunt is present and if there is no significant reduction in pulmonary blood flow: for example, mild PS. The pulmonary vasculature may, however, look normal on the conventional radiograph even in the presence of significant CHD.

294

SECTION A  The Chest and Cardiovascular System

A

B Fig. 13.2  Perioperative Chest X-Ray. (A) A 3-year-old patient following total cavo-pulmonary connection surgery, postoperative chest x-ray demonstrating tube positions in intensive care. Note two chest and one mediastinal drains, endotracheal tube and veno-venous collateral occluder device (right upper zone). (B) Third postoperative day following extubation and removal of mediastinal drain. Note change in cardiomediastinal contour caused by large pericardial clot, requiring evacuation.

B

A

Fig. 13.3  Physiological Assessment Using Chest X-Ray. (A) Pulmonary plethora in a patient with a ventricular septal defect. Note the increased number and size of discrete vessels without haziness. (B) Pulmonary oedema in a supine patient with cor triatriatum (membranous obstruction to left atrium outflow) resulting in increased pulmonary venous pressure. Note cardiomegaly, perihilar alveolar haziness/consolidation and peribronchial cuffing.

Increased pulmonary perfusion (pulmonary plethora) is recognised by enlarged central and peripheral pulmonary arteries and veins in all zones (Fig. 13.3A), as occurs in situations with increased pulmonary blood flow: ASD, VSD and PDA with large left-to-right shunts (Table 13.1). Decreased pulmonary perfusion (oligaemia) (Fig. 13.4) is caused by a reduction in pulmonary blood flow and is typically a phenomenon of cyanotic CHD. Dark lungs and sparse pulmonary vascular markings

suggest the diagnosis. Image acquisition must be optimal because overexposure will significantly confound correct interpretation. Pulmonary blood flow may be impaired by obstruction to normal flow through the right heart: for example, tricuspid atresia, TOF and PS (Table 13.2). Pulmonary venous congestion and oedema (see Fig. 13.3B) in CHD is caused by functional or anatomical obstruction to pulmonary venous

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

295

B

A

Fig. 13.4  Pulmonary Oligaemia. (A) Supine anterior-posterior (AP) chest x-ray (CXR) in an 8-week-old patient with tetralogy of Fallot with severe pulmonary stenosis and cyanosis. Note black lungs with sparse, small-calibre vessels. (B) Supine AP CXR in the same patient following construction of a right modified Blalock–Taussig shunt on the next day. Note the increased size of the left cardiac contour caused by increased left ventricle filling, increased pulmonary vascular markings, now plethoric, suggestive of high pulmonary blood flow arising from the shunt. Indeed, the patient had compromised systemic perfusion caused by redistribution of cardiac output to the lungs, necessitating clipping the shunt to reduce its calibre.

TABLE 13.1  Increased Pulmonary

TABLE 13.2  Neonatal Pulmonary

Level of Shunt

Cardiac Lesion

Level

Cardiac Lesion

Atria

Ostium secundum ASDa Ostium primum ASD (partial AVSD)* Sinus venosus defect Anomalous pulmonary venous drainage  (partiala; totalb) Complete AVSD Partial AVSDa VSDa Double outlet ventricleb Single ventricleb PDAa Aortopulmonary window Common arterial trunkb Coronary artery-RV fistula Transposition of great arteriesb Systemic to pulmonary artery shunts   (unrestrictive BT-shunt) Vein of Galen malformation

Tricuspid valve

Tricuspid atresia Tricuspid stenosis Ebstein anomaly Pulmonary infundibular stenosis (severe) Pulmonary valvar stenosis (severe) Tetralogy of Fallot Pulmonary artery atresia Right or left pulmonary artery interruption (differential lung oligaemia) Peripheral pulmonary artery stenosis (regional oligaemia) Transposition of great arteries with pulmonary valve stenosis

Perfusion (Plethora)

Atrioventricular valves Ventricle

Great vessels

Other a

Common cause of plethora without cyanosis. Common cause of plethora with cyanosis. ASD, Atrial septal defect; AVSD, atrioventricular septal defect; BT, Blalock–Taussig; PDA, patent ductus arteriosus; RV, right ventricle; VSD, ventricular septal defect. b

return. In addition to oedema formation caused by increased transvascular pressure gradients, consideration should be given to other pathological processes such as increased vessel leakiness caused by acute lung injuries, for example (Table 13.3). The usual adult pattern of basal oedema, resulting in alveolar hypoxia and constriction of lower

Oligaemia

Right ventricular outflow

Pulmonary artery

pulmonary vasculature and redirection to the apices does not apply to the supine infant. As pulmonary venous pressure increases, there is progressive accumulation of radiological signs, beginning with redistribution (in older children/adults), progressing to interstitial oedema (perivascular haziness, peribronchial cuffing, Kerley B lines, subpleural effusions) and, finally, migration of extravasated fluid centrally, resulting in perihilar alveolar consolidation. Systemic to pulmonary collateral vessels. Abnormal systemic arterial connections to the pulmonary vasculature may occur as an adaptive mechanism to inadequate pulmonary blood flow. This usually occurs in the setting of pulmonary atresia associated with VSD, in which the RV and pulmonary arteries are not in continuity; instead, discrete MAPCAs (major aortopulmonary collateral arteries) and non-discrete networks of bronchial arteries are the source of pulmonary blood flow.

296

SECTION A  The Chest and Cardiovascular System

TABLE 13.3  Pulmonary Oedema and

Venous Congestion Level

Cardiac Lesion

Pulmonary veins

Obstructed TAPVD Pulmonary vein stenosis Cor triatriatum Mitral valve stenosis/atresia Left atrioventricular valve regurgitation Hypoplastic left ventricle LV endocardial fibroelastosis Cardiomyopathy LV ischaemia-aberrant left coronary artery from pulmonary artery (ALCAPA) Aortic stenosis/atresia Coarctation/interruption of the aorta Asphyxia Acute lung injury Intravenous overhydration

Left atrium

Left ventricle

Aorta Non-cardiac pulmonary oedema

LV, Left ventricle; TAPVD, total anomalous pulmonary venous connection.

It may also occur during staged management of the single ventricle. They may be recognisable by a nodular lung pattern in the central third of the lung parenchyma, with many small, rounded, opacities representing enlarged bronchial arteries seen end-on. Pulmonary arterial hypertension may complicate unrepaired CHD. Increased pulmonary blood flow caused by left-right shunting in unrepaired ASD, VSD or PDA gradually causes changes in the pulmonary vasculature, which, over time, leads to increased PVR and overt hypertension. The central pulmonary arteries enlarge and the peripheral pulmonary arteries become smaller than normal. In cases where pulmonary pressure exceeds systemic pressure, shunt reversal occurs, resulting in cyanosis—as occurs in Eisenmenger syndrome.

Heart Size, Shape and Position Abnormalities of the position of the cardiac apex, aortic arch, liver and stomach may be determined from examination of the CXR. The presence of situs inversus and left aortic arch may be discerned; however, this may or may not be associated with underlying CHD. Some assessment of global and regional heart size is possible (see Fig. 13.3B) and should be described; however, the limitations of CXR in this regard should be considered. In a study comparing echocardiographical assessment of cardiac enlargement in 95 consecutive paediatric outpatients, the sensitivity of the CXR to identify cardiomegaly was only 58.8% (95% confidence index (CI): 32.9 to 81.6), specificity was 92.3% (95% CI: 84.0 to 97.1).

SPECIFIC LESIONS In the following discussion lesions have been classed as acyantoic and cyanotic for convenience. It is important to understand, however, that in various situations a lesion typically described in this way may present in the opposite manner, perhaps caused by the presence or absence of a particular morphological feature or the imposition of altered haemodynamics such as elevated PVR. For example, TOF with minimal outflow tract obstruction may have no cyanosis or a VSD that is so large as to facilitate complete intracardiac mixing may produce cyanosis. Furthermore, certain lesions do not fit easily into either category: for example, Ebstein anomaly of the tricuspid valve when mild is acyanotic but in its severe form is cyanotic. Similarly, congenitally corrected TGA, although acyanotic, is better understood when discussed alongside its cyanotic

relative, simple TGA. For further illustrations and images, the reader is referred to the imaging atlas on CHD by Sridharan et al (see Further Reading).

ACYANOTIC LESIONS Septal Defects Atrial Septal Defects ASDs are the most common congenital heart defect detected in adults. Irrespective of their type and location, isolated ASDs cause left-to-right shunting at the atrial level. This leads to atrial dilation, predisposing to tachyarrhythmias, and RV volume overload. The degree and direction of atrial shunting can be modified by AV valve function and ventricular compliance. The presence of an ASD is an independent risk factor for thromboembolic stroke. This is caused by the ability of thromboemboli, originating either in the right atrium or venous vasculature, to pass through the ASD into the systemic circulation. ASDs are, anatomically and developmentally, a heterogeneous group of lesions (Fig. 13.5A). The specific nature of the ASD influences the natural history and management of this disease. Ostium secundum defects make up 80% of ASDs and are located in the fossa ovalis (see Fig. 13.5B). These defects are caused by failure of the septum secundum to form closure of the ostium secundum. Other forms of ASD are more properly termed interatrial communications because they do not occur in the true morphological atrial septum. The ostium primum defect is actually a component of a common AV junction, also known as an AVSD. This defect usually occurs together with some degree of AV valve abnormality. The sinus venosus defect is found at the junction of the right atrium and either one of the caval veins (see Fig. 13.5C). This type of ASD is less common and is always associated with partial anomalous pulmonary venous drainage. The least common type of ASD occurs in the coronary sinus and is termed an unroofed coronary sinus. In this case, there is deficiency of the coronary sinus wall as it passes behind the left atrium, allowing shunting from left to right through the coronary sinus itself. The management of ASDs has changed in recent years, particularly with the increasing use of transcatheter ASD closure devices. Previously, surgical closure was only considered when a large left-to-right shunt led to RV volume overload, atrial dilation and symptoms; however, with the advent of transcatheter techniques, management has become more aggressive. Transcatheter techniques are only viable in patients with small-to-medium-sized ostium secundum defects that have adequate margins with which to anchor the device. Deficiency of the anterior or posteroinferior rim of the defect usually precludes transcatheter closure. Patients with large ostium secundum defects, or defects with deficiency of the anterior or posteroinferior rim. or with sinus venosus lesions, usually require operative repair. The clinical aim is to complete ASD closure before the development of cardiac failure or atrial dilation and timing of intervention depends on the haemodynamic status of the patient; thus, evaluation of ASDs requires definition of type and location of the defect, quantification of the net shunt (pulmonary flow: systemic flow (Qp:Qs)), detection of any intra-atrial thrombus, assessment of RV volume and systolic function and visualisation of the pulmonary venous anatomy. Visualisation of most interatrial communications is possible by transthoracic echocardiography, although sinus venosus or coronary sinus defects are challenging without a high level of suspicion. In addition, detection of pulmonary venous abnormalities is technically difficult using the transthoracic approach. Transoesophageal echocardiography is the main imaging technique used to assess ASDs (particularly at the time of catheter and surgical closure); however, transoesophageal echocardiography cannot be used to accurately quantify the shunt (Qp:Qs)

297

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

Superior sinus venosus defect

Primum ASD

Secundum ASD Inferior sinus venosus defect

Coronary sinus defect

A

B

600 Aorta Main PA

Flow (mL/s)

400

200 100

-200

0

300

600 Time (m/s)

C

D Fig. 13.5  Atrial Septal Defects. (A) Schematic drawing of atrial septal defect (ASD) positions. (B) Balanced steady-state free-precession (b-SSFP) cardiac magnetic resonance (CMR) image. Four-Chamber view showing a large secundum ASD with posterior extension. The absence of a posterior rim (arrow) precludes insertion of an ASD closure device. Note the dilated right atrium (RA) and right ventricle (RV) and flattened interventricular septum. (C) b-SSFP CMR image. Axial view showing a large superior sinus venosus defect with partial anomalous pulmonary venous connection/drainage of the right upper and right middle pulmonary veins straddling the deficient atrial septum (arrow). LA, Left atrium; RA, right atrium. (D) Plot of instantaneous flow (measured by velocity-encoded phase-contrast magnetic resonance imaging) as a function of time showing a left-to-right shunt through an ASD. Note increased pulmonary blood flow. PA, pulmonary artery.

900

298

SECTION A  The Chest and Cardiovascular System

and it can be difficult to delineate pulmonary venous anatomy. CMR has, therefore, a significant role in the diagnosis and pre-interventional assessment of ASDs. Three-dimensional whole-heart techniques, with isotropic resolution, allow accurate multiplanar reformatting with no loss of resolution. These techniques allow 3D rendering of the atrial anatomy. Multislice 2D gradient-echo techniques can be used to assess the dynamic 3D anatomy of the defect, and phase-contrast through-plane flow techniques can accurately size the cross-sectional dimensions of the defect. Multiple or fenestrated defects may also be diagnosed. Haemodynamic assessment is also an important part of the evaluation of ASDs. Invasive catheterisation has historically been used to quantify left-to-right shunts. Quantification of left-to-right shunts using velocityencoded phase-contrast MRI compares well to invasive catheterisation results (see Fig. 13.5D). It has the benefit of being non-invasive and does not require exposure to ionising radiation. Ventricular overload can also be accurately assessed using multislice b-SSFP short-axis imaging and can give important information influencing the timing of intervention. Key imaging goals • Assess defect location, diameter and margin size—suitability for device anchorage. • Quantify right heart volume and function—assess volume overload. • Quantify shunt (see Fig. 13.5D). • Look for sinus venosus defect, which has an associated partially anomalous pulmonary venous drainage. • Look for signs suggestive of elevated PVR—RV hypertrophy, systolic flattening of the interventricular septum and notching of the pulmonary artery flow curve.

Atrioventricular Septal Defects An atrioventricular septal defect (AVSD) is a lesion caused by a deficiency of the tissues that normally interpose the atrial and ventricular chambers (Fig. 13.6A and C). The involved tissues include the atrial primum septum, the AV valves and the inlet portion of the ventricular septum. The feature shared by all AVSDs is a common AV junction guarded by a common AV valve, which may have either one or two orifices (see Fig. 13.6A).

RAS

The common AV junction can be discerned by the loss of the usual ‘offset’ of the tricuspid and mitral valves in the normal heart. The valve, even when it has two orifices, is no longer referred to as a mitral and tricuspid valve; instead they are called left and right AV valves. The common valve typically has five leaflets, referred to as the superior bridging, right anterosuperior, right inferior/mural, inferior bridging and left mural leaflets (see Fig. 13.6AB). The relative deficiency of the septal structures and the number of valve orifices give rise to the classification as complete (both ASD and VSDs and single valve orifice), intermediate/incomplete (VSD with two valve orifices) and partial (ASD with two valve orifices, also called an ostium primum ASD). Another clinically useful description is the relative size of the ventricular chambers, allowing for classification as balanced (equal-sized ventricles) or unbalanced (disproportionate ventricles). AVSD can be associated with other cardiac abnormalities, including TOF, subaortic stenosis, atrial isomerism and ventricular hypoplasia, which modify the presentation, prognosis and surgical management. The diagnosis of AVSD is made in the neonatal period on the basis of a transthoracic ECG. Other imaging techniques are usually not required. Surgical repair is carried out at approximately 3 to 4 months of age and certainly before 6 months of age to prevent the development of pulmonary vascular disease. The repair involves closing the septal defects and creating competent left and right valves from the common AV valve tissue. The association of AVSD with trisomy 21 is well known; repair in this group is associated with lower mortality than non-trisomy 21. Additional imaging techniques including CMR may be useful in the long-term management of patients with repaired AVSD, including surveillance for important late complications such as AV valve regurgitation (see Fig. 13.6C). Key imaging goals • Assess ventricular proportion—unbalanced ventricle may not be suitable for biventricular repair. • Assess valve structure. • Identify associated abnormalities—isomerism of the atrial appendages. • Quantify ventricular volume and function. • Quantify shunt. • Evaluate AV valve regurgitation.

SB

LM

A

RI

IB

B

C

Fig. 13.6  Atrioventricular Septal Defects. (A) Schematic drawing of orthogonal views of a common atrioventricular valve: Short-axis view from below (left), Long-axis (top right), four-chamber (bottom right). (B) Valve view showing a complete atrioventricular septal defect (AVSD) in a patient with right atrial isomerism and double outlet right ventricle. Valve leaflets: IB, inferior bridging leaflet; LM, left mural leaflet; RAS, right anterosuperior leaflet; RI, right inferior (mural) leaflet; SB, superior bridging leaflet. (C) Balanced steady-state free-precession cardiac magnetic resonance image showing four-chamber view of a balanced complete AVSD. There are large atrial and ventricular components. Note the VSD (arrow) and moderate left AV valve regurgitation (arrowhead).

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

Ventricular Septal Defects The ventricular septum is a complex, almost helical 3D structure. VSDs are the commonest form of CHD in childhood (Fig. 13.7). Physiologically, the defect causes left-to-right shunting and the magnitude of the shunt determines the signs and symptoms. The volume of the shunt depends on the size of the defect and the relative resistances of the systemic and pulmonary circulations; at birth the PVR is high, reducing the magnitude of the shunt (it also remains high for longer in patients with left-to-right shunts, explaining why infants may initially be asymptomatic). The degree of shunting is usually estimated by measuring the velocity of blood crossing the defect by ECG; if PVR is normal, higher-velocity jets indicate that the expected pressure difference between chambers is preserved and are termed ‘restrictive’ defects; low-velocity jets suggest the LV and RV have similar pressures and that the defect is ‘unrestrictive’. The exact volume of the shunt cannot be determined accurately by ECG, but can be qualitatively inferred by relative chamber dilatation. As described above for ASDs, shunt volume can be measured using velocity-encoded phase-contrast CMR (see Fig. 13.5D). The commonest location is the perimembranous region, accounting for 80% of VSDs; many small perimembranous VSDs close spontaneously. The rest of the ventricular septum is muscular and has three components: inlet, outlet (subarterial) and midmuscular regions (see Fig. 13.7A). The appropriate management depends on the type and size of the defect. ECG is the mainstay of diagnosis but CMR can provide accurate 2D and 3D images, which are particularly useful in complex defects. Multislice 2D gradient-echo techniques can be used to assess the dynamic 3D anatomy of the defect; however, multislice techniques suffer from poor through-plane resolution. Three-dimensional b-SSFP techniques with isotropic resolution allow accurate multiplanar reformatting, permitting 3D rendering of the ventricular anatomy. Image acquisition during the diastolic period is useful in assessing the anatomy of a VSD and its relationship to valvular structures (see Fig. 13.7B). Key imaging goals • Assess position and size of defect(s). • If outlet VSD, describe commitment to a particular vessel: subaortic, subpulmonary (see Fig. 13.7B).

• Quantify shunt (note LV stroke volume contributes to the PA forward flow during systole). • Look for signs suggestive of elevated PVR—RV hypertrophy, systolic flattening of the interventricular septum and notching of the pulmonary artery flow curve. • Quantify ventricular volume. • Assess aortic valve regurgitation and associated abnormalities. • Post repair, assess integrity of VSD patch (see Fig. 13.7C).

Abnormalities of the Great Vessels Patent Ductus Arteriosus

The arterial duct, in fetal life, connects the pulmonary artery to the aortic arch, allowing blood ejected from the RV to bypass the highresistance pulmonary circulation and enter the descending aorta. The ductal tissue constricts after birth in response to changes in the blood gas composition. If the duct fails to close beyond the first few days of life, it is termed a PDA. The PDA permits a left-to-right shunt from the aorta into the pulmonary artery throughout the cardiac cycle. This leads to increased pulmonary blood flow and dilation of the left heart. The length and diameter of the PDA and the relative resistances of the pulmonary and systemic circulations determine the volume of the shunt, which, in turn, determines the signs and symptoms of the lesion. A common neonatal problem is the failure of the ductus arteriosus to close in the premature infant. In this situation, a PDA can confound the management of the lung disease associated with prematurity and can prolong ventilation. It can be comprehensively assessed using an ECG in infancy but in older patients the duct may not be easily demonstrated. CMR may be required to visualise and quantify the shunt in the same way as described for ASD and VSD. Key imaging goals • Assess position and size of defect(s). • Quantify shunt. • Assess for signs of elevated PVR—RV hypertrophy, systolic flattening of the interventricular septum and notching of the pulmonary artery flow curve. • Quantify volume overload. • Identify any associated intracardiac abnormalities.

Perimembranous VSD Doubly committed subarterial VSD Muscular outlet VSD

A

Muscular inlet VSD

Apical muscular VSD

B

299

C

Fig. 13.7  Ventricular Septal Defects. (A) Schematic drawing of ventricular septal defect (VSD) positions viewed from the right ventricular aspect. (B) Balanced steady-state free-precession image of a VSD (arrow) with over-riding aorta in a patient with tetralogy of Fallot. (C) Coronal oblique view following correction with VSD patch (arrowhead).

300

SECTION A  The Chest and Cardiovascular System

Coarctation of the Aorta

Interrupted Aortic Arch

Coarctation of the aorta is a luminal narrowing of a short section of the aorta (Fig. 13.8A–D). It occurs most commonly at the site of insertion of the ductus arteriosus and is thought to develop because of the presence of excessively integrated ductal tissue around the aortic isthmus, which contracts along with the ductus arteriosus at the time of birth. In severe coarctation, systemic perfusion will be compromised with ductal closure in infancy caused by increased luminal narrowing and the loss of the anatomical bypass provided by the duct itself. In less severe coarctation, the body maintains perfusion by renal mechanisms, resulting in systemic hypertension manifested proximal to the coarctation. Collateral arterial vessels develop over time to maintain lower body perfusion as the patient grows (see Fig. 13.8A and C). Patients may present with unexplained hypertension as a teenager or adult and are at increased risk of the attendant micro- and macrovascular complications. Treatment in infancy is usually by surgical excision of the narrowing but in older subjects, balloon angioplasty may be undertaken. Patients remain at increased risk of hypertension even if repaired in infancy. An ECG is used in the initial diagnosis of infants, children and adults. Typical ECG features include increased systolic and diastolic velocities across the stenosis (see Fig. 13.8D). CMR or CT may be required in the postoperative phase to establish if there is re-coarctation (up to 35% of patients in some series), aneurysmal dilatation or LV hypertrophy secondary to hypertension. CMR is preferred to CT if there are no contraindications. Imaging is crucial to establish the location and degree of stenosis, length of coarctation segment (see Fig. 13.8B), associated aortic arch involvement (such as tubular hypoplasia), the collateral pathways (internal mammary and posterior mediastinal arteries), presence and relationship to an aberrant subclavian artery, post-stenotic dilatation and LV hypertrophy. Three-dimensional contrast-enhanced MRA can show the severity and extent of involvement (see Fig. 13.8C). An assessment of collateral flow by measuring flows in the proximal and descending aorta can be performed. Reassessment of collateral flow following treatment can also be used to assess the success of the treatment. In patients with metal stents, high flip angle gradient-echo sequences can be used to overcome metal artefact and assess luminal narrowing. CMR can also be used to assess secondary pathology in patients with coarctation: for example, aortic root for dilatation associated with a bicuspid aortic valve (frequency in coarctation of 15%), aortic valve incompetence and stenosis and ventricular function and LV mass (an indirect indicator of increased LV afterload). Late after repair, imaging is used to assess for re-coarctation (especially in hypertensive patients), pseudoaneurysm and dissection. Thoracic aorta morphology is highly variable and represents the combination of repair-type residual hypoplasia, stenosis and dilatation (Fig. 13.9). Key imaging goals • Describe location and degree of stenosis (MR flow mapping) and length of coarctation segment. • Look for aortic root involvement and post-stenotic dilatation. • Assess aortic valve—often bicuspid. • Delineate collateral vessels. • Describe head and neck anatomy. • Assess ventricular function, volume and LV mass. • After repair also look for re-coarctation or pseudoaneurysm. • Assess calibre of stented vessels (high flip angle gradient-echo or CT). • Echocardiography—quantify Doppler-derived gradient across stenosis and identification of ‘diastolic tail’ on continuous-pulsed-wave Doppler flow profile (see Fig. 13.8D).

Interrupted aortic arch results from a structural discontinuity between the ascending and descending aorta. The site of interruption relative to the brachiocephalic arteries forms the basis of classification. There is a high incidence of DiGeorge syndrome, which is also associated with variable thymic hypoplasia, the presence of which can be examined. Physiologically, systemic blood flow is provided distal to the interruption by deoxygenated blood via a patent arterial duct. The lower body will be cyanosed and circulation may be compromised following duct closure. The site and length of interruption and any associated anomalies can be demonstrated well by CMR. Following repair, assessment is similar to that of coarctation. It is important to interrogate the repair site for residual narrowing, assess for the presence of LV outflow tract obstruction due to posterior deviation of the outlet septum and look for residual intracardiac shunts. Three-dimensional MRA and 3D whole-heart imaging are particularly useful. Cine imaging will identify regions of flow acceleration. Key imaging goals • Describe the site of interruption and relationship of head and neck vessels: • Type A—interruption distal to the left subclavian artery (29% of cases). • Type B—interruption between the left common carotid and the left subclavian (70% of cases). • Type C—interruption between the innominate artery and the left common carotid (1% of cases). • Evaluate additional arch hypoplasia. • Measure aortic cross-sectional area proximal and distal to the interruption. • Measure the distance of the interruption gap. • Identify associated anomalies—VSD, LVOT obstruction, TGA, common arterial trunk. • Look for thymus—absence supports a diagnosis of DiGeorge syndrome

Abnormalities of the Aortic Arch and Vascular Rings The aortic arch connects the ascending aorta to the descending aorta. Abnormalities of this vascular section include disorders of sidedness. Arch sidedness refers to the side of the trachea that the aortic arch passes as it crosses a mainstem bronchus: namely, left, right and double. In certain circumstances, the morphological pattern of the aortic arch and related structures (its branches or the ductus/ligamentum arteriosus) may produce a vascular ring, which can compress the trachea or oesophagus, producing symptoms of stridor or dysphagia. This usually involves the retro-oesophageal course of either the descending aorta or an aberrant subclavian artery combined with a ligamentum arteriosus on the opposite side of the arch, although non-ring structures such as anomalous origin of the left pulmonary artery from the right pulmonary artery or ‘pulmonary artery sling’ may also cause vascular compression. Cross-sectional imaging (CMR or CT) can be regarded as the gold standard for the assessment of the aortic arch. For CMR, imaging begins with a simple transverse stack from the level of the larynx to the diaphragm. This information is augmented by 3D MRA. Using this information, patent vascular structures compressing the trachea/ oesophagus can easily be identified. It is important to remember that some important components of a vascular ring may not be patent and thus remain invisible on imaging—for example, the ligamentum arteriosus; however, clues to these structures often remain and include dimples opposite the side of the aortic arch, a diverticulum opposite the side of the arch or if the proximal descending aorta descends on the opposite side of the arch.

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

A

B

C

D

301

Fig. 13.8  Severe Coarctation of the Aorta. (A) Posteroanterior chest x-ray showing characteristic bilateral rib-notching (arrow) secondary to the development of collateral circulation. (B) Black blood, spin-echo oblique sagittal image through the aorta showing a tight discrete coarctation (arrow). (C) Volume-rendered 3D reconstruction of magnetic resonance angiography showing a tight coarctation (arrowhead), and multiple enlarged collateral vessels. (D) Echocardiographic continuous-wave Doppler profile of the coarctation region, demonstrating increased velocity across the stenosis, 4.18 m/s (blue cross), corresponding to a pressure gradient of 70 mm Hg from the simplified Bernoulli equation. There is also markedly increased diastolic velocity, characteristic in coarctation, termed ‘diastolic tail’ (red star).

Key imaging goals • Describe side of aortic arch. • Delineate location and course of branches. • Identify presence of associated PDA, or dimple suggestive of ligamentum arteriosus. • Assess presence and level of compression of oesophageal/tracheal compression. • Identify associated abnormalities.

Valvular Heart Disease Aortic Valve Disease

Congenital aortic valve disease is predominated by stenosis, which may occur at subvalvular, valvular or supravalvular levels. The haemodynamic consequence of AS is pressure loading of the LV and the development of secondary concentric hypertrophy. Aortic regurgitation (AR) is usually a manifestation of treated AS (e.g. balloon angioplasty) or secondary to

302

SECTION A  The Chest and Cardiovascular System

Fig. 13.9  Collage of Volume-Rendered Aortas from Patients with Repaired Coarctation of the Aorta. Note the highly variable morphology of the root, aortic arch and isthmus. Clinically and haemodynamically relevant features include hypoplasia of the arch and re-coarctation. Patients are likely to have non-compliant vessels, which predispose to hypertension.

pathological dilatation of the aortic root, which can occur in connective tissue disease (e.g., Marfan syndrome, Fig. 13.10). The haemodynamic consequence is volume loading of the LV and eccentric hypertrophy (dilatation). Doppler ECG assessment of transvalvular pressure gradient is the commonest technique to assess severity of AS; however, transvalvular pressure gradients are flow dependent and measurement of valve area represents, from a theoretical point of view, the ideal way to quantify AS. Inaccuracies in both gradients and valve area, however, require consideration of a combination of flow rate, pressure gradient and ventricular function, as well as functional status that may require multimodality contributions. AS with a valve area less than 1.0 cm2 is considered severe; however, in patients with either unusually small or large body surface area, indexed areas, with a cut-off value of 0.6 cm2/m2, is helpful. The presence of valvular stenosis can be identified by loss of signal on CMR cine images. Velocity mapping can be used to establish an accurate peak velocity across the stenosis, and planimetry can assess the aortic valve area. An ECG is also used to routinely assess AR, in particular using colour Doppler (to determine extension and width of regurgitant jet) and continuous-wave Doppler (to assess the rate of decline of aortic regurgitant flow and holodiastolic flow reversal in the descending aorta); however, these are semiquantitative measures. CMR permits precise assessment of the regurgitant volume and assessment of the volume and function of the eccentrically hypertrophied LV. It also allows assessment of the effective regurgitant orifice area. Subvalvular AS is the least common form of AS. It may be an isolated lesion, or associated with hypertrophic cardiomyopathy or, occasionally, following repair of AVSD. Valvular AS covers a broad spectrum of anomalies, including critical AS presenting with compromised systemic perfusion in infancy sometimes associated with the hypoplastic left-heart syndrome, through to mild AS caused by a bicuspid aortic valve. Supravalvular AS is a rare lesion that is typically associated with underlying Williams (Williams–Beuren) syndrome, a genetic disorder of the connective tissue protein elastin. Elastin is responsible for the normal distensibility of the aorta during systole and its subsequent recoil during diastole. In Williams–Beuren syndrome, the reduced net deposition of arterial wall elastin leads to increased proliferation of

Fig. 13.10  Dilated Sinuses of Valsalva in a Patient With Marfan Syndrome and Aortic Regurgitation.

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging arterial wall smooth muscle cells and multilayer thickening of the medial of large arteries, resulting in the development of obstructive hyperplastic intimal lesions. A characteristic hourglass narrowing of the aorta develops at the sinotubular junction but in approximately 30% of cases there is a diffuse, tubular narrowing of the ascending aorta, often extending to the arch and the origin of the brachiocephalic vessels. Key imaging goals • Assess level and severity of stenosis, subvalvular, valvular, supravalvular—MR flow mapping. • Assess valve leaflet structure (bicuspid valve). • Quantify aortic regurgitation. • Evaluate aortic root dilatation/hypoplasia. • Quantify LV volume and systolic function. • Measure effective orifice area. • Identify other arterial stenosis—head and neck vessels and renal arteries. • Echocardiography—quantify Doppler-derived gradient across stenosis or width and extent of regurgitant jet in AR.

Pulmonary Valve Disease Obstructive lesions dominate congenital pulmonary valve disease, similar to the aortic valve, whereas significant pulmonary regurgitation (PR) is most often iatrogenic following surgical or catheter-based interventions for obstructive lesions. Trivial PR is commonly discerned on an ECG and can be considered physiological. Important congenital PR can occur in the absent pulmonary valve syndrome. According to the site, PS is classified as valvular, subvalvular (infundibular) or supravalvular. It can occur as isolated finding or in constellation with other lesions such as VSDs or more complex lesions (TGA, TOF), which may significantly alter the clinical presentation. PS and PR have physiological consequences analogous to aortic valve stenosis and regurgitation. In the former, RV pressure rises to overcome the obstruction and maintain stroke volume. Compensatory mechanisms include RV hypertrophy. In PR, volume loading of the RV results in progressive dilatation and dysfunction, which is associated with adverse clinical outcomes. In pulmonary valve stenosis, the pressure gradient across the valve is used to assess severity of the lesion more so than in left-sided valve conditions due in part to the difficulty of obtaining an accurate assessment of pulmonary valve area. The systolic pressure gradient is derived from the transpulmonary velocity flow curve using the simplified Bernoulli equation (pressure gradient = 4 × velocity2). Mild stenosis is defined by a peak velocity under 3 m/s on continuous-wave Doppler, which corresponds to a peak gradient under 36 mm Hg; moderate stenosis is defined by a peak velocity from 3 to 4 m/s, corresponding to a peak gradient between 36 and 64 mm Hg; severe stenosis is characterised by a peak velocity above 4 m/s, corresponding to a peak gradient above 64 mm Hg. CMR assessment of PS can be performed in a similar fashion as outlined for AS, above. CMR is the gold standard for the assessment of PR. As described for AR, quantification of the regurgitant volume and its effect on the RV can be precisely determined. RV volume and function cannot be accurately assessed by 2D ECG but CMR measurements have been shown to be associated with adverse outcomes and can aid decision making for timing of interventions. Key imaging goals • Assess level and severity of stenosis, subvalvular, valvular, supravalvular—MR flow mapping. • Assess valve leaflet structure. • Quantify PR. • Evaluate main pulmonary artery dilatation/hypoplasia. • Quantify RV volume and systolic function.

303

• Measure effective orifice area. • Echocardiography—quantify Doppler-derived gradient across stenosis.

Ebstein Anomaly of the Tricuspid Valve Ebstein anomaly is a congenital abnormality of the tricuspid valve and right ventricle with the following components: (1) adherence of the tricuspid leaflets to the underlying myocardium (failure of delamination); (2) anterior and apical rotational displacement of the functional annulus; (3) dilation of the ‘atrialised’ portion of the right ventricle with variable degrees of hypertrophy and thinning of the wall; (4) redundancy, fenestrations and tethering of the anterior leaflet; (5) dilation of the right AV junction (the true tricuspid annulus); and (6) variable ventricular myocardial dysfunction. The degree of displacement determines the clinical presentation. In severe cases there is gross right atrial enlargement and raised right atrial pressure. The anomaly is usually associated with an ASD and, therefore, right-to-left shunting at the atrial level and subsequent cyanosis may occur. Ebstein anomaly results in gross enlargement of the cardiac contour with a prominent curved right atrial border on the plain chest radiograph. Treatment is problematical, although expert surgical repair of the tricuspid valve is possible in some centres. Imaging should assess the valve morphology, quantify ventricular function and volume and quantify right atrial enlargement. Key imaging goals • Describe apical displacement of the septal leaflet of the tricuspid valve. • Assess mobility of anterosuperior and inferior tricuspid valve leaflets. • Note eccentric coaptation. • Quantify tricuspid regurgitation. • Quantify right atrium dilatation and size of atrialised RV. • Assess RV and LV volume, function and mass. • Quantify right-to-left shunt. • Exclude right ventricular outflow tract (RVOT) obstruction.

Coronary Artery Abnormalities Anomalous Coronary Arteries

Coronary artery abnormalities are rare. They involve anomalous proximal and epicardial courses of the left coronary artery (LCA) and right coronary artery (RCA) (Fig. 13.11) or, rarely, anomalous origin of the LCA from the pulmonary artery (ALCAPA), Fig. 13.12. Anomalous course is increasingly important when interventions are carried out in close proximity to a coronary artery—for example, percutaneous pulmonary valve implantation into the pulmonary trunk and compression of the LCA. Similarly, the course of the coronary arteries are important during surgical repair of TOF; a transannular patch repair may not be possible if the LCA arises from the RCA and passes anterior to the RVOT tract. ALCAPA results in the LCA territory being supplied with low-pressure deoxygenated blood; blood must therefore be supplied by collateralisation from RCA. Patients experience myocardial ischaemia and usually present approximately 4 to 5 months of age when PVR drops and LCA blood flow is reduced. Patients with sufficient collateralisation may survive to adulthood. Treatment usually involves surgical reimplantation of the coronary artery using a button transfer technique or coronary artery bypass grafting. From an anatomical point of view, coronary anomalies are classified according to the coronary artery involved, the origin of the anomalous coronary artery and the anatomical course of the proximal segment (see Fig. 13.11). From a clinical point of view, the anomalies are divided into ‘benign’ and ‘malignant’ lesions. The latter, especially those of ALCAPA and in cases where the LCA arises from the RCA and passes between

304

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E

F

Fig. 13.11  Coronary Artery Anomalies. Schematic diagram of the coronary arteries viewed in the axial oblique plane on CMR. (A) Anomalous LCX from RCA. (B) Anomalous RCA from left main stem (LMS), with interarterial course between pulmonary artery and aorta. (C) Anomalous RCA from LMS passing posteriorly between the aorta and atria. (D) Anomalous left coronary artery arising from RCA with interarterial course between the pulmonary trunk and aorta. (E) Anomalous left coronary artery arising from RCA passing anterior to pulmonary trunk. (F) Anomalous left coronary artery arising from RCA passing posteriorly between aorta and atria. Ao, Aorta; LA, left atrium; LAD, left anterior descending artery; LCX, left circumflex artery; LV, left ventricle; RA, right atrium; RCA, right coronary artery; RVOT, right ventricular outflow tract.

the aortic root and RVOT tract or pulmonary artery, have an increased risk for developing myocardial ischaemia and sudden cardiac death. Even with multiple projections, the precise location of the proximal course of the vessel in a patient with an abnormal origin of a coronary artery can be difficult to depict with conventional angiography and echocardiography; however, CMR and CT angiography provide reliable visualisation of the root of the arteries and the coronary artery tree. Key imaging goals • Delineate coronary artery anatomy—consider CT may be better, Fig. 13.12A and B. • Look for myocardial perfusion defects—consider adenosine stress. • Identify areas of regional wall motion defects. • Assess ventricular function, volume and LV mass. • Identify mitral regurgitation—secondary to ischaemia. • Perform late gadolinium enhancement—to determine infarct size if present, Fig. 13.12D.

CYANOTIC CONGENITAL HEART DISEASE Tetralogy of Fallot TOF is the most common cyanotic congenital heart defect, with an incidence of approximately 420 per million live births. It is caused by

malalignment of the muscular outlet septum, which leads to RVOT obstruction, a subaortic VSD with aortic override and RV hypertrophy (Fig. 13.13A–C). This produces the physiological pattern of low pulmonary blood flow and right-to-left shunt as described above. Current management consists of early single-stage reconstructive surgery, with closure of the VSD, and relief of the RVOT obstruction, usually by the placement of a transannular patch (across the pulmonary valve annulus) to enlarge the RVOT. Staged reconstruction is still required if there is severe cyanosis caused by a very narrow RVOT or significant hypoplasia of the central pulmonary arteries. In such cases, a systemic-to-pulmonary anastomosis called a modified Blalock–Taussig (BT) shunt is placed (usually) between the innominate artery and the right pulmonary artery (see Fig. 13.13D). This shunt is then taken down during subsequent definitive repair. A transthoracic ECG is the imaging technique of choice for the initial diagnosis and assessment of paediatric patients; however, CMR imaging does have a role in untreated or shunt-palliated patients in delineating pulmonary artery anatomy or excluding significant pulmonary artery distortion. While early surgical mortality from complete repair of TOF is very low in the modern surgical era, residual anatomical and haemodynamic abnormalities are almost universal. These include right ventricle dilatation from PR, RVOT aneurysm, RVOT obstruction,

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

A

B

C

D

305

Fig. 13.12  Anomalous Left Coronary Artery from Pulmonary Artery. (A) Computed tomography image showing left coronary artery (LCA) origin from pulmonary artery (PA). (B) Volume-rendered image (arrow to anomalous LCA from the pulmonary artery (ALCAPA). LPA, left pulmonary arteries; RV, right ventricle. (C and D) Short-axis balanced steady-state free-precession cardiac magnetic resonance image and corresponding late gadolinium enhancement image from the same patient showing anteroseptal and papillary muscle enhancement.

pulmonary artery stenosis, residual atrial or VSD, tricuspid valve regurgitation and aortic root dilatation. CMR has emerged as the gold standard for the assessment of the right ventricle in patients with repaired TOF. Two-dimensional b-SSFP sequences can be used to define RVOT anatomy and quantitatively assess RVOT dilatation or stenosis. Velocity-encoded phase-contrast MR can accurately quantify the degree of PR and can be used to measure peak velocities at the level of RVOT obstruction, as well as differential regurgitation in the branch pulmonary arteries. MR assessment of RV function/volumes with multislice short-axis b-SSFP imaging is particularly important when determining the timing and evaluating the impact of invasive therapeutic strategies. Key imaging goals • Describe the RVOT and pulmonary trunk anatomy. • Identify RVOT aneurysm. • Identify RVOT obstruction. • Quantify PR. • Assess branch PA anatomy and measure flow split. • Assess biventricular volume and function. • Exclude residual shunt (Qp:Qs). • Measure aortic root for dilatation. • Delineate coronary artery anatomy (important for surgical or percutaneous pulmonary valve replacement).

Transposition of the Great Arteries TGA is the second-commonest cyanotic CHD diagnosed in the first year of life, with an incidence of 315 per million live births. In this condition, the aorta arises from the right ventricle, and the pulmonary artery from the right ventricle (ventriculo-arterial discordance) (Fig. 13.14A). This produces the physiological pattern of parallel circulations described above, which is incompatible with life. Treatment for TGA patients was revolutionised with the introduction of the Senning procedure, in which an intra-atrial baffle was used to divert blood from the right atrium to the left ventricle, and the left atrium to the right ventricle. A variation to the Senning procedure, the Mustard procedure uses a pericardial patch or prosthetic material to construct the intra-atrial baffle; however, although both these procedures produce a physiologically normal circulation, the patient is still left with a systemic RV. Patients surviving with these repairs may have unique complications, including pulmonary venous pathway or baffle obstruction, baffle leaks and failure of the systemic RV. Currently, the arterial switch operation has become the procedure of choice. In this operation, the great vessels are transected above the valve sinuses and sutured to their appropriate ventricle; the coronary arteries arising below this transection level must also be transferred separately (see Fig. 13.14B). In cases of TGA associated with a VSD and sub-PS, the Rastelli procedure (where blood from the LV is channelled through the VSD to the aorta) is preferred.

306

SECTION A  The Chest and Cardiovascular System

A

B

C

D Fig. 13.13  Tetralogy of Fallot. (A) and (B) Right ventricular outflow tract, morphological specimen and corresponding black-blood spin-echo image in coronal view. The deviated outlet septum (asterisk), aortic root (arrowhead) and hypertrophied septoparietal trabeculations (arrow) are shown. (C) Balanced steady-state free-precession image of unrepaired tetralogy of Fallot. Inflow/outflow view of the left ventricle shows a ventricular septal defect with overriding aorta. Note the severe hypertrophy of the right ventricle. Ao, Aorta; LV, left ventricle; RV, right ventricle. (D) Black-blood, spin-echo image of right modified Blalock–Taussig shunt; 3.5 mm gortex tube from innominate artery to right pulmonary artery (arrow).

Key imaging goals • Arterial switch repair: • Identify any RVOT or LVOT obstruction. • Assess branch pulmonary arteries, which may be narrowed as they straddle the aorta (see Fig. 13.14C and D) (measure differential branch flow). • Identify aortic root dilatation.

• Assess coronary arteries for kinking, ostial stenosis. • Mustard/Senning: • Assess patency of baffle and pulmonary venous pathway. • Assess atrial baffles for leak (presence of shunt). • Quantify systemic RV volume and function. • Perform late gadolinium enhancement for ventricular fibrosis/ scar.

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

A

B

C

D Fig. 13.14  Transposition of the Great Arteries. (A) Balanced steady-state free-precession (b-SSFP) cardiac magnetic resonance (CMR) image showing an oblique sagittal outlet view of the aorta arising from the right ventricle (RV) and pulmonary artery arising posteriorly from the left ventricle (LV). (B) Schematic drawing of the arterial switch repair of transposition of the great arteries, showing the Le Compte manoeuvre with the translocation of the aorta and pulmonary artery. Note sites of coronary artery ‘button’ removal and subsequent re-implantation into the neo-aortic root. (C) b-SSFP CMR image showing the pulmonary arteries straddling the aorta following the arterial switch procedure with Le Compte manoeuvre. (D) Volume-rendered 3D reconstruction of a contrast-enhanced magnetic resonance angiography showing bilateral proximal branch pulmonary artery narrowing.

307

308

SECTION A  The Chest and Cardiovascular System

Congenitally Correct Transposition of the Great Arteries

Pulmonary Atresia

Congenitally corrected transposition of the great arteries (CCTGA) is a rare disorder characterised by both AV discordance and ventriculoarterial discordance (right atrium to left ventricle to pulmonary artery to lung (Fig. 13.15A) and left atrium to right ventricle to aorta). Therefore, although the heart is anatomically abnormal, it is physiologically normal in terms of the pulmonary and systemic circuits. This lesion does not usually cause cyanosis; however, many of the problems are similar to those experienced by patients with TGA, particularly those treated by an atrial switch operation, and thus a systemic RV. CCTGA may be asymptomatic and in some patients is an incidental finding; however, the majority of patients with CCTGA have associated cardiac lesions (see Fig. 13.15B), the most common being VSD. PS is present in approximately 50% of cases and tricuspid valve abnormalities (e.g. Ebstein abnormality) are found in 20% of cases. Even without associated abnormalities, most patients with CCTGA eventually develop systemic ventricular failure. The main role of imaging is in evaluation of associated lesions, quantification of ventricular function and assessment of postoperative complications. Key imaging goals • Describe atrioventricular and ventriculo-arterial connections. • Identify associated VSD and tricuspid valve abnormalities. • Identify outflow tract obstruction. • Perform late gadolinium enhancement to assess for systemic RV fibrosis or scarring.

Pulmonary atresia is the lack of luminal continuity and absence of blood flow from the RV to the pulmonary artery. Pulmonary atresia can be separated into two groups depending on the presence of a VSD. As the diagnosis and subsequent management of these two groups is different, it is useful to consider them separately. Pulmonary atresia with a ventricular septal defect.  This is the more common variant and is considered by some to be a severe form of TOF, with a subaortic VSD and overriding aorta. Pulmonary blood flow is supplied via MAPCAs. Surgical repair aims to establish RVOT to the pulmonary artery continuity with a homograft (RV-PA conduit), repair the VSD and bring the aortopulmonary collaterals into the pulmonary circulation. As with TOF, the main role of imaging in patients with pulmonary atresia and a VSD is assessment of postoperative complications. The most common long-term complication is homograft failure: usually mixed stenosis and regurgitation leading to RV dysfunction. Conduit stenosis is often secondary to calcification of the non-viable homograft and although calcified tissue is difficult to visualise using CMR, it is clearly seen on CT. Other long-term complications are similar to those found in TOF. Key imaging goals • Characterise the presence or absence of the central PAs. • Characterise the source of pulmonary blood flow (PDA or multifocal MAPCAs). • Quantify distance between RVOT and PA confluence. • Post-op (as for TOF): Assess RV-PA conduit function.

Aorta

Ebstein’s malformation

Pulmonary artery

Stenotic pulmonary valve Fibrous tissue tags A

B

Morphological LV

VSD Fibrous shelf

Fig. 13.15  Congenitally Corrected Transposition of the Great Arteries. (A) Balanced steady-state freeprecession cardiac magnetic resonance image of congenitally corrected transposition of the great arteries (CCTGA) showing the discordant atrioventricular connection with anterior left ventricle (LV). RV, Right ventricle. Note the apical offset of the left-sided tricuspid valve. (B) Schematic drawing of CCTGA and frequent associated lesions. LV, Left ventricle; RV, right ventricle; VSD, ventricular septal defects.

Morphological RV

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

309

Pulmonary atresia with an intact ventricular septum.  This is the less common variant of pulmonary atresia. There is complete atresia of the pulmonary valve in conjunction with a variable degree of hypoplasia of the tricuspid valve and RV cavity. The type of surgical repair depends on the size and shape of the RV cavity. The presence of an RV infundibulum allows a biventricular repair. If the RV cavity is small, then single ventricular physiology is established and angiography demonstrates large venous sinusoids in the RV wall. Imaging assessment is considered in the section below on the single ventricle.

Double Outlet Right Ventricle The term double outlet right ventricle (DORV) refers to any cardiac anomaly in which both the aorta and the pulmonary artery originate, predominantly or entirely, from the right ventricle. In this situation the LV has no direct outlet to either great vessel and must eject blood into the RV through a VSD. Rarely there may be no VSD and then the LV is very hypoplastic. The physiological picture and type of surgical correction depend on the arrangement of the great vessels and the anatomy of the VSD. Cyanosis is not invariably present and depends on the degree of PS. The most common variant is a normal arrangement of the great vessels and a subaortic VSD. This variant is often referred to as the ‘Fallot’s type’ because it is often associated with PS and has a similar presentation. DORV may also be associated with an anterior aorta and a subpulmonary VSD known as the ‘Taussig–Bing’ anomaly. Surgical correction for the ‘Fallot type’ variant consists of patch closure of the VSD, which redirects blood to the aorta, and correction of any PS. For the ‘Taussig–Bing’ anomaly the surgical approach depends on the presence of pulmonary obstruction. In the absence of pulmonary obstruction, correction consists of patch closure of the VSD and arterial switch. In the presence of obstruction, LV flow is tunnelled through the VSD to the aorta and an RV-PA pathway is established, known as the Rastelli procedure. Imaging plays an important role in preoperative assessment. The 3D anatomy of the VSD and the arrangement of the great vessels are well visualised by CMR and are particularly important when deciding the type of surgery. Black-bood spin-echo CMR of the VSD has been shown to compare well with surgical findings and is able to indicate the optimal type of repair. Key imaging goals • Describe arrangement of great vessels—normal with subaortic VSD (Fallot type), anterior aorta with subpulmonary VSD (Taussig– Bing type). • Describe VSD size, position and commitment. • Identify associated outflow tract obstruction. • Quantify shunt (may be left to right or right to left). • Quantify ventricular volume, function and mass. • Postoperatively evaluate for TOF or TGA, described above, depending on type of DORV.

Common Arterial Trunk Common arterial trunk (truncus arteriosus) is defined as a single arterial trunk arising from both ventricles, which overrides a large misaligned VSD, Fig. 13.16. The pulmonary, systemic and coronary arteries all originate from the trunk. This produces the physiological pattern of cyanosis caused by intracardiac mixing as the simultaneous ejection of both ventricles into the common trunk merges streams. The classification of common arterial trunk relies on the branching pattern of the pulmonary artery. In type I, a short main pulmonary artery arises from the common trunk and subsequently divides (see Fig. 13.16). In type II, the right and left pulmonary arteries originate from the posterior wall of the common trunk and, in type III, the right and left pulmonary

Fig. 13.16  Unrepaired Truncus Arteriosus Presenting in a 17-Year-Old Boy. A very short main pulmonary artery (MPA). Arises from the common trunk before dividing into left and right pulmonary arteries. The patient was cyanosed due to intracardiac mixing. There was pulmonary arterial hypertension with an estimated pulmonary vascular resistance of 17 woods units.m2. VSD, Ventricular septal defect.

arteries emerge from the lateral wall of the common trunk. The truncal valve is often abnormal, with varying degrees of stenosis and insufficiency. Approximately 40% of truncal arches are on the right side, with most truncal arches rising higher in the mediastinum than the normal aortic arch. Surgical repair consists of reconstruction of the common trunk to produce a systemic vessel from the left ventricle, patch closure of the VSD and establishment of a right ventricle-to-pulmonary artery conduit. The main role of CMR is in assessment of postoperative complications (homograft failure, truncal valve regurgitation and VSD patch leak). Imaging can be used to better delineate the vascular anatomy before surgery. Key imaging goals • Define morphological subtype: • Type 1, RPA and LPA arise from a main PA segment. • Type 2, RPA and LPA arise from the posterior wall of the common trunk. • Type 3, RPA and LPA arise from lateral wall of the common trunk. • Describe truncal valve morphology (tricuspid, quadricuspid). • Assess truncal valve function, stenosis or regurgitation. • Describe position of VSD. • Look for associated abnormalities, coarctation or interrupted aortic arch. • If unrepaired, assess shunt and examine for pulmonary vascular disease.

Anomalous Pulmonary Venous Connection/Drainage Pulmonary veins may connect abnormally to a site other than the left atrium—usually the right atrium, systemic vein or coronary sinus. If

310

SECTION A  The Chest and Cardiovascular System

all the veins connect abnormally, then it is described as total anomalous pulmonary venous connection/drainage (TAPVD), and if less than all the veins connect abnormally (usually only 1) then it is termed partial anomalous pulmonary venous connection/drainage (PAPVD). In TAPVD, a complete left-to-right shunt causes all of the pulmonary venous return to mix with systemic venous return. Survival is dependent on obligatory right-to-left shunting of the mixed pulmonary venous and systemic venous blood, usually at atrial level. PAPVD is an acyanotic condition, which results in a physiology similar to an ASD. The number of veins involved determines the magnitude of the shunt and the clinical symptoms. It is associated with superior and inferior sinus venosus ASD (see Fig. 13.5C). In TAPVD the pulmonary veins coalesce posterior to the left atrium but do not drain into it. Drainage from this venous confluence to the right atrium may be: (a) via either an ascending vein to the innominate vein and then to the superior vena cava (SVC) (supracardiac, 50% of patients, Fig. 13.17); (b) the coronary sinus directly into the right atrium (cardiac, 15% of patients); (c) via a descending vein, which passes

A

through the diaphragm into either the IVC or portal venous system (infracardiac, 25% of patients, Fig. 13.18); or (d) a mixture of these routes may coexist (mixed, 10% of patients). The infracardiac type is usually associated with a degree of obstruction as the descending vein passes through the diaphragm (see Fig. 13.18); thus, unlike the supracardiac and cardiac variants, which present with a left-to-right shunt, intracardiac mixing and cardiac failure, the clinical picture for infracardiac TAPVD is potentially more severe, with superimposed pulmonary venous hypertension, resulting in tachypnoea, tachycardia, liver enlargement, cyanosis, pulmonary oedema and respiratory distress. Symptoms usually appear within 24 to 36 hours of birth. Importantly, the diagnosis of infracardiac TAPVD can be missed on a ECG and must always be considered in the differential diagnosis of pulmonary oedema on the neonatal chest radiograph (see Fig. 13.18A). Associations of TAPVD are complex cardiac anomalies such as isomerism of the atrial appendages, AVSD, PS, DORV, hypoplastic left-heart syndrome (HLHS), common arterial trunk, TGA and aortic

B

Fig. 13.17  Supracardiac Total Anomalous Pulmonary Venous Drainage. (A) Postero-anterior chest x-ray in a teenager with a late presentation of supracardiac total anomalous pulmonary venous connection. (B) Volume-rendered gadolinium-enhanced magnetic resonance angiogram from the same patient. Note the dilated superior vena cava (blue arrow) and right-sided pulmonary veins, which had evidence of obstruction. The patient subsequently underwent complete repair.

A

B

Fig. 13.18  Infracardiac Total Anomalous Pulmonary Venous Drainage. (A) Anterior-posterior chest x-ray in newborn infant presenting with cyanosis and respiratory distress showing pulmonary oedema. (B) Volumerendered 3D reconstruction of magnetic resonance angiography showing total anomalous infracardiac drainage of the pulmonary veins in the same patient. Note the narrowing of the veins as they pass through the diaphragm (white arrow) before draining into the portal vein (blue arrow).

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging coarctation. The anomalous pulmonary venous connection is easily visualised with ultrasound but in patients with poor ECG acoustic windows, cross-sectional imaging with CMR and CT is very useful, often avoiding the need for a diagnostic x-ray catheterisation. Key imaging goals • Total anomalous pulmonary venous drainage: • Describe pulmonary venous connection type: supracardiac, cardiac, infracardiac or mixed. • Identify areas of pulmonary venous obstruction. • Confirm obligatory right-to-left shunt. • Identify associated abnormalities—isomerism of the atrial appendage. • Partial anomalous pulmonary venous connection: • Describe pulmonary venous connection. • Assess branch pulmonary arteries. • Quantify ventricular volume, function and assess RA dilatation. • Quantify shunt. • Consider if shunt measured but no ASD or VSD seen. • Consider PAPVC-associated conditions such as Turner’s syndrome.

SINGLE VENTRICLES The term single ventricle covers a wide range of different cardiac morphologies: for example, HLHS. (Fig. 13.19A), tricuspid atresia or pulmonary atresia with intact ventricular septum; however, pragmatically, the term can be used to describe a group of patients who, following surgical ‘correction’, have a circulation supported by one ventricle. Modern palliative management has resulted in long-term survival for many patients who would otherwise have died as infants. This division extends the narrower morphological classification to include also functionally single ventricle: for example, where two ventricles are connected by a large VSD, which cannot be surgically septated because of AV valve apparatus straddling the ventricular septum (chordae tendinae attached to the septal crest or opposite ventricle) or when there is significant imbalance of the ventricular chambers, such as in double inlet left ventricle. The ultimate management goal is the creation of a single ventricle circuit with separation of the systemic and pulmonary circulations—a Fontan or total cavopulmonary circulation (TCPC)—such that the single ventricle pumps blood into the systemic circulation and systemic venous return is directed to the pulmonary circulation without a ventricular pump. This is performed in a step-wise surgical fashion involving two or three stages.

Systemic to Pulmonary Artery Shunt In patients with inadequate pulmonary blood flow caused by a hypoplastic RV (tricuspid atresia or pulmonary atresia with intact ventricular septum) and in patients with HLHS as part of the Norwood procedure, the first stage involves the creation of a systemic-to-pulmonary artery shunt such as a modified BT shunt (see Fig. 13.13D). This is a temporising procedure in infants, in whom PVR is not low enough to proceed immediately with a bidirectional superior cavopulmonary connection (BCPC), known as the Glenn procedure. Key imaging goals following stage 1: pre-bidirectional cavopulmonary connection (pre-BCPC) • Assess branch PA anatomy and exclude deformation at site of shunt insertion. • In HLHS, assess aortic arch for residual obstruction or coarctation. • Assess systemic ventricular volume and function. • Identify any AV valve regurgitation. • Quantify aortic/neo-aortic valve regurgitation. • Evaluate adequacy of atrial communication.

311

• Look for systemic-pulmonary or veno-venous collateral vessels. • Look for bilateral SVCs.

Bidirectional Glenn Circulation The second stage in the creation of a single ventricular circulation is the BCPC. It is usually performed at approximately 3 to 9 months of age. In this procedure, an anastomosis is created between the SVC and the right pulmonary artery (such that the SVC blood flows into both arteries), and the SVC-RA junction oversewn (see Fig. 13.19B). Any previous surgical systemic-to-pulmonary artery shunts are also taken down at this time. Imaging should be used to assess branch pulmonary artery narrowing and pulmonary venous obstruction; otherwise. The circulation may fail. Key imaging goals following stage 2: pre-total cavopulmonary connection (pre-TCPC) • Assess branch PA anatomy and SVC-PA connection. • Describe systemic venous return (SVC, IVC, hepatic veins). • Quantify ventricular function and volume. • Identify AV valve regurgitation. • Describe outflow tracts and aortic arch. • Evaluate adequacy of atrial communication. • Look for veno-venous (usually SVC to IVC territory) or systemicto-pulmonary collateral vessels (common after BCPC) and quantify collateral flow. • At time of general anaesthetic, assess pulmonary arterial pressure by transducing the pressure in the internal jugular vein. • Look for pulmonary vein stenosis.

Fontan Circulation The Fontan circulation can be completed by a number of different surgical methodologies, but in the current era it is performed by the creation of an intracardiac (lateral tunnel) (see Fig. 13.19C) or extracardiac conduit (Fig. 13.20A) between the IVC and the pulmonary arteries. This is known as a total cavopulmonary connection (TCPC). It is usually performed between 18 months and 5 years of age. The classical Fontan operation involved connection of the SVC to the RPA (non-bidirectional Glenn procedure) and connection of the right atrium to the LPA (see Fig. 13.20B). Right atrial dilatation is the major complication of the classical Fontan procedure and may cause arrhythmias, thrombosis or pulmonary vein compression, which can lead to failure of the Fontan circulation. CMR allows evaluation of the branch pulmonary arteries and their systemic venous connections and can be used to accurately assess ventricular function. Despite optimal medical and surgical management, the intrinsic shortcomings of the TCPC invariably manifest as late attrition during follow-up. This situation, known as a ‘failing Fontan’, is particularly difficult to manage because systemic venous hypertension and low cardiac output may result in peripheral oedema and unusual clinical syndromes of plastic bronchitis and protein-losing enteropathy (see Fig. 13.19D). Ultimately, this may result in death or necessitate high-risk cardiac transplantation. Key imaging goals following stage 3: post-total cavopulmonary connection (post-TCPC) • Assess branch PA anatomy and SVC-PA, IVC-PA connections. Exclude stenosis. • Quantify ventricular volume and function. • Identify AV valve regurgitation. • Evaluate flow in fenestration between conduit and RA if present. • Evaluate adequacy of atrial communication. • Look for veno-venous or systemic-to-pulmonary collateral vessels (common after BCPC) and quantify collateral flow. • Look for thrombus in RA if classical Fontan. • Identify any pulmonary vein stenosis.

312

SECTION A  The Chest and Cardiovascular System

A

B

C

D Fig. 13.19  Single Ventricle. (A) Balanced steady-state free-precession (b-SSFP) cardiac magnetic resonance (CMR) image showing hypoplastic left heart syndrome, with severe hypertrophy of the systemic right ventricle (RV). LV, Left ventricle. Note the large interatrial communication (arrowed) that allows mixing of systemic and pulmonary venous return. (B) Volume-rendered 3D reconstruction of a magnetic resonance angiogram showing the Glenn, bidirectional cavopulmonary anastomosis (arrow) and (C) a lateral tunnel total cavopulmonary anastomosis (arrow) to the right pulmonary artery (arrowhead). (D) b-SSFP CMR image showing severe ascites (arrow) and right pleural effusion (arrowhead) in a patient with a failing total caval pulmonary connection circulation and protein-losing enteropathy.

Recent Developments in Cardiac Magnetic Resonance Imaging Hybrid Catheter/Cardiac Magnetic Resonance Imaging Laboratory

In a hybrid catheter/CMR lab (XMR), a cardiac catheterisation laboratory is connected to the CMR scanner via a shielded sliding door.

A mobile patient table facilitates seamless transfer between the two techniques. As a diagnostic tool, XMR is particularly helpful in the evaluation of haemodynamic problems in which pressure-only or flow-only measurements result in an incomplete characterisation of the problem: for example. In the assessment of PVR for pulmonary hypertension. XMR can also be used to guide structural intervention and electrophysiological

CHAPTER 13  Congenital Heart Disease: General Principles and Imaging

313

B

A

Fig. 13.20  (A) Total cavopulmonary connection using an extracardiac conduit from the IVC to the pulmonary arteries in a patient with hypoplastic right heart caused by pulmonary atresia and intact ventricular septum. (B) Volume-rendered magnetic resonance angiogram of atriopulmonary Fontan in a patient with tricuspid atresia. Note the anastomosis between the right atrial appendage and the main pulmonary artery, MPA, (blue arrow). CS, coronary sinus; IVC, inferior vena cava; LPA, left pulmonary artery; RA, right atrium; RPA, right pulmonary artery; SVC, superior vena cava.

studies and in these cases would reduce the radiation exposure, which is particularly important in children and infants.

Fetal Cardiac Magnetic Resonance Imaging Fetal CMR is an emerging technique that may prove complementary to a fetal ECG. As with other ultrasound techniques, a fetal ECG can be affected by maternal body habitus or other situations that reduce acoustic windows. Fetal CMR is affected by several technical challenges related to the fetal cardiac dimensions and fast heart rate, fetal motion and the lack of fetal ECG trace for gating.

Post-Mortem Cardiac Magnetic Resonance Imaging Perinatal and paediatric cardiac autopsies have an important role in the counselling of parents with regard to the cause and implications of death of their child. Recently, less invasive post-mortem MR has been proposed as an alternative for conventional autopsy and could be used as the first-line assessment for structural heart disease in this situation. Post-mortem imaging allows an opportunity to investigate the heart in situ before dissection and both post-mortem CT and postmortem MRI have shown excellent accuracy in detecting most clinically significant cardiac lesions in the perinatal and paediatric population. If all non-diagnostic and positive post-mortem CMR scans were referred for conventional autopsy, very few diagnoses would be missed based upon currently available evidence.

Three-Dimensional Printing Three-dimensional-engineered replicas of different anatomical structures have been used extensively in different fields of medicine over the past 20 years. As the manufacturing techniques, referred to as ‘rapid prototyping’, have become more refined over the years, medical researchers

have used such 3D models for pre-surgical planning personalisation of prostheses or testing of novel devices. One of the other benefits of anatomical 3D models is being able to visualise the location and dimensions of the area of interest as an aid in communication, both within a surgical team and, crucially, between the physician and the patient. Any 3D cardiac images (CMR, CT or echo) can be used to create a 3D model. The 3D data needs to be extracted from the images and then converted into an .stl file that can be input into a 3D printer. The 3D piece can be manufactured using a wide variety of materials to meet the needs of the model—rigid for patient explanation, flexible for device pre-procedural planning and tissue-like for surgical practice with suturing (Fig. 13.21).

CONCLUSION CHD is a complex area of medicine that requires a well-integrated understanding of anatomy and cardiovascular physiology. Using the principles illustrated in this chapter, cardiovascular imaging can be successfully utilised to guide medical and surgical management of patients with CHD. An ECG remains the first-line imaging technique; however, when an ECG cannot provide a complete diagnosis, cross-sectional imaging with CMR and CT is the next line of investigation. Catheter angiography is usually reserved for problem solving, coronary artery assessment, haemodynamics and for catheter-guided therapeutic procedures. It is hoped that the complementary nature of imaging techniques has been demonstrated and that when used in combination, ECG, CMR, CT and x-ray catheter angiography can provide a comprehensive assessment of patients with CHD.

314

SECTION A  The Chest and Cardiovascular System

A

B

C

D

Fig. 13.21  Three-Dimensional Printing of a Variety of Congenital Anatomies Made From Different Materials. (A) Great vessels in rigid material for patient education. (B) Multicolour print: right heart printed in purple, left heart in red. (C) Transparent material for visualisation of device deployment in coarctation. (D) Right ventricular outflow tract and pulmonary trunk for testing deployment of novel pulmonary valve device.

FURTHER READING Anderson, R., Baker, J., Penny, D., et al., 2010. Paediatric Cardiology, third ed. Churchill Livingstone, Edinburgh. Bogaert, J., Dymarkowski, S., Taylor, A., et al., 2012. Clinical Cardiac MRI. Springer, New York. Han, B.K., Rigsby, C.K., Hlavacek, A., et al., 2015. Computed tomography imaging in patients with congenital heart disease part I: rationale and utility. An expert consensus document of the Society of Cardiovascular Computed Tomography (SCCT): endorsed by the Society of Pediatric Radiology (SPR) and the North American Society of Cardiac Imaging (NASCI). J. Cardiovasc. Comput. Tomogr. 9, 475–492.

Han, B.K., Rigsby, C.K., Leipsic, J., et al., 2015. Computed tomography imaging in patients with congenital heart disease, part 2: technical recommendations. An expert consensus document of the Society of Cardiovascular Computed Tomography (SCCT): endorsed by the Society of Pediatric Radiology (SPR) and the North American Society of Cardiac Imaging (NASCI). J. Cardiovasc. Comput. Tomogr. 9, 493–513. Lai, W.W., Mertens, L., Cohen, M., et al., 2009. Echocardiography in Pediatric and Congenital Heart Disease: From Fetus to Adult. Wiley-Blackwell, Oxford. Sridharan, S., Price, G., Tann, O., et al., 2010. Cardiovascular MRI in Congenital Heart Disease: An Imaging Atlas. Springer, Berlin; London.

14  Nonischaemic Acquired Heart Disease Luigi Natale, Veronica Bordonaro

CHAPTER OUTLINE Role of Imaging, 315 Cardiomyopathies, 316 Valvular Heart Disease, 334

Nonischaemic heart diseases (NIHDs) account for nearly half of the cardiac deaths. This group of diseases is extremely heterogeneous, including cardiomyopathies (CMPs), valvular problems, cardiac masses and pericardial disease. Modern noninvasive imaging techniques have increased diagnostic accuracy for all these diseases, with consequent decrease in the number of invasive procedures.

ROLE OF IMAGING In the past, NIHD diagnosis was based on chest radiography and invasive angiography; the introduction of echocardiography has deeply modified the diagnostic approach, as both myocardium and heart chambers are visualised noninvasively, in real time and with the same examination. Furthermore, magnetic resonance has increased the role of noninvasive imaging, with a wider field of view, higher contrast resolution and tissue characterisation capabilities, coupled with extremely accurate, operator-independent functional assessment. Finally, multidetector electrocardiographic (ECG)-gated computed tomography (CT) has had a deep impact on noninvasive coronary artery imaging; technological improvements have also made CT effective in the assessment and therapeutic planning of NIHDs, particularly in valve diseases and cardiac masses.

Chest Radiography (CXR) It still remains the first-line examination in both ischaemic and NIHD. Its advantages are low cost, noninvasiveness, wide availability and unique information on pulmonary haemodynamics; its downside of course is that it is neither specific nor sensitive, particularly if disease is not yet at an advanced stage. Chest x-ray (CXR) interpretation is based on a stepwise procedure: the chest wall anatomy may explain modification in heart contours; pleural or parenchymal disease may cause nonspecific symptoms such as chest pain. Analysis of vessel size and distribution which reflect the haemodynamic status is fundamental in assessing heart disease: size and distribution are strictly related to capillary wedge pressure and pulmonary venous pressure, that equalises left atrial pressure and, consequently, left ventricular end-diastolic pressure (LVEDP). So, depending on acute or chronic development of the disease, it is possible to obtain a noninvasive qualitative assessment of LVEDP.

Tumours of the Heart, 348 Malignant Cardiac Tumours, 356 Pericardial Diseases, 358

Furthermore, a general assessment of heart size determines whether there is cardiomegaly or not; this will help to create a list of possible diseases, with or without cardiomegaly, according to signs, symptoms and other clinical data. The next step is the analysis of modifications of cardiac contours in both the frontal and lateral views, which may be helpful in identifying specific chamber enlargement.

Echocardiography Echocardiography is a noninvasive, portable ultrasound technique that allows high-resolution, two- and three-dimensional (2D and 3D) views of the cardiac chambers, valves and pericardium. These techniques, either using a transthoracic or transoesophageal approach, can assess cardiac anatomy and ventricular function. When combined with Doppler and colour Doppler techniques, valvular regurgitation and stenosis and transvalvular pressure gradients can also be assessed. Echocardiography is the most commonly performed imaging examination in the assessment of NIHDs.

Magnetic Resonance Imaging Cardiac magnetic resonance (CMR) imaging is rapidly becoming very useful in the assessment of NIHDs. Its role is most valuable in (A) serial measurement of ventricular function in patients with CMP (considered superior to echocardiography especially with respect to reproducibility); (B) evaluation and quantification of valve function, including stenosis and regurgitation; (C) morphology and extent of involvement in cardiac tumours; and (D) value of postcontrast delayed enhancement (also called late gadolinium enhancement [LGE]) in assessing diagnosis and determining prognosis in many diseases, as hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), as well as many types of infiltrative myocardial diseases, including sarcoidosis and myocarditis. Recently introduced new CMR parametric techniques allow for a quantitative evaluation of myocardial tissue, based on changes in T1 (native and postcontrast agent), T2 and T2* (star) relaxation times. Mapping sequences use different techniques to obtain a series of images at various inversion times, from which a recovery curve is derived; the result is a parametric image that shows the T1 or T2 relaxation values pixel by pixel. Changes of T1 relaxation times, although not specific for single disease, can be grouped into different patterns which reflect intracellular alterations of the cardiomyocyte (such us iron overload or glyco­sphingolipid accumulation in Anderson–Fabry disease [AFD]), extracellular changes in the myocardial interstitium (such us collagen or amyloid proteins

315

316

SECTION A  The Chest and Cardiovascular System

accumulation, as seen in myocardial fibrosis or cardiac amyloidosis, respectively) or both (myocardial oedema with increased intracellular and/or extracellular water). In particular, native T1 lengthens with an increase of tissue water (as in cases of oedema in acute infarction or inflammation) or an increase of interstitial space (fibrosis or amyloid deposition) and conversely shortens in presence of an iron overload or an intracellular lipid accumulation (as in AFD). Moreover, T1 mapping performed before and after injection of an extracellular T1-shortening contrast agent is used for calculating a parameter called the extracellular volume (ECV), a measure of the proportion of extracellular space within the myocardium. Several pathophysiological processes can alter the ECV: it may increase with fibrosis, oedema or other protein deposition, such as amyloid. Conversely, low ECV values occur in thrombus and fat/lipomatous metaplasia. Regional myocardial T2 mapping has emerged to directly quantify local myocardial inflammation and oedema. T2 parametric mapping overcomes some of the limitations of the qualitative evaluation of T2 weighted images such as image quality, low reproducibility and subjective assessment. Finally, T2* mapping, derived from gradient echo (GRE) sequences, provides an accurate quantification of myocardial tissue iron overload, which is extremely important for clinical management of transfusiondependent haemoglobinopathies (e.g. β-thalassaemia major).

Computed Tomography Until recently, conventional CT had a limited role in the evaluation of NIHDs. However, the increasing use of ECG-gated multidetector (>16 slice) CT techniques currently makes cardiac CT a viable alternative in assessing cardiac function. More recently, cardiac CT has also been shown to be useful in the assessment of valvular function. However, as yet, CT has no demonstrable advantage over echocardiography and CMR in the evaluation of NIHDs. Furthermore, despite recent reconstruction advances, radiation dose remains an issue.

CARDIOMYOPATHIES According to the 2006 American Heart Association (AHA) definition, CMPs are ‘a heterogeneous group of diseases associated with mechanical and/or electrical dysfunction, usually exhibiting inappropriate

hypertrophy or dilatation, due to a variety of causes, often genetic, confined to the heart or part of systemic disorder’. This classification divided CMPs into primary and secondary; primary CMPs are subdivided into genetic, mixed or predominantly nongenetic and acquired, while secondary CMPs are a variety of diseases that can affect the myocardium (Table 14.1). In 2008, the European Society of Cardiology (ESC) proposed another classification based on different phenotypes (hypertrophic, dilated, etc.) that are subclassified into familial-genetic and nonfamilial-nongenetic (Table 14.2). The AHA classification is more based on pathology, whereas the ESC classification is more clinical, as, for example, a hypertrophic phenotype can be due to many diseases that cause myocardial thickening such as HCM, hypertensive CMP, Fabry disease, amyloidosis, etc. In this chapter, the phenotype approach will be used, according to the ESC classification.

Hypertrophic Pattern Increased myocardial thickness can be due to a variety of disease, both familial and nonfamilial. Among the familial, the classical HCM is the most common; it is autosomal dominant and defined as a sarcomere disease (with a number of different mutations) that is characterised by an excessive hypertrophy of the myocardium (not explained by other causes) in a nondilated left ventricle. Eleven mutations have been recognised, with the most common affecting the β-myosin heavy chain. Pathologically there is disarray of myocytes with a variable amount of interstitial fibrosis, caused by microvascular bed damage. Typically, increased septal thickness, exceeding 15 mm, is seen at echocardiography; in the presence of ECG abnormalities and symptoms (e.g. chest pain, shortness of breath and dizziness, but also presyncope, syncope and arrhythmias), increased thickness of the interventricular septum is sufficient for diagnosis. Another echocardiographic criterion is a ratio between septal thickness and inferior wall of the left ventricle at midventricular level exceeding 1.3. Chest x-ray is often unhelpful: in concentric hypertrophy there may be a rounded third left cardiac contour; this is different from that due to aortic stenosis and systemic hypertension, which both can cause concentric hypertrophy of the left ventricle. Echocardiography can easily assess myocardial thickness (Fig. 14.1); however, there are many

TABLE 14.1  Summary of Familial Diseases Causing Cardiomyopathies, by American Heart Association FAMILIAL HCM

DCM

ARVC

RCM

Unclassified

• Familial (unknown gene) • Sarcomeric protein mutation • Glycogen storage disease • Lysosomal storage disease • Disorders of fatty acid metabolism • Carnitine deficiency • Phosphorylase B kinase deficiency • Mitochondrial cytopathies • Syndromic HCM • Other

• Familial (unknown gene) • Sarcomeric protein mutation • Cytoskeletal genes • Nuclear membrane • Mildly dilated CM • Intercalated disc protein mutations • Mitochondrial cytopathy

• Familial (unknown gene) • Intercalated disc protein mutations • Cardiac ryanodine receptor • Transforming growth factor-β3

• Familial (unknown gene) • Sarcomeric protein mutation • Familial amyloidosis • Desminopathy • Pseudoxanthoma elasticum • Haemochromatosis • Anderson–Fabry disease • Glycogen storage disease

• Left ventricular noncompaction • Barth syndrome • Lamin A/C • ZASP • α-Dystobrevin

ARVC, arrhythmogenic right ventricular cardiomyopathy: DCM, dilated cardiomyopathy; HCM, hypertrophic cardiomyopathy; RCM, restrictive cardiomyopathy. Modified from Maron, B.J., Towbin, J.A., Thiene, G., et al., American Heart Association; Council on Clinical Cardiology, Heart Failure and Transplantation Committee, 2006. Contemporary definitions and classification of the cardiomyopathies: an American Heart Association Scientific Statement from the Council on Clinical Cardiology, Heart Failure and Transplantation Committee; Quality of Care and Outcomes Research and Functional Genomics and Translational Biology Interdisciplinary Working Groups; and Council on Epidemiology and Prevention. Circulation 113 (14), 1807–1816.

CHAPTER 14  Nonischaemic Acquired Heart Disease

317

TABLE 14.2  Summary of Nonfamilial Diseases Causing Cardiomyopathies, by American

Heart Association HCM

DCM

ARVC

RCM

Unclassified

• Obesity • Infants of diabetic mothers • Disorders of fatty acid metabolism • Athletic training • Amyloid

• Myocarditis • Kawasaki disease • Eosinophilic • Viral persistence • Drugs • Pregnancy • Endocrine • Nutritional • Alcoholic • Tachycardiomyopathy

• Inflammation?

• Amyloid • Scleroderma • Endomyocardial fibrosis • Carcinoid heart disease • Metastatic cancers • Radiation • Drugs

• Takotsubo cardiomyopathy

ARVC, arrhythmogenic right ventricular cardiomyopathy: DCM, dilated cardiomyopathy; HCM, hypertrophic cardiomyopathy; RCM, restrictive cardiomyopathy. Modified from Maron, B.J., Towbin, J.A., Thiene, G., et al., American Heart Association; Council on Clinical Cardiology, Heart Failure and Transplantation Committee, 2006. Contemporary definitions and classification of the cardiomyopathies: an American Heart Association Scientific Statement from the Council on Clinical Cardiology, Heart Failure and Transplantation Committee; Quality of Care and Outcomes Research and Functional Genomics and Translational Biology Interdisciplinary Working Groups; and Council on Epidemiology and Prevention. Circulation 113 (14), 1807–1816.

A

B

C Fig. 14.1  M-mode (A) and B-mode (B, C) echocardiography in hypertrophic cardiomyopathy; M-mode allows measurements of left ventricle diastolic diameter (44 mm) and systolic diameter (28 mm), as well as thickened interventricular septum (28 mm). B and C show diastolic and systolic short-axis images, with clear evidence of hypertrophic septum.

318

SECTION A  The Chest and Cardiovascular System

patterns of hypertrophy distribution, not all of them easily recognisable by ultrasound. Hypertrophy can be asymmetrical or septal, with or without left ventricle obstruction, symmetrical, apical, midventricular, mass-like and noncontiguous. In asymmetrical or septal forms, 25% of cases show a systolic obstruction of the left ventricle outflow tract (Fig. 14.2), due to the movement of the anterior leaflet of the mitral valve toward the hypertrophic interventricular septum, caused by the Venturi effect. This obstruction can be present at rest or only during/following physical exercise; it can easily be recognised at echocardiography and echo-Doppler. Echocardiography has limitations particularly in apical forms close to the low-frequency probe, and in mass-like forms, where differential diagnosis with tumours can be difficult. In all these cases, and especially when the acoustic window is poor, magnetic resonance imaging (MRI) is extremely useful and accurate; it provides precise measurement of wall thickness (Fig. 14.3), is more

accurate in left ventricle mass quantification and can easily detect and quantify right ventricle involvement. Information on wall thickness can also be easily obtained by cardiac CT (Fig. 14.4); modern CT equipment provides this information at a very low dose (1–3 mSv). However, the most relevant information that only MRI provides is the presence of interstitial fibrosis; due to the increased extravascular bed of collagen and its impaired wash-out delayed enhancement after administration of contrast agent delineates fibrotic tissue as an area of ‘bright’ myocardium compared with normal ‘dark’ myocardium. The distribution of delayed enhancement can be either diffuse or with selective localisation in the septum and relative sparing of the subendocardial layer (differently from myocardial infarct scar), at the anterior and inferior septal insertion, or patchy with large foci of intramural enhancement (see Fig. 14.3B).

A

B

C

D Fig. 14.2  B-mode (A, B) and Doppler echocardiography (C, D) horizontal long-axis views in hypertrophic cardiomyopathy; B-mode images show basal septal hypertrophy. (C) Doppler interrogation in outflow tract demonstrates a rest systolic gradient of 30 mm Hg, while transmitral flow evaluation (D) shows an impairment of diastolic function, with reduced E wave (E) equalised to A wave (A).

CHAPTER 14  Nonischaemic Acquired Heart Disease

A

319

B Fig. 14.3  Hypertrophic Cardiomyopathy. (A) Cine–magnetic resonance imaging (MRI) frame showing typical localisation at basal-septum. (B) Late gadolinium enhancement MRI, short axis: thickened interventricular septum with large amount of fibrosis (hyperintense intramural foci).

Fig. 14.4  Cardiac computed tomography in hypertrophic cardiomyopathy: short-axis midventricular image shows a diffuse left ventricle myocardial hypertrophy, with predominant involvement of anterior wall.

Detection of fibrosis is extremely important because it is strictly related to prognosis: a variety of published papers reported the incidence of severe arrhythmias, due to reentry mechanisms, and sudden cardiac death in young patients (40 years) with HCM and severe fibrosis demonstrated at MRI. MRI findings suggest that the fibrotic tissue probably develops due to impairment of intramural myocardial blood supply: in patients presenting with acute chest pain and focal intramural areas of oedema (acute damage) with a nonischaemic pattern, vasodilator stress perfusion MRI often reveals a reduced myocardial flow reserve, corresponding to areas of fibrosis during late enhancement (Fig. 14.5). Finally, MRI permits recognition of associated findings, such as mitral regurgitation, and more sophisticated functional evaluation may reveal impairment of diastolic filling and radial or circumferential strains.

Histologically, fibrosis is often global and diffuse in these patients and it may be undetected on LGE imaging because of the absence of normal reference myocardium; furthermore, the evaluation of microscopic interstitial fibrosis is often hampered by the spatial resolution of LGE images; T1 mapping can be helpful in these cases. Native T1 values are elevated in HCM (Fig. 14.6A) and correlate with wall thickness, suggesting potential clinical utility as a marker of disease severity. In these patients, ECV is usually elevated, due to extracellular matrix expansion and myocardial disarray (see Fig. 14.6B and C); furthermore, this parameter can differentiate between HCM and athletic remodelling in athlete’s heart as the latter shows reduced ECV values, probably due to an increase in healthy myocardium by cellular hypertrophy. In addition to its role in primary diagnosis, T1 mapping can be a valid tool for a quantitative follow-up of the degree of fibrosis, enabling an accurate stratification of risk for adverse events. The hypertrophic phenotype, characterised by increased myocardial thickness, is also seen in other CMPs; delayed enhanced MRI is particularly useful for determining the differential diagnosis, which includes storage diseases such as AFD, amyloidosis and noninfectious granulomatous diseases (e.g. sarcoidosis). In these cases, delayed enhancement patterns differ from those of HCM (Fig. 14.7); in AFD it is typically subepicardic in the inferolateral wall, with different presentations in amyloidosis. Due to the diffuse infiltration by amyloid and its link to gadolinium compounds, it is difficult to null myocardial signal, resulting in diffuse intermediateto-bright signal intensity; alternatively, a patchy intramural pattern can be present, with small bright spots. Sarcoidosis can also present with this hypertrophic phenotype; again, patchy pattern of distribution is more frequently seen with small foci, reflecting interstitial distribution of granulomas.

Dilated Phenotype DCM is defined as a left ventricular dilatation with systolic dysfunction not caused by abnormal loading (as hypertension or valve disease) or coronary artery disease.

320

SECTION A  The Chest and Cardiovascular System

A

B Fig. 14.5  Gadolinium-Enhanced Magnetic Resonance Imaging Short Axis. (A) First-pass frame showing intramural perfusion defect in lateral wall (arrow); (B) late enhancement due to fibrosis is evident in the same segment.

A

B

C

Fig. 14.6  Hypertrophic Cardiomyopathy. Hypertrophic cardiomyopathy: (A) Native T1 colour map shows increased T1 values (in green) within septum, anterior and anterolateral walls of left ventricle; (B) extracellular volume colour map demonstrates higher values (in blue) within anterior septum; both abnormalities are related to interstitial fibrosis; (C) late enhancement shows no gadolinium uptake in the myocardium.

A dilated phenotype can be due to many different causes; among familial forms, autosomal dominant ones are more frequent, due to mutations of cytoskeletal, sarcomeric and other protein genes. Other forms are X-linked, as muscular dystrophia. Nonfamilial DCM include end-stage inflammatory diseases (infective and noninfective myocarditis), nutritional deficiencies, endocrine dysfunctions and drug toxicity. Chest x-ray has limited sensitivity, as cardiac contour abnormalities can be observed only in advanced stages; typically, on frontal view the third left cardiac arch is enlarged, heart size is increased and, eventually, there are signs of left atrium enlargement (carina widening and double contour of second right cardiac arch) (Fig. 14.8). However, the most relevant role of chest x-ray is the evaluation of pulmonary vasculature; due to increased left ventricular end-diastolic pressure (LVEDP), left atrium, pulmonary veins and capillary wedge pressures increase, with consequent balanced distribution or caudocranial redistribution of pulmonary vessels. In case of high-pressure values, further evolution

causes pulmonary oedema (interstitial to alveolar). This is the only noninvasive tool to estimate LVEDP. Echocardiography is considered the first-line examination in clinically suspected DCM; echocardiographic criteria are increased left ventricle end-diastolic diameter (Fig. 14.9), exceeding normal values of 112% after age and body surface area correction, ejection fraction lower than 45% and fractional shortening lower than 25%. Another useful parameter is the spherical index that correlates the left ventricular end-diastolic volume with the long-axis diameter and is increased in DCM. Furthermore, echocardiography is able to demonstrate regional diffuse hypokinesis, sometimes restricted to apical segments, decreased forward flow velocities across all the valves and dominant E wave (early diastolic filling) at mitral flow interrogation, as well as complications such as mitral and/or tricuspid regurgitation and left ventricle thrombi. In patients with a poor acoustic window, cardiac MRI is extremely useful, because of the high intrinsic contrast resolution between blood and myocardium (Fig. 14.10); planimetric and volume measurements

CHAPTER 14  Nonischaemic Acquired Heart Disease

321

Fig. 14.7  Different Late Gadolinium Enhancement in Hypertrophic Phenotypes. First column: large septal intramural late enhancement in hypertrophic cardiomyopathy (HCM; arrows). Second column: intramural lateral wall late enhancement in Fabry disease (arrow). Third column: subendocardial diffuse late enhancement in amyloidosis (black arrows in left ventricle, white arrows in right ventricle). Fourth column: upper row T2 weighted image showing intramural hyperintense foci and streaks due to oedema, lower row late enhancement in the same areas.

A

B Fig. 14.8  Primary Dilated Cardiomyopathy. (A) Frontal view and (B) lateral view show overall increased cardiac size, with signs of left atrial enlargement (black arrows in A, white arrows in B) and left ventricle enlargement (thick black arrows in A, thick white arrow in B).

are extremely accurate and more reproducible than echocardiographic ones even if, in clinical settings, ultrasound measurements are commonly used. One of the major contributions of MRI is the differential diagnosis between ischaemic and nonischaemic CMP that is crucial for decision

making and treatment planning (revascularisation versus medical therapy and/or transplant); echocardiography in fact has a low specificity and this differential diagnosis is not always easy. Using delayed enhancement technique, MRI is able to easily differentiate an ischaemic DCM (Fig. 14.11), demonstrating subendocardial or transmural scars, whereas in

Fig. 14.9  Dilated Cardiomyopathy. Echocardiography shows enlarged left ventricle.

Fig. 14.10  Black-blood fast spin-echo T1 weighted image of dilated cardiomyopathy: left atrial and left ventricle enlargement is clearly evident.

A

C

B

D Fig. 14.11  Ischaemic Versus Nonischaemic Dilated Cardiomyopathy. Upper row: (A) diastolic frame of two chambers cine–magnetic resonance imaging (MRI) showing dilated left ventricle with inferoapical myocardial thinning; (B) late enhancement in subendocardial layer of basal, mid and apical inferior segments, due to previous infarct in right coronary artery territory. Lower row: (C) diastolic frame of two-chamber cine–MRI showing dilated left ventricle; (D) short-axis late enhancement image, showing no gadolinium uptake.

CHAPTER 14  Nonischaemic Acquired Heart Disease nonischaemic DCM, late enhancement is absent or faint and limited to mesocardial layers, usually diffuse or septal (Fig. 14.12). Another important indication for MRI is the evaluation of the left ventricle before resynchronisation therapy. In clinical practice ECG and echocardiography with Doppler interrogation are mostly used, but the significant percentage of patients not responding to implantation— considering the high cost of the procedure—have increased the role of MRI in the preprocedural assessment of the presence of scar tissue in the inferior wall. Finally, MRI is extremely accurate in thrombus detection (Fig. 14.13); contrast-enhanced images showed the highest sensitivity, also in areas where echocardiography has false negatives (e.g. close to the apex).

323

Differential diagnosis of nonischaemic forms is only partially feasible with delayed enhancement technique: in general, intramural or mesocardial enhancement is more frequent in postmyocarditis DCM that in the idiopathic form, but this has to be further investigated and confirmed. Histology is still mandatory in these cases; in other secondary forms, MRI is extremely useful: for example, in dilated end stage of HCM, the pattern and distribution of late enhancement help to establish the correct diagnosis. The prognostic role of MRI in DCM is still under investigation; data are few and less robust than in HCM for risk assessment; left ventricular remodelling, ventricular tachycardia and sudden cardiac death seem to be related to the presence and amount of late enhancement.

Fig. 14.12  Late gadolinium enhancement images in short-axis and four-chamber view, showing septal intramural contrast uptake in idiopathic dilated cardiomyopathy.

324

SECTION A  The Chest and Cardiovascular System

A

B Fig. 14.13  Magnetic resonance imaging of dilated cardiomyopathy complication: thrombi in left ventricle are evident in late enhancement on four-chamber (A) and two-chamber (B) views.

A

B

C

Fig. 14.14  Dilated Cardiomyopathy. (A) Native MRI T1 colour map demonstrates increased T1 values (in green) within interventricular septum and lateral septum; (B) extracellular volume colour map shows diffuse increased values (in green), due to interstitial fibrosis; (C) late enhancement shows no abnormalities.

Recently, several studies have investigated the value of T1 mapping. Similarly to hypertrophic phenotype, DCMs are associated with the development of diffuse myocardial fibrosis during disease progression. On MRI native T1 is increased in DCM (Fig. 14.14) and inversely correlated with wall thickness. Furthermore, native T1 values were found to be increased not only in areas corresponding to LGE but also in areas without LGE (see Fig. 14.14C), which suggests that this parameter can also detect interstitial myocardial fibrosis, allowing for initiation of timely management.

ECV measurement reflects myocardial collagen content in DCM and correlates with LV dysfunction, supporting its use as a noninvasive imaging biomarker to monitor therapy response. ECV values in DCM have been shown to be increased (see Fig. 14.14B) similar to hypertrophic phenotype; however, this overlap in ECV between the two phenotypes is to be considered clinically irrelevant because DCM and HCM can be usually distinguished by their different morphological features and ventricular geometry. When echocardiography proves difficult (e.g. poor acoustic window), cardiac CT can be used as an alternative to MRI to assess ventricular

CHAPTER 14  Nonischaemic Acquired Heart Disease volumes and ejection fraction, but most important, it is useful in differential diagnosis between ischaemic and nonischaemic forms by means of its capability to exclude coronary artery disease (Fig. 14.15). CT is also highly accurate in left ventricle thrombus detection, whereas late enhancement technique is not quite so useful in CT due to its lower contrast resolution compared with MRI.

Restrictive Phenotype In this phenotype, the increased wall stiffness causes a rapid pressure increase with only small volume increase; this restrictive pattern can occur in a wide range of diseases; by definition, both diastolic and systolic are normal or reduced and wall thickness is normal. Familial restrictive cardiomyopathy (RCM) is very rare and mostly autosomal dominant; nonfamilial forms are caused by systemic disorders, as amyloidosis, sarcoidosis, haemochromatosis, AFD, carcinoid heart disease, anthracycline toxicity, endomyocardial diseases, with or without hypereosinophilia (as endomyocardial fibrosis) and endocardial fibroelastosis. Chest x-ray is frequently unremarkable; only at an advanced stage does left atrial enlargement become evident with signs of increased pulmonary venous pressure, as in mitral stenosis (MS). Echocardiography often shows normal-sized ventricles, enlarged atria and normal or decreased contractile function; in some cases, as

325

in Löffler syndrome, or endomyocardial fibroelastosis or carcinoid syndrome, endocardial thickening is evident (Fig. 14.16). Doppler evaluation of mitral flow is particularly important, showing elevated early diastolic velocity, short deceleration time and low and shortened atrial velocity. However, these abnormal parameters can be present also in case of constrictive pericarditis, where pericardial stiffness does not allow ventricular filling; in this case, it is important to evaluate pericardial thickness, the interventricular septal kinetics and inferior vena cava flow during deep inspiration and expiration. The pericardium is not easily assessed by ultrasound, and here either cardiac CT or MR plays a major role; a cut-off value of 4 mm is highly predictive of pericardial constriction, but it is important to remember that pericardial constriction without pericardial thickening can also occur. MRI can easily assess morphological and functional abnormalities of restriction (Fig. 14.17); T1 and T2 weighted images can help in tissue characterisation (Fig. 14.18). As previously described, unenhanced and delayed enhancement can be useful in differentiating some diseases, as amyloidosis (Fig. 14.19), sarcoidosis and AFD (see Fig. 14.7). Tissue characterisation with parametric mapping methods has also a promising value in the differential diagnosis of these forms.

Fig. 14.15  Cardiac Computed Tomography in Dilated Cardiomyopathy. (A) Left ventricle dilatation in four-chamber view; (B, C) Maximum intensity projection (MIP) reformatted images of right coronary artery, common trunk and left descending coronary artery showing no atherosclerotic lesions.

A

B

C

326

SECTION A  The Chest and Cardiovascular System

Fig. 14.16  Two-dimensional echocardiography in restrictive cardiomyopathy: end-diastolic frame on left, end-systolic frame on right. In both images endocardial thickening with thrombotic layer and calcification are evident. Thickening of mitral valve is also present.

A

B Fig. 14.17  Restrictive Cardiomyopathy. (A) Cine–magnetic resonance imaging frame in four-chamber view, showing tubular shape of ventricles and enlarged atria. (B) Late enhancement short-axis image does not show any myocardial signal abnormality.

CHAPTER 14  Nonischaemic Acquired Heart Disease

A

327

B Fig. 14.18  Restrictive Cardiomyopathy Due to Fibroelastosis. (A) Spin-echo T1 weighted MR image shows an apparent thickened myocardium. (B) Short tau inversion recovery image shows a thickened hyperintense endocardium (arrows) with normal myocardium.

A

B

C

Fig. 14.19  Cardiac Amyloidosis. (A and B) Contiguous T1 weighted axial MR slices show marked myocardium thickening of both ventricles, with heterogeneous signal intensity; a small amount of pericardial effusion is present (arrows). (C) Late gadolinium enhancement short-axis view shows no myocardial suppression due to increased extracellular space and interstitial amyloid accumulation.

Cardiac amyloidosis can be due mainly to transthyretin deficit Amyloid Transthyretin (ATTR) or to light chain disease (AL); LGE patterns are different between them, as light chain amyloidosis shows more frequently a global subendocardial distribution, whereas ATTR amyloidosis is characterised more frequently by a global transmural distribution with failure in myocardial nulling at any given inversion time. This explains why native T1 and ECV may have more predictive power than LGE, providing a quantitative assessment of the diffuse extracellular expansion due to interstitial deposition of amyloid proteins. In cardiac amyloidosis, abnormal proteins accumulate within the myocardial interstitial space modifying the composite relaxation time of the tissue, and resulting in elevated native T1 values that can be markedly increased in AL disease. ATTR amyloidosis is more frequently associated with a higher ECV value compared with AL amyloidosis.

Furthermore, native T1 and ECV are also elevated when normal conventional clinical testing and LGE imaging suggested no cardiac involvement, representing a potential early disease marker. In AFD, the low native T1 of fat can serve as an early biomarker of myocardial glycol-sphingolipid storage, even before the development of LV hypertrophy. Conversely, the ECV is typically normal, because this parameter reflects an extracellular/interstitial disease, whereas AFD is an intracellular (lysosomal) storage disease. The inferolateral wall, which is typically characterised by LGE in these patients, can show a pseudonormalisation or even elevation of T1 value, due to the effects of extensive fibrotic replacement which exceed the fatty-related T1 decrease. Cardiac sarcoidosis is characterised by noncaseating granulomatous infiltration that produces a patchy disorder which often involves small amounts of myocardial tissue, and in early stages it may not cause LV

328

SECTION A  The Chest and Cardiovascular System

dysfunction. CMR appearance largely depends on timing of disease: in the acute phase a patchy increased T2 signal intensity can be due to oedema and inflammation and the LGE can show a patchy midmyocardial, subepicardial or epicardial pattern; in chronic disease, nodular foci of LGE without oedema on T2 weighted images are due to fibrosis and scar formation. Although LGE imaging is a powerful predictor of risk in patients with sarcoidosis, the earlier stages of the disease can go undetected on LGE images; in these cases, abnormal myocardial T2 times can reveal an active inflammation before the development of myocardial scar, suggesting that T2 mapping may improve the detection of granulomatous

active inflammation when it is potentially reversible with appropriate treatment. Finally, cardiac MRI is extremely important in the assessment of iron overload of the myocardium (Fig. 14.20) as the measurement of T2* (transverse relaxation time in gradient-echo sequence) closely correlates with iron overload; consequently, it is possible to modulate chelation therapy in these patients, which helps to reduce the mortality for heart failure, the leading cause of death in this disease. Cardiac CT, as well as MRI, can be used in differentiating RCM from constrictive pericarditis, by means of pericardium thickness measurement (Fig. 14.21); a potential advantage over MRI is the assessment of pericardial calcification.

Fig. 14.20  Iron Overload in Major Thalassaemia. Multiecho fast gradient-echo sequence for T2* quantification: from a echo time (TE) (1.1 ms, top left) to a long TE (18 ms, bottom right), a rapid decay of myocardium signal intensity.

A

B Fig. 14.21  Constrictive Pericarditis. (A) This axial black-blood fast spin-echo MR image shows a diffuse pericardial thickening (5 mm), more evident anteriorly. Note also incomplete blood suppression in right atrium due to slow flow. (B) Axial unenhanced cardiac computed tomography confirms pericardial thickening, but also the presence of calcification.

CHAPTER 14  Nonischaemic Acquired Heart Disease

Arrhythmogenic Right Ventricular Cardiomyopathy Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a relatively uncommon familial disease, usually autosomal dominant, characterised by various desmosome proteins mutations. These result in right ventricle dysfunction, global or regional, with or without left ventricle involvement, in the presence of histopathological evidence of right ventricular myocardial replacement with fatty and fibrous tissue in various amounts. There is preferential localisation in the so-called triangle of dysplasia (outflow tract, inflow tract and apex). There are distinctive electrocardiographic abnormalities, according with previously published and recently revisited criteria. ARVC is a frequent cause of sudden cardiac death in young people due to severe ventricular arrhythmias. For this reason, early diagnosis is crucial. In 1994 a number of criteria were proposed: familial history, ECG alterations, repolarisation abnormalities, arrhythmias (ventricular tachycardia with left bundle branch block, extrasystoles >1000/24 hours), right ventricle systolic dysfunction, fibrofatty replacement at endomyocardial biopsy. Based on the severity of the alterations, major and minor criteria are distinguished and a certain combination of findings is needed to establish (or rule out) the diagnosis. In doubtful cases, it is important to strictly follow the patient. In 2010 these criteria were revised, and, particularly, quantitative parameters for right ventricle enlargement and ejection fraction were introduced by means of ultrasound or cardiac MRI measurements with the aim to increase sensitivity (Table 14.3). MRI has been used in the past quite exclusively to demonstrate fatty infiltration of the right ventricle myocardium, particularly in the so-called triangle of dysplasia, the region between the subtricuspid area, the apex and the free wall of the outflow tract. However, the capability to assess the presence of fat is strictly related to the amount of the fat, because of the limited spatial resolution: for this reason, it is possible to recognise only cases with extensive replacement (Fig. 14.22). Another issue is the relative difficulty in differentiating intramyocardial from subepicardial fat accumulation, partly because of limited spatial resolution but also because subepicardial fat accumulation is also seen in other physiological and pathological conditions such as obesity, steroid therapy and old age (Fig. 14.23). For the revised criteria, MRI has been included only for the assessment of ventricular size and global/segmental functional assessment because it has been demonstrated to be more sensitive than ultrasound for right ventricular systolic function evaluation; it is superior for volumes measurements as it uses a 3D approach (unlike ultrasound) that needs a geometrical assumption (not available for the right ventricle) to measure volumes. In addition, regional function is superiorly evaluated by MRI: systolic bulging and aneurysms of the right ventricle anterior wall are easily detected (Fig. 14.24) and can be also seen in the free wall, which is not always assessable by echocardiography. Recent data have been published on the use of delayed enhancement technique to demonstrate presence of fibrotic tissue, although this is not easy to visualise in the thin myocardium of the right ventricle.

329

Despite the fact that the typical fibrofatty replacement of ARCV should lead to alterations of both native T1 and ECV, the thin RV free wall hinders the use of parametric mapping techniques. Cardiac CT is not important considering the pre-eminent role of MRI; however, in small series, CT has detected small foci of fat deposition/ substitution even in the right ventricle due to its high spatial resolution and good contrast resolution for fat. However, it is less easy to assess regional wall motion abnormalities of the right ventricle at CT, although new CT techniques may improve this.

Fig. 14.22  Arrhythmogenic Right Ventricular Cardiomyopathy. Black-blood axial MR image shows a complete fatty substitution of right ventricle free wall (high signal intensity tissue); similar foci are evident in left ventricle apex and basal lateral wall (arrows).

TABLE 14.3  Secondary Cardiomyopathies Secondary Cardiomyopathies Include: • Infiltrative • Neuromuscular/neurological • Storage • Nutritional deficiency • Toxicity • Autoimmune/collagen • Endomyocardial • Electrolyte imbalance • Inflammatory (granulomatous) • Consequence of cancer therapy • Endocrine

Fig. 14.23  T1 weighted axial MR image in a patient under chronic steroid treatment; increased amount of fat in the mediastinum, in prepericardial and in subepicardial spaces, but normal right ventricle myocardium is visible.

330

SECTION A  The Chest and Cardiovascular System

A

B Fig. 14.24  Arrhythmogenic Right Ventricular Cardiomyopathy. Cine–magnetic resonance (A, axial; B, short-axis) images show huge right ventricle dilatation with small free wall bulgings (arrows in A).

Unclassified Cardiomyopathy In this group, two entities are included: left ventricular noncompaction (LVNC) and takotsubo CMP. LVNC is characterised by prominent left ventricular trabeculae and deep intertrabeculae recesses; this results in segments of thickened myocardium, composed of thin compacted epicardium with a thick endocardial layer. It is unclear if LVNC is a separate CMP, a congenital or acquired trait shared by other phenotypes; it can be isolated or associated with congenital anomalies (complex cyanotic, Ebstein) or muscular dystrophies. It is frequently familial, with approximately 25% of asymptomatic relatives showing a wide range of ultrasound abnormalities. Echocardiography is the first and often unique diagnostic tool to assess LVNC, especially in paediatric patients. Increased trabeculation is usually detected in midapical segments, both on lateral wall and septum (Fig. 14.25), with the latter being normally nontrabeculated. In difficult cases, MRI can be useful, using a cut-off value of 2.3 for the ratio between noncompacted and compacted myocardium (Fig. 14.26). Furthermore, delayed enhancement is able to demonstrate the presence of fibrotic tissue in the compacted myocardium, more often in forms associated with muscular dystrophies (Fig. 14.27). The native T1 values in LVNC patients with and without LGE were significantly higher than in the normal controls; furthermore, LVNC patients without LGE finding may show elevated native T1, suggesting that native T1 mapping can be used earlier than LGE imaging to detect myocardial fibrosis. Takotsubo CMP or transient left ventricular apical ballooning syndrome is characterised by a reversible regional systolic dysfunction, associated with chest pain and a negative invasive coronary angiography; typically, patients present with an acute coronary syndrome, mostly seen in postmenopausal women after a physical or emotional stress with norepinephrine acting as a neuromediator. Functional recovery usually occurs within days or a couple of weeks.

Fig. 14.25  Echocardiography in Left Ventricular Noncompaction. Two-chamber view, shows thickened endocardium and increased trabeculation of left ventricle apex (arrows).

CHAPTER 14  Nonischaemic Acquired Heart Disease

A

B Fig. 14.26  Left Ventricular Noncompaction: Magnetic Resonance Imaging. (A) T1 weighted black-blood vertical long-axis MR image shows increased number and thickness of myocardial trabeculae in mid and apical left ventricle regions. (B) Cine–magnetic resonance frame horizontal long-axis with measurement of noncompacted and compacted myocardium.

Fig. 14.27  Left Ventricular Noncompaction: Late Enhancement MR Imaging. Show contrast uptake in compacted myocardium (mostly at septal level), with subendocardial sparing, due to fibrotic changes. As an ancillary finding, multiple thrombi are also visible.

331

332

SECTION A  The Chest and Cardiovascular System

A

B Fig. 14.28  Takotsubo Cardiomyopathy. (A) T2 weighted MR image in vertical long axis shows diffuse hyperintensity of the myocardium, due to oedema (arrows). (B) Late enhancement image in the same plane shows no contrast uptake, demonstrating absence of any irreversible lesion.

Echocardiography is usually able to detect the abnormal function and the ballooning but, because of the acute presentation, invasive coronary angiography is required to exclude obstructive coronary artery disease according to guidelines. In these cases, after negative coronary angiography, there is a strong indication for cardiac MRI, with the aim to differentiate a myocardial infarction with normal coronary arteries, from an acute myocarditis and a takotsubo CMP. MRI is indeed able to detect the functional ballooning but, more importantly, the oedema of the myocardium without any delayed enhancement, indicating the absence of necrosis and the reversibility of the damage (Fig. 14.28). T1 and T2 mapping can be used to confirm the presence of increased T1 and T2 relaxation times in the hypokinetic regions, and they show higher diagnostic accuracy than T2 short tau inversion recovery (STIR) images to detect oedema, because of their quantitative nature.

Myocarditis Myocarditis is not included in the ESC CMP classification as a single definite category but is included in the acquired group as inflammatory CMP; in the ESC classification, chronic myocarditis is described as demonstrating the dilated phenotype, whereas acute myocarditis does not show a precise phenotype. Consequently, myocarditis has to be treated apart from the classification, also considering its increasing incidence and the recently increased possibilities to define the diagnosis. Myocarditis is an acute or chronic inflammatory process affecting the myocardium; underlying causes can be toxic, infective (viral, bacterial, rickettsial, fugal, parasitic) or a hypersensitivity reaction. It typically evolves through an active, healing and healed stage with progression of inflammatory infiltrates, interstitial oedema, myocyte necrosis and finally scarring. In some cases, subclinical forms of viral myocarditis can trigger an autoimmune reaction causing immunological damage to the myocardium and/or cytoskeletal disruption leading to

a DCM phenotype with LV dysfunction. In these cases, viral persistence and chronic inflammatory infiltrates have been demonstrated. Histology obtained through an endomyocardial biopsy is still considered the ‘gold standard’ for the diagnosis, based on the combination of leucocyte infiltration and necrosis defined by the so-called Dallas criteria, even if these criteria have been recently debated. Substantial refinement of the diagnosis can be reached by using molecular analysis of the specimens, such as DNA–RNA extraction and polymerase chain reaction. Acute presentation can often mimic an acute coronary syndrome with chest pain, exertion dyspnoea, ECG abnormalities and mild enzyme elevation. In this case there is an indication to invasive coronary angiography to exclude obstructive coronary artery disease, even if the presence of a recent viral infection or unexplained fever is indicative of the correct diagnosis. Echocardiography is usually performed to assess and quantify left ventricle systolic dysfunction; in case of associate pericardial effusion, ultrasound can suggest the suspicion of an inflammatory process, but need for further examination remains. In the acute presentation, after negative coronary angiography, there is a strong indication to perform cardiac MRI, which can demonstrate oedema and delayed enhancement in the subepicardial layer, most frequently in the lateral and/or inferior wall (Fig. 14.29). This pattern easily allows differential diagnosis from acute myocardial infarction, where delayed enhancement is located in the subendocardial layer or is transmural. In takotsubo CMP, delayed enhancement is typically absent. However, a negative delayed enhancement study does not exclude an acute myocarditis, because there is not always macroscopic detectable necrosis; in this case it is important to acquire MR images before and immediately after administration of contrast agent (early enhancement) to demonstrate inflammatory hyperaemia. The combination of oedema imaging (T2 weighted sequences), early and late enhancement imaging constitutes the cornerstone for the

CHAPTER 14  Nonischaemic Acquired Heart Disease

A

333

B Fig. 14.29  Acute Myocarditis. (A) T2 weighted short tau inversion recovery short-axis MR image shows a subepicardial hyperintense area (arrow) in inferior wall. (B) Late enhancement image in corresponding plane shows contrast uptake with a nonischaemic pattern. Patient presented 36 hours before in an emergency unit with chest pain and slightly elevated cardiac enzymes; emergency coronary angiography was negative.

A

B Fig. 14.30  Chronic Myocarditis in Patients Presenting with New-Onset Severe Tachyarrhythmia. (A) T2 weighted short tau inversion recovery short-axis MR image shows no evidence of oedema. (B) Late enhancement image in the same plane shows septal intramural contrast uptake. Septal endomyocardial biopsy demonstrated a lymphocytic infiltrate with interstitial fibrosis.

diagnosis according to the so-called Lake Louise criteria. Cardiac MRI is of course useful and accurate to evaluate biventricular global and regional systolic function and eventually associated pericardial effusion, with or without pericardial enhancement, in case of myopericarditis. Acute fulminant forms or cardiogenic shock can represent other rare but possible acute presentations. Clinical presentation can also be more subtle, e.g. as new-onset cardiac failure or with tachyarrhythmias; these presentations are more often associated with chronic myocarditis.

In these situations, echocardiography is useful to exclude other diseases: for example, a new-onset heart failure. Cardiac MRI is particularly helpful to demonstrate fibrosis with late enhancement technique typically located in the mesocardium. Enhancement is generally less intense than in the acute forms, diffuse rather than patchy and located in lateral wall or in the septum, the latter being more frequent in tachyarrhythmias (Fig. 14.30). The presence of tissue oedema on MRI is usually detected visually on T2 weighted-STIR images or using a semiquantitative method such

334

SECTION A  The Chest and Cardiovascular System

as the ratio of myocardial to skeletal muscle (T2 ratio) by drawing two distinct regions of interest in the same slice. However, these methods can be prone to errors due to a highly variable distribution of inflammatory foci into the myocardium, which range from solitary small infiltrates to a diffuse involvement of myocardial tissue. Moreover, the contemporary presence of skeletal muscle inflammation with oedema, which is possible in coexisting myositis, can lead to false-negative results for myocardial oedema. In this scenario, T1, ECV and T2 mapping have a crucial clinical value and may be used in conjunction with the Lake Louise Criteria. In particular, myocardial T2 mapping provides a noncontrast quantitative assessment of myocardial oedema in patients with suspected myocarditis without the limitations associated with T2 weighted imaging such us signal heterogeneity and the need of reference tissue for signal normalisation. Furthermore, native T1 values are significantly elevated in these patients (Fig. 14.31A), and it has been proven to be superior to T2 weighted imaging and LGE, providing a high level of diagnostic accuracy with high positive and negative predictive values. The elevated T1 relaxation time is likely due to both the increased water content in the intracellular and extracellular space, as well as an impaired electrolyte distribution in the inflamed tissue. Lastly, ECV is also increased in acute myocarditis, reflecting both myocardial oedema and myocyte necrosis with subsequent myocardial fibrosis (see Fig. 14.31B–E); this parameter, together with

LGE significantly improves the diagnostic accuracy compared with the Lake Louise CMR criteria.

VALVULAR HEART DISEASE Compared with the past, the aetiology of acquired valve disease is currently usually degenerative rather than due to rheumatic or infective causes; the prevalence of mitral regurgitation and aortic stenosis is higher than MS and aortic insufficiency. In addition, the age of onset is increasing. Findings on plain radiography are often minimal and nonspecific, and the assessment is based on echocardiography.

Mitral Valve Disease In Western countries, nonrheumatic mitral valve disease is the most common manifestation of mitral disease; in non-Western countries, rheumatic heart disease is still prevalent. Among nonrheumatic diseases, mitral regurgitation is the most common, while nonrheumatic MS is very rare. Many conditions can result in significant mitral regurgitation, affecting the mitral leaflets (prolapse, endocarditis, mucopolysaccharidosis, lupus, rheumatoid arthritis) or the subvalvular apparatus (annular dilatation, chordae tendinae rupture, annular calcification, myocardial infarction, HCM). In mitral regurgitation, a portion of the left ventricular stroke volume is directed retrogradely into the atrium during systole, returning to the left ventricle during diastole with consequent left ventricular volume

C

B

A

D

E

Fig. 14.31  Myocarditis MR Images. (A) Native T1 colour map shows markedly increased T1 value (in red) within lateral segments. (B and D) Extracellular volume (ECV) colour maps in short- and horizontal long-axes views demonstrate increased ECV values (in green) within anterior, anterolateral, inferolateral and inferior segments. These abnormalities are due to diffuse inflammatory oedema. (C and E) Late enhancement corresponding planes demonstrate patchy and subepicardial gadolinium accumulation only in lateral segments.

CHAPTER 14  Nonischaemic Acquired Heart Disease loading. To maintain an adequate stroke volume, both the left ventricular stroke volume and the ejection fraction increase. Acute regurgitation may result from infective endocarditis or rupture of chordae tendineae/papillary muscles. Sudden volume loading into the noncompliant left atrium may result in markedly elevated atrial pressure, causing acute pulmonary oedema and symptoms of heart failure. Rupture or elongation of chordae tendineae, ischaemic and nonischaemic CMP, hypertrophic obstructive CMP and rheumatic heart disease are all causes of chronic mitral regurgitation. In response to the chronic volume load, both the left atrium and left ventricle dilate, thus serving as a reservoir for the regurgitant volume without necessarily increasing pulmonary vascular pressure. However, if the left ventricle decompensates and the forward stroke volume decreases, heart failure is the result. The appearances on chest radiography depend on the duration and severity of the mitral regurgitation and any other associated heart disease. Acute, severe mitral regurgitation may present with pulmonary oedema but with a virtually normal heart size and shape (Fig. 14.32). After an interval, the heart usually decompensates by developing marked left ventricular dilatation. Selective left atrial enlargement may be absent, slight or moderate, with or without left atrial appendage enlargement (Fig. 14.33). Pulmonary vascular changes reflect the haemodynamic derangement and the effects of treatment. As previously described, mitral regurgitation can be due to many different causes such as mitral prolapse, chordae tendineae rupture (during bacterial endocarditis, collagen diseases, acute myocardial infarction) or be functional (DCM, ischaemic).

335

The diagnosis is typically made by echocardiography (Fig. 14.34), but is also possible with CMR with two- and three-chamber views preferred; associated imaging findings include leaflet thickening (thickness of >5 mm) and flail leaflet. CT has also a good sensitivity for detecting mitral valve prolapse, small vegetations and ruptured chordae; furthermore, cardiac CT may be useful to assess valvular and subvalvular calcifications. Prolapse of the mitral valve may be associated with chest pain and ECG changes, which may suggest ischaemic heart disease.

Chordal Rupture It may complicate bacterial endocarditis or, less frequently, myocardial infarction or connective tissue diseases, leading to flail of part of a

Mitral Valve Prolapse The most common cause of severe (nonischaemic) mitral regurgitation is the mitral valve prolapse syndrome. Mitral valve prolapse is defined as systolic bowing of the mitral leaflet more than 2 mm beyond the annular plane into the atrium, caused by rupture or elongation of the chordae tendineae. The middle scallop of the posterior leaflet is most often affected. Mitral valve prolapse is due to elongation of the chordae tendineae associated with myxomatous degeneration of the valve leaflets, occurring alone or in association with Marfan syndrome and in patients with atrial septal defect.

Fig. 14.33  Chest x-ray shows left atrial appendage enlargement (arrow) in mitral regurgitation; subcarinal opacity with slight elevation of left main bronchus is also evident due to left atrium enlargement.

Fig. 14.32  Chest x-ray, frontal view, in intensive care unit shows signs of alveolar pulmonary oedema without cardiac enlargement.

Fig. 14.34  Mitral Prolapse. Echocardiography (parasternal long axis) shows wide anterior leaflet prolapse of the mitral valve. Ao, Aorta; LA, left atrium; LV, left ventricle.

336

SECTION A  The Chest and Cardiovascular System

leaflet; as a result there will be eversion of the mitral leaflet tip into the atrium during systole, preventing proper closure in systole and producing severe mitral regurgitation. In case of an acute event, acute pulmonary oedema is detected at CXR usually without cardiomegaly (see Fig. 14.32).

Functional Mitral Regurgitation This may occur in DCM or ischaemic cardiac failure. In this case the valve is normal, but regional wall motion abnormalities, left ventricular dilatation, tethering of chordae tendinae or annular dilatation, alone or in combination, lead to dysfunction of the mitral apparatus with ‘functional’ mitral regurgitation. Mitral regurgitation in these situations is very common and severe (Fig. 14.35). Mitral regurgitation detection and severity are usually assessed by 2D Doppler echocardiography. This can detect left atrial dilatation, increased atrial emptying volume and gradual closure of aortic valve during systole, coupled with visualisation of the systolic regurgitant, colour-coded flow within the left atrium. Pulsed Doppler interrogation is extremely sensitive also to small amounts of regurgitant flow; the extent of penetration and the area of the regurgitant jet within the left atrium allow a semiquantitative estimation of the disease (mild, moderate, severe) (Fig. 14.36). However, in case of asymmetrical jets (flail) this

method underestimates the severity. In addition, continuous Doppler is very sensitive, giving a characteristic high-velocity, parabolic, systolic spectrum of the flow. MRI identifies mitral regurgitation in cine, as a retrograde jet through the mitral orifice from the left ventricle into the left atrium, due to turbulent flow and consequent spin dephasing (Fig. 14.37). Mitral valve regurgitation can be quantified as the difference between left ventricle stroke volume (LVSV) and right ventricle stroke volume (RVSV). Every difference between the two measured stroke volumes indicates the amount of blood which comes back through the insufficient valve during diastole. This estimation is valid only if the tricuspid valve is competent (tricuspid regurgitation is reported in up to 50% of patients with significant mitral regurgitation); moreover, the calculation of RVSV is less reproducible compared with LVSV due to the extensive trabeculation of the right ventricle. In cases with combined aortic and mitral regurgitation, the difference represents the sum of regurgitant volume. Phase-contrast MRI can discriminate between anterograde and retrograde flow during the cardiac cycle; the mitral regurgitant volume (MRV) can be measured by MRI as the difference between the LV stroke volume and the aortic forward flow. The regurgitant fraction (RF) is the ratio of the MRV divided by the LVSV. It may also be possible to directly measure mitral flow by phase-contrast velocity flow mapping at the tips of the mitral valve leaflets, but this is superiorly obtained with a specialised imaging sequence which tracks the motion of the mitral valve annulus during the cardiac cycle to adjust the plane of velocity encoding for diastolic mitral valve motion. With cine–phase-contrast MRI, mitral regurgitation can be calculated as left ventricular inflow through the mitral valve minus left ventricular outflow in the ascending aorta. This approach is also applicable in patients with mitral and aortic regurgitation, because diastolic left ventricular inflow (= left ventricular mitral inflow and aortic regurgitation volume) is equal to the systolic left ventricular outflow (= aortic outflow and MRV).

Fig. 14.35  Mitral Regurgitation. Colour Doppler ultrasound shows severe regurgitation in the left atrium and marked left ventricular enlargement (mitral annulus dilatation)

Fig. 14.36  Mitral Regurgitation. Colour Doppler shows severe mitral regurgitation (apical 4C). Mosaic effect is evident with complete occupation of the left atrium.

Fig. 14.37  Cine–magnetic resonance imaging frame of functional mitral regurgitation (black jet directed from left ventricle left atrium, arrow) in in dilated cardiomyopathy.

CHAPTER 14  Nonischaemic Acquired Heart Disease The AHA and the American College of Cardiology (ACC) have established echocardiographic criteria for grading the severity of mitral regurgitation. Even in the absence of intersociety established criteria for MRI, the measurements of regurgitant volume and fraction derived from LV volume and ascending aortic flow, allow to grade the regurgitation as: mild = RV 50%. For a comprehensive assessment of patients with mitral regurgitation, three components are mandatory: (1) quantification of regurgitation, (2) assessment of left ventricular adaptation to volume overload and (3) anatomy of mitral valve and subvalvular apparatus. Whereas MR satisfies points 1 and 2, the method of choice for assessment of valve anatomy is echocardiography, although improvements in MR imaging strategies allow detection of morphological abnormalities such as flail mitral valve leaflets.

Mitral Stenosis MS is a structural abnormality of the mitral valve, which prevents proper opening during diastolic filling of the left ventricle. Increased left atrial pressure is necessary to move blood across the stenotic mitral valve and into the left ventricle. Chronic elevation of left atrial pressure causes atrial dilatation and pulmonary venous hypertension. Atrial fibrillation (due to atrial dilatation) and dyspnoea (due to pulmonary venous hypertension) are common symptoms of MS. Prolonged pulmonary venous hypertension may also lead to right ventricular dilatation and failure, as well as tricuspid regurgitation. MS is highly prevalent in developing countries because of its association with rheumatic fever but is also seen in developed countries. The most common cause of MS is rheumatic fever. Isolated MS is twice as common in women as in men; it occurs in 40% of all patients presenting with rheumatic heart disease; a history of rheumatic fever can be elicited from approximately 60% of patients presenting with pure MS. Other causes of MS are very rare and include congenital anomalies, prior exposure to chest radiation, mucopolysaccharidosis, severe mitral annular calcification, ball valve thrombus and left atrial myxoma. In cor triatriatum, the left atrium is divided by a membrane into two chambers; blood flow may be restricted before it reaches the mitral valve, mimicking MS. The main features of a stenotic mitral valve are leaflet thickening, nodularity and commissural fusion, with narrowing of the valve to the shape of a fish mouth. Leaflets might be calcified; chords may be fused and shortened. The normal mitral valve cross-sectional area is 4–6 cm2 and a gradient is rare unless the valve is less than 2 cm2. Symptoms correlate with increasing mean left atrial pressure, when mitral valve area (MVA) reduces to 1.5 cm2. The increase in left atrial pressure from obstruction across the mitral valve is transmitted to the pulmonary circulation, causing dyspnoea and leading to pulmonary oedema. Haemoptysis may occur. In chronic severe MS, secondary pulmonary hypertension may cause right ventricular failure and tricuspid regurgitation. Occasional embolic episodes are related to the atrial fibrillation. Death is mainly due to heart failure or systemic embolism. The standard for diagnosis and determination of the severity of MS is 2D/Doppler echocardiography. Two-dimensional echocardiography evaluates the morphology of the mitral valve leaflets and the subvalvular apparatus. Leaflet thickening (Fig. 14.38) or calcification (hockey stick deformity of the mitral valve leaflets is typical), leaflet mobility, commissural or subvalvular fusion can be seen. Narrowing of the MVA can be appreciated.

337

Fig. 14.38  Mitral Stenosis. Echocardiography (parasternal long axis) shows marked thickening of mitral leaflets with restricted mitral valve orifice (doming anterior leaflet). Left atrial (LA) enlargement is evident.

Two-dimensional echocardiography helps assessment of the suitability for mitral valve valvotomy: a pliable noncalcified valve could be suitable for balloon valvuloplasty or commissurotomy; a calcified fibrotic valve with subvalvular fusion that may preclude valvotomy. Criteria for determining the severity of mitral valve obstruction are the mean mitral valve gradient and the MVA: mild MS is present when the area exceeds 1.5 cm2 and the mean gradient is less than 5 mm Hg; moderate stenosis is seen with an area 1.0 to 1.5 cm2 and a gradient of 5 to 10 mm Hg; severe stenosis occurs when the area is less than 1.0 cm2 and the gradient exceeds 10 mm Hg. Mitral valve area (MVA): • The most reliable method to calculate the valve area is planimetry from the short-axis view at the tip of the mitral valve leaflets; an even higher reliability might be achieved with 3D echocardiography. • MVA is inversely related to diastolic half-time ( T 1 2 ): that is the time it takes for the maximal mitral gradient to decrease by 50%. This can be obtained from the rate of velocity decrease during early and mid-diastole, as assessed on the transmitral velocity curve: T 12 = DT × 0.29; MVA = 220 T 12 where T 1 2 = half-time and DT = deceleration time. • MVA can be calculated with the continuity equation when the area derived from the half-time does not correlate with the mean transmitral gradient MVA = (LVOT TVI × LVOT area) MV TVI where LVOT = left ventricular outflow tract, MV = mitral valve and TVI = time-velocity integral. Cine–MRI can be helpful in selected cases, however, with good visualisation of the restricted mitral leaflets and the anterograde jet due to turbulent flow across stenotic valve orifice, particularly on the two chambers (Fig. 14.39) and LV outflow tract views. MS that is caused by rheumatic disease may have distinctive morphological features: restricted opening, thickened leaflets, commissural fusion, valve calcification, a ‘fish-mouth’ appearance on short-axis images. Bowing of a thickened and fibrotic anterior leaflet during diastole results in a ‘hockey stick’ appearance that is best seen on two- or fourchamber images. Direct measurement of the orifice area can be performed in the same way as for aortic stenosis, by placing an imaging plane perpendicular

338

SECTION A  The Chest and Cardiovascular System

Fig. 14.39  Cine–Magnetic Resonance Imaging Frame of Mitral Stenosis. A small flow void directed from left atrium (LA) to left ventricle (LV) is visible (arrows), due to mild mitral stenosis. Left atrium is enlarged.

to the direction of flow at the mitral valve tips during diastole and drawing a contour around the smallest valve orifice. The technique has good agreement with echocardiography, but care needs to be taken in positioning the plane at the tips to obtain an accurate valve area, and multiple parallel thin slices may be helpful. Diastolic flow and velocity can also be measured in this image plane; velocity-encoded cardiovascular magnetic resonance can be used as a robust tool to quantify MVA via mitral flow velocity analysis with the pressure-half-time (PHT) method, although the frequency of atrial fibrillation in severe MS reduces the accuracy of the flow measurements. MVA planimetry can be also determined by multidetector computed tomography (MDCT); can also be determined by CT which may provide reliable quantification of mitral valve stenosis (MVS) and allow accurate assessment of severity. Mitral valve leaflet calcification on MDCT indicates mitral valve sclerosis or stenosis. Cardiac catheterisation and angiocardiography is used in those rare situations where echocardiography has failed to elucidate the contribution of each valve lesion or when coexistent coronary artery disease needs assessment. Nonrheumatic causes of MS usually produce nonspecific imaging features such as valve thickening or leaflet fixation. However, CT and MR imaging characteristics occasionally are suggestive of the cause of stenosis. For example, radiation-induced valve disease is associated with mitral premature calcification of the apparatus, lung fibrosis and focal vertebral abnormalities. Calcification of the mitral leaflets, a cause of senescent MS, may be depicted at CT. Nonvalvular disease (e.g. ball-valve thrombus, left atrial myxoma) may also produce signs and symptoms of MS.

Rheumatic Mitral Valve Disease Acute rheumatic disease can cause a pancarditis. During the acute phase, mitral regurgitation can be present; this is usually reversible. Chronic rheumatic mitral valve disease often leads to stenosis, due to fusion of the commissures, thickening and shortening of the chordae tendineae and fibrosis of the papillary muscles. Severe mitral regurgitation results from leaflet destruction. However, there is usually a combination of MS and regurgitation, with the first limiting the amount of the second, so that both cannot be severe at the same time.

Fig. 14.40  Tricuspid Regurgitation. Echo colour Doppler demonstrates severe tricuspid insufficiency with mosaic effect occupying the entire right atrium.

Tricuspid Valve Disease Tricuspid stenosis is generally rheumatic, more common in females than in males, and usually associated with tricuspid regurgitation and MS. Most commonly, tricuspid regurgitation is functional and secondary to marked dilatation of the tricuspid annulus due to RV enlargement in the presence of pulmonary hypertension, mitral valve disease or mitral valve replacement, ischaemic heart disease or DCM. Tricuspid valve regurgitation may be directly caused by rheumatic disease. Severe tricuspid regurgitation can also occur in endomyocardial fibrosis and carcinoid heart syndrome (also responsible for stenosis). In Ebstein anomaly, the insertion of the septal cusp of the tricuspid valve is displaced towards the apex of the right ventricle. Endocarditis can also cause tricuspid regurgitation. The clinical recognition of tricuspid valve disease can be difficult. On CXR, the main radiological sign is right atrial enlargement that can be appreciated on frontal view as an increase of the second right cardiac; in the lateral view, an enlarged right appendage is seen as increased retrosternal opacity between the aortic arch and the outflow tract of the right ventricle. Again, echocardiography is the most important diagnostic tool. Colour Doppler can easily detect retrograde flow in the right atrium; by measuring the depth and the area of penetration of the jet, regurgitation can be semiquantitatively graded (Fig. 14.40). At pulsed Doppler, regurgitation appears as a pansystolic turbulent signal in the right atrium;

CHAPTER 14  Nonischaemic Acquired Heart Disease

Fig. 14.41  Cine–magnetic resonance imaging frame of tricuspid regurgitation: a retrograde black jet directed from right ventricle to right atrium is evident (arrowheads).

if retrograde flow is appreciated also in inferior vena cava, regurgitation can be graded as severe. The estimation of the peak velocity of regurgitant flow through the valve at continuous Doppler is not useful to grade the regurgitation but allows estimation of the peak systolic pressure in the right ventricle and pulmonary artery, applying the modified Bernoulli formula. This parameter is extremely useful for evaluating the haemodynamic effect of all left chambers and myocardial diseases. In tricuspid stenosis, echocardiography allows the visualisation of thickened valve leaflets and their limited motion, whereas Doppler permits visualisation and measurement of the jet. Cardiac MRI, with steady-state free-precession cine sequences, is used to delineate abnormal valvular morphology such as in Ebstein anomaly or carcinoid heart disease. Tricuspid regurgitation signal void seen in the right atrium is less evident than mitral regurgitation for its lower velocity and turbulence (Fig. 14.41). The regurgitant volume can be quantified combining phase-contrast flow measurements at the pulmonary valve and RV stroke volume.

Aortic Valve Disease Aortic Stenosis

The predominant cause of aortic stenosis in Western countries is degenerative calcific disease in middle-aged or elderly patients. Compared with MS, where calcium is deposited on an already stenotic valve, calcific aortic stenosis results from calcification of relatively normal aortic cusps, which then cause obstruction. Rheumatic disease is a rare cause nowadays. Clinical presentation may include breathlessness, chest pain or syncope, with ECG signs of left ventricular hypertrophy. CXR can detect rounding of the cardiac apex, suggesting left ventricular hypertrophy (Fig. 14.42). Prominence of the ascending aorta (first right cardiac arch) may indicate poststenotic dilatation, but in older patients the whole thoracic aorta may be widened from atherosclerosis. In lateral view, it is easier to demonstrate the presence of aortic valve calcification. TTE is the first-line examination, assessing valve morphology (leaflets number and thickening morphology, calcification) and function

339

Fig. 14.42  Chest x-rays of aortic stenosis shows rounded profile of left ventricle (white arrows), with slight enlargement of ascending aorta (arrowheads).

(leaflet mobility). The degree of calcification is related to the severity of stenosis and progression of the disease. Doppler echocardiography allows stenosis quantification, in terms of maximal velocity, pressure gradient and aortic valve area (AVA); particularly, pressure gradient is obtained by the modified Bernoulli equation (Δp = 4v2, where v is maximal velocity), while AVA is functionally obtained from continuous Doppler. The ‘anatomical’ orifice area (AOA) is not equivalent to the valve effective orifice area (EOA). The latter reflects the cross-sectional area of the transvalvular flow jet and is generally smaller than the valve area because there is a contraction of the flow downstream of the valve orifice. The EOA is given by the continuity equation: EOA = SV VTIAo where VTI is the velocity-time integral. This estimation is based on the principle that the flow in the left ventricle outflow tract area must equal the flow in the subsequent valve leaflet area, in absence of shunts. According to maximal velocity, pressure gradient and valve area, aortic stenosis is graded as mild, moderate and severe, with subtle differences between the guidelines of the ESC and the ACC–AHA. To overcome acoustic limitations of the transthoracic approach, a transoesophageal echocardiography can be performed, which allows a planimetric measurement of the valve orifice at maximal opening (mid-systole), especially if 3D technique is applied that overcomes off-axis 2D measurements at the tips of the valve (with consequent overestimation of the area). Coupled with quantitative stenosis grading, echocardiography assesses LV size and function that is mandatory for several reasons: firstly, surgical or interventional corrections are indicated in asymptomatic patients with reduced function (EF 25–30 mm Hg to alveolar oedema (grade 3). The most common cause of PV hypertension, by far, is left-sided heart disease (Table 16.1) due to left ventricular failure, mitral valve disease or aortic valve disease. The severity of mitral valve stenosis can be non-invasively gauged by assessing the PV pressure. In cases of aortic valve disease, however, the degree of PV hypertension is more indicative of myocardial failure than severity of stenosis. It has to be noted that an increased left ventricular pressure load does not immediately result in PV hypertension. Only an elevated end-diastolic left ventricular pressure leads to elevation of the left atrial pressure and subsequently to PV hypertension. The PV pressure can be estimated from the pulmonary artery wedge pressure (PAWP) using a Swan–Ganz catheter and is usually < 12 mm Hg. The radiological findings can be thought of as a progressive series of changes that occur in response to the underlying changes in physiology. Three grades of severity of pulmonary congestion are differentiated (Table 16.2).

Vascular Redistribution (Grade 1)

(thus reversing the normal ‘gravity-dependent’ pattern). This is described as ‘upper lobe venous diversion’ and is often the first recognised radiological sign of PV hypertension (Fig. 16.1). Similar calibres of upper and lower lobe veins do not indicate increased PV pressure if seen in a bedside supine radiograph. Patients suffering from their first episode of acute PV pressure elevation tend to immediately develop an interstitial or alveolar oedema. Only recurrent periods or chronically increased PV pressure result in distended veins.

Interstitial Oedema (Grade 2) If the PV pressure continues to rise and exceeds the plasma oncotic pressure, fluid will begin to accumulate in the lung interstitium. This is known as interstitial pulmonary oedema. Typical radiological signs of interstitial oedema are interstitial (Kerley) lines (Fig. 16.2) caused by thickening of the interlobular septa as a result of fluid accumulation. Kerley B lines are the most obvious ones and are short (1 cm or less) interlobular septal lines, found predominantly in the lower zones peripherally, and parallel to each other but perpendicular to the pleural surface. Kerley A lines are deep septal lines (lymphatic channels), radiating from the periphery (not reaching the pleura) into the central portions of the lung and approximately 4 cm long. Their presence normally indicates a more acute or severe degree of oedema. Septal lines can be differentiated from blood vessels, as the latter are not visible in the outer 1 cm of the lung. In addition, deep septal lines do not branch and are seen with a greater clarity than a blood vessel of similar calibre, as they represent a sheet of tissue (Fig. 16.3A). Under normal circumstances septal lines caused by interstitial fluid overload would be expected to resolve after suitable reduction in PV pressure. Exceptionally, however, they may persist, e.g. in long-standing PV hypertension, where haemosiderin deposition or fibrosis has occurred. Alternative causes of persistent septal lines include idiopathic interstitial fibrosis, lymphangitis carcinomatosa and pneumoconiosis. Other signs of interstitial fluid overload include perihilar haze (loss of visible clarity of the lower lobe and hilar vessels), peribronchial cuffing (apparent thickening of proximal bronchial walls as a result of

As PV pressure rises, the upper lobe veins distend. They initially reach the size of, and eventually become larger than, the lower lobe vessels

TABLE 16.1  Causes of Pulmonary Venous

Hypertension and Pulmonary Oedema

• Left ventricular outflow obstruction (aortic stenosis, aortic coarctation, hypoplastic left heart) • Left ventricular failure • Mitral valve disease • Left atrial myxoma • Fibrosing mediastinitis • Pulmonary veno-occlusive disease

TABLE 16.2  Patterns of Oedema and Corresponding Pulmonary Venous Hypertension

Acute Chronic

Vascular Redistribution Grade 1 (mm Hg)

Interstitial Oedema Grade 2 (mm Hg)

Alveolar Oedema Grade 3 (mm Hg)

12–19 15–25

20–25 25–50

> 25 > 30

399

Fig. 16.1  Upper Lobe Venous Diversion.

400

SECTION A  The Chest and Cardiovascular System

A

A

B Fig. 16.2  Interstitial Oedema. Thickened interlobular septa (Kerley B) at the base of the right lower lobe, bronchial wall thickening (cuffing), overall distended veins and a small pleural effusion with fluid in the interlobar fissures.

interstitial fluid accumulating around their walls) and thickening of the interlobar fissure due to thickened subpleural interstitium (to differentiate from interlobar pleural effusion).

B Fig. 16.3  Interstitial Oedema. PA (A) and lateral (B) chest radiographs show thickened interlobular septa in (A) versus the opacification ranging from ground-glass to dense consolidation in (B).

Alveolar Oedema (Grade 3) As the PV pressure continues to increase, fluid begins to accumulate in the alveolar spaces. This is termed alveolar oedema. Kerley B lines, airspace nodules, bilateral symmetric consolidation in the mid and lower lung zones and pleural effusions may be seen. Depending on the amount of alveolar fluid overload, there are many variations of increased lung density, ranging from subtle haziness to dense consolidation with air bronchograms (see Fig. 16.3B). Certain patterns of opacification may suggest particular diagnoses. The often-cited ‘perihilar bat’s wing’ pattern of airspace consolidation is seen most commonly in left ventricular and renal failure, whereas alveolar oedema localised to the right upper zone is suggestive of severe

mitral regurgitation (Fig. 16.4). The latter is thought to be a result of predominantly regurgitant blood flow in the right upper lobe pulmonary vein, from the superiorly and posteriorly positioned mitral valve. A predominantly upper lobe oedema is seen in patients with a severe head trauma (neurogenic oedema). Alveolar fluid accumulation changes with patient position and gravity: asymmetric consolidations mimicking a pneumonia may be the result of the left- or right-sided position of the patient. Computed tomography (CT) findings of oedema are similar to the radiographic findings, although very atypical patterns are possible, causing differential diagnostic difficulties.

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

401

A

Fig. 16.4  Alveolar Pulmonary Oedema. Bilateral hilar consolidation due to alveolar fluid accumulation (bat’s wing oedema).

Interstitial oedema is characterised by smoothly thickened interlobular septa that do not follow the hydrostatic gradient (Fig. 16.5A). The peribronchovascular interstitium is thickened, which is best seen in the perihilar area. Commonly, there is also, at least subtly, increased parenchymal density due to alveolar fluid overload. Alveolar oedema may initially be recognised as peribronchovascular airspace nodules progressing to diffuse ground-glass or dense airspace consolidation (see Fig. 16.5B). Increased density may follow a ventrodorsal gradient but can also be localised predominantly in the perhilar region or in a patchy distribution, the latter causing sometimes quite atypical patterns (Fig. 16.6). In chronic pulmonary venous hypertension, signs of pulmonary arterial hypertension (PAH) may also develop. In addition, a fine nodular ‘interstitial’ pattern may appear throughout both lungs. These nodules represent haemosiderin deposition. This pattern was previously most commonly seen in patients with long-standing severe mitral stenosis. Pre-existing underlying lung disease influences the pattern and distribution of oedema, Patients with extensive emphysema do not develop homogeneous consolidation; even though degree of PV pressure or fluid overload would otherwise lead to an alveolar oedema, the fluid remains within the interstitium, leading to thickened septa and a rather ‘interstitial’ fluid distribution, meaning that the radiographic appearance may lead to underestimation of the severity of oedema in these patients (Figs 16.7 and 16.8). Although most cases of PV hypertension are associated cardiomegaly resulting from valvular and/or myocardial dysfunction, the presence of cardiomegaly is not universal. An important example of this is in the first 24–48 hours post myocardial infarction. This is due to an acute decrease in myocardial compliance, which essentially resolves in the first week after infarction. Other situations, where signs of pulmonary oedema may be seen associated with a normal heart size, are in patients with

B Fig. 16.5  Interstitial oedema (A) is characterised by thickened interlobular septa and focal areas of increased density. Alveolar oedema (B) causes dense consolidations which can be diffuse or more patchy in distribution. The subpleural area may be spared.

non-cardiogenic pulmonary oedema, in patients with acute overhydration or drug-induced lung oedema (e.g. heroin, aspirin, nitrofurantoin). Cardiogenic oedema has to be differentiated from other underlying diseases that result in an imbalance of hydrostatic pressure, colloid osmotic pressure or capillary permeability, all of them resulting in pulmonary oedema (Table 16.3).

Pulmonary Arterial Hypertension PAH is defined by a mean pulmonary artery pressure of ≥ 25 mm Hg at rest, as measured invasively at right heart catheterisation. This can result from a broad spectrum of disease processes originating in the lungs, pulmonary vasculature or heart diseases with different pathophysiologies, treatments and prognoses. Irrespective of its underlying cause, PAH is a progressive disease leading to substantial morbidity and mortality. Because symptoms are non-specific and evaluation of pulmonary artery pressure is relatively inaccessible, there is often considerable delay between the onset of symptoms and the diagnosis of PAH. Imaging, such as high-resolution computed tomography (HRCT), computed tomography pulmonary angiography (CTPA), magnetic resonance imaging (MRI) and echocardiography, plays a crucial role in the diagnostic work-up of patients with known or suspected pulmonary hypertension. Imaging also plays an important role in raising the possibility of PAH, both in those at risk of developing PAH due to comorbidities and in those presenting with non-specific cardiorespiratory

402

SECTION A  The Chest and Cardiovascular System

A

A

B Fig. 16.6  Two Patients With Alveolar Oedema. The distribution of fluid is very variable, causing different patterns of opacifications, ranging from ground glass to consolidation. Note the subpleural sparing on the right side and the ventrodorsal gradient on the left side in (A), both relatively typical features of oedema. The distribution, however, can be also very asymmetric and sharply demarcated, as in (B).

symptoms where a de novo diagnosis may be suggested. Imaging is particularly important for identifying patients with recurrent or chronic pulmonary thromboembolism where it is fundamental in assessing disease extent and distribution and evaluating the technical feasibility of pulmonary thrombendarterectomy. PAH is a clinical and haemodynamic syndrome that results from increased vascular resistance in the pulmonary circulation. This may be secondary to raised PV pressures (left heart disease), chronic hypoxia resulting either from disease or altitude, or primary disease of either the large vessels (e,g, chronic thromboembolic disease) or the small vessels of the lung. In each of these disease processes exposure of the pulmonary circulation to persistently raised pressures results in ongoing small vessel remodelling and progressive PH. Increase in pulmonary vascular resistance requires high driving pressure to maintain CO, which in turn leads to RV hypertrophy. Ultimately, if untreated, the RV will

B Fig. 16.7  Patient with severe chronic obstructive pulmonary disease without (A) and with (B) signs of pulmonary fluid overload and cardiac decompensation. Note the only discrete increase of vascular markings in (B) while the pleural effusion and the enlarged heart indicate the severity of disease.

fail initially during exercise, later at rest, with death from RV failure occurring a medium of 2.8 years after diagnosis. There are several different causes of PAH. The current PAH classification (Tables 16.4 and 16.5) was developed at the Fifth World Symposium on Pulmonary Hypertension in Nice, France, in 2013 and represents a

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

403

TABLE 16.3  Differential Diagnosis of

Pulmonary Oedema

A

B Fig. 16.8  Corresponding computed tomography images in the same patient as Fig. 16.7 show severe centrilobular emphysema without (A) and with (B) increased septal thickening due to fluid overload but no consolidation.

modification of the previous Dana Point classification. The classification groups together diseases sharing similar pathophysiological mechanisms, clinical presentation and therapeutic options. Group 1 comprises diseases primarily affecting the pulmonary arterioles with vascular remodelling resulting in progressive luminal obliteration. While collectively described as Group 1 or PAH, this group encompasses idiopathic and heritable PAH, PAH associated with drugs, connective tissue diseases, HIV infection, congenital heart disease, portal hypertension and schistosomiasis (Fig. 16.9). PAH is characterised clinically by the presence of pre-capillary PAH (a pulmonary capillary wedge pressure (PCWP) 15 mm Hg DPG < 7 mm Hg and/ or PVR ≤ 3 WU DPG ≥ 7 mm Hg and/ or PVR > 3 WU

All 1—Pulmonary arterial hypertension 3—PH due to lung diseases 4—Chronic thromboembolic PH 5—PH with unclear and/or multifactorial mechanisms 2—PH due to left heart disease 5—PH with unclear and/or multifactorial mechanisms

Definition

DPG, Diastolic pressure gradient; PAPm, mean pulmonary arterial pressure; PAWP, pulmonary artery wedge pressure; PH, pulmonary hypertension; PVR, pulmonary vascular resistance; WU, Wood units.

The ventilation–perfusion (VQ) lung scintigram is a nuclear medicine test widely performed in patients with suspected PAH to exclude CTEPH (Fig. 16.11). According to the European Society of Cardiology and the European Respiratory Society guidelines published in 2009, it is recommended in all patients as the screening method of choice to diagnose or rule out CTEPH. A normal or low-probability VQ result effectively excludes CTEPH with a sensitivity of 90%–100% and a specificity of 94%–100%. Unmatched perfusion defects may also be seen in PVOD and sometimes in patients with PAH. In such cases. CTPA may be used as a complementary investigation. As CTPA is more frequently being used as a first-line test in patients with unexplained symptoms, VQ may not be required if CT has already proven the presence of chronic thromboembolic disease. Traditional pulmonary angiography is still considered the reference standard in the anatomical assessment of the pulmonary arteries, though non-invasive imaging techniques are increasingly utilised. In some centres, angiography is considered mandatory for the work-up of CTEPH to identify patients who may benefit from thrombendarterectomy. Non-invasive imaging in many centres is playing an increasing role in the surgical evaluation or CTEPH. However, with increasing recognition of a potential role for balloon pulmonary angioplasty in CTEPH, a greater emphasis is being placed on angiography in inoperable cases (Fig. 16.12). Angiography may also be helpful in the evaluation of possible vasculitis or pulmonary arteriovenous malformations (PAVMs). Right heart catheterisation is indicated in all patients with suspected PAH to confirm the diagnosis, to evaluate the severity and when PAHspecific drug therapy is considered. Vasoreactivity testing is indicated in patients with idiopathic pulmonary arterial hypertension (IPAH), heritable PAH and other types of PAH; however, it should only be performed in specific centres and under controlled conditions. This testing is not recommended for other PAH groups. MRI plays an important role in diagnosis of PAH because of its ability to assess RV dysfunction caused by increased afterload. Magnetic

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

405

Fig. 16.10  Pulmonary Arterial Hypertension With Massively Enlarged Pulmonary Arteries.

Fig. 16.9  Computed tomography shows large patent ductus arteriosus between left pulmonary artery and aorta (red arrow) resulting in a leftto-right shunt and severe pulmonary hypertension with dilated and hypertrophied (blue arrow) right ventricle.

resonance angiography (MRA) of the pulmonary vasculature (Figs 16.13 and 16.14) and parenchymal perfusion can be combined with dynamic quantitative assessment of ventricular volumes and function (Fig. 16.15). Cardiac MRI provides direct evaluation of the RV size, morphology and function, and allows non-invasive assessment of parameters including stroke volume, CO and RV mass, as well as evaluation of left ventricular and valvular function. Phase contrast MRI permits evaluation of flow in the PA and aorta and, therefore, quantification of any intracardiac shunt as well as dynamic evaluation of the distensibility of the PA. The lack of ionising radiation and objective nature of cardiac MRI assessment of right heart haemodynamics are particularly valuable for follow-up purposes. MRI data can be of prognostic importance—decreased

stroke volume, increased RV end-diastolic volume and decreased LV end-diastolic volume being poor prognostic indicators. Using temporally resolved MR perfusion techniques combined with spatially resolved MRA patients with CTEPH can be accurately differentiated from those with PAH. Pulmonary arterial obstruction or stenosis leads to wedge-shaped perfusion defects or perfusion delay, while in PAH, the perfusion is reduced and heterogeneous. Quantitative evaluation of perfusion in patients with PAH has shown a significantly reduced pulmonary blood flow (PBF) and prolonged mean transit time (MTT) in patients when compared with healthy volunteers, but this did not correlate with the severity of pulmonary hypertension. Computed tomography plays a major role in assessing patients with suspected or known PAH and may suggest the presence of PAH, even when clinically unsuspected. CT is widely available, non-invasive, inexpensive and well tolerated. CTPA permits comprehensive evaluation of the pulmonary vasculature, heart and lung parenchyma. Systematic evaluation of each component is key to full pathophysiological understanding. While CTPA’s role in diagnosing acute and chronic PE is indisputable (see below), the comprehensive evaluation that it provides also demonstrates the impact on the right ventricle, identifies any left heart comorbidity and evaluates the lung parenchyma for signs of chronic lung disease and features of vasculopathy or hypoperfusion (mosaic perfusion). CT may suggest the presence of PAH, independent of its cause when the ‘generic signs’ of PAH are present. The generic signs can include both vascular and cardiac features, which are described below. In the presence of known PAH, or CT signs suggesting its presence, a detailed evaluation of its potential cause should be sought. This will include a systematic evaluation of cardiac causes (left ventricular dilatation, signs of prior infarct, presence of valvular disease or an intracardiac shunt), large vessel obstruction (by chronic thromboembolic disease or rarely tumour or vasculitis) and the lung parenchyma (for primary lung disease, signs of vasculopathy/mosaic perfusion).

Vascular Signs Dilatation of the pulmonary arteries is a useful sign which may suggest pulmonary hypertension. The transverse diameter of the MPA at its bifurcation or the ratio of MPA and aortic diameter at the same level

406

SECTION A  The Chest and Cardiovascular System

Fig. 16.11  Perfusion scintigram in a patient with chronic thromboembolic pulmonary hypertension showing patchy, irregular distribution of the tracer with multiple small and medium-sized perfusion defects.

are the most commonly used parameters. The upper limit of normal for the diameter of the pulmonary trunk is variously described as 29–33 mm and PA:Ao ratio less than 1. A PA:AA ratio greater than 1 carries a high specificity (around 90%) and acceptable sensitivity (around 70%) for the diagnosis of PAH > 25 mm Hg (Fig. 16.16). The dilatation of segmental vessels also aids PAH diagnosis. If the segmental arterialbronchial ratio is greater than 1 in at least three lobes it suggests PAH with a specificity increased to 100% when seen in combination with a dilated main PA. A wide variation of the upper limit of the PA diameter is reported in the literature. Although it is likely that the PA diameter is correlated to patent size or stage in the cardiac cycle, neither the ratio of PA:AA nor normalisation of the PA to the body mass index (BMI) increased the correlation with the mean PAP. As the PA diameter can be considerably increased in patients with interstitial lung disease (e.g. up to 4 cm) in the absence of PAH, it should be interpreted with caution in this setting.

From a practical point of view, the PA:AA ratio has become the most widely accepted sign for many radiologists. Dilatation of bronchial arteries (>1.5 mm) is common in CTEPH when bronchial blood flow may account for over 20% of total blood flow to the lungs and may contribute to oxygenation (Fig. 16.17). Dilated bronchial arteries are most frequently seen in patients with CTEPH but may also be seen in patients with Eisenmenger syndrome or less frequently in IPAH. The enlarged bronchial collaterals are considered a good prognostic sign in CTEPH patients undergoing pulmonary endarterectomy, perhaps reflecting preserved vascular beds distal to pulmonary arterial obstruction.

Cardiac Signs Straightening and later bowing of the interventricular septum towards the left ventricle are signs of RV dysfunction/strain and are common findings in PAH independent of cause (Fig. 16.18A). RV dilatation is

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

407

A Fig. 16.13  Maximum intensity projection of spatially resolved magnetic resonance demonstrating laminated thrombus (yellow arrows), causing severe stenosis of the right main pulmonary artery and occlusion of most left-sided segments with tight webs (blue arrows) in the remaining perfused segments of the left lower lobe. Occlusion of a hepatic vein (red arrow) is noted in keeping with tricuspid regurgitation.

B Fig. 16.12  Pulmonary angioplasty in a patient with chronic thromboembolic pulmonary hypertension. (A) TEPHSubsegmental web causing significant stenosis (blue arrow), which (B) post balloon angioplasty has resolved (green arrow).

considered to be present when the maximal short-axis diameter of RV is greater than that of the LV (RV:LV >1). This is also widely used in the setting of acute PE as a sign of RV dysfunction. • Reflux of contrast medium into the inferior cava and hepatic veins is commonly seen in association with tricuspid regurgitation secondary to PAH but is not specific. While reflux into the inferior vena cava can be seen in normal patients with injection rates exceeding 3 mL/s, reflux into the peripheral hepatic veins is considered abnormal (see Fig. 16.18B). • Pericardial effusions are common in patients with moderate or severe pulmonary hypertension; greater than 10–15 mm in fluid depth in the anterior pericardial recess has been reported in association with increased RV strain in PAH.

Parenchymal Signs Mosaic perfusion is a hallmark of CTEPH, reflecting hypoperfusion of obstructed vascular beds and normal/hyperperfusion of patent beds

Fig. 16.14  Maximum intensity projection image from spatially resolved magnetic resonance showing occlusion (yellow arrow) of the left lower lobe and complex web (blue arrow) in the right lower lobe in a patient with proximal chronic thromboembolic pulmonary hypertension.

(Fig. 16.19). When seen in the presence of generic signs of PAH and the absence of signs of airways disease, it is highly suggestive of CTEPH, even on unenhanced images. Small vessel vasculopathy in PAH may be associated with the presence of subtle diffuse centrilobular ground-glass nodular opacities similar

408

SECTION A  The Chest and Cardiovascular System

A Fig. 16.16  Computed tomography in a patient with chronic thromboembolic pulmonary hypertension demonstrating enlarged main pulmonary artery when compared with the ascending aorta. There are calcified organised chronic thromoemboli in the proximal pulmonary arteries.

B Fig. 16.15  Magnetic Resonance Imaging of the Heart Showing the Regurgitation Jet (Arrows) in Both Atria Due to Tricuspid and Mitral Valve Insufficiency. Note the enlargement of all chambers, the pericardiac and pleural effusion in (A) and the enlarged pulmonary trunk (>3 cm) compared with the aorta in (B). (Courtesy of N. Prakken, MMC, Amersfoort, Netherlands.)

to those seen in non-PAH patients with hypersensitivity pneumonitis or respiratory bronchiolitis. Imaging plays an important role in suggesting the diagnosis of PVOD and PCH, which are considered part of the spectrum in group 1’. Because the obstruction in PCH/PVOD is on the capillary/venous side of the circulation, the use of PAH vasodilator therapy targeting the arterioles can result in life-threatening oedema. In PCH/PVOD, lung parenchymal findings are usually much more pronounced than in PAH; lymphadenopathy and pleural effusion are frequent additional findings.

Fig. 16.17  Coronal maximum intensity projection from computed tomography in a patient with chronic thromboembolic pulmonary hypertension showing unilateral occlusion of the right pulmonary artery. Note enlarged bronchial arteries (arrows) showing the same opacification as the aortic arch.

In PCH, diffuse centrilobular nodular opacities are more dense and may be associated with patchy ground-glass opacity and septal thickening (Fig. 16.20A). In PVOD, interlobular septal thickening is usually a prominent finding (see Fig. 16.20B). As smooth septal thickening is a feature most commonly seen in interstitial oedema, a review of the

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

409

A Fig. 16.19  Mosaic perfusion in the left upper lobe in a patient with computed tomography in a patient with chronic thromboembolic pulmonary hypertension. Note the differences in calibre of the pulmonary arteries in the areas of hypoperfusion (black) and normal perfusion (white).

B Fig. 16.18  Marked dilatation of the right heart chambers (A) and reflux of contrast medium into the inferior vena cava and hepatic veins (B) in a patient with severe computed tomography in a patient with chronic thromboembolic pulmonary hypertension.

PCWP, cardiac morphology and function is required to exclude Group 2 disease. In patients with Eisenmenger syndrome and IPAH, tiny serpiginous intrapulmonary vessels may be seen (so-called neovascularity) arising from centrolobular arterioles not conforming to the usual pulmonary artery anatomy.

Pulmonary Arteriovenous Malformations PAVMs may be diagnosed on clinical grounds and/or by familial screening in patients with hereditary haemorrhagic telangiectasia (HHT). When

acquired, they may be seen in conjunction with liver cirrhosis, schistosomiasis and metastatic thyroid carcinoma. Clinically, they may produce systemic arterial desaturation and give rise to signs of dyspnoea, hypoxia, cyanosis and heart failure. If they rupture, massive haemoptysis and haemothorax occur. Direct communication between a pulmonary artery and vein predisposes one to paradoxical embolism, which is responsible for two-thirds of neurological symptoms in patients with HHT. Although 10% of cases present in the first decade, most do not manifest clinically until the third or fourth decade. Multiple lesions are seen in up to 50% of cases. PAVMs can be treated non-invasively by embolotherapy or by surgery; the preference goes for the former in most cases. Precise understanding of the angioarchitecture, which is necessary before interventional procedures, can be achieved using modern computed tomography angiography (CTA) techniques and three-dimensional (3D) reconstructions. Two types of PAVMs can be differentiated: 1. Simple PAVMs with a single feeding artery and one or several draining veins (80%). 2. Complex PAVM with more than one feeding artery and one or more draining veins (20%). Radiographically, they may appear as round, oval or lobulated opacities with associated often serpiginous feeding and draining vessels, but if small and discrete, they may not be detected on plain chest radiography. They occur most frequently in the lower lobes. Although pulmonary angiography has been considered the ‘gold standard’ for the diagnosis of PAVMs, it has been largely replaced by CTA using thin-slice acquisition and 3D reconstructions such as volume rendering technique (VRT) and maximum intensity projection (MIP), which permits excellent demonstration of the angioarchitecture (Fig. 16.21): CT is more sensitive (approaching 100%) than pulmonary angiography (around 60%). MRI can also be used, but it has lower spatial resolution compared with the multidetector computed tomography (MDCT) technique.

410

SECTION A  The Chest and Cardiovascular System

A

A

B

B

Fig. 16.20  Computed tomography of two patients with severe pulmonary hypertension and parenchymal findings typical for pulmonary capillary haemangiomatosis (PCH) (A) and pulmonary veno-occlusive disease (PVOD) (B). Prominent centrilobular ground-glass nodularity and smooth interlobular septal thickening (A) and widespread ground-glass opacity with centrilobular accentuation and mild smooth interlobular septal thickening (B). Both patterns are considered to represent two spectra of the same disease.

Fig. 16.21  Arteriovenous Malformation (Simple Type) With One Feeding Artery and Draining Vein. Axial maximum intensity projection (A) and coronal volume rendering technique (B) demonstrating the angioarchitecture. (Courtesy of M. Prokop, RUMC, Nijmegen, Netherlands.)

SUMMARY BOX: Pulmonary Hypertension  • Pulmonary hypertension is defined by a mean pulmonary artery pressure measured at invasive right heart catheterisation of ≥ 25 mm Hg. • Pulmonary hypertension independent of cause carries a poor prognosis and is typically diagnosed late when treatment is less effective. • Imaging is key to suggesting the diagnosis, potentially permitting earlier diagnosis and helping identify a specific cause of pulmonary hypertension.

• Pulmonary hypertension is classified according to the pathophysiological mechanism of disease, which helps guide therapy. The most common causes of pulmonary hypertension are secondary to chronic left heart disease and lung disease. Chronic thromboembolism is generally under-recognised, with imaging being fundamental to diagnosis.

411

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

PULMONARY THOMBOEMBOLIC DISEASE ACUTE PULMONARY THROMBOEMBOLISM Background Venous thromboembolism (VTE) encompasses deep venous thrombosis (DVT) and its most serious complication PE. Acute PE is the third most common cause of acute cardiovascular disease after coronary artery disease and stroke, with an annual clinically detected incidence of around 50 per 100,000. As the incidence of acute PE increases with age, and both the size of the population and their life expectancy is increasing, the number of patients diagnosed is likely to significantly increase over time. PE may result in significant morbidity both in the acute phase as well as later in the form of chronic thromboembolic disease. PE is also a potentially lethal disease, directly causing or contributing to the patient’s death in one-third of those diagnosed with the condition. If treated with anticoagulants, mortality directly due to PE is reduced to 8%. The causes of undetected PE are twofold: PE may remain clinically silent, or the non-specific clinical signs and symptoms fail to arouse suspicion of PE. Most patients present with pleuritic chest pain, tachypnoea and/or dyspnoea. The classic clinical triad of sudden chest pain, dyspnoea and haemoptysis is present in only a minority of cases. Other symptoms and signs include cough, syncope, tachycardia, fever and signs of DVT, but patients may also present with shock, hypotension, or even cardiac arrest indicating severe PE. Obstruction of the pulmonary artery may have several physiological effects which interfere with both the circulation and gas exchange, the severity being related to the extent of obstruction. In order to raise the PAP significantly, it is estimated that at least 30%–50% of the pulmonary vascular bed needs to be obstructed. Other factors resulting in vasoconstriction, such as humoral agents and reflex mechanisms, are considered to play an additional role. The abrupt increase in pulmonary vascular resistance results in dilatation of the RV. The increase in RV pressure and volume may eventually lead to RV failure, which is considered the primary cause of death in severe PE. Factors predisposing one to the development of PE include: increasing age (the risk of VTE almost doubles with each decade after the age of 40 years); previous VTE; instrumentation (e.g. indwelling intravenous catheters); neoplastic disease; immobilisation; (orthopaedic) surgery; hypercoagulable state; and hormonal treatment (including oral contraceptives and pregnancy). In some cases, the underlying cause remains unknown. Over 90% of pulmonary emboli originate from thrombus in the deep veins of the legs or pelvis which becomes detached and migrates via the systemic veins to the right side of the heart and into the pulmonary arteries. Emboli normally lodge either at the bifurcation of branching pulmonary arteries (a few are situated at the bifurcation of the MPA, so-called ‘saddle emboli’) or in the peripheral small pulmonary branches. Once an embolus has lodged in a pulmonary artery, it is normally either lysed by the patient’s fibrinolytic system (although this is frequently incomplete) or occasionally becomes organised with recanalisation. The degree to which each of these processes occurs depends, to some extent, on the patient’s fibrinolytic system, the amount of thrombus deposited on the embolus, and the degree of organisation of the embolic material itself. In cases of repeated thromboembolism without lysis of the embolic material, arterial hypertension (CTEPH) can develop. Acute embolus results in either a reduction or a cessation of the distal perfusion. Because of the collateral circulation offered by the bronchial arteries, which increases in the case of PE, lung viability is preserved in the majority of cases and pulmonary infarction usually

does not occur in patients without pre-existing cardiovascular disease. However, if an impaired circulation exists, e.g. due to chronic congestion, the presence of PE may result in local hypoxia, capillary damage, exudation, haemorrhage and coagulation necrosis. Pulmonary infarction develops in only around 15% of thromboembolic events and is seen most commonly in the lower lobes. When a part of the visceral pleura is involved in this process, ischaemia may result in inflammation, which can irritate or adhere to the sensitive parietal pleura, resulting in pleuritic pain. Pulmonary infarcts become revascularised from the periphery, leading to either complete resolution or the development of small scars.

Diagnosis Accurate diagnosis of acute PE is essential, because if untreated, it carries high morbidity and mortality; furthermore, unnecessary anticoagulation may also result in increased morbidity and mortality due to the increased bleeding risk. Presentation is often non-specific—of all the patients with clinically suspected PE, only around one-third are eventually diagnosed with thromboembolic disease.

Clinical (Pre-Test) Probability Estimate and D-Dimer Testing There are no reliable bedside tests available to diagnose PE. Electrocardiogram (ECG) and measurement of arterial pO2 are not diagnostic for PE, and are more useful in suggesting other causes for the patient’s symptoms, e.g. myocardial infarction. A D-dimer blood test (a degradation product of fibrin in the process of fibrinolysis; an increased D-dimer suggests activation of the coagulation and fibrinolytic systems) has a very high negative predictive value (NPV). However, the NPV is not 100%, meaning that a negative test cannot exclude PE. On the other hand, the positive predictive value of a positive D-dimer test is low as there are many conditions in which the thrombolytic system is activated, such as inflammation, recent surgery, trauma, bleeding, neoplastic disease, during pregnancy, and in hospitalised patients. In addition, the specificity of the D-dimer test decreases with increasing age. A recent meta-analysis has shown that using age-adjusted cut-off values of the test (age × 10 µg/L above the age of 50 years) instead of the standard value of 500 µg/L increases the specificity while the sensitivity is maintained. The D-dimer test generally is performed in combination with a clinical probability estimate (CPE), usually a clinical prediction rule. Several clinical prediction rules, such as the Wells and revised Geneva rules, have been developed, of which, the original and the simplified Wells rule is the best evaluated and most frequently used (Table 16.6).

TABLE 16.6  Wells Clinical Prediction Rule

Previous PE or DVT Heart rate >100 beats/min Surgery or immobilisation 4

≤1 >1

DVT, Deep venous thrombosis; PE, pulmonary embolism.

412

SECTION A  The Chest and Cardiovascular System

However, the performance of these clinical prediction rules varies between different patient groups and in different clinical settings. The combination of a negative D-dimer test and a low or intermediate CPE reliably rules out PE and no further testing is indicated. This combination is found in up to 51% of outpatients or those seen via the emergency ward. In all other circumstances (high clinical suspicion and/or positive D-dimer testing) further diagnostic work-up using imaging is warranted. There is insufficient evidence that it is safe to rule out PE in pregnant patients based on D-dimer testing and CPE alone as this strategy is not validated in this patient category.

Imaging Findings

Plain Chest Radiography The chest x-ray may be normal (up to 40% of patients with PE) or show non-specific findings, even in extensive PE. The chest x-ray is performed not to diagnose PE but to exclude other causes of the symptoms, such as pneumonia, pleuritis or pneumothorax. Several signs related to PE (and therefore suggestive of the diagnosis) have been described. However, they are infrequently present and non-specific: i.e. its presence does not confirm the diagnosis of PE. The most important ones are: • Hampton hump. This is a pleural-based, wedge-shaped opacity with the apex of the triangle pointing towards the occluding vessel/hilum. It is typically not seen in the first 24 hours after the embolus has lodged in the pulmonary artery and it represents a parenchymal infarction (Fig. 16.22). This condition may take from 3–5 weeks up to months to resolve, and a band-like opacity, due to scarring or focal pleural thickening, may remain (see Fig. 16.22C). The opacity may not always be triangular as the infarction may be surrounded by haemorrhage. If the infarcted area becomes secondarily infected, cavitation or abscess formation may occur. The latter may also be caused by septic emboli. These opacities may not only be seen in case of infarction but can also be the result of oedema and haemorrhage. The latter are usually found in the lower lobes, from 12 hours to several days after the thromboembolic event, and show relatively rapid resolution (7–10 days). • Westermark sign, defined by a hyperlucent area with decreased vascularity due to oligaemia of the involved part of the lung. Although this finding is not specific for PE, it should be considered, especially if newly found.

A

B

• ‘Knuckle’ or ‘sausage’ sign, describing a dilatation of a central pulmonary artery due to occlusion by the embolus with collapse or constriction of the distal arteries, resulting in an abrupt tapering of these arteries. Other secondary findings that may be present are: plate-like atelectasis, (haemorrhagic) pleural effusion, and an elevation of the diaphragm, either due to pleuritic pain or as a result of decreased pulmonary compliance. If PE is severe, signs of RV failure may be encountered, such as dilatation of the right heart, the superior vena cava and the azygos vein.

Transthoracic or Transoesophageal Ultrasound (Echocardiography) This examination is the first imaging method of choice in cardiorespiratory-unstable patients, in whom massive PE is suspected. The RV function can be assessed and central pulmonary emboli be detected. As this technique has a low sensitivity for peripheral emboli, its use is not recommended in stable patients. The advantage of this technique is the assessment can be performed at the bedside and can diagnose other cardiovascular diseases that may explain the patient’s symptoms, such as cardiac tamponade or acute myocardial infarction.

Conventional Pulmonary Angiography Until recently, pulmonary angiography was considered the gold standard for the diagnosis of PE. For several reasons (e.g. costs, limited availability and invasiveness of the procedure), it has not gained general acceptance. In experienced hands it remains a valuable examination with a low complication rate (mortality 1% and non-fatal complications 95%) and the patient can be treated for PE without further investigation. All other V/Q results, i.e. subsegmental perfusion defects or ventilation defects matching the segmental perfusion defects (occurring in 60%–70% of the V/Q-scans), are called non-diagnostic or indeterminate, as other diseases, such as asthma and chronic obstructive pulmonary disease, may also cause these defects. Further investigation therefore is indicated,

Fig. 16.23  Static perfusion scintigram in six standard views (anterior, posterior and left/right posterior oblique) in a patient with suspected acute pulmonary embolism reveals a wedge-shaped segmental perfusion defect in the left lung. Chest x-ray was normal (not shown).

414

SECTION A  The Chest and Cardiovascular System

as only up to 30%–40% of the patients with an indeterminate result eventually have PE. Recently, it was proposed to use, instead of the probability-based interpretation of V/Q scinitigraphy, an alternative reporting system similar to CTPA examinations and more clinically useful. V/Q scintigraphy results are reported as either PE positive or PE negative, with only a minority designated as indeterminate or nondiagnostic for PE. As with CT, if the pre-test probability is discordant with the V/Q result, further testing should be considered. Perfusion scintigraphy is a good initial imaging investigation in patients with a normal chest x-ray and no history of pulmonary disease; in this setting a normal result or a segmental perfusion defect is a reliable finding that leads to few non-diagnostic results. Perfusion scintigraphy should also be considered when a relative contraindication for CT (see below) exists, such as severe renal impairment.

Computed Tomography Pulmonary Angiography In 1992 the first publication on the use of CTPA for the diagnosis of acute PE appeared. Since then, CTPA has become the investigation of choice in the work-up of patients with suspected PE. Its preference is due to the continuous improvement of the CT technique, resulting in substantial improvement in the acquisition speed, spatial resolution, image quality and, most importantly, diagnostic accuracy. Together with the broad availability, low cost and minimal invasiveness of this technique have led to a broad acceptance of it in clinical practice. With modern MDCT reported sensitivities and specificities of 83%–100% and 89%–97% are comparable to those of the former gold standard of PE diagnosis, invasive pulmonary angiography of 98% and 97%, respectively. This means that with a qualitatively good CT, PE can be ruled out safely, at least in (out)patients without a high clinical probability of PE and further testing for PE is not warranted. The advantage of CT (and to a lesser extent MRI) over other imaging techniques is the direct visualisation of the emboli as well as the other structures of the chest, including the lung parenchyma, mediastinum and chest wall, resulting in an additional or alternative diagnosis (e.g. pneumonia, pleuritis, aortic dissection, pneumothorax and lung tumours) in a significant proportion of patients with clinical symptoms suggestive of PE. In patients with PE, CTPA can also provide parameters considered to be related to clinical outcome, such as right ventricular dysfunction (RVD), the total amount of thrombus present (although contradicting results for the latter are found) and, if available, pulmonary perfusion (see below). A RV/left ventricular (LV) diameter ratio greater than 1.0 measured on standard axial views has been shown to be directly correlated with RVD and to predict adverse outcome and early death (Fig. 16.24). Quantitative assessment of the thrombus load, for which several scoring systems have been proposed, is very time-consuming, although this may be overcome with the use of computer-assisted detection (CAD) software in the future. With the improvement of the CT technique, very small PEs are frequently detected that are at or even beyond the subsegmental level (Fig. 16.25). At the moment, it is uncertain whether the risk of having a solitary very small PE outweighs the risk of anticoagulant treatment (minor to severe bleeding). Currently, treatment of a solitary small PE is considered indicated if patients have limited cardiopulmonary reserve, coexistent DVT or recurrent small PEs. In a subgroup of patients, especially those who have a good cardiopulmonary reserve, limited timeframe of risk factors and in whom DVT is excluded, treatment for a shorter time period with anticoagulants or even withholding treatment may be considered. Computed tomography pulmonary angiography protocol. Data acquisition is performed during one breath-hold, preferably at total lung capacity. The standard CT parameters are 100–140 kV, depending

Fig. 16.24  Significant dilatation of the right ventricle (blue arrow) and right atrium as compared with the left ventricle (yellow arrow) in a patient with massive pulmonary embolism.

A

B Fig. 16.25  Small clots (arrows) in a subsegmental branch (A) and subsubsegmental (B) branch of the right lower lobe pulmonary artery.

on patient habitus. There is a general trend to decrease dose by lowering the kV in slim patients to 70– 80 kV. The additional advantage of choosing lower kV is the increase in vascular enhancement (in HU) due to increased absorption of iodine at lower kV. Dose modulation techniques should always be used to further reduce the dose. Regardless of the

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

415

number of detector rows, between 300 and 450 0.9- to 1.0-mm-thick slices using overlapping reconstruction should be obtained. The injection protocol has to provide a constant and high degree of pulmonary arterial enhancement during the complete data acquisition, which has become a challenge with the very short acquisition times. For an adequate assessment, an attenuation of at least 300 to 350 HU (i.e. 250 to 300 HU net contrast enhancement) in the pulmonary arteries is considered optimal. Suboptimal vascular opacification is one of the major causes of non-diagnostic images, and it compounds the effect of concomitant problems such as partial volume or movement artefacts. Combined protocols: one-stop-shop procedure.  CT venography has been considered as a part of a one-stop-shop procedure in order to diagnose VTE. After administering one bolus of contrast medium, first the pulmonary arteries are investigated followed by additional late-phase imaging of the deep venous system from the calves up to the inferior vena cava to detect DVT. Although this combined procedure is feasible and has the advantage of detecting DVT in pelvic veins and inferior vena cava (IVC), which is not possible with CUS, studies have shown that in comparison with CTPA alone, this combined technique results in only limited increase in sensitivity with a comparable specificity. A major drawback of CT venography is the significant increase of radiation, which at this moment does not justify its routine use in patients with suspected PE (Fig. 16.26). Alternatively, CTA with ECG gating can be performed in patients presenting with acute chest pain without significant increase in radiation dose. During one data acquisition, information can be obtained on the most important vascular diseases causing acute chest pain: acute coronary syndrome, aortic dissection and acute PE. The use of dual-source CT or systems with high numbers of detector rows may overcome the initial limitations of ECG-gated CTPA, providing faster acquisition times, better image quality in patients with abnormal cardiac rhythms, and lower radiation dose. The downside of such a protocol is the increase in complexity both for the technician and the reader, with an increase in post-processing and interpretation time. Computed tomography pulmonary angiography during pregnancy.  During pregnancy and puerperium, the incidence of VTE is two- to fourfold higher and is one of the most important causes of maternal mortality. As diagnosing DVT in patients with suspected PE justifies PE treatment, leg ultrasound is considered the first diagnostic test of choice as no radiation is used, at least if signs or symptoms of DVT are present. Because of the concern about radiation, there is little agreement about optimal imaging during pregnancy when leg ultrasound

is normal. Either CTPA or perfusion scintigraphy can be obtained, the latter being useful if the chest radiograph is normal. As an alternative, MRI has been proposed (see below). At any time during pregnancy (including the first 3 months) the radiation dose to the unborn child delivered by either V/Q scintigraphy or CTA is considered negligible, and the risks of a potential fatal ending due to undiagnosed PE are substantial. If CTA is performed, the CT protocol should be adapted to reduce radiation dose and to the hypercirculatory state of pregnant women to limit the number of inconclusive examinations. Furthermore, it is advised that thyroid function should be checked in the first week after birth, as the iodine in the CT contrast medium may decrease thyroid function. In young women, either pregnant or not, the radiation exposure to the radiation-sensitive breast tissue (in the order of 10 to 70 mGy) is an important issue, and to a lesser extent the exposure to the lung. Perfusion scintigraphy produces a lower breast dose (1 mGy). During pregnancy this would be at the expense of a slightly increased dose for the fetus as the radiopharmaceutical agent is excreted by the kidneys and may give radiation from the bladder to the neighbouring uterus. Flushing the bladder with saline via a catheter has been proposed to reduce the exposure. Computed tomography pulmonary angiography assessment.  The diagnosis of PE on CTA is based on the direct visualisation of the thrombus that may result in a partial filling defect or complete obstruction of the pulmonary artery (Figs 16.27 and 16.28). Secondary findings of acute PE may help to point attention to a certain area, such as the presence of pulmonary infarcts. Signs of acute pulmonary hypertension can also be present, usually as a result of extensive obstruction of the arterial bed, such as dilatation of the pulmonary trunk and/or right heart. The direct and indirect signs of both acute and chronic PE are listed in Table 16.7. Artefacts and pitfalls frequently occur and one should be aware of them to avoid false-positive (see Fig. 16.28) and false-negative results (Fig. 16.29). They are listed in Table 16.8. Computed tomography perfusion.  Introduction of the dual-energy CT (DECT) technique has led to several potential advantages, including novel image interpretation concepts without an increase in radiation dose. The DECT technique uses the different absorption characteristics of iodine at different voltage levels. Two data sets are (nearly) simultaneously acquired at different tube voltages, or using a relatively new

Fig. 16.26  Computed tomography venography demonstrating thrombus in the left popliteal vein (arrows) with contrast enhancement of the vessel wall (indirect sign), not seen on the right side.

Fig. 16.27  Patient with massive pulmonary embolism (asterisks). Eccentric thrombi with acute angles to the vessel wall as well as completely occluding thrombi are present.

416

SECTION A  The Chest and Cardiovascular System

A

B

C

Fig. 16.28  (A) Central thrombus (yellow arrow) with smooth margins, centrally located and acute angles with the vessel wall, and lymphadenopathy (red arrows). Coronal reconstructions (B and C) may aid in the differentiation between endoluminal clots (yellow arrows) and lymphadenopathy (red arrows). Note the central position of the clot surrounded by contrast medium.

TABLE 16.7  Direct and Indirect Signs of Acute and Chronic PE at CTA Acute

Chronic

Direct features

= or ↑ (expansion by impacted thrombus) Central (on by impacted thrombus) (distal) d Complete obstruction with concave filling defect, or Eccentric, acute angle with vessel wall

  Vessel diameter

Sharp margin No recanalisation

  Filling defect

No calcification 25–40 HU Local hypoperfusion (oligaemia) Wedge-shaped, pleural-based consolidations, or ground-glass opacity with fine reticular changes (pulmonary infarction with or without haemorrhage) Plate-like atelectasis Linear densities Volume loss Small pleural effusion Thin RV wall RV dilatation, bowing of interventricular septum (severe PE) Tricuspid regurgitation (severe PE) Dilatation SVC/IVC and azygos vein Backflow of contrast agent into IVC and hepatic veins due to increased cardiac pressure No dilatation of systemic arteries No tortuosity of pulmonary arteries No vessel wall calcification

↓, 5–40 H-stenotic dilatation or aneurysms Complete (abrupt narrowing or cut-off), or Eccentric, obtuse angle with vessel wall, or Intraluminal filling defect with the morphology of acute PE present for >3 months. Often irregular margin An abrupt tapering of a vessel, usually the consequence of recanalisation Calcification 65–105 HU Intimal irregularities, webs, bands Mosaic patterna Small subpleural scars or irregular lines, cavitaties Cylindrical bronchial dilatation in the involved areas of vascular obstruction (less frequent)

Indirect features  Parenchyma

 Pleura  Cardiac  Vessels

a

Focal pleural thickening Thick RV wall, wall thickness > 4 mm (CTEPH) RV enlargement (CTEPH), often with concomitant RA dilatation Tricuspid regurgitation (CTEPH) Backflow of contrast into IVC and hepatic veins Dilatation of the pulmonary trunk (>29 mm) central pulmonary arteries (CTEPH) Hypertrophy of bronchial arteries and non-bronchial systemic collateral arteries Tortuous pulmonary arteries Vessel wall calcification

Mosaic pattern is defined as sharply demarcated areas of relatively increased and decreased lung attenuation, in the case of CTEPH due to hypoperfusion of the involved parenchymal bed and redistribution of blood flow to the non-involved areas. CTA, Computed tomography angiography; CTEPH, chronic thromboembolic pulmonary hypertension; PE, pulmonary embolism; RA, right atrium; RV, right ventricle; SVC/IVC, superior/inferior vena cava.

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

417

TABLE 16.8  Causes of False-Positive and False-Negative Computed Tomography

Angiography Results in the Assessment of Pulmonary Embolism Causes of False-Positive Results • Artefacts due to: • Pulsation (mainly lingula and left lower lobe) • Volume averaging • Flow artefacts (non- or only partially opacified arteries) • Lymph nodes (esp. at the right hilum) and lymphatic tissue • Slow flow (due to increased resistance such as in presence of atelectasis or pleural fluid) • Tumours (e.g. pulmonary artery sarcoma) • Complete obstruction of a vessel without visualisation of the proximal part of the thrombus (non-specific) • Mucus-filled bronchi • Pulmonary veins

Fig. 16.29  Computed Tomography. Beam hardening artefact in the right upper lobe pulmonary artery (yellow arrow) due to dense contrast medium in the superior caval vein. Thrombus (blue arrow) more peripherally. Red arrows, non-opacified pulmonary veins.

technique with a single tube and multilayer detectors. Fusion of the two data sets results in a ‘standard’ CTA. Subtraction of the lower kV data set from the higher kV data can produce colour-coded CT regional iodine density maps of the lung which act as a surrogate for perfusion (Fig. 16.30). The quantitative assessment of perfusion defects of the lung parenchyma has been found to be an important determinant for patient outcome. In addition, the presence of perfusion defects may help in the detection of PE and therefore increase sensitivity, which can be beneficial, especially for less experienced readers. However, this technique also has several pitfalls and artefacts, caused by underlying pulmonary disease such as emphysema, cardiac motion or diaphragm movement or beam hardening effects caused by dense contrast material in the thoracic veins. Another post-processing technique possible with DECT is virtual monoenergetic imaging, which is the result of mixing the high- and low-energy data sets in a particular ratio that results in images at a specified virtual monoenergetic photon energy. The potential of this reconstruction technique is optimisation of image contrast, improvement

Causes of False-Negative Results • Artefacts due to: • Motion (breathing, pulsation) • High contrast (beam hardening) • Suboptimal enhancement (e.g. protocol related, intracardiac shunts) • Inadequate image quality or reconstruction parameters (image noise, window settings) • Isolated (sub)subsegmental pulmonary embolism

A

B Fig. 16.30  Coronal (A) and sagittal (B) perfusion maps using dual-energy computed tomography technique fused with computed tomography angiography showing multiple thrombi with perfusion defects and a consolidation (asterisk) due to infarction (Hampton hump).

418

SECTION A  The Chest and Cardiovascular System

A

Fig. 16.32  Patient With Bilateral Perfusion Defects on MRI Due to Lobar Pulmonary Embolism. (Courtesy of M.P. Revel, Georges Pompidou European University Hospital, Paris, France.) B Fig. 16.31  (A) Post-gadolinium three-dimensional time-of-flight (3D TOF) magnetic resonance angiography image (slice thickness is 2.4 mm) demonstrates right lower lobe segmental pulmonary embolism. (B) Corresponding unenhanced steady-state free-precession (SSFP) sequence (FIESTA) image. (Courtesy of M.P. Revel, Georges Pompidou European University Hospital, Paris, France.)

of the contrast-to-noise ratio and reduction of beam hardening. Both virtual monoenergetic and perfusion imaging may improve reader confidence and diagnostic accuracy. Although these techniques seem promising, their potential benefit for diagnosis, prognosis and therapy monitoring still needs to be determined.

Magnetic Resonance Imaging MRI is an attractive alternative to CTA as no ionising radiation is used and modern MRI contrast media yield less risk for the development of contrast-related nephropathy and contrast agent reactions as compared with iodinated contrast media. Due to the long data acquisition and limited availability, especially in the acute setting, MRI is still not widely performed. On the other hand, recent improvements with respect to spatial resolution and new sequences with shorter acquisition times make MRA a feasible technique for the acute setting. Various MRI techniques are available (e.g. unenhanced and post-gadolinium angiography sequences, with or without MR perfusion sequences) (Fig. 16.31). According to the literature, the accuracy of MRA is comparable to CTPA for central and segmental pulmonary arteries, but still limited for PE in the peripheral pulmonary vessels. Reported overall sensitivity compared with MDCT and V/Q is 78%–85% and specificity 99%–100%. However, the rate of inconclusiveness ranges from 25% to 30% and sensitivity drops down to 21%–33% for subsegmental PE. As with the newest CT techniques, MRI offers the option to obtain additional functional information such as on cardiac function and lung perfusion (Fig. 16.32).

At the moment, the technique is, however, less robust as compared with CTA and examination times remain long (up to 10–25 minutes for a combined technique protocol). However, further improvement in technique and acquisition speed is to be expected and, together with more experience, may result in better accuracies and a decrease in inconclusive results. As the specificity of MRI is high, it should be considered as a viable alternative to CTPA, especially in patients with (relative) contraindications to CTA such as young and/or pregnant women.

Diagnostic Strategies The most straightforward diagnostic approach as recommended in the 2014 ESC guidelines for patient with suspected acute PE is presented in Fig. 16.27. This diagnostic algorithm is based on integrating clinical data and laboratory and imaging tests. It should be noted that the choice of technique is also dependent on local availability and experience, user preference, and last but not least the patient studied.

CHRONIC PULMONARY THROMBOEMBOLISM PE becomes chronic if the clots in the pulmonary arteries do not resolve adequately. This may result in a progressively increased pulmonary arterial blood pressure and CTEPH, which is a serious and life-threatening complication, and if it occurs it is usually in the first 2 years after the thromboembolic event. The estimated incidence of CTEPH after an episode of acute PE varies, but figures as high as 4% have been described. On the other hand, about half of the patients with CTEPH do not have a clinical history of acute PE. Chronic PE is one of the few causes of pulmonary hypertension that effectively can be treated by surgical resection of the thrombus (pulmonary thromboendarterectomy) if the thrombi are localised in central vessels, up to the proximal segmental arteries. Percutaneous balloon angioplasty is emerging as a less invasive alternative that has been shown to be feasible and effective for a subgroup of patients, but its exact role in the treatment of CTEPH is still to be determined. Medical treatment is recommended in patients with persistent or recurrent CTEPH after surgical treatment, or who are considered inoperable.

CHAPTER 16  Pulmonary Circulation and Pulmonary Thromboembolism

Fig. 16.33  Chronic Thromboembolic Pulmonary Hypertension Shown by CT. Eccentric thrombus with obtuse angles in the right pulmonary artery and a web in the left lower lobe pulmonary artery.

419

Fig. 16.34  Chronic Thromboembolic Pulmonary Hypertension. CT shows Dilatation of the right side of the heart with thickening of the ventricular wall.

SUMMARY BOX: Pulmonary Thromboembolism  CTEPH should be ruled out in patients who were treated for acute PE and suffer from persistent dyspnoea. In addition, chronic PE should be evaluated in every patient with pulmonary hypertension with a known history of PE or in whom the cause is unclear. According to international guidelines, V/Q scintigraphy is the imaging method of choice to rule out suspected CTEPH, as a normal perfusion scintigram confidently excludes the disease whereas multiple, bilateral segmental perfusion defects are suggestive but not specific for chronic PE. DECT and perfusion MRI both permit maps of regional perfusion to be generated and show promise as a ‘one-stop shop’ to evaluate both the pulmonary arteries and parenchymal perfusion. Currently, pulmonary angiograpy is still considered the gold standard for the diagnosis of distal chronic PE, but as with acute PE, CTA has established a prominent role in many centres in the work-up of patients with pulmonary hypertension in general and suspected chronic PE in particular. CTA may effectively diagnose or rule out other causes, such as parenchymal disease, and diagnose and assess the location and extent of chronic thrombi. Although, with MRI, important morphological and functional cardiac information can be obtained, it is not performed routinely, because of the known limitations: cost, availability and examination time. As for acute PE, at CT both direct (Figs 16.33 and 16.36) and indirect (see Figs 16.16, 16.34, 16.35, 16.36) signs of chronic PE—which are partly related to the presence of pulmonary hypertension—have been described. The signs are summarised in Table 16.7. It should be noted that in patients with recurrent PE both chronic and acute PE can coexist (see Fig. 16.36).

• Acute pulmonary embolism (PE) is the third most common acute cardiovascular disease and consequently, a major cause of mortality, morbidity, and hospitalisation. • In up to 5% of patients with acute PE, the emboli do not completely resolve and become chronic. Chronic PE may lead to chronic thromboembolic pulmonary hypertension (CTEPH). • Clinical signs and symptoms of both acute and chronic PE are non-specific. Therefore, when there is a clinical suspicion of acute/chronic PE, further objective testing is warranted. • In acute PE, clinical prediction rules help to categorise patients into low-, moderate- and high-probability risk groups. PE can be safely excluded when there is a low or moderate pre-test probability in combination with a negative D-dimer test. • Multidetector computed tomography pulmonary angiography (CTPA) is the first imaging test of choice for acute PE. If positive for PE, right ventricular dilatation (RV/LV ratio ≥ 1) should be assessed as this is an independent predictor for an adverse clinical outcome. • Lung scintigraphy is most useful in outpatients suspected of having acute PE without known cardiopulmonary disease and normal chest x-ray. Especially in younger patients, lung scintigraphy should be considered as first imaging test of choice. • Proximal deep vein thrombosis at compression ultrasound (CUS) of the legs in patients with suspected acute PE is sufficient to warrant anticoagulant treatment without further testing. • V/Q scintigraphy is the screening method of choice in suspected CTEPH. The diagnosis of CTEPH is based on specific signs on either CTPA, magnetic resonance imaging or conventional pulmonary angiography.

CONCLUSION The investigation of changes in pulmonary vascular physiology has many facets. The initial clinical assessment may significantly influence further investigations. Chest radiography is still the primary method for assessing effects of PV hypertension. CTA is the first imaging investigation in the diagnostic work-up of patients with suspected PE. In addition, CT may provide an alternative diagnosis in a significant percentage of patients in which PE is excluded. Pulmonary hypertension is a very serious disease that can be caused by a number of different diseases requiring

different treatment and having varying prognosis. CT and MRI play a major role in determining the different types of PAH.

ACKNOWLEDGEMENTS This chapter is based on the previous versions prepared by Dr. Paras Dalal, Michael B. Rubens and Cornelia M. Schaefer-Prokop. Their contribution and material are gratefully acknowledged.

420

SECTION A  The Chest and Cardiovascular System

FURTHER READING

Fig. 16.35  Chronic Thromboembolic Pulmonary Hypertension. Tortuous pulmonary vessels and constrictions. CT: Coronal maximum intensity projection.

Fig. 16.36  Coronal CT Reconstruction. Patient with both acute (yellow arrow) and chronic pulmonary embolism (blue arrows).

Albrecht, M.H., Bickford, M.W., Nance, J.W., Jr., et al., 2017. State-of-the-art pulmonary CT angiography for acute pulmonary embolism. AJR Am. J. Roentgenol. 208, 495–504. Aluja Jaramillo, F., Gutierrez, F.R., Duti Telli, F.G., et al., 2018. Approach to pulmonary hypertension: from CT to clinical diagnosis. Radiographics 38, 357–373. Benson, D.G., Schiebler, M.L., Repplinger, M.D., et al., 2017. Contrast-enhanced pulmonary MRA for the primary diagnosis of pulmonary embolism: current state of the art and future directions. Br. J. Radiol. 26, 145–151. Castañer, E., Gallardo, X., Ballesteros, E., et al., 2009. CT diagnosis of chronic pulmonary thromboembolism. Radiographics 29, 31–50. Devaraj, A., Hansell, D.M., 2009. Computed tomography signs of pulmonary hypertension: old and new observations. Clin. Radiol. 64, 751–764. Gali7, N., Humbert, M., Vachiery, J.L., et al., 2015. 2015 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension: the Joint Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS): endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT). Eur. Respir. J. 46, 903–975. Gluecker, T., Capasso, P., Schnyder, P., et al., 1999. Clinical and radiologic features of pulmonary edema. Radiographics 19, 1507–1531. Konstantinides, S.V., Torbicki, A., Agnelli, G., et al., 2014. 2014 ESC guidelines on the diagnosis and management of acute pulmonary embolism. Task Force for the Diagnosis and Management of Acute Pulmonary Embolism of the European Society of Cardiology (ESC). Endorsed by the European Respiratory Society (ERS). Eur. Heart J. 35, 3033–3069. Metter, D., Tulchinsky, M., Freeman, L.M., 2017. Current status of ventilation-perfusion scintigraphy for suspected pulmonary embolism. AJR Am. J. Roentgenol. 208, 489–494. Ruggiero, A., Screaton, N.J., 2017. Imaging of acute and chronic thromboembolic disease: state of the art. Clin. Radiol. 72, 375–388. Schulman, S., Ageno, W., Konstantinides, S.V., 2017. Venous thromboembolism: past, present and future. Thromb. Haemost. 117, 1219–1229. Simonneau, G., Gatzoulis, M.A., Adatia, I., et al., 2013. Updated clinical classification of pulmonary hypertension. J. Am. Coll. Cardiol. 62, D34–D41. Wittram, C.1., Kalra, M.K., Maher, M.M., et al., 2006. Acute and chronic pulmonary emboli: angiography-CT correlation. AJR Am. J. Roentgenol. 186, S421–S429.

15  Ischaemic Heart Disease Jan Bogaert, Rolf Symons

CHAPTER OUTLINE Introduction, 368 Pathophysiology of Ischaemic Heart Disease, 368 Coronary Artery Imaging, 370 Functional Imaging, 375 Stress Imaging, 379 Myocardial Infarct Imaging, 383 Myocardial Viability Imaging, 389

INTRODUCTION Ischaemic heart disease (IHD) is a complex, heterogeneous and incompletely understood disease that is usually caused by underlying coronary artery disease (CAD). Worldwide, it is the single most common cause of death and its frequency is increasing. Although the mortality associated with IHD has declined, due to therapeutic improvements and prevention campaigns reducing the incidence of fatal and non-fatal myocardial infarction, the prevalence of IHD will continue to increase. For instance, in patients with STEMI (i.e. ST-segment elevation myocardial infarction), mortality remains substantial, with in-hospital mortality and 1-year mortality rates up to 10%. Moreover, survivors of a first myocardial infarction are thought to die of IHD at later ages due to heart failure and late cardiac deaths. Other contributing factors are an increasing prevalence of type 2 diabetes, physical inactivity and obesity. Although catheter-based coronary angiography is the diagnostic procedure of choice to diagnose and treat CAD, non-invasive cardiac imaging—that is, echocardiography, nuclear medicine, cardiac computerised tomography (CCT) and cardiovascular magnetic resonance (CMR)—are key imaging techniques in unravelling the intricate relationship between CAD and IHD, in preclinical detection of CAD and in the assessment of patient prognosis.

PATHOPHYSIOLOGY OF ISCHAEMIC HEART DISEASE CAD, that is the process of atherosclerotic plaque formation, is the usual cause of IHD. Symptoms of myocardial ischaemia occur when the coronary blood flow is significantly impaired. This may happen when the coronary artery lumen is slowly and progressively impinged by an evolving atherosclerotic plaque (chronic stable plaque), or when a coronary artery plaque ruptures—or, less frequently, plaque erosion— and a thrombus is formed with a sudden occlusion of the lumen, causing an acute coronary syndrome (Fig. 15.1). Moreover, less common—or

368

Imaging of Complications Related to Ischaemic Heart Disease, 393 Prognosis Assessment in Ischaemic Heart Disease, 394 Role of Conventional Chest Radiography in Ischaemic Heart Disease, 394 Differential Diagnosis in Ischaemic Heart Disease, 394

superimposed—causes of myocardial ischaemia are coronary artery spasm and microcirculatory dysfunction. Acute coronary occlusion triggers in the myocardial perfusion territory distal to the occlusion, that is the jeopardised myocardium or myocardium at risk, an ischaemic cascade starting with metabolic disturbances followed by regional dysfunction, electrocardiographic (ECG) changes and, finally, onset of anginal symptoms (Fig. 15.2). Systolic contraction typically ceases within seconds after coronary occlusion. After approximately 20 to 30 minutes of sustained ischaemia, irreversible myocardial damage (i.e. myocardial infarction) occurs with myocardial cell swelling and apoptosis, ultimately leading to myocyte necrosis. Cellular necrosis always initiates at the endocardial side of the myocardium with the lateral boundaries of infarction closely corresponding to the myocardium at risk, and follows a transmural wave front progression taking 3 to 6 hours to reach the subepicardium (Fig. 15.3). As a consequence, the amount of necrosis is mainly determined by extent of myocardium at risk and degree of transmural progression. Current therapeutic strategies aim to timely restore coronary flow by percutaneous coronary intervention or thrombolysis, which will stop the transmural progression of necrosis, and salvages the ischaemic, but still viable, myocardium; however, in spite of the beneficial effects of reperfusion, the process of cell death may continue during the first hours of reperfusion, a phenomenon called ‘myocardial reperfusion injury’. This occurs in the infarct core and is characterised by a lack of reperfusion at myocardial capillary level (i.e. microvascular obstruction or no-reflow phenomenon) despite an effective recanalisation of the infarct-related artery. Myocardial infarction can be recognised by clinical features, including ECG findings, elevated values of biochemical markers (biomarkers) of myocardial necrosis, and by imaging, or may be defined by pathology. Infarctions are usually classified by (a) location (anterior—inferior— lateral); (b) size (focal necrosis, small (30% of LV myocardium)); and (c) temporally as evolving (28 days). Although myocardial infarctions usually affect the left ventricle,

CHAPTER 15  Ischaemic Heart Disease

369

Fig. 15.1  Electron Microscopy of a Fresh Thrombus Extracted From Four Patients With an Acute Myocardial Infarction (3000x Magnification). Presence of red blood cells and platelets entrapped in the fibrin clot (left upper and lower panel) and a variable amount of white bloods (right lower panel). (Courtesy Dr. J Zalewski, Cracow, Poland.)

Anginal chest pain ECG abnormalities Regional wall motion abnormalities Contractility abnormalities Relaxation disturbances Reversible myocardial changes

Metabolic abnormalities (ATP/PCr)

Irreversible myocardial damage

20–30 min Coronary artery occlusion

Fig. 15.2  The Ischaemic Cascade. Series of events occurring following a coronary artery occlusion. While changes are initially reversible, after 20 to 30 minutes irreversible myocardial damage occurs.

extension toward the right ventricle may occur. Isolated right ventricular infarctions, conversely, are seldom. In the days, weeks and months following the acute event, the heart undergoes a remodelling with changes in ventricular size, shape and function, with changes not limited to the infarcted myocardium but involving the remote myocardium as well.

In an early phase, tissue oedema, haemorrhage and acute inflammation lead to an expansion of the infarct size. This may lead to weakening of the myocardial wall, eventually resulting in an early ventricular rupture or rapid evolution towards an aneurysm. During infarct healing, an opposite phenomenon occurs, with a progressive replacement of the necrotic myocardium by a collagen-rich fibrous scar causing a thinning of the affected myocardial wall. Depending on the extent of the infarct, part of the ventricle loses contractile force, and as a compensatory mechanism, the ventricle usually dilates to maintain stroke volume. This, however, is a potentially adverse event because it may trigger the evolution towards a dilated ischaemic cardiomyopathy and, ultimately, ischaemic heart failure (IHF). If the duration of coronary occlusion is brief (3– 6 h

Reperfusion

Consequences of ischaemia/ reperfusion

• Stunning • Preconditioning • Tissue viability (no necrosis)

• Subendocardial necrosis (salvage of outer layers)

• Necrosis extends into midmyocardium, subepicardium

• Near transmural infarction (no salvage of tissue but may lead to negative LV remodelling) Normal myocardium Myocardium at risk Necrosis

Fig. 15.3  Effects of Ischaemia and Reperfusion on Myocardial Tissue Viability and Necrosis. Studies in anaesthetised canine model of proximal coronary artery occlusion. (Adapted from Kloner RA, Jennings RB. Consequences of brief ischemia: stunning, preconditioning and their implications. Circulation 2001;104:2981–2989.)

typically present with moderate to severely reduced ventricular function, presence of several stenotic lesions on their coronary angiogram, and symptoms of heart failure. Because of the cost of the revascularisation procedure and the inherent risk related to these interventions, in particular when performed in patients with poor cardiac function, it is crucial to determine preoperatively the potential benefit of a revascularisation procedure. It should be emphasised that IHD patients may present with a mixture of different ischaemic substrates (ischaemic, stunned, hibernating, necrotic/scarred myocardium), urging for accurate myocardial tissue characterisation to choose the best therapeutic option.

CORONARY ARTERY IMAGING The process of atherosclerotic plaque formation usually involves the epicardial part of the coronary artery system. The diagnosis of CAD is made at conventional coronary angiography by impingement with narrowing or occlusion of the coronary artery (CA). Significant CAD is considered in the presence of a diameter stenosis of ≥70% in a major vessel or ≥50% in the left main, and usually results in referral for intervention. Although coronary angiography provides valuable information regarding the severity and length of stenosis, coronary artery (CA) occlusions, number of vessels affected, stenosis configuration (smooth, ulcerated), presence of thrombus, collateral vessels, CA anatomy and variants, this technique faces several limitations. Mild or non-stenotic CAD is not visualised and no or limited information is provided regarding the plaque composition or degree of vascular remodelling; thus a normal coronary angiogram does not exclude CAD, and a stenotic plaque may

be just the tip of the iceberg in some CAD patients. It should be emphasised that in most patients presenting with an acute myocardial infarction, it is caused by rupture of non- or minimally stenotic plaque. Because of its invasive nature, the need to administer iodinated contrast agent, and radiation issues, the use of conventional coronary angiography should be limited to symptomatic patients with high pre-test likelihood of obstructive CAD. Finally, the relationship between myocardial ischaemia and CAD is complex, and many patients fulfilling the criteria of significant CAD turn out not to have a flow-limiting stenosis when measuring the fractional flow reserve (FFR). While treatment should be reserved to patients with myocardial ischaemia, the oculo-stenotic reflex yields the risk of overuse of revascularisation. Other issues such as collateral vessels and the number and length of plaques should also be considered in decision-making. In a minority of patients presenting with typical angina and ST-segment depression on exercise testing, no obvious abnormalities are found on conventional coronary angiography. Diffuse subendocardial perfusion defects in these patients during stress perfusion imaging suggest that microvascular dysfunction rather than CAD is the causative mechanism of myocardial ischaemia. Thanks to rapid technological advances in the field of CCT and CMR in recent years, non-invasive coronary angiography has become a reality and these novel techniques are now becoming integrated into daily clinical care. Current state-of-the-art multidetector computed tomography (CT) (at least 64 or more slices) affords coronary artery imaging with sufficient spatial and temporal resolution for clinical use. A typical clinical examination consists of unenhanced CCT for detection and quantification of coronary calcium followed by contrast-enhanced

CHAPTER 15  Ischaemic Heart Disease CCT for coronary artery imaging, detection of coronary artery plaques and, to some extent, characterisation of the non-calcified plaques (Figs 15.4 and 15.5). The introduction of single x-ray source 256- and 320-slice volumetric scanners and dual x-ray source 2×128- and 2×192-slice CT systems has opened the door to ‘single heartbeat’ CCT in which the entire heart is imaged within one heartbeat. Modern, single x-ray-source CT systems require approximately 125 ms to image the entire heart because only 180 degrees of gantry rotation (plus the fan beam angle) are required to reconstruct a CT image (the remaining 180 degrees are a mirror image of the first 180 degrees). Dual x-ray-source CT systems have an even higher temporal resolution of up to 66 ms because only 90 degrees of gantry rotation are necessary (the two x-ray sources are mounted at a 90-degree angle and, therefore, only 90 degrees of gantry rotation are required to obtain a 180-degree view of the heart). Using these state-of-the-art CT systems, high-quality CCT can be acquired in heart rates of up to 65 to 70 beats per minute, which are obtainable in the vast majority of subjects by administering β-blockers (Fig. 15.6). Patients are optimally suited if they have a regular heart rate, a body mass index below 40 kg/m2 and a normal renal function. The examination is performed following intravenous injection of iodinated contrast agent (∼50–80 mL). Coronary vasodilatation can be achieved using

A

sublingual nitroglycerin administration. Single heartbeat CCT has not only improved image quality but has also significantly reduced CCT radiation dose. Where older CT systems used so-called retrospective gating and acquired images throughout the cardiac cycle of multiple heartbeats, single heartbeat CT systems use prospective gating, where images are only acquired during a small fraction of the cardiac cycle (typically during end-diastole when cardiac motion is minimal). Further reduction in radiation dose can be achieved with iterative reconstruction algorithms. Using state-of-the-art CCT hardware with iterative reconstruction, routine, high-quality CCT examinations can be acquired with a dose of approximately 1 mSv (Fig. 15.7). Coronary MR angiography is achieved using high-resolution imaging targeted to each coronary artery separately, or using a whole-heart approach. Images are acquired during repeated breath-holds or during free breathing using a respiratory trigger algorithm (Fig. 15.8). Sequences are available for luminal and for wall imaging. Despite enormous efforts of CMR experts worldwide, coronary MR angiography has still not been incorporated into daily clinical care, mainly because of long acquisition times, lack of reliable image quality, and the comparative ready availability and ease of use of CCT; however, appropriate indications for coronary MR angiography include imaging of congenital anomalies of the coronary arteries,

B

C

Fig. 15.4  Cardiac Computed Tomography of the Left Coronary Artery. Normal findings. Volume-rendered view (A); curved reformatted view (B and C). LAD, Left anterior descending coronary artery; LCx, left circumflex coronary artery; LM, left main coronary artery.

A

371

B

Fig. 15.5  Cardiac Computed Tomography of the Right Coronary Artery. Normal findings. Volume-rendered view (A); curved reformatted view (B). LAD, Left anterior descending coronary artery; RCA, right coronary artery.

372

SECTION A  The Chest and Cardiovascular System

A

B

C

D

Fig. 15.6  Coronary Artery Calcium (CAC) Score and Coronary Computed Tomography (CT) Angiography (CCTA) Images in a 49-Year-Old Male Recreational Cyclist With T-Wave Inversion in V1-V4 During Exercise Testing. Axial CAC score image (A) of the proximal and mid-left anterior descending coronary artery (LAD) does not show calcified plaque. Corresponding axial (B) and curved multiplanar reconstruction (C) CCTA images demonstrate an extensive non-calcified plaque with severe mid-LAD stenosis of 90% (arrows). Three-dimensional cinematic rendering (D) may facilitate understanding of complex coronary and cardiac anatomy. If no CCTA is performed, significant non-calcified lesions may be missed. With state-of-the-art CT machines, high-quality CCTA images can be acquired at relatively high heart rates at low radiation doses (heart rate, 66 beats/min; CAC sore radiation dose, 0.38 mSv; CCTA radiation dose, 0.82 mSv).

A

B

Fig. 15.7  Axial Coronary Computed Tomography (CT) Angiography (CCTA) Images in a 53-Year-Old Male Subject With Scleroderma and Atypical Chest Pain. Compared with filtered back projection (FBP) (A), ADvanced Modelled Iterative REconstruction (ADMIRE) (B) reduced image noise with 38%, resulting in excellent CCTA image quality at a radiation dose of 0.81 mSV. There is a small, calcified coronary plaque in the mid-left anterior descending coronary artery (LAD) with minimal less than 30% stenosis (arrows).

CHAPTER 15  Ischaemic Heart Disease

A

B

373

C

Fig. 15.8  Coronary Magnetic Resonance Angiography in a Healthy Volunteer. Curved reformatted views in three different views using Soap-Bubble Tool (Philips, Best, The Netherlands). (A) view showing RCA, LM, and LCx. (B) view showing LM and LAD. (C) view showing LM and LAD (perpendicular to B). Ao, Aorta; LAD, left anterior descending coronary artery; LCx, left circumflex coronary artery; LM, left main coronary artery; RCA, right coronary artery.

Fig. 15.9  Benign Form of Congenital Anomaly of the Left Circumflex Coronary Artery. Common origin of RCA and LCx coronary artery with retroaortic course of LCx. The abnormal retro-aortic course of the LCx can be diagnosed on the axial cardiac computed tomography images (arrows, A). The origin and course can be well appreciated on the virtually rendered views (B–D). ao, Aorta; LAD, left anterior descending coronary artery; LCx, left circumflex coronary artery; RCA, right coronary artery.

A

B

C

D

coronary artery imaging in (postoperative) patients with congenital heart disease and diagnosis (and follow-up) of patients with coronary aneurysms in Kawasaki disease (Figs 15.9–15.11). Novel approaches such as coronary artery positron emission tomography/magnetic resonance (PET/MR) imaging and PET/CT imaging may open perspectives towards plaque inflammation imaging in the coronary arteries. Calcification in the coronary arteries occurs, with exception of patients with advanced chronic kidney disease, almost exclusively in patients

with coronary atherosclerosis (Fig. 15.12). As the amount of coronary calcium roughly correlates to the atherosclerotic plaque extent, detection and quantification of coronary calcium is of interest for patient risk stratification. Most patients with an acute coronary syndrome (ACS) show coronary calcium, and the amount of calcium in these patients is substantially greater than in age- and gender-matched subjects without CAD; however, coronary calcification is not related to plaque stability/ instability and only weakly related to the severity of luminal stenosis

374

SECTION A  The Chest and Cardiovascular System

Fig. 15.10  Malignant Form of Congenital Anomaly of the Right Coronary Artery. Coronary magnetic resonance angiography. The right coronary artery (RCA) arises from the left aortic cusp and has a proximal inter-arterial course (between aortic root and pulmonary trunk). Ao, Aorta.

(Fig. 15.13). In young symptomatic patients, negative coronary calcium findings do not exclude coronary artery stenoses. The Agatston score, less frequently volume or mass scores, is used to quantify the amount of calcium. Reference data sets stratified by age and gender are available for interpreting coronary calcium scans. In the 2010 Appropriate Use Criteria for CCT, coronary calcium scoring was considered appropriate in asymptomatic patients without known CAD in low-risk patients with a family history of premature IHD, and to risk stratify intermediaterisk patients. Where previously CCT was typically not performed above a certain coronary artery calcium (CAC) score threshold (>400 or >1000 Agatston units) because the risk of a non-diagnostic CCT examination exposing the patient to a significant amount of radiation was too high, this strategy has been abandoned. Using state-of-the-art CCT equipment with iterative reconstruction algorithms, diagnostic CCT can be obtained in most patients with high CAC scores, yielding incremental prognostic value over CAC score alone. Coronary CT angiography offers high accuracy for the detection of and, especially, for ruling out significant CAD (Fig. 15.14). In two recent multicentre trials, a sensitivity of 95%–99%, specificity of 64%–83%, negative predictive value of 97%–99% and a positive predictive value of 64%–86% to identify patients with at least one coronary artery stenosis among individuals at low-to-intermediate risk for CAD. The lower positive predictive value is explained by the tendency to overestimate the degree of stenosis by coronary CT angiography due to partial volume or blooming artefacts because the spatial resolution of current CT systems is limited to 0.5 × 0.5 mm2 or 0.6 × 0.6 mm2. These systems

A

B

C

D

Fig. 15.11  Aneurysm of the Left Anterior Descending Coronary Artery in a Young Patient With Kawasaki Disease. Echocardiography (A and B) and cardiac computed tomography (CCT) (C and D). Presence of a fusiform aneurysm (*) in the proximal part of the left anterior descending coronary artery. The ultrasound and CCT findings match well. No evidence of thrombus formation in the aneurysm. ao, Aorta; LV, left ventricle.

CHAPTER 15  Ischaemic Heart Disease

375

Fig. 15.12  Visualisation of Calcified Non-Stenotic Plaques by Cardiac Computed Tomography. Presence of three calcified plaques in the right coronary artery (RCA) and left anterior descending coronary artery (LAD). The calcified plaques do not impinge on the coronary artery lumen but cause an enlargement of the external border (‘positive remodelling’). ao, Aorta.

use so-called energy-integrating detectors (EIDs) with an indirect conversion of x-ray photons into light photons by scintillator crystals and subsequent conversion of these light photons into an electric signal by light sensors. To prevent crossover of the light photons to adjacent pixels, optically isolating septa separate EID pixels; however, because these septa have a finite thickness, the geometric dose-efficiency of EIDs is reduced with smaller pixels. Therefore, there is a trade-off between EID spatial resolution and dose-efficiency. New photon-counting detectors (PCDs) may overcome this limitation. PCDs directly convert incident x-ray photons into an electric signal and measure their energy. This direct conversion eliminates the need for optically isolating septa and allows for higher spatial resolution CT without compromising dose-efficiency. Prototype PCD CT systems with a spatial resolution of 0.25 × 0.25 mm2 have demonstrated the potential of this technology to improve coronary stent visualisation and improve the diagnostic accuracy of CCT in coronary stenosis assessment (Fig. 15.15). Additionally, the lower sensitivity of PCDs to electronic noise may be used to improve CCT image quality at the same radiation dose or to maintain current image quality at reduced radiation doses. Coronary CT angiography performs best in symptomatic patients with a low-to-intermediate likelihood of CAD. In such patients, according to the 2010 Appropriate Use Criteria for CCT, coronary CT angiography was deemed appropriate to detect CAD in symptomatic patients without known heart disease, and also in those presenting with a clinical suspicion of ACS but having normal ECG and cardiac biomarkers, uninterpretable/ non-diagnostic ECG or equivocal biomarkers. Coronary CT angiography yields promise to determine and quantify the coronary plaque burden and, to a certain extent, to characterise the plaque composition. Using

intravascular ultrasound (IVUS) as reference technique, lipid-rich plaques yielded result in attenuation values between 11 and 99 Hounsfield units (HU) versus 77–121 HU for fibrous plaques. With similar intra-reader, inter-reader, and inter-scan reproducibility to IVUS, serial CCT may be used as a non-invasive alternative to IVUS for longitudinal plaque follow-up. Finally, CCT is excellent to depict anomalous coronary arteries and myocardial bridging.

FUNCTIONAL IMAGING Assessment of cardiac function is essential in the diagnostic work-up of IHD patients. For instance, in acute myocardial infarction patients, the dead myocardium ceases to contribute to the expulsion of blood, shifting the workload to the non-affected (‘remote’) myocardium. Although a compensatory increase in contractility in the remote myocardium has been described, the net result is usually impairment in ventricular performance. In patients clinically presenting with angina pectoris, the myocardium supplied by a haemodynamically significant stenosis may become dysfunctional under stress conditions. Also, in ischaemic heart failure patients, low-dose stress functional imaging may provide essential information with regard to the viability in dysfunctional parts of the heart. A series of imaging techniques are available to assess function at the chamber (or global) or myocardial (or regional) level: that is, echocardiography, CMR, nuclear medicine (planar radionuclide ventriculography, gated blood pool single-photon emission computed tomography [SPECT]), catheter angiography and CCT. Requirements are that these techniques are accurate, reproducible and preferably

376

SECTION A  The Chest and Cardiovascular System

A

C

E

B

D

F

Fig. 15.13  The Calcium Paradox. Unenhanced cardiac computed tomography (CCT) (A and B), contrastenhanced CCT (C and D) and coronary angiography of left (E) and right (F) coronary artery. Presence of diffuse calcified coronary atherosclerosis (total calcium score 1251) and suspicion of several coronary artery stenoses on contrast-enhanced CCT (C and D); however, as clearly shown by coronary angiography, except for a mild (40%) stenosis in proximal left anterior descending coronary artery, no flow-limiting stenoses are found.

A

B

C

Fig. 15.14  Flow-Limiting Stenosis in Mid-Left Anterior Descending (LAD) Coronary Artery. Cardiac computed tomography (A), coronary angiography before (B) and after percutaneous coronary intervention (PCI) (C). There is an atherosclerotic plaque with mixed density in mid LAD impinging on coronary artery lumen (arrow, A) and several non- or minimally stenotic calcified plaques. At coronary angiography, a 70% excentric stenosis is found in mid LAD (arrow, B). The patient was treated with a 3-mm bare metal stent. No residual stenosis was present post-PCI (arrow, C).

CHAPTER 15  Ischaemic Heart Disease

A

B

C

D

E

F

Fig. 15.15  Example of radiation dose-matched standard resolution (0.5 × 0.5 mm2) (A, C and E) and high resolution (0.25 × 0.25 mm2) (B, D and F) photon-counting images of a Synergy Monorail coronary stent (Boston Scientific) made of a platinum chromium alloy with a nominal diameter of 2.75 mm and a length of 20 mm. Axial (A and B), coronal (C and D) and coronal maximum intensity projection reconstructions (E and F) demonstrate substantially improved stent lumen and strut visibility with high-resolution images.

non-invasive. Global ventricular performance looks at the amount of blood expulsed by the ventricle in relation to the size of the ventricle. Normalisation to body surface area enables comparison with gender- and age-matched normal subjects. Typically, ventricular volumes are measured (or estimated) at end diastole (i.e. maximal filling) and at end systole (i.e. maximal emptying) (Fig. 15.16). From the volumes, the ventricular stroke volume (i.e. end-diastolic volume minus end-systolic volume), ejection fraction (i.e. stroke volume divided by end-diastolic volume) and cardiac output (i.e. stroke volume multiplied by heart rate) can be calculated. In clinical practice two approaches are used: (a) assumption

377

comparing the ventricle to a geometrical model and (b) volumetric techniques. Whereas geometric assumption allows fast estimation of ventricular size, it intrinsically depends on how well the model fits with the ventricle, which may not be the case in diseased ventricles. Moreover, the complex geometry of the right ventricle impedes easy use of geometric assumption. In contrast, volumetric techniques use a slice summation, cutting the ventricle in a set of (contiguous) slices. Summing up the volume of each of these slides yields an accurate estimate of the ventricular volume. The downside is that this approach is more time demanding for acquisition and analysis During systole, a complex deformation of the myocardium results in expulsion of approximately two-thirds of the ventricular blood volume. This myocardial deformation (also called ‘myocardial strain’) is a direct consequence of the intricate myofibre anatomy, the interaction between deep and superficial myofibre layers and of a repositioning of fibre sheets during systole, resulting in a ventricular shortening in long- and short-axis direction, wall thickening and, on top, a wringing motion of the ventricle in longitudinal direction—a phenomenon called ventricular torsion or twisting. This deformation can be visually assessed and graded, and/or quantified and described. Visual assessment describes (systolic) wall motion as normokinetic (preserved motion), hypokinetic (decreased but still present), akinetic (completely absent), dyskinetic (wall moving outward during contraction) and hyperkinetic (increased). Systolic wall thickening can be visually graded as normal, diminished, absent, as wall thinning, or as increased. Strain imaging is used to quantify myocardial deformation, analysing deformation (or strain) in longitudinal, circumferential and radial directions. Quantification of ventricular torsion necessitates adapted approaches such as MR myocardial tagging. Although beyond the scope of this chapter, it should be noted that the second part of the cardiac cycle, i.e. diastole, consisting of myocardial relaxation with subsequent ventricular filling, is also an essential of the cardiac performance. Diastolic dysfunction may be the cause of heart failure or superimposed on systolic heart failure in IHD patients. To describe regional myocardial morphology and function in a standardised and imaging-modality independent way, the American Heart Association has promulgated a 17-segment model to map the left ventricle. Because of conical shape of the left ventricle, it is divided in longitudinal direction into four equal levels (or rings): that is a basal, mid and apical level, and a final level representing the LV apex. Next, the basal and mid-level are divided in six equiangular segments, and the smaller apical level in four equiangular segments. Finally, the segments are numbered starting with the anterobasal segment and following a counter-clockwise direction (viewing the LV from apically): that is, basal level: segments 1 to 6, mid-level: segments 7 to 12, apical level: segments 13 to 16, and LV apex, that is segment 17. Moreover, segments can be attributed to a coronary artery perfusion territory. The left anterior descending (LAD) coronary artery typically supplies segments 1, 2, 7, 8, 13, 14, 17; the right coronary artery (RCA) segments 3, 4, 9, 10, 15; and the left circumflex (LCx) coronary artery segments 5, 6, 11, 12, 16 (Fig. 15.17). Several variations, however, may occur depending on the coronary artery anatomy and dominance. Despite the intrinsic advantages of using a standardised segmentation model, application in a myocardial infarction patient is often not ideal as the segments often only partially fit with infarct location, thus ending with segments consisting in a mixture of infarcted and viable myocardium. Alternatively, a compartment model using the location and extent of the myocardial infarction may be more appealing to study the regional function as well as the interplay between the infarcted, peri-infarct (or adjacent) and remote myocardium. As right ventricular (RV) dysfunction in infarct patients portends poor outcome, assessment of the ventricular performance is not complete without RV assessment.

378

SECTION A  The Chest and Cardiovascular System

A

D

B

E

C Fig. 15.16  Imaging Techniques Used for Assessment of Cardiac Function. In the left part of the figure are shown five different imaging technique: that is, echocardiography (A), nuclear medicine (B), cardiac computed tomography (C), cardiac catheterisation (D) and cardiac magnetic resonance (E), which are used daily for assessment of ventricular volumes and function. The graph shows the time–volume changes of the left ventricular (LV) cavity during a cardiac cycle. At time 0 corresponding to end diastole, the LV has its largest volume. After aortic valve opening (AVO), part of this volume (stroke volume (SV)) is ejected into the thoracic aorta during ventricular contraction (‘ejection’). Between aortic valve closure (AVC) and mitral valve opening (MVO), the LV volume remains constant (isovolumic relaxation phase). At the moment of MVO, LV filling occurs and is characterised by three phases: early diastasis, atrial contraction and then, the last phase of the cardiac cycle, the isovolumic contraction between mitral valve closure (MVC) and AVO. Most techniques measure the changes in LV volume at end diastole (maximal filling) and end systole (maximal emptying) to express LV volumes and ejection fraction.

Cardiac ultrasound is the first-line technique to assess cardiac function in IHD patients. It can be performed at the bedside, provides valuable information regarding cardiac structure and ventricular and atrial volumes and function, and allows the visualisation of complications such aneurysm or thrombus formation post-infarction. Novel techniques such as speckle tracking echocardiography allow myocardial strain imaging with good feasibility in the clinical setting; however, in many patients, image quality is suboptimal, and geometric assumptions are used in clinical routine to assess ventricular volume and function. In recent years, CMR has emerged as an interesting alternative to echocardiography. Compared with competing techniques, such as CCT or nuclear medicine, no iodinated contrast material nor radioactive tracer needs to be injected for volumetric/functional cardiac imaging. Moreover, as explained in more detail below, CMR offers a comprehensive view on the heart. For volumetric and functional imaging, CMR relies on bright-blood CMR, nowadays using the balanced state-state free-precession sequence

yielding a high intrinsic contrast between blood and surrounding myocardium. Dynamic information of the cardiac function is obtained when multiple images are acquired over the cardiac cycle. These can be played in cine-loops allowing to appreciate dynamic phenomena such as myocardial/valve motion, and to visualise infarct-related complications such as aneurysm formation and thrombus formation. CMR images are typically acquired at breath-hold. In uncooperative patients or in patients with atrial fibrillation, non-ECG-gated real-time cine imaging may be a valuable alternative. Functional CMR imaging is typically performed using a combination of short- and long-axis planes: that is horizontal and vertical long-axis and three-chamber view (Fig. 15.18). For ventricular volumetric and functional imaging, the ventricles are completely encompassed by a set of contiguous slices, usually acquired in short-axis direction. In analogy to speckle-tracking echocardiography, CMR-based strain analysis is nowadays possible, based on optical flow technology (feature-tracking) or application of more complex elastic

CHAPTER 15  Ischaemic Heart Disease

379

Fig. 15.17  Seventeen-Segment Approach for Left Ventricular (LV) Segmentation as Proposed by the American Heart Association. The left ventricle is divided in longitudinal direction (left images) into a basal, mid-cavity and an apical short-axis ring (middle images). Subsequently, these short-axis rings are divided into 6 basal, 6 mid-cavity and 4 apical segments, with segment 17 being the apex seen on the long-axis views (left images). Basal: 1, anterior; 2, anteroseptal; 3, inferoseptal; 4, inferior; 5, inferolateral; 6, anterolateral. Mid-cavity: 7, anterior; 8, anteroseptal; 9, inferoseptal; 10, inferior; 11, inferolateral; 12, anterolateral. Apical: 13, anterior; 14, septal; 15, inferior; 16, lateral. Bull’s eye plot (right image) representation of all segments of the left ventricle. The segment numbers refer to the same segments.

registration algorithms. Furthermore, magnetisation preparation pulses enable to non-invasively create tag lines (‘myocardial tagging’) on the myocardium, allowing the unravel the intramyocardial deformation patterns in IHD patients. CCT is an alternative to echocardiography and CMR for ventricular volumetric assessment. The administration of contrast agent should be adapted to obtain enhancement of both ventricular cavities. Routine use of CCT for this purpose, however, is hampered by the need to use radiation over the entire cycle. In many

hospitals, planar radionuclide ventriculography is an established technique to assess LV volumes and function. Alternatively, gated blood pool SPECT can be used to assess wall motion and regional ejection fraction.

STRESS IMAGING In patients with chronic stable CAD, treatment goals are threefold: (a) relief of symptoms and ischaemia; (b) prevention of premature

380

SECTION A  The Chest and Cardiovascular System

A

B

D

E

C Fig. 15.18  MR Approach for Quantification of Left Ventricular (LV) Function. A standardised approach is used for assessment of LV volumes, function and mass. Cardiac magnetic resonance imaging is started with axial images (A). The vertical long-axis plane (B) is aligned from the axial plane through the mitral valve and LV apex, which may be on a separate more inferior slice. The horizontal long-axis plane (C) is aligned from the vertical long-axis plane through the mitral valve and LV apex. The short-axis (E) is aligned from the vertical long-axis and horizontal long-axis planes—perpendicular to both. For quantification of LV volumes, function and mass, the set of short-axis images is positioned completely encompassing the left ventricle (D). Contouring the endocardial (green line, E) and epicardial border (yellow, E) of the short-axis images at end diastole and end systole enables calculation of the LV end-diastolic and end-systolic volume, stroke volume, ejection fraction and LV myocardial mass.

cardiovascular death; and (c) prevention of progression of CAD leading to myocardial infarction, LV dysfunction and congestive heart failure. Management of CAD, however, remains highly challenging as several studies have shown that revascularisation fails to improve mortality over medical treatment in randomised trials. The explanation of this paradox lies most likely in the poor relation between stenosis severity in diffuse CAD and coronary flow physiology. Whereas anatomical techniques (conventional coronary angiography, CCT) provide limited information regarding the impact of a stenosis on the coronary flow, stress testing can be recommended to assess the extent of myocardial ischaemia before coronary angiography. There is substantial evidence that a moderate-to severe ischaemic burden greater than 5%–10%, with or without angina, is an indication for revascularisation, whereas those patients without clear evidence of myocardial ischaemia likely benefit from an optimum medical treatment (e.g. high-dose statins—risk factor

modification) to alter the natural history of CAD. At cardiac catheterisation, the functional severity of a stenosis can be determined by the FFR expressing the maximum achievable blood flow to the myocardium supplied by a stenotic artery as a fraction of normal maximum flow. A normal value is 1.0 and a value of 0.75 reliably identifies stenoses associated with inducible ischaemia. The diagnostic accuracy of FFR is greater than 90%. Although FFR may be helpful in determining in patients presenting with diffuse CAD at cardiac catheterisation which stenoses may benefit from percutaneous coronary intervention (PCI), non-invasive testing for reversible (or inducible) ischaemia is warranted to optimally stratify patients with stable CAD. Non-invasive testing for reversible ischaemia is achieved by stressing the heart and evaluating whether, during stress, symptoms of angina, ECG signs of myocardial ischaemia, ischaemia-induced myocardial wall motion abnormalities (WMAs) or myocardial perfusion disturbances

CHAPTER 15  Ischaemic Heart Disease occur. Exercise ECG test (EET), although widely used in daily practice, has a low sensitivity (approximately 60%) and moderate specificity (nearly 80%), and a normal test does not exclude CAD. In particular, EET is a poor diagnostic test in low-risk populations (such as women) owing to its low positive value in a population with a low prevalence of the disease. The limited accuracy of EET in diagnosing CAD is also due in part to its position near the bottom of the ischaemic cascade. Therefore, the diagnosis of CAD may be improved by using non-invasive tests higher up in the ischaemic cascade than EET, assessing abnormalities in myocardial function or in myocardial perfusion during stress conditions. While there is most experience with myocardial perfusion scintigraphy and stress echocardiography for these purposes, several other single or hybrid techniques have emerged in the field of stress imaging, such as stress perfusion CMR, stress function CMR, stress perfusion CT, combined coronary and stress myocardial perfusion imaging by CCT, and hybrid cardiac SPECT/CT or cardiac PET/CT. Nuclear medicine is a cornerstone in the assessment of myocardial perfusion in CAD patients, and it has an established role in risk

A

C

381

stratification for major adverse cardiac events. Most often, SPECT is used to diagnose and evaluate the severity of CAD, while PET is more accurate but also more expensive and less available. SPECT measures the relative myocardial distribution of radionuclides, such as thallium-201 (201Tl), technetium 99m (99mTc) sestamibi (MIBI). Study protocols are specific for the different tracers: for instance, MIBI SPECT is performed using an injection of tracer during stress and a second injection at rest (or vice versa), while 201Tl is injected during stress, and the redistribution of tracer is measured at rest after a delay (e.g. 4 hours). In regions with impaired myocardial perfusion, the number of counts is lower than normally perfused myocardium, resulting in a ‘defect’. Reversible defects (i.e. present on stress but absent at rest) are caused by flowlimiting stenoses and should be differentiated from fixed defects (i.e. present at rest/redistribution at rest), reflecting myocardial scarring (Fig. 15.19). The severity of the defect (i.e. reduction in counts) is related to stenosis severity, while the extent of the defect is related to the myocardium supplied by the stenotic artery. Although SPECT is widely used in clinical practice, yielding good sensitivity (approximately

B

D

Fig. 15.19  Myocardial Ischaemia Testing, Sestamibi (MIBI) Single-Photon Emission Computed Tomography (SPECT) Versus Cardiac Magnetic Resonance (CMR). A 49-year-old man with type 1 diabetes, no symptoms, no anginal pain. MIBI SPECT shows reversible perfusion defect in lateral left ventricular (LV) wall with an estimated ischaemia of 19% of LV myocardium* (A and B). Stress perfusion CMR shows extensive stress-induced perfusion defect in lateral LV wall (arrows, C) and subendocardial perfusion in anterior LV wall and septum (arrowheads, C). ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR showed partial infarction of both papillary muscles (images not shown). At coronary angiography (D) severe two-vessel coronary artery disease (CAD) is shown with distal left circumflex (LCx) 80% stenosis, 1st lateral 80% stenosis, and mid anterior left anterior descending (LAD) 70% stenosis and distal LAD 80% stenosis. Coronary artery bypass graft surgery was performed with left internal mammary artery (LIMA) to LAD and free LIMA from LIMA to LCx. (MIBI SPECT courtesy O. Gheysens M.D., Department of Nuclear Medicine, UZ Leuven, Leuven, Belgium.)

382

SECTION A  The Chest and Cardiovascular System

90%) and moderate specificity (around 65%), certain pitfalls need to be mentioned. Radiation exposure of the injected isotopes, depending on the protocol used, ranges between 8 and 20 mSv. Subendocardial and small infarcts may be missed by SPECT because of the lack in spatial resolution (Fig. 15.20). To avoid false-positive readings and to improve the test specificity, use of attenuation correction methods and gated analysis of wall motion is recommended. Finally, in patients with multivessel CAD, hypoperfusion of the entire myocardium may mask regional abnormalities. PET is very useful for assessing myocardial perfusion and metabolism. Assessment of myocardial perfusion can be performed with 13N-ammonia, 15O-H2O, 82Ru or 11C-acetate. PET has several advantages over SPECT, such as a higher spatial resolution and the possibility to measure absolute myocardial blood flow. This is advantageous in patients with balanced ischaemia caused by left main or three-vessel CAD in which maximal myocardial blood flow is reduced in all regions of the left ventricle, or in patients in whom the myocardial ischaemia is caused by microvascular dysfunction. The reported sensitivity (and specificity) of PET for detecting angiographic

A

B

stenosis ≥50% is approximately 90%. Drawbacks of PET are patient exposure to radiation (although less than SPECT), availability and limited half-life of PET tracers, cost and availability of PET scanners, false-positive myocardial perfusion defects due to misregistration, and the limited spatial resolution when compared with CMR. Perfusion CMR uses the changes in myocardial signal intensity during first pass of a contrast media bolus through the myocardium to assess myocardial perfusion. This necessitates fast imaging sequences with saturation pre-pulses to suppress myocardial signal and obtain T1 weighting. Sufficient coverage of the left ventricle is obtained with three short-axis slices (possibly in combination with a long-axis view), enabling segmental perfusion assessment using the 16-segment model (see Fig. 15.17). Similar to nuclear imaging, myocardial hyperaemia (induced by a vasodilator such as dipyridamole or adenosine) enables depiction of haemodynamic significant stenoses. Hypoperfused myocardium during stress conditions appears as a non- or slow-enhancing parts of the myocardium (‘perfusion defect’) (Fig. 15.21). The defect, typically, obeys anatomical borders as well as the boundaries of the coronary artery

C

Fig. 15.20  Small Inferolateral Myocardial Infarct, Sestamibi (MIBI) Single-Photon Emission Computed Tomography (SPECT) Versus Cardiac Magnetic Resonance. History of percutaneous coronary intervention (PCI) with stent placement in left anterior descending (LAD) coronary artery in 45-year-old man. MIBI SPECT shows reversible defect in mid/apical left ventricular (LV) anterior wall (±10% of LV myocardium) and decreased tracer activity in mid/basal inferolateral wall. Rest perfusion (A) and stress perfusion (B) cardiac magnetic resonance (CMR) shows focal hypointense appearance of the mid/basal LV inferolateral wall during first pass of contrast (arrow, A and B). On ‘late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR, this area shows a focal, almost completely transmural enhancement (arrow, C), compatible with healed myocardial infarction. No evidence of myocardial ischaemia in LAD territory in stress perfusion CMR. (MIBI SPECT courtesy O. Gheysens M.D., Department of Nuclear Medicine, UZ Leuven, Leuven, Belgium.)

CHAPTER 15  Ischaemic Heart Disease

A

B

383

C

Fig. 15.21  Stress Perfusion Cardiac Magnetic Resonance (CMR) in Patient With Severe Three-Vessel Coronary Artery Disease (CAD). A 62-year-old man with a history of coronary artery bypass graft presented with increasing complaints of angina chest pain post-surgery. Medical treatment yielded unsatisfactory results. Stress perfusion CMR shows extensive perfusion defect in the anteroseptal left ventricular (LV) wall (arrows) and inferolateral wall (arrowheads); these findings would make surgical re-intervention appropriate. (A) basal short-axis. (B) middle short-axis. (C) apical short-axis.

perfusion territories, and the extent is determined by the position of the stenosis along the coronary artery. Semiquantitative measures show a decline in the upslope during first pass of contrast, and when related to the upslope during resting conditions, the myocardial perfusion reserve (MPR) ratio is typically decreased in the hypoperfused areas. Similar to PET, absolute myocardial blood flow can also be quantified with CMR. In a landmark paper by Greenwood et al. studying 752 prospectively included patients, stress perfusion CMR yielded a higher accuracy than SPECT. This, and other recently published studies, emphasise the intrinsic value of CMR in assessing patients with stable CAD. As the spatial resolution of CMR is superior to SPECT, smaller, subendocardial perfusion defects missed at SPECT can be shown at CMR. Moreover, CMR is not hampered by soft-tissue and attenuation artefacts. In analogy to CMR, CCT can also be used to study myocardial perfusion under stress conditions; in particular, the combination with morphological depiction of the plaque and plaque stenosis severity at coronary CT angiography, this is potentially a very appealing application. The downside of this approach remains the substantial radiation exposure. In analogy to stress echocardiography, stress function studies can be performed safely in an MR environment. Dobutamine, a β-agonist, increases oxygen consumption by increasing myocardial contractility and heart rate. A stepwise dose increment of dobutamine allows for the evaluation of the myocardial response at each stress level. Whereas normally supplied myocardium shows a progressive increase in myocardial contractility, myocardium supplied by a flow-limiting coronary stenosis becomes ischaemic when the compensatory increase in coronary blood supply is insufficient to match the increased demand in oxygen (therefore when the coronary flow reserve is superseded). This, in turn, will cause a decrease in regional contractility and lead to WMAs. Using a combination of long- and short-axis cine CMR sequences, regional contractility can be assessed in all segments of the left ventricle. Imaging commences in resting conditions, and the same set of sequences is repeated for each stress level. Dobutamine infusion is started at low dose (5 µg/kg body weight per minute). Cine images are analysed for new (or worsening) WMA. If normal, the dose of dobutamine is increased using steps of 5 to 10 µg/kg, and the above approach is repeated. The test is considered positive for obstructive CAD when WMAs develop or worsen, or when the patient develops chest pain at a certain stress level. If the target heart rate of the patient, defined as (220−age) × 0.85, is not reached at 40 µg/kg dobutamine, intravenous atropine can be

additionally administered up to 1 to 2 mg using fractionate doses of 0.25 mg every minute (Fig. 15.22). Dobutamine administration needs to be stopped on patient request, and is also discontinued if the systolic blood pressure decreases greater than 20 mm Hg below the baseline systolic blood pressure; the systolic blood pressure decreases greater than 40 mm Hg from a previous level; the blood pressure increases to above 240/120 mm Hg; or when severe arrhythmias occur. As this is a potentially dangerous examination, haemodynamic parameters such as heart rate and blood pressure should be closely monitored, cardiac resuscitation material should be available, and teams should be trained in case of cardiac complications for fast patient evacuation from the magnet; however, in experienced hands, high-dose dobutamine-atropine stress MRI is a safe examination with minimal side effects. Besides detecting flow-limiting coronary stenoses in patients suspected of obstructive CAD, low-dose stress function imaging (≤20 µg/kg dobutamine) enables differentiation between viable and non-viable dysfunctional myocardium in patients with chronic CAD. Also, in patients with a recent myocardial infarction, low-dose stress imaging can differentiate between stunned and irreversibly damaged myocardium. High-dose dobutamine/atropine stress CMR yields good sensitivity (approximately 90%) and specificity (also approximately 90%) for detection of significant CAD. Novel CMR techniques such as sensitivity encoding (SENC)-CMR, which allows strain visualisation, provide incremental value for the detection of CAD compared with conventional wall motion readings, and have the strength to detect CAD at lower stress levels. SUMMARY BOX: Biomarkers of Acute Myocardial Infarction • ST segment abnormalities on the electrocardiogram. • Temporal raise in cardiac enzymes (e.g. troponin). • Fixed perfusion defect at gated single-photon emission computed tomography. • Subendocardial/transmural ‘late’ or ‘delayed’ gadolinium-enhanced (LGE) at cardiac magnetic resonance in a coronary artery perfusion territory

MYOCARDIAL INFARCT IMAGING Assessment of the electrical cardiac activity using 12-lead electrocardiography and analysis of cardiac biomarkers are central diagnostic techniques in patients presenting with an ACS. Firstly, patient’s triage

384

SECTION A  The Chest and Cardiovascular System

Fig. 15.22  Practical Scheme for Dobutamine Stress Cardiac Magnetic Resonance. For each stress level, four short-axis and two long-axis cine studies are obtained in three consecutive breath-holds. Per stress level images are analysed for new wall motion abnormalities (WMA). If termination criteria are not met at the highest dobutamine dose, atropine can be additionally administered. SBP, systolic blood pressure.

and treatment are, to a large extent, based on ECG changes indicative of new ischaemia (new ST-T changes or new left bundle branch block), and development of new pathological Q waves (Fig. 15.23). The ECG allows the clinician to suggest the infarct-related artery, to estimate the amount of myocardium at risk, and to detect prior (healed) myocardial infarction. In infarcts presenting an ST-segment elevation, the degree of ST resolution post-reperfusion reflects the success of reperfusion. Secondly, as myocardial cell death is characterised by a release of different proteins into the circulation from the damaged myocytes, increased cardiac blood biomarkers are indicative of recent myocardial necrosis. The preferred biomarker at present is cardiac troponin (I or T), which has a nearly absolute myocardial tissue specificity as well as high clinical sensitivity. If troponin assays are not available, the best alternative is creatine kinase-MB. Because of the complex release characteristics of these cardiac proteins, it is still unclear whether the peak value, or a single point measurement of cardiac biomarkers, provides the best estimate of myocardial infarct size. Moreover, neither blood biomarkers nor ECG provides a good insight in the evolving processes in the jeopardised myocardium. Imaging techniques such as cardiac ultrasound and cardiac catheterisation visualise an acute myocardial infarction indirectly by the impact of the infarction on wall motion or wall strain, but do not visualise the necrotic myocardium. Gated SPECT shows a myocardial infarction as a fixed defect with loss in regional function; however, small (subendocardial) infarcts are frequently missed because of the lack in spatial resolution. In recent years, CMR has become the in vivo reference technique for myocardial infarction imaging, providing an in-depth, comprehensive view (Figs 15.24 and 15.25). Although several groups have explored the use of CCT to assess acute myocardial infarctions, this technique has not yet entered the clinical arena. One

major limitation of current CCT in the setting of myocardial infarction is the relatively low soft-tissue contrast between normal and infarcted myocardium. Dual-energy and photon-counting spectral CT may overcome this limitation by imaging the contrast-to-noise ratio (CNR) of late iodine contrast material enhancement in the infarcted myocardium. Additionally, photon-counting CT may allow for the simultaneous assessment of multiple phases of contrast enhancement (i.e. first-pass perfusion and late enhancement) during CCT, with a substantial reduction in radiation exposure. Myocardial tissue characterisation is probably the key feature where CMR excels and outperforms all other imaging techniques. As T1- and T1 and T2 relaxation times are tissue specific, alterations in relaxation times are indicative of myocardial pathology. Firstly, CMR sequences can be ‘weighted’, comparing the signal intensity of tissue deemed abnormal to the signal intensity of tissue deemed normal. In the jeopardised myocardium, myocardial-free water increases as a consequence of prolonged myocardial ischaemia, resulting in a prolongation of both T1 and T2 relaxation times. Myocardial oedema is an early phenomenon that can be depicted as soon as 30 minutes after onset of myocardial ischaemia. In particular, using T2-weighted sequences, the jeopardised myocardium will appear as an abnormally bright (hyperintense) area in clear contrast to the grey normal myocardium (Fig. 15.26). Currently, different optimised dark-blood and bright-blood T2-weighted sequences are available. Although well suited to depict areas of myocardial oedema, these sequences lack specificity, image quality may be suboptimal, and differentiation with hyperintense signal from stagnant blood adjacent may be challenging. Moreover, in the bright-appearing jeopardised myocardium, differentiation between irreversibly damaged and reversible (viable) myocardium is not possible. Fortunately, this issue can be solved

CHAPTER 15  Ischaemic Heart Disease

A

B

385

C

Fig. 15.23  Extensive Acute Anteroseptoapical Myocardial Infarction. Electrocardiographic (ECG)-cardiac magnetic resonance (CMR) correlate. Successful percutaneous coronary intervention of occluded proximal left anterior descending (LAD) coronary artery. Both CMR and ECG were obtained 3 days after the acute event. The ECG shows anterior necrosis with QS complex in V1 to V4 and persistent ST-segment elevation with positive T-waves in the same leads. These findings suggest lack of myocardial tissue perfusion (i.e. microvascular obstruction). ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR in horizontal long-axis (A), vertical long-axis (B) and mid-ventricular short-axis (C) images show transmural enhancement in segments 1, 2, 7, 8, 13, 14, 15 and 17, reflecting extensive necrosis in the LAD perfusion territory (arrows). Presence of microvascular obstruction inside the infarct area as suggested on ECG. The extent of microvascular obstruction was more prominent on CMR images early after contrast administration.

by intravenously administering gadolinium chelates. As these contrast agents have extracellular distribution properties, the distribution volume is the largest in the necrotic part of the jeopardised myocardium, resulting in a greater T1 shortening compared with the less-damaged, viable, parts of the jeopardised myocardium. The area of enhancement—in case of acute myocardial infarction—is typically subendocardial, located in the distribution territory of one of the coronary arteries, and the transmural

spread of enhancement is variable (Fig. 15.27). Using the relation of the extent of enhancement to the extent of myocardial oedema, a ratio can be calculated reflecting the myocardial salvage index. For example, if no necrosis has occurred then no enhancement is present in the oedematous myocardium. In this condition, also called aborted myocardial infarction, the myocardial salvage equals 100%. Assessment of myocardial salvage is important because it yields prognostic value. Over the years,

386

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E

F

Fig. 15.24  Comparison of Imaging Techniques in Assessing Patients With an Acute Myocardial Infarction. Conventional chest radiography (A), cardiac catheterisation (B), cardiac ultrasound (C) and cardiac magnetic resonance (CMR) using ‘late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR (D), and cine CMR in vertical long-axis (E) and horizontal long-axis (F) images. Conventional chest (bedside) radiography shows moderate cardiomegaly without evidence of pulmonary oedema. Left ventricular (LV) contrast ventriculography (RAO position, end-systolic time frame) shows extensive area of decreased contractility involving the anterior wall, apex and the apico-inferior LV wall (arrows, B). Cardiac ultrasound (longitudinal parasternal view) reveals similar information (arrow, C). LGE CMR shows extensive myocardial infarction involving most of the ventricular septum, apical two-thirds of the anterior wall, apex, and apical inferolateral wall (arrows, D). While the periphery of the infarct is strongly enhanced, centrally an extensive zone of microvascular obstruction remains on LGE CMR, reflecting severe microvascular damage. The functional consequences of the infarction can be well appreciated on cine CMR (arrows, E and F).

the sequence design for T1-enhanced imaging has been modified and constantly improved, especially with the introduction of the inversionrecovery sequence, which has led to a paradigm shift in myocardial infarct imaging. In brief, the difference in longitudinal relaxation time between normal and infarcted myocardium can be exploited to create, and to improve, the (differential) tissue contrast. Moreover, one should be familiar with the pharmacokinetics of gadolinium chelates. Within the infarcted myocardium, the optimum timing for accurate infarct imaging is approximately 10 to 25 minutes following administration of contrast agent. Therefore, this sequence is called ‘late’ or ‘delayed’ gadolinium-enhanced CMR (LGE). This sequence has been extensively validated, and areas of irreversible myocardial damage as small as 1 mL can be depicted (Fig. 15.28). Moreover, the centre of the infarction may occlude the microvasculature—the so-called microvascular obstruction or no-reflow phenomenon—and is found in approximately 50% of successfully reperfused STEMI. This can be visualised at contrast-enhanced MRI as lack of enhancement (‘dark area’) in the centre of the area of enhancement. Another sign of reperfusion injury is extravasation of red bloods (‘haemorrhagic infarction’) secondary to severe

capillary damage. As deoxyhaemoglobin yields paramagnetic properties, intramyocardial haemorrhage can be recognised at T2-weighted imaging as a central dark area (Fig. 15.29). Both are considered to represent reperfusion damage incurred in the ischaemic myocardium and bear prognostic value. In recent years, quantitative (‘parametric’) imaging has entered the front line for infarct imaging, with T1, T2 and T2* mapping providing a more objective means to tissue characterise the myocardium (Fig. 15.30). Mapping sequences acquire several images at different stages of the relaxation, allowing a pixel-wise fitting of the relaxation behaviour. Colour-coded representation facilitates interpretation. In a patient with an acute myocardial infarction, both T1 and T2 values increase in the jeopardised myocardium, and the extent of abnormal increase can be used to calculate the area at risk. T1 mapping can be repeated following contrast administration (measurement performed 10 to 25 minutes post-contrast administration). Lowest post-contrast T1 values will be found in the necrotic myocardium, less severe in the reversible damaged part of the jeopardised myocardium. Text continued on p. 389

CHAPTER 15  Ischaemic Heart Disease

A

B

C

Fig. 15.25  Comprehensive Cardiac Magnetic Resonance (CMR) in Acute Myocardial Infarction. ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR (A), short-axis cine CMR (end-systolic frame) (B) and radial strain curve (C). A 69-year-old was admitted with acute chest pain and raised cardiac enzymes. LGE CMR in the mid-anterolateral left ventricular (LV) wall area of subendocardial enhancement (75% transmural enhancement) (arrow, A), reflecting acute myocardial necrosis. Cine imaging shows hypokinesia with impaired contractility in the anterolateral LV wall (arrow, B). This is confirmed by strain analysis (using CMR feature tracking) showing diminished radial strain (i.e. decreased systolic wall thickening in the anterolateral LV wall (yellow curve) (arrow, C).

A

C

B

D

Fig. 15.26  Acute Laterobasal Myocardial Infarction A 37-year-old man was admitted with retrosternal chest pain irradiating to the mandibula. Positive cardiac enzymes and ST-segment elevation in anterior and lateral ECG leads. Coronary angiography shows proximal occlusion of the first lateral branch of the left circumflex coronary artery. T2-weighted short-axis (A) and horizontal long-axis (B) imaging. ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) cardiac magnetic resonance (CMR) in short-axis (C) and horizontal long-axis (D) imaging. Sharply defined zone of myocardial oedema in anterior and lateral wall (segments 1, 6) (arrows, A and B). LGE CMR shows strong enhancement in anterior and lateral wall (segments 1, 6) (arrows, C and D). The extent of enhancement coincides very well with the extent of myocardial oedema, which means that the major part of the jeopardised myocardium has been irreversibly damaged. In other words, myocardial salvage is very low.

387

388

SECTION A  The Chest and Cardiovascular System

B

A

C

Fig. 15.27  Typical Presentation of Acute Transmural Anterior Myocardial Infarction. Successful reperfusion of completely occluded mid-left anterior descending coronary artery. ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) cardiac magnetic resonance in horizontal long-axis (A), vertical long-axis (B) and short-axis (C) images. Transmural myocardial enhancement is shown in segments 7, 8, 13, 14 and 17 (arrows, A–C).

A

B

C

D

Fig. 15.28  Myocardial Damage and Ischaemic Cerebrovascular Accident. A 47-year-old patient was admitted with retrosternal chest pain and left hemiparesis. Slightly increased cardiac enzymes (troponin I: 0.5 µg/L) and no evidence of obstructive coronary artery disease (CAD) on coronary angiography. ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) cardiac magnetic resonance (CMR) in cardiac short-axis (A), horizontal long-axis (B), and vertical long-axis imaging (C). Presence of a small spot of LGE (10 years) showed a significant benefit for patients in the bypass group compared with the medical treatment group. These above findings underscore the complexity of the myocardial viability issue as well as the quest for the ideal imaging technique to

A

391

depict hibernating myocardium. Chronic dysfunctional myocardium recovery following successful coronary revascularisation is deemed to reflect hibernating myocardium, although reversed cellular differentiation has never been documented, but only assumed. Hibernating myocardium is therefore a post-hoc observation, and as such not useful to select those patients which might benefit from revascularisation. A first approach is use of end-diastolic wall thickness. As scar formation in myocardial infarction causes wall thinning, preserved wall thickness is considered to reflect preserved myocardial viability and most workers use a cut-off of 5.5 to 6 mm; however, many segments with preserved wall thickness fail to functionally recover post-revascularisation. Conversely, thinned wall segments may functionally improve after revascularisation. A second approach is based on myocyte membrane integrity and is used by both nuclear imaging and contrast-enhanced CMR. Both thallium-201 and technetium-99m sestamibi/tetrofosmin SPECT techniques compare the uptake of radiotracer in the dysfunctional myocardium relative to remote (normal) myocardium. A tracer activity greater than 50% of the maximum tracer uptake is used as threshold for viable tissue. Although generally accepted and widely used, the downside is the low spatial resolution (i.e. 7 mm), limiting its ability to depict subendocardial infarcts. A third approach is based on the principle of increased interstitial space (or extracellular volume) in scarred myocardium. In particular, LGE-CMR is a well-validated, highly accurate and reproducible technique for sizing healed infarcts and has become the reference technique to depict infarct-related myocardial scarring (Fig. 15.33). In a landmark paper by Kim et al. (2000), the likelihood of functional improvement after revascularisation was predicted by the transmural extent of enhancement. In dysfunctional segments without CMR evidence of scar, 78% of segments improved contractility post-revascularisation versus 2% of segments with a scar involving greater than 75% of wall thickness. A fourth approach is imaging of myocardial metabolism using quantitative PET imaging. Combining a metabolism tracer (fatty acid or glucose analogs (e.g. 18F-fluorodeoxyglucose (18F-FDG) with a perfusion tracer (13N-ammonia), the myocardial viability can

B

C

D

E

F

G

Fig. 15.32  Adverse Ventricular Remodelling Following Acute Myocardial Infarction. Patient with extensive anterior myocardial infarction. Short-axis ‘late’ or ‘delayed’ gadolinium-enhanced (LGE) cardiac magnetic resonance (CMR) image (A) early post-infarction shows extensive transmural enhancement (arrows, A) (segments 1, 2, 7, 8, 13, 14, 15, 17) with large zone of microvascular obstruction. Cine CMR early (B–D) and 6 months (E–G) post-infarction. Note the important wall thinning of the involved segments with aneurysm formation at 6 months follow-up (arrows, E–G). The end-diastolic volume increased from 257 mL at baseline to 432 mL at follow-up, while the ejection fraction decreased from 36% to 18%.

392

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E

F

G

H

l a t

s e p

post

apex

13

14

15

16

17

12

13

14

15

16

12

13

14

15

16

11

12

13

14

15

J

l a t

s e p

base

I

K

Fig. 15.33  Viability Imaging. Viability imaging in a 57-year-old patient presenting ischaemic cardiomyopathy and increasing dyspnea (NYHA III). Chest radiography (A and B) shows moderate cardiomegaly with redistribution of the pulmonary vascularisation to the upper lung fields, reflecting increased pulmonary venous pressures. Coronary angiography shows complete occlusion of the proximal left circumflex coronary artery (arrow, C), and mid right coronary artery (arrow, D). Cine cardiac magnetic resonance (CMR) in horizontal long-axis (E) and short-axis (F) images shows severely dilated left ventricle (end-diastolic volume 453 mL, ejection fraction 36%) with severe thinning of the entire inferolateral wall (arrows, E and F). Presence of a severe mitral regurgitation owing to mitral valve enlargement secondary to left ventricular (LV) dilatation. ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR shows presence of transmural enhancement in the inferolateral wall, reflecting a healed extensive inferolateral myocardial infarction (arrows, G and H). Positron emission tomography (PET) with perfusion imaging (NH3 as tracer) and myocardial metabolism imaging (fluorodeoxyglucose (FDG) as metabolic tracer). Reconstructed slices in short-axis and horizontal long-axis (I), NH3 (J) and FDG (K) polar maps. Presence of an extensive perfusion defect in the entire inferolateral wall (*, J) that matches perfectly with the lack of metabolism on FDG PET (*, K). This match pattern reflects irreversibly damaged myocardium. The PET and CMR abnormalities correlate perfectly. (FDG/NH3 PET courtesy O. Gheysens M.D., Department of Nuclear Medicine, UZ Leuven, Leuven, Belgium.)

CHAPTER 15  Ischaemic Heart Disease be accurately studied. PET yields excellent sensitivity (approximately 90%) and moderate specificity (approximately 60%). PET compared with SPECT has a superior spatial resolution, lower radiation burden and allows absolute quantification of myocardial blood flow. Drawbacks are availability and short half-life of current available PET perfusion tracers, cost and the lower spatial resolution compared with CMR. A fifth approach is assessment of contractile reserve. Dysfunctional myocardium that contracts (or improves contractility) is deemed viable if stimulated: for example, using dobutamine ± atropine. Non-viable myocardium, in contrast, shows no functional improvement or even a worsening in wall motion. Myocardial contractility reserve is usually performed at echocardiography but can also be assessed by CMR. The accuracy of viability imaging can be improved by combining approaches. In particular, the likelihood of functional recovery is uncertain in patients showing dysfunctional myocardium with preserved end-diastolic wall thickness and in those presenting with intermediate grades of scar transmurality (i.e. 25% to 75%): for instance, by performing additional low-dose dobutamine stress imaging. Also, the emergence of hybrid PET–MR scanners may further improve diagnostic accuracy. As many ischaemic heart failure patients have evidence of ventricular dyssynchrony, they may benefit from cardiac resynchronisation therapy (CRT) as part of an effective therapy for heart failure. Although not yet included in the guidelines for CRT, imaging has an increasingly important role in determining those patients that might benefit from CRT. Goals of imaging are threefold: (a) assessment of the degree of mechanical dyssynchrony, (b) myocardial scar imaging and (c) coronary venous imaging. Ventricular performance can be improved if the contraction of the different parts of the ventricle is synchronised. The degree of mechanical dyssynchrony can be quantified using speckle-tracking echocardiography, strain-based cine CMR as well as more advanced techniques such as myocardial tagging, DENSE (displacement encoding with stimulated echoes) and tissue velocity mapping. Secondly, correct placement of the CRT leads is crucial to improve mechanical dyssynchrony. Thirdly, ventricular scar mapping can be achieved using LGE CMR, enabling to determine the presence, location and extent of myocardial scarring. Information regarding the coronary vein anatomy and patency, which may be important for CRT lead placement, can be achieved with CCT.

A

393

IMAGING OF COMPLICATIONS RELATED TO ISCHAEMIC HEART DISEASE A series of potentially lethal complications is related to IHD, necessitating timely recognition and treatment (Fig. 15.34). Although echocardiography remains first in line and can easily be performed at the bedside, CMR and CCT are important when findings are equivocal. Acute myocardial rupture is a rare, but often lethal, complication in patients with extensive transmural myocardial infarction. If the rupture is contained by the pericardium, a false aneurysm is formed and the patient may survive the event. Extensive thinning of the myocardial wall may lead to a true aneurysm. Differentiation between false and true aneurysms is not always straightforward. In false aneurysms, the orifice is usually smaller than the maximal internal diameter, whereas in true aneurysms the dimensions are similar. Moreover, pericardial enhancement is frequently found in false aneurysms while it rarely occurs in true aneurysms. Ventricular thrombus formation is a frequent complication post myocardial infarction, and may be incidentally found in patients with ischaemic-dilated cardiomyopathy. Early thrombus detection is of paramount importance to avoid neurological and peripheral embolic events, and to initiate anticoagulation therapy. Small-sized thrombi are easily missed by transthoracic echocardiography, particularly when located in the apex or when trapped in the endocardial trabeculations. CCT and LGE CMR facilitate the diagnosis of thrombi, as the blood pool is enhanced by the injection of contrast material, therefore improving the detection of intraluminal masses such as thrombi. The myocardial damage (i.e. necrosis) incurred by the coronary artery occlusion—in particular in transmural infarctions—may cause pericardial inflammation early post-infarction, a condition that should be differentiated from late post-infarction pericarditi: that is, Dressler syndrome. Pericardial abnormalities at CMR include enhancement of pericardial layers and pericardial effusion at the area of myocardial infarction or more diffusely involving the pericardium. Imaging biomarkers of pericardial injury are closely related to blood biomarkers of inflammation (i.e. C-reactive protein [CRP]). Finally, mitral valve regurgitation in infarct patients can be caused by valve ring dilatation due to adverse LV remodelling and/or infarction of the papillary muscle(s).

B

Fig. 15.34  Post-Infarct Aneurysm With Large Thrombus. A 55-year-old-man with a history of inferior myocardial infarction. Cine cardiac magnetic resonance (CMR) (A), and ‘late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR (B) in vertical long-axis direction. The cine images were acquired following contrast administration. Presence of a saccular aneurysm arising from the basal inferior left ventricular (LV) wall (arrows, A and B) with large thrombus (*, A and B) almost completely filling the aneurysm.

394

SECTION A  The Chest and Cardiovascular System

PROGNOSIS ASSESSMENT IN ISCHAEMIC HEART DISEASE In asymptomatic patients as well as in patients with suspected or known CAD risk, assessment and prediction of future cardiac events is important. Traditional risk assessment classifies patients into those at high risk, intermediate risk and low risk. Although this classification helps patients to adapt their lifestyle (primary prevention) or medication (e.g. statins), most high-risk patients will never experience a cardiac event, while patients belonging to the intermediate or low-risk group will not be event-free; hence the need for approaches improving prognosis assessment. Coronary calcium score assessment is well known, well validated and widely used, reflecting the atherosclerotic burden in a patient. In asymptomatic patients without evidence of CAD, coronary calcium scoring adds prognostic information beyond clinical risk factors. In particular, asymptomatic adults belonging to the intermediate-risk group (i.e. 10% to 20% 10-year risk of events) can be ‘upgraded’ or ‘downgraded’, depending on the Agatston calcium score. In symptomatic patients suspected of CAD, the severity of CAD and/or presence of myocardial ischaemia are important prognosticators for future events even though a negative coronary calcium score in symptomatic patients does not exclude obstructive CAD. The severity of CAD assessed by CCT, together with LV ejection fraction, predicts all-cause mortality. In the absence of myocardial ischaemia on SPECT, future cardiac events are highly unlikely. In patients without a history of myocardial infarction but a clinical suspicion of CAD, evidence of ischaemia-related myocardial scarring carries an increased risk for future major adverse cardiovascular event (MACE) independent of the extent of LGE. In patients presenting an ACS in whom myocardial infarction is excluded by cardiac biomarkers and ECG, stress perfusion MRI is an accurate and independent predictor of future cardiac events. Moreover, a substantial number of MRI studies have shown that several parameters other than ejection fraction are important in predicting adverse remodelling and patient outcome following an acute myocardial infarct; these include microvascular obstruction, post-infarction myocardial haemorrhage and myocardial salvage. In patients with ischaemic cardiomyopathy, the extent of myocardial enhancement is a strong and independent predictor of all-cause mortality,

A

B

even in the presence of traditional well-known prognosticators such as ejection fraction, congestive heart failure and age.

ROLE OF CONVENTIONAL CHEST RADIOGRAPHY IN ISCHAEMIC HEART DISEASE Even though ‘advanced’ imaging techniques are nowadays central in the diagnosis of heart diseases, the contribution of conventional chest radiography in evaluating IHD patients should not be neglected. Valuable information can be provided regarding the cardiac size, enlargement of a specific cardiac chamber, or pulmonary filling status, and the chest radiograph can help exclude some pulmonary, pleural or aortic disease such as aortic aneurysm. In ill cardiac patients, bedside radiography can readily be performed. Left-sided cardiac decompensation in patients with recent myocardial infarction or ischaemic cardiomyopathy leads to an apical redistribution of pulmonary vascularisation, onset of pulmonary interstitial and alveolar oedema, and pleural effusion. Chest radiography can closely monitor the effects of therapy, and to demonstrate concomitant pulmonary disease such as infection or acute respiratory distress syndrome (ARDS). It also serves to check the correct positioning of devices such as endotracheal tubes, central venous catheters, pulmonary artery catheters and pacing leads. Infarct-related complications, such as pericardial effusion/hematoma or aneurysm formation, can be detected on chest radiography, although echocardiography, CMR and CCT are definitely superior (Fig. 15.35).

DIFFERENTIAL DIAGNOSIS IN ISCHAEMIC HEART DISEASE In patients suspected of having an ACS, the current American College of Cardiology (ACC)/American Heart Association (AHA)/Unstable Angina (UA)/STEMI guidelines recommend a classification into (1) ‘definite’ ACS, (2) ‘possible’ ACS, (3) chronic stable CAD and (4) non-cardiac cause of chest pain. This classification is based on the patient’s history, physical examination, 12-lead ECG and initial cardiac biomarkers. In patients with normal/non-diagnostic ECG or normal initial biomarkers, however, the question arises whether the symptoms arise from unstable angina pectoris, which is characterised by ischaemia without myocardial damage to release detectable quantities of markers of myocardial injury.

C

Fig. 15.35  Severe Pericardial Effusion. Chest radiography (A), transthoracic echocardiography (B) and cardiac magnetic resonance (CMR) (cine imaging) (C). Presence of severe cardiomegaly caused by lateral displacement of the left heart border. The abnormalities are caused by a moderate to severe pericardial effusion as clearly visible on cardiac ultrasound (*, B). CMR confirms the cardiac ultrasound findings (*, C). Much of the pericardial effusion is left-sided located (maximal width 38 mm), explaining the chest radiograph findings.

CHAPTER 15  Ischaemic Heart Disease

A

C

E

B

D

F

395

Fig. 15.36  Fulminant Myocarditis. Seventeen-year-old man admitted with severe respiratory-related retrosternal chest pain. Increased serum biomarkers (troponin I: 83 µg/L). Coronary angiography shows normal coronary arteries. Short-axis T2-weighted cardiac magnetic resonance (CMR) (A). Short-axis cine imaging post-contrast administration (B). ‘Late’ or ‘delayed’ gadolinium-enhanced (LGE) CMR in cardiac short-axis (C and D), horizontal long-axis (E), and vertical long-axis (F). Presence of diffuse subepicardially located myocardial oedema (arrows, A). LGE CMR shows strong subepicardial enhancement in left ventricle (arrows, C, D, and F), and focal strong enhancement in righ ventricl (arrowhead, D). The subepicardial enhancement is nicely visible on cine imaging post-contrast administration (arrows, B). CMR findings of severe form of acute myocarditis. Myocardial biopsy shows lymphohistiocytic infiltrate.

These patients with ‘possible’ ACS are usually admitted for observation 12 hours or more from symptom onset and stress testing is usually performed to provide evidence of myocardial ischaemia. As an alternative in these patients, CCT can be recommended to demonstrate or exclude significant CAD in those with low or intermediate pre-test probability of CAD, while the role of CCT in patients with a high pre-test likelihood is uncertain. A negative CCT, defined as no CAD or stenosis less than 50%, yields an excellent negative predictive value for ACS or MACE, while in those patients having a positive CCT, ischaemia testing can be subsequently performed. Cardiac CT is also of interest to rule out other causes of chest pain related to pathology of the pulmonary arteries (i.e. pulmonary embolism) and thoracic aorta (i.e. aortic dissection). This so-called ‘triple rule out’ approach needs an adaptation of the administration of contrast agent to assure sufficient enhancement of pulmonary arteries, coronary arteries and thoracic aorta during CT data acqisition. Although promising, the value of triple rule-out CCT in the emergency department is still uncertain. In a small but important group of patients presenting with chest pain and elevated cardiac biomarkers, subsequent coronary angiography reveals normal appearances or non-flow limiting CAD, questioning the underlying cause of the clinical presentation. Possible causes include non-cardiac aetiologies, myocardial infarction with a recanalised coronary

artery and acute myocarditis. In these patients, CMR is now recommended. If patients have experienced an ischaemic event, T2-weighted imaging will show myocardial oedema while myocardial enhancement on LGE CMR is suggestive of myocardial necrosis and the functional consequences can be evaluated with cine imaging. Not infrequently, smaller coronary artery branches are affected that were not initially recognised at coronary angiography. The same CMR approach is of great help in depicting patients with acute myocarditis. These patients show a different pattern of myocardial enhancement on LGE CMR than acute myocardial infarction patients: i.e. midwall/subepicardial enhancement instead of subendocardial enhancement with variable transmural spread (Fig. 15.36). CMR is also of help in patients with takotsubo cardiomyopathy (also called stress cardiomyopathy).

FURTHER READING Achenbach, S., Raggi, P., 2010. Imaging of coronary atherosclerosis by computed tomography. Eur. Heart J. 31, 1442–1448. Anderson, J.L., Adams, C.D., Antman, E.M., et al., 2007. ACC/AHA 2007 Guidelines for the management of patients with unstable angina/non ST-elevation myocardial infarction: a report of the American College of Cardiology / American Heart Association task force on practice guidelines

396

SECTION A  The Chest and Cardiovascular System

for the management of patients with unstable angina/non ST-elevation myocardial infarction): developed in collaboration with the American College of Emergency Physicians, the Society for Cardiovascular Angiography and Interventions, and the Society of Thoracic Surgeons: endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation and the Society for Academic Emergency Medicine. Circulation 116, e148–e304. Bogart, J., Gheysens, O., Dymarkowski, S., et al., 2014. Comprehensive evaluation of hibernating myocardium: use of noninvasive imaging. J. Thorac. Imaging 29, 134–146. Bogart, J., Masci, P.G., Symons, R., et al., 2015. Impact of pericardial injury on inflammatory biomarkers early post myocardial infarction: a cardiovascular magnetic resonance (CMR) study. Int. J. Cardiol. 186, 139–140. Bonow, R.O., Maurer, G., Lee, K.L., et al., 2011. Myocardial viability and survival in ischemic left ventricular dysfunction. N. Engl. J. Med. 364, 1617–1625. Cerqueira, M.D., Weissman, N.J., Dilsizian, V., et al., 2002. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 105, 539–542. Chow, B.J.W., Small, G., Yam, Y., et al., 2011. Incremental prognostic value of cardiac computed tomography in coronary artery disease using CONFORM. COroNary computed tomography angiography evaluation for clinical outcomes: an inteRnational Multicenter registry. Circ. Cardiovasc. Imaging 4, 463–472. Detrano, R., Guerci, A.D., Carr, J.J., et al., 2008. Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N. Engl. J. Med. 358, 1336–1345. Einstein, A.J., Elliston, C.D., Arai, A.E., et al., 2010. Radiation dose from single-heartbeat coronary CT angiography performed with a 320–detector row volume scanner. Radiology 254, 698–706. Eitel, I., Gehmlich, D., Amer, O., et al., 2013. Prognostic relevance of papillary muscle infarction in reperfused infarction as visualized by cardiovascular magnetic resonance. Circ. Cardiovasc. Imaging 6, 890–898. Eitel, I., von Knobelsdorff-Brenkenhoff, F., Bernhardt, P., et al., 2011. Clinical characteristics and cardiovascular magnetic resonance findings in stress (takotsubo) cardiomyopathy. JAMA 306, 277–286. Greenwood, J.P., Maredia, N., Younger, J.F., et al., 2012. Cardiovascular magnetic resonance and single-photon emission computed tomography for diagnosis of coronary heart disease (CE-MARC): a prospective trial. Lancet 379, 453–460. Greenwood, J.P., Ripley, D.P., Berry, C., et al., 2016. Effect of care guided by cardiovascular magnetic resonance, myocardial perfusion scintigraphy, or NICE guidelines on subsequent unnecessary angiography rates. The CE-MARC2 randomized clinical trial. JAMA 316, 1051–1160. Hsu, L.Y., Ingkanisorn, W.P., Kellman, P., et al., 2006. Quantitative myocardial infarction on delayed enhancement MRI. Part II: clinical application of an automated feature analysis and combined thresholding infarct sizing algorithm. J. Magn. Reson. Imaging 23, 309–314. Ibanez, B., James, S., Agewall, S., et al., 2017. 2017 ESC guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation. The task force for the management of acute myocardial infarction in patients presenting with ST-segment elevation of the European Society of Cardiology (ESC). Eur. Heart J. 39, 119–177. Joshi, N.V., Vesey, A.T., Williams, M.C., et al., 2014. 18F-fluoride positron emission tomography for identification of ruptured and high-risk coronary atherosclerotic plaques: a prospective clinical trial. Lancet 383, 705–713. Kim, R.J., Wu, E., Rafael, A., et al., 2000. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N. Engl. J. Med. 343, 1445–1453.

Leschka, S., Koepfli, P., Husmann, L., et al., 2008. Myocardial bridging: depiction rate and morphology at CT coronary angiography— comparison with conventional coronary angiography. Radiology 246, 754–762. McCollough, C.H., Leng, S., Yu, L., et al., 2015. Dual-and multi-energy CT: principles, technical approaches, and clinical applications. Radiology 276, 637–653. Messroghli, D.R., Moon, J.C., Ferreira, V.M., et al., 2017. Clinical recommendations for cardiovascular magnetic resonance mapping of T1,T2,T2* and extracellular volume: a consensus statement by the Society for Cardiovascular Magnetic Resonance (SCMR) endorsed by the European Association for Cardiovascular Imaging (EACVI). J. Cardiovasc. Magn. Reson. 19, 75. Min, J.K., Labounty, T.M., Gomez, M.J., et al., 2014. Incremental prognostic value of coronary computed tomographic angiography over coronary artery calcium score for risk prediction of major adverse cardiac events in asymptomatic diabetic individuals. Atherosclerosis 232, 298–304. Morais, P., Marchi, A., Bogart, J.A., et al., 2017. Cardiovascular magnetic resonance myocardial feature tracking using a non-rigid, elastic image registration algorithm: assessment of variability in a real-life clinical setting. J. Cardiovasc. Magn. Reson. 19, 24. Pfisterer, M.E., Zellweger, M.J., Gersh, B.J., 2010. Management of stable coronary artery disease. Lancet 375, 763–772. Raff, G.L., Hoffmann, U., Udelson, J.E., 2017. Trials of imaging use in the emergency department for acute chest pain. JACC Cardiovasc. Imaging 10, 338–349. Rahimtoola, S.H., 1989. The hibernating myocardium. Am. Heart J. 117, 211–220. Symons, R., Claus, P., Marchi, A., et al., 2018. Quantitative and qualitative assessment of acute myocardial injury by CMR at multiple time points after acute myocardial infarction. Int. J. Cardiol 259, 43–46. Symons, R., Masci, P.G., Goetschalckx, K., et al., 2015. Effect of infarct severity on regional and global left ventricular remodeling in patients with successfully reperfused ST segment elevation myocardial infarction. Radiology 274, 93–102. Symons, R., Morris, J.Z., Wu, C.O., et al., 2016. Coronary CT angiography: variability of CT scanners and readers in measurement of plaque volume. Radiology 281, 737–748. Taylor, A.J., Cerqueira, M., Hodgson, J., et al., 2010. ACCF/SCCT/ACR/AHA/ ASE/ASNC/SCAI/SCMR 2010 Appropriate use criteria for cardiac computed tomography. A report of the American College of Cardiovascular foundation appropriate use criteria task force, the Society of Cardiovascular Computed Tomography, the American College of Radiology, the American Heart Association, the American Society of Echocardiography, the American Society of Nuclear Cardiology, the Society for Cardiovascular Angiography and Interventions, and the Society for Cardiovascular Magnetic Resonance. Circulation 122, e525–e555. Wagner, A., Mahrholdt, H., Holly, T.A., et al., 2003. Contrast-enhanced MRI and routine single photon emission computed tomography (SPECT) perfusion imaging for detection of subendocardial myocardial infarcts: an imaging study. Lancet 361, 374–379. Wahl, A., Gollesch, A., Paetsch, I., et al., 2003. Safety and feasibility of high-dose dobutamine-atropine stress MRI for diagnosis of myocardial ischemia: experience in 1000 consecutive cases. J. Cardiovasc. Magn. Reson. 5, 51. Wang, Y., Qin, L., Shi, X., et al., 2012. Adenosine-stress dynamic myocardial perfusion imaging with second-generation dual-source CT: comparison with conventional catheter coronary angiography and SPECT nuclear myocardial perfusion imaging. AJR Am. J. Roentgenol. 198, 521–529. Yellon, D.M., Hausenloy, D.J., 2007. Myocardial reperfusion injury. N. Engl. J. Med. 357, 1121–1135.

17  The Thoracic Aorta: Diagnostic Aspects Rossella Fattori, Luigi Lovato, Vincenzo Russo

CHAPTER OUTLINE The Normal Aorta, 421 Diagnostic Aspects, 421

THE NORMAL AORTA The aorta is the main artery delivering oxygenated blood from the left ventricle to all parts of the body. In common with other arteries, it has three histologically distinct layers: an intima consisting of a thin endothelial layer; a media containing an elastic lamella, smooth muscle and connective tissue; and a thin outer adventitia made of connective and elastic tissues also containing nerves, lymphatics and the vasa vasorum. The aortic root begins at the upper part of the left ventricle and is approximately 3 cm in diameter. A normal aorta passes superiorly and slightly to the right for approximately 5 cm, then arches posteriorly over the root of the left lung, descending within the thorax beside the vertebral column, gradually achieving the median plane, and becoming the abdominal aorta, after it passes through the aortic hiatus in the diaphragm. The abdominal aorta is approximately 2 cm in diameter; it ends slightly to the left of the median plane at the lower border of the fourth lumbar vertebra by dividing into the right and left common iliac arteries. The aortic root and most of the ascending aorta are contained within the pericardium. The root consists of three sinuses: the right coronary artery arising from the right coronary sinus, the left coronary artery from the left coronary sinus and a noncoronary sinus which is usually located on the right posterolateral aspect. The ascending aorta forms the right mediastinal border on a posteroanterior (PA) chest radiograph. It becomes the aortic arch at the origin of the innominate artery and also gives rise to the left common carotid and left subclavian arteries (LSAs). Approximately three-quarters of people show this ‘normal’ branch pattern of the supra-aortic arteries, but in 20% the innominate and left common carotid arteries have a common origin and in 6% the left vertebral artery arises directly from the aortic arch. The aortic arch ends and the descending thoracic aorta begins immediately beyond the origin of the LSA. At this site the ligamentum venosum (the embryological ductus arteriosus, which closes within a few days of birth) joins the inferior concavity of the aortic arch to the main pulmonary artery. The aorta is fixed at this point. Occasionally the duct may persist as a short diverticulum.

DIAGNOSTIC ASPECTS The last few decades have seen an increasing recognition of thoracic aortic disease among Western people, partly due to greater longevity

Acquired Aortic Abnormalities, 424 Congenital Aortic Abnormalities, 450

and an increased awareness of its clinical importance. Recent technological advances in computed tomography (CT) and magnetic resonance imaging (MRI) have greatly contributed to the increased recognition and pathological understanding of aortic disease. There are at least three main goals of imaging concerning thoracic aortic diseases: disease recognition, preoperative evaluation and imaging follow-up. The appropriate imaging technique depends on which of these aspects is pre-eminent (Table 17.1). Aortic disease often presents as a clinical emergency, with patients becoming rapidly haemodynamically unstable over time. Accordingly, noninvasiveness, diagnostic accuracy and speed are the main properties requested in this setting, together with a comprehensive evaluation of the thoracic aorta (crucial for an accurate assessment before any intervention). Thus the choice of the optimal imaging technique should consider these aspects and various patient-related factors (namely, acute or chronic presentation).

Chest X-Ray and Echocardiography Evaluation of the thoracic aorta has always been very difficult with first-line imaging techniques such as chest x-ray (CXR) and transthoracic ultrasound (except for proximal aortic segments) due to its anatomical location. A CXR may identify only indirect signs of aortic aneurysm or dissection such as a widening of the upper mediastinum or an abnormal aortic contour increase, but it lacks sufficient sensitivity and cannot exclude significant aortic disease, especially in high-risk patients. Therefore additional imaging is almost invariably required for clarification. Moreover, a CXR does not give any information about anatomical details for surgical or endovascular planning. Transthoracic echocardiography (TTE) is limited by its restricted field of view, further reduced by acoustic window limitations in adult patients (e.g. due to chronic pulmonary diseases, surgical scars or obesity), and has a typical ‘blind spot’ at the level of the proximal aortic arch due to superposition of air in the right bronchus. Transoesophageal echocardiography (TOE) is superior to TTE for thoracic aorta evaluation, and it can be easily performed at the bedside. However, it is partially invasive and not well tolerated by patients. Although echocardiography is routinely used for follow-up of aortic root and proximal ascending aorta aneurysms in chronic diseases, its narrow field of view prevents a comprehensive evaluation of the thoracic aorta. Furthermore, operator dependence limits the overall accuracy. However, ultrasound still plays the main role for valvular assessment in thoracic aortic diseases (coexisting valvular

421

422

SECTION A  The Chest and Cardiovascular System

TABLE 17.1  Diagnostic Goals of Aortic

Diseases and Corresponding Features Requested to an Imaging Test Diagnostic Aspects

Appropriate Features

Disease recognition

Diagnostic accuracy Noninvasiveness Rapidity Logistic convenience Availability Multiplanarity Wide field of view High spatial resolutions 2D/3D reformatting images Measurement reproducibility Noninvasiveness and low radiation burden Availability Diagnostic accuracy

Preoperative evaluation

Imaging follow-up

disease or valvular involvement in type A aortic dissection or ascending aortic aneurysm), whereas intraoperative TOE is often fundamental for endovascular or surgical aortic treatment. In summary, CXR and ultrasound, although easy to perform, noninvasive and inexpensive, provide some helpful information, but alone they cannot provide comprehensive information about aortic disease.

Angiography For many decades, angiography was the only available imaging technique for diagnosis and preoperative evaluation of aortic diseases. It was intrinsically invasive, relatively costly and needed a well-organised and experienced team. It provided only limited information about the aortic wall, because angiography provides only ‘luminographic’ data. Thus it provided limited information about an intramural haematoma (IMH) and could be misleading in aortic dissection with a complete false lumen thrombosis. Over the past 30 years, with the advent of CT and MRI and their technological evolutions, angiography has become progressively abandoned by physicians for diagnostic purposes.

Computed Tomography and Magnetic Resonance Imaging CT and MRI combine noninvasiveness with high spatial and temporal resolution and can provide information about the entire length of the thoracic or thoracoabdominal aorta. Multiplanar images allow precise measurements of aortic diameters, preferably taken perpendicular to the longitudinal axis in more than one plane to avoid source of errors. In fact, the aorta has such variable geometry that it cannot usually be entirely visualised in a single plane. MR angiography (MRA) and CT angiography (CTA) have further enhanced the noninvasive visualisation of vascular structures with a high degree of spatial and contrast resolution in all three dimensions. Different two-dimensional (2D) and three-dimensional (3D) processing techniques, such as multiplanar reformation (MPR), maximum intensity projection (MIP) and volume rendering (VR), play an important role for preoperative planning. Although thin-slice MPRs in any arbitrary plane provide high anatomical resolution, they cannot visualise the entire aorta in a single plane. MIP images of appropriate slab thickness, while demonstrating the whole aorta, yield information only of perfused lumina similar to digital subtraction angiography (DSA); as a threshold

technique, MIP images may not discriminate lower-density intraluminal structures such as thrombus, plaque or IMH. MIP images also do not provide any information about adjacent structures and their important relationships. VR is a different 3D reconstruction technique where all tissues can be simultaneously represented. Using adequate filters, metallic stents or clips do not create artefacts, and aortic wall lesions can be differentiated from the lumen. VR provides 3D anatomical information displaying the spatial relationship between aortic lesions and branching vessels. Finally, curved reformations can reconstruct even the most tortuous vascular structure in a single plane. All these processing and display options make CT and MRI measurements highly reproducible and less operator-dependent than ultrasound. CT and MRI currently form the backbone of thoracic aortic imaging. Both techniques show comparable results in terms of diagnostic accuracy, measurement reproducibility and anatomical detail definition. MRI does not use ionising radiation, and gadolinium is less nephrotoxic than iodinated contrast agents. Consideration should only be given to patients with severe renal dysfunction (creatinine clearance 8 days), methaemoglobin shows high signal intensity. However, when the signal intensity is medium to low, it can be difficult to distinguish IMH from mural thrombus. T2 weighted spin-echo sequences may help in differentiating the two entities: signal intensity is high in recent haemorrhage but low in chronic thrombosis (Fig. 17.14).

Computed Tomography CT, like MRI, has proven to be highly accurate in the diagnosis of IMH, with comparable sensitivity and specificity. In the suspect of IMH, it is important to perform unenhanced CT: in the acute phase the haematoma appears as a crescent-like aortic wall thickening typically hyperdense

on unenhanced CT with respect to the aortic lumen, whereas after enhancement the density of wall and lumen are reversed, with the IMH remaining unenhanced, unlike the false lumen in aortic dissection (Fig. 17.15). The differentiation between IMH and a completely thrombosed false lumen may be very difficult and the following findings are useful for differential diagnosis. IMH maintains a constant circumferential relationship to the wall (subintimal lesion), whereas the thrombosed false lumen tends to longitudinally spiral around the aorta. Secondly, IMH does not reduce the lumen, which maintains its regular shape, whereas the false lumen can variably compress the true lumen (Fig. 17.16). MDCT is highly accurate for the detection of small circumscribed intimal defects that can appear at multiple levels of the IMH; they can enlarge over time, evolving towards aneurysmal dilation, and may eventually represent a patient subgroup with worse prognosis. The diagnosis of IMH is mainly based on axial images, but 2D reformatted images may be useful to evaluate the extent of IMH and its relationships with aortic branches. MRA, like conventional angiography, has poor value in IMH diagnosis because it provides only luminal information. An appropriate adaptation of the window level in reformatted images can help to identify the wall haematoma.

Penetrating Atherosclerotic Ulcer An aortic ulcer is generated by erosion of an atheromatous plaque disrupting the internal elastic lamina, exposing the media to pulsatile

CHAPTER 17  The Thoracic Aorta: Diagnostic Aspects

A

B

Fig. 17.16  Differentiation Between False Lumen Thrombosis and Intramural Haematoma (IMH). Magnetic Resonance Angiography of a type B aortic dissection (A): the true lumen is significantly reduced along the descending aorta. Sagittal oblique MDCT image (B) shows no reduction of the aortic lumen along the iMH in descending aorta.

A

431

arterial flow and subsequent haematoma formation. This is distinguished from an atheromatous plaque by the presence of a focal, contrast medium–filled outpouching surrounded by an IMH. An atheromatous plaque does not extend beyond the intima, is frequently calcified and lacks an IMH. The extension of the ulceration to the medial layer can also evolve in localised dissection or even break through into the adventitia, creating an aortic pseudoaneurysm. If the adventitia ruptures, only the mediastinal tissue can contain the haematoma; otherwise, the rupture is complete. Penetrating atherosclerotic ulcers (PAUs) are mainly located in the descending aorta but may be also seen in the aortic arch. Ulcers can be multiple and are frequently associated with a severe atherosclerotic aortic wall. The imaging diagnosis of PAUs is based on the visualisation of a crater-like, contrast-filled outpouching with jagged edges, of variable extension, which may result in a large pseudoaneurysm (Fig. 17.17). Mural thickening can be associated (localised haematoma), as well as aortic dissection. Differently from IMH, conventional angiography has a good sensitivity for PAUs, but both CT and MRI are better suited to evaluate the presence of associated lesions like atherosclerotic disease extent, localised haematoma and dissection. Unenhanced CT has the advantage of visualising intimal calcification displacement. MRA and CTA, including 2D and 3D reformatted images, are important for analysing the often-complex spatial relationships between the ulcers and the aortic branches Fig. 17.18).

B

Fig. 17.17  Penetrating Atherosclerotic Ulcers. Multidetector computed tomography sagittal oblique multiplanar reformation image (A) shows diffuse multiple finger-like ulcers of the descending thoracic aorta. A severe and diffuse aortic wall atheroma is present. A sagittal oblique black-blood fast spin-echo (BBFSE) image (B) demonstrates a large ulcer of the descending aorta developed in a pseudoaneurysm (arrows).

432

SECTION A  The Chest and Cardiovascular System

A

B

Fig. 17.18  Penetrating Ulcers of the Aortic Arch. A multidetector computed tomography volume-rendering reconstructed image (A) clearly displays the relationships of the ulcer with the epiaortic vessels. A computed tomography axial image (B) defines the aortic wall alterations at the level and on both sides of the ulcers.

The various aortic diseases as described so far are related to each other: PAU and IMH are both potential precursors of dissection. Because they are lesions that involve a more external portion of the aortic wall, they are more prone to rupture than aortic dissection and need careful monitoring in the acute phase. Moreover, IMH can evolve into PAU (Fig. 17.19) and even dissection.

Traumatic Aortic Injury Traumatic aortic injury (TAI) can result from both penetrating and blunt chest injuries. Motor vehicles accidents are one of the main causes of TAI. In the United States there are approximately 40,000 motor vehicle deaths yearly, and it has been estimated that 20% of deaths are caused by aortic rupture. A number of mechanisms with an increased risk for TAI have been proposed, of which the most important is rapid and sudden deceleration during the impact (especially at speeds greater than 30 mph). The aortic segment subjected to the greatest strain is the isthmus, where the relatively mobile thoracic aorta is fixed by the ligamentum arteriosum: 90% of traumatic aortic ruptures occur here. Another mechanism leading to TAI consists of torsion caused by displacement of the heart to the left during anteroposterior (AP) compression, which typically involves the ascending aorta close to the innominate artery or immediately superior to the aortic valve; this is seen with vertical deceleration caused by falls from large heights (especially greater than 10 ft). Other aortic segments are less commonly involved like the distal descending (diaphragmatic) aorta or the abdominal infrarenal segment, suggesting underlying mechanisms other than the sudden deceleration strain: one process refers to the ‘osseous pinch’, leading to compression of the heart and aorta between the sternum and vertebral column with a trauma extended from the intima to the adventitia (Table 17.3). In most patients (80% to 90%) there is complete rupture of the aorta, with death occurring at the scene of the accident. Those patients that reach the hospital alive have injuries that vary from a simple intimal lesion, an IMH to a false aneurysm when the laceration extends through the media into the adventitia (which may be the only layer maintaining

aortic integrity). Periaortic haemorrhage is frequently seen, irrespective of the type of lesion.

Imaging The clinical diagnosis of TAI can be difficult due to the lack of specific symptoms or signs in many patients. CXR gives only indirect signs of an aortic lesion like haemomediastinum, which is insufficiently specific and is more likely the consequence of venous bleeding relative to the thoracic trauma. Moreover, chest radiographs in patients with suspected TAI are taken in the supine position, so the interpretation of mediastinal widening can be problematic—especially in obese patients. Upper limits for a normal mediastinal width or the ratio of the mediastinal width to chest width at the level of the aortic arch (M/C ratio) have been proposed (8 cm and 25%, respectively) but have a wide range of reported sensitivities and specificities. In any case the initial CXRs performed as part of a trauma series may suggest an aortic involvement with satisfactory sensitivity (80% to 90%), showing the displacement of the nasogastric tube by the haematoma (Fig. 17.20). This makes chest radiography a useful screening tool for mediastinal haemorrhage, even though a normal mediastinum does not exclude a significant aortic injury. Thoracic aortography is no longer the preferred diagnostic test: first because it provides less information than CT and MRI about wall alterations and anatomical preoperative evaluation and, secondly, because of its lower accuracy, with sensitivities ranging from 84% to 96%, with false positives caused by prominent ductus diverticulum, severe aortic atheroma or double densities from overlapping adjacent vessels and false negatives due to poor opacification of the aorta or small intimal defects. Examples are shown in Figs 17.21 and 17.22. TOE has a sensitivity of 91% and specificity of 98% for demonstration of isthmic aortic injuries. TOE has the advantage that can be performed at the patient’s bedside in 15 to 20 minutes, even in highly unstable patients. However, it may be contraindicated in the presence of severe facial injuries or unstable cervical spine fractures. The entire aortic circumference may not be adequately visualised in approximately 30%

CHAPTER 17  The Thoracic Aorta: Diagnostic Aspects

A

B

C

D

E

F

433

Fig. 17.19  Multidetector computed tomography multiplanar reformation axial (A to D) and sagittal oblique (E and F) images of an intramural haematoma of the descending thoracic aorta (A, C and E). After 8 days from symptom onset the intramural haematoma has partially reabsorbed but, at the level of the isthmic and midthoracic descending aorta, has evolved into a penetrating ulcer (arrows in B, D and F).

of patients, while the aortic arch is not easily displayed, thus limiting preoperative evaluation. On the other hand, TOE is extremely useful as an intraoperative assistant tool for guiding endovascular intervention, providing excellent visualisation of the proximal aorta for accurate placement of stent grafts in relation to aortic branch vessel origins.

With the exception of extremely unstable patients, CT and MRI are the ideal imaging investigations for TAI, with a diagnostic accuracy approaching 100%. They can demonstrate both indirect signs such as mediastinal haematoma and direct signs of aortic trauma, especially giving high-definition images of the aortic wall alterations such as IMH,

434

SECTION A  The Chest and Cardiovascular System

TABLE 17.3  Presley Trauma Center CT

Grading System: Grades and CT Findings Grade

Subgrade

CT Findings

Grade I Normal aorta

Ia

Normal thoracic aorta No mediastinal haematoma Normal thoracic aorta Mediastinal haematoma (para-aortic) Small (50% within 48 hours) if untreated. Fatal complications include aortic rupture, cardiac tamponade, acute aortic regurgitation and acute myocardial infarction. Involvement of the arch vessels harbours a high risk for neurological complications. Type B dissection.  The current treatment of acute type B dissections is based on a complication-specific approach. In uncomplicated type B dissection (no evidence of rupture or branch vessel ischaemia), medical treatment is initiated because both medical and emergency surgical management are associated with similar mortality rates. Patients who fail under medical treatment (persistent pain and/or progression of dissection) or develop complications are referred for either surgical or endovascular intervention. Early surgery is recommended for patients with Marfan syndrome.

Endovascular treatment of type B dissection.  Surgery for type B dissection is associated with mortality exceeding 50% with high risk for organ ischaemia. Endovascular techniques have been successfully used for the treatment of type B dissection with reduced morbidity and mortality. The three techniques used are stent insertion, stent-graft insertion and/or fenestration of the intimal flap. The indications for stent or stent-graft placement in type B dissection are twofold: 1. Contained rupture. Persistent flow in the false lumen is associated with aneurysmal dilatation and increased risk of rupture of the false lumen (20% to 50% of patients within 1 to 5 years). Placement of a stent graft covering the site of the entry tear can promote thrombosis of the false lumen and stabilisation of the dissection, thereby reducing the risk of rupture. Acute type B dissections are the most appropriate group to treat as the dissection flap is thin and mobile, and will readily fuse with the aortic wall. In chronic dissections, the flap becomes thickened and rigid and thus endovascular stenting is less likely to result in complete false lumen exclusion. 2. Occlusion of branching vessels. The treatment in patients with impeding branch vessel ischaemia is dependent on the cause. Dynamic obstruction results from true lumen collapse. Sealing the entry tear with stent-graft placement directs blood flow back into the true

446

SECTION A  The Chest and Cardiovascular System

A

B

C

D

E

F

G

H

Fig. 17.38  Magnetic Resonance (MR) Imaging Follow-Up After Endovascular Treatment for Type B Aortic Dissection. MR black-blood fast spin-echo (BBFSE) images (A and B) identify a bright signal heterogeneity (*) near the stent graft, but only steady-state free-precession (SSFP) (C) and Contrast Enhancement Magnetic Resonance Angiography (CEMRA) (D) images can diagnose the endoleak (arrows). However, only multidetector computed tomography, through the optimal visualisation of stent-graft material (E to H), may identify the full nature of endoleak on multiplanar reformation (E and F) and volume-rendering (G and H) images: a type III due to segment disconnection (arrows). This can result from long-term wear and tear and changes in aneurysm morphology causing dislocations of modular components. Repeat stent grafting or open surgery is required.

lumen, increasing the size of the true lumen, moving the dissection flap away from the branch vessel and in that way relieving branch vessel ischaemia. Static obstruction can be successfully treated by direct stent insertion in the compromised vessel via the true lumen. In cases where the dissection extends distally to cause lower limb ischaemia, direct access to the true lumen is gained by directly accessing the involved side with weakened or absent femoral pulse. Currently, stent grafts are mostly used in patients who have indications for surgical intervention. Given that the majority of patients undergo thrombosis of the false lumen following stent-graft placement, it is not inconceivable that, in the future, this may become the treatment of choice for all patients with acute type B dissection. More recent data on middle- and long-term results of endovascular treatment versus medical therapy for stable type B dissection seem to favour endovascular treatment. However, more prospective and randomised comparative follow-up data are required about long-term patient outcome and device durability. Since the introduction of stent grafts, percutaneous fenestration has been less frequently used in the management of branch vessel ischaemia due to true lumen compression. Patients who are managed conservatively require long-term follow-up with either CT or MR to monitor the false lumen for aneurysmal dilatation and distal extension of the dissection over time. Equally, all patients who undergo endovascular treatment require long-term follow-up to ensure integrity of the device.

Inflammatory Diseases of the Aorta and Midaortic Syndrome A number of conditions can lead to an inflammatory aortitis, including ergotism, radiation fibrosis, syphilis, tuberculosis, giant cell arteritis and Buerger, Behçet, Cogan and Kawasaki disease. There are also congenital inflammatory diseases that affect the aorta, such as Ehlers– Danlos syndrome and Marfan syndrome, as well as neurofibromatosis. In the acute phase, aortitis may mimic an acute aortic syndrome, especially an IMH. Differential diagnosis can be difficult if based on clinical symptoms and laboratory tests alone. CT and MR imaging are very helpful for differentiating these two pathological entities: IMH has a typical crescent-like morphology, whereas in aortitis concentric wall thickening is usually observed (Fig. 17.40). Tissue characterisation with T1 and T2 BB sequences may aid to discern haemorrhagic products from mere inflammation.

Midaortic Syndrome Midaortic syndrome is characterised by segmental narrowing of the proximal abdominal aorta and ostial stenosis of its major branches (Fig. 17.41). It is usually diagnosed in young adults but can present in childhood. Clinical presentation and radiological findings are dependent on the underlying disease, but hypertension is a common feature in all patients.

CHAPTER 17  The Thoracic Aorta: Diagnostic Aspects

A

B

C

D

E

F

447

Fig. 17.39  Computed tomography (CT) volume-rendering (A and D), CT maximum intensity projection (B and E) and angiography (C and F) images of a patient with acute traumatic aortic injury before (A to C) and after (D to F) endovascular treatment. Complete thrombosis of the excluded lesion is demonstrated with almost complete shrinkage at follow-up CT (E). The left subclavian artery was covered by the stent graft (*).

Congenital aortic coarctation is a very uncommon cause of midaortic syndrome, in which the aortic narrowing occurs in the thoracic or abdominal aorta. It may be seen in fetal alcohol syndrome and then associated with intellectual disability. Granulomatous vasculitis (Takayasu disease).  It is a chronic inflammatory disease that involves the aorta, its branches and the pulmonary arteries, causing varying degree of stenosis, occlusion or dilatation of the involved vessels; aetiology and precise pathogenesis are unknown. It is more common in parts of the world with a high incidence of tuberculosis but also occurs more frequently in Japan. It is predominantly a disease of young adults but may also affect children. It is very rare in infancy. The female-to-male ratio varies from 9 : 1 in reports from Japan to 1.3 : 1 in India. The pattern of vessel involvement also varies in different parts of the world. The involvement of the aortic arch and its branches is common in Japan, whereas the thoracoabdominal aorta is mainly

involved in patients from Korea and India. It is not known whether this variation reflects differing causes of the disease or differing Human Leucocyte Antigen (HLA)-associated genetic subtypes. Racial variation also occurs, the disease being relatively uncommon in Caucasians. The initial site of inflammation is around the vasa vasorum in the media and adventitia but, later, nodular fibrosis in all layers of the arterial wall is seen and the intima can obliterate the lumen. The diagnosis depends on typical angiographic morphology, a history or presence of constitutional symptoms suggestive of a systemic illness and the differential diagnosis of other, similar conditions as listed previously. Atherosclerosis of the aorta is distinguished on clinical and morphological grounds, but secondary atherosclerotic changes may occur in older patients with Takayasu arteritis. The radiological features occur late in the course of the disease and include luminal irregularity, vessel stenosis, occlusion, dilatation or aneurysms in the aorta or its primary branches (Fig. 17.42).

448

SECTION A  The Chest and Cardiovascular System

A

B

C

D

Fig. 17.40  Axial computed tomography (CT) images of descending thoracic aorta: imaging aspects of aortitis (A and B) and intramural haematoma (IMH) (C). Note the circumferential thickening and enhancement of aortic wall in acute aortitis (A), hyperdense on unenhanced CT (B), while IMH appears as a typical crescent-like thickening of the aortic wall (C).

Fig. 17.41  Longitudinal Abdominal Ultrasound Examination in a 12-Year-Old Girl Presenting With Hypertension. There is clearly a stenosis at the origin of the superior mesenteric artery and narrowing and irregularity of the aorta. This is characteristic of midaortic syndrome. The patient had café au lait spots consistent with the diagnosis of neurofibromatosis.

Neurofibromatosis of the abdominal aorta and some other causes of midaortic syndrome may produce an identical angiographic picture in children. Based on angiographic morphology, Takayasu arteritis is divided into type I (involving the aortic arch and its branches), type II (thoracoabdominal aorta and its branches) and type III (involving lesions of both types I and II). Involvement of pulmonary arteries, in addition to any of the aforementioned types, is grouped as type IV. The infrarenal aorta and the iliac vessels are usually not involved in Takayasu arteritis. Similarly, the inferior mesenteric artery (IMA) is rarely involved. Unlike coarctation of the aorta, intercostal collaterals rarely occur as the diffuse intimal disease in the aorta also involves the ostia of these intercostal vessels. Aortic intimal calcification may be seen. Saccular or fusiform aneurysms of the aorta occur in 2% to 26% of cases and usually coexist with stenotic lesions. Aneurysms without stenosis occur in 1% to 2% of cases. Pseudoaneurysm or dissection of the aorta is extremely rare. Imaging.  CT, ultrasound and particularly contrast-enhanced MRI and MRA provide information on mural changes of the vessels and have largely replaced angiography for diagnostic and monitoring purposes (Fig. 17.43). FDG PET may be useful, especially in cases of fever of unknown origin, by illustrating increased FDG uptake in the involved vessels. Prognosis and treatment.  Takayasu arteritis in children has a mortality between 10% and 30%. The prognosis has significantly improved due

CHAPTER 17  The Thoracic Aorta: Diagnostic Aspects

449

A

Fig. 17.42  Multidetector computed tomography volume-rendering reconstruction of a 20-year-old woman presenting with right arm ischaemia and hypertension with diagnosis of Takayasu disease. Thickening of all layers of the proximal and mid descending thoracic aorta wall produced segmental lumen reduction. B

to interventional procedures for the treatment of renal and aortic stenosis. Long-term follow-up data on children are not available. Five- and ten-year survival in adults is 91% and 84%, respectively. Severe hypertension, aortic regurgitation, retinopathy, aneurysms or cardiac involvement are predictors of poor outcome. In the absence of these complications, 80% of patients remain stable for years, but approximately 20% show progression. In the acute phase, treatment with corticosteroids leads to clinical remission in 60% of cases. Cytotoxic drugs can also be used in resistant cases. The major morbidity and mortality of Takayasu arteritis results from stenosis and occlusion of the aorta, renal and carotid arteries. Interventional radiological techniques for stenosed segments have revolutionised the treatment of Takayasu arteritis. Surgical treatment is not preferred for Takayasu arteritis because of the diffuse, inflammatory and possibly progressive nature of the disease, except for otherwise therapy-resistant, symptomatic, stenotic lesions and large aneurysms. Von Recklinghausen disease (type 1 neurofibromatosis).  It can be distinguished from other causes of midaortic syndrome by the presence of café au lait skin lesions and neurofibromas. Approximately 2% of patients develop vascular abnormalities, including renal, aortic and mesenteric stenoses (Fig. 17.44). Vessels are surrounded by neurofibromatous or ganglioneuromatous tissue in the adventitia. Alagille syndrome (a multisystem autosomal dominant disorder caused by mutations in the JAG1 gene on chromosome 20p12) and Williams syndrome (a rare genetic condition estimated to occur in 1 in 20,000 births) are both associated with aortic coarctation (thoracic or abdominal).

Fig. 17.43  Magnetic Resonance (MR) Imaging of Aortitis. Black-blood fast spin-echo axial image after medical therapy for aortitis (A) shows slight concentric thickening of the ascending aorta. Low signal of the aortic wall indicates absence of active inflammation. MR angiography of visceral vessels (B) shows a residual focal stenosis at the origin of coeliac trunk.

Aortic Occlusive Disease Atherosclerosis is the predominant cause of chronic aortic occlusive disease (more than 90% of cases), with Takayasu disease (see earlier) accounting for the rest; acute occlusion may result from an aortic bifurcation ‘saddle’ embolus or in situ aortic thrombosis.

Chronic Aortic Occlusive Disease Atherosclerotic aortic occlusive disease affects a younger population than this, generally presenting with lower limb arterial disease. Patients are typically female, heavy smokers with hyperlipidaemia; they have a small infrarenal aorta and hypoplastic iliofemoral arteries (hypoplastic aortoiliac syndrome). The infrainguinal arteries are ‘protected’ by the aortic lesion and are typically disease free. Patients present with chronic lower limb ischaemia and are graded according to the severity of their disease. Aortic occlusive disease is largely associated with Fontaine grade I or II symptoms. Symptoms of critical limb ischaemia (grade III or IV) are unusual at initial presentation as these are associated with both suprainguinal and infrainguinal disease. Although the management of grade I and IIa patients is based solely on risk factor modification, patients with grades IIb–IV are investigated

450

SECTION A  The Chest and Cardiovascular System vary from angioplasty alone, angioplasty with selective stenting, to primary stenting. Angioplasty alone is used in short focal stenoses (5 cm, presence of haustra

Small-Bowel Dilatation Small-bowel dilatation may be due to mechanical obstruction or paralytic ileus/pseudo-obstruction. In paralytic ileus, characterised by lack of an obstructing lesion, small and large bowel may be dilated. Peritonitis and a postoperative abdomen are common causes but other causes include small-bowel ischaemia, metabolic disturbances, renal failure, drugs such as morphine and general debility. CT is valuable in identifying an abrupt calibre change—the transition point—proximal to which bowel is dilated and distal to which bowel is collapsed. Such features indicate small-bowel obstruction (SBO). Careful inspection of the transition point at CT usually aids with defining the aetiology of SBO. The causes of SBO are myriad but can be largely divided into mural lesions (e.g. tumour, stricture due to Crohn’s disease, irradiation), luminal (bezoar, gallstone, Ascaris lumbricoides bolus, intussusception) and extrinsic (adhesions, hernia, volvulus, abdominal malignancy). In the developed world, the most common cause of SBO is adhesions from previous surgery, with SBO resulting from strangulated hernia occurring in less than 10% of cases in the developed world compared with 75% in the underdeveloped world. Mechanical obstruction of the small bowel normally causes small-bowel dilatation, with an accumulation of both gas and fluid and a reduction in the calibre of the large and/or small bowel, beyond the site of obstruction. On the plain radiograph, it is important to look for a hernia, which may be identified as a gas-filled

B Fig. 18.11  Small-Bowel Obstruction. (A) Supine abdominal radiograph (AXR). There are distended loops of small bowel, identified by its central position, multiple loops and thin and frequent valvulae conniventes. Large-bowel obstruction (B). AXR showing significantly dilated bowel loops with visible haustration.

viscus below the level of the inguinal ligament (Fig. 18.12). Visualisation of a hernia does not always mean that it is the cause of the SBO. Dilated loops of small bowel are readily identified if they are gas-filled on the supine radiograph. The string of beads sign, caused by a line of gas bubbles trapped between the valvulae conniventes, is seen only when very dilated small bowel is almost completely fluid filled, and is virtually diagnostic of SBO. In a proportion of patients with SBO, the plain radiograph appears normal or only equivocally abnormal, since the dilated loops are mainly fluid filled. Completely fluid-filled loops are not easily identified on plain radiographs, but are readily seen on CT (Fig. 18.13). Overall, the plain AXR has a sensitivity of approximately 66% for SBO. Where a gallstone is the cause of obstruction, this is known as gallstone ‘ileus’ (technically an obstruction, not an ileus). The gallstone passes into

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract the duodenum by eroding the inflamed gall bladder wall, thereby bypassing the common bile duct. This condition is well known to clinicians but relatively rare, accounting for approximately 2% of patients presenting with SBO. The diagnosis is frequently delayed or missed even though the characteristic radiological features of gallstone ileus are present in 38% of cases. Over half the patients will have plain radiographical evidence of intestinal obstruction and about one-third will have gas present in the biliary tree (Fig. 18.14A). Gas in the biliary tree (pneumobilia) can be recognised by its branching pattern, with gas being more prominent

Fig. 18.12  Small-Bowel Obstruction: Strangulated Right Inguinal Hernia, Supine Abdominal Radiograph. Small-bowel dilatation with a grossly dilated loop passing down into the right inguinal region. Patient was an 80-year-old woman with abdominal pain and vomiting for 5 days.

A

B

469

centrally. Gas in the portal vein, from which it must be distinguished, tends to be peripherally located in small veins around the edge of the liver. The obstructing gallstone, which is frequently located in the pelvic loops of ileum, can be identified in approximately one-third of patients on plain AXRs. Pneumobilia is more commonly caused by previous sphincterotomy or biliary surgery and may also be seen with perforation of a peptic ulcer into the common bile duct and a malignant fistula. Pneumobilia is also seen in the elderly due to a lax sphincter of Odi, often due to the previous asymptomatic passage of a gallstone. CT has a much higher sensitivity for pneumobilia of 95%–100%. In gallstone ileus, obstructing gallstones tend to be 2–3 cm in size and are frequently found in the terminal ileum; they may be difficult to differentiate from bowel content as less than 15% are calcified (see Fig. 18.14B). Intussusception in adults nearly always develops as a result of a neoplasm of the bowel. Neoplasms arising in the submucosa such as lipoma, lymphoma and melanoma metastases may act as lead points for intussusception. An intussusception may show on a plain radiograph as a soft-tissue mass, possibly part-outlined by gas (Fig. 18.15). If the intussusception is orientated end-on, a target sign may be seen comprising two concentric circles of fat density alternating with soft-tissue density. In adults, intussusception when seen on CT (sometimes with the help of multiplanar 3D reconstruction) may show the characteristic feature of the intussusceptum bringing mesenteric fat into the lumen of the intussuscipiens. An underlying cause for the intussusception may also be apparent. Intussusception is occasionally diagnosed on CT during investigation of abdominal pain, where the diagnosis has not been suspected (Fig. 18.16). For the more common paediatric intussusception, the reader is referred to Section G of this book. US can detect fluid-filled loops of small bowel but is not usually definitive in diagnosing the cause of the obstruction; therefore, it is not usually recommended. If complete SBO has been diagnosed and the patient is to undergo surgery, further radiological examination is not strictly necessary, although most surgeons now find a preoperative diagnosis by CT very useful. If there is clinical doubt, partial obstruction

C

Fig. 18.13  Small-Bowel Obstruction: Computed Tomography. (A) Multiple dilated small-bowel loops are filled with fluid. (B) The small-bowel obstruction is caused by strangulation of a right inguinal hernia. (C) Computed tomography scout image. Only a few of the distended loops are visible because most are filled with fluid.

470

SECTION B  Abdominal Imaging

A

A

B B Fig. 18.14  Gallstone Ileus. (A) Abdominal radiograph showing free air within linear air in right upper quadrant consistent with pneumobilia. (B) Contrast-enhanced CT showing evidence of small-bowel obstruction. In addition, gas can be identified within the common bile duct (black arrow). A gallstone is seen in the distal ileum with laminar calcification (white arrow).

Fig. 18.15  Intussusception. (A) Supine abdominal radiograph. A softtissue mass is demonstrated extending across the upper abdomen and a crescent of gas is identified surrounding the head of the intussusception. (B) Barium enema. Huge intussusception identified in a 40-year-old man who had suffered abdominal pain for several months. A carcinoid tumour 10 mm in diameter was found to be responsible for the intussusception.

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract

471

A

A

B

B Fig. 18.16  Ileocolic Lymphoma Leading to Ileocolic Intussusception. (A) US image of the right iliac fossa showing the pseudotumour sign. The ileum can be seen centrally (arrow), surrounded by mesenteric fat that is hyperechoic, all within the thickened ascending colon. (B) Computed tomography showing oral contrast medium in the ileal lumen, the surrounding mesenteric fat accompanying the intussusceptum and the thickened ascending colon, which is the intussuscipiens (arrow).

or when conservative management is planned, CT has a valuable role. CT will demonstrate dilated small-bowel loops, whether fluid or gas filled, and will add further information regarding the site and level of obstruction, and frequently the cause (Fig. 18.17). Importantly, CT can add information on strangulation of a smallbowel loop, a sign that surgery is urgently required. The mortality of SBO with strangulation rises from 5%–8% to 20%–37% in comparison to SBO without strangulation, and mortality rises with treatment delay. CT signs of a closed loop include small-bowel dilatation, V-shaped or radial fluid-filled loops, mesenteric vessels converging towards the point of obstruction and a triangular loop with or without a whirl or beak. Where there is strangulation, the bowel wall becomes thickened and may be of high attenuation due to haemorrhage. Gas may be seen within the bowel wall and mesentery and there may be congestion of the mesentery attached to the loop. There may be free peritoneal fluid whether strangulation is present or not.

Fig. 18.17  Small-Bowel Obstruction Due to Jejunal Phytobezoar. (A) Computed tomography through the upper pelvis demonstrating a filling defect within dilated jejunum (arrow). The bowel calibre was normal distal to this. (B) The phytobezoar after surgical removal.

Where there is malignancy, additional staging information is gained by CT with the detection of lymphadenopathy, peritoneal deposits and other metastatic lesions. Adhesions are not visualised with certainty by any imaging technique, and are usually diagnosed on the basis of clinical history and exclusion. However, CT may demonstrate angulated and tethered loops, which suggest the presence of adhesions. CT and MR enterography or enterocolysis represent more recent techniques used to evaluate the small bowel, and are more sensitive than conventional CT in detecting intraluminal lesions such as polyps, neuroendocrine lesions such as carcinoid tumours or metastatic lesions to the mucosa or submucosa.

Large-Bowel Dilatation There are numerous causes of large-bowel dilatation without obstruction, including paralytic ileus and pseudo-obstruction.

Pseudo-Obstruction Pseudo-obstruction usually occurs in elderly patients. It mimics intestinal obstruction clinically and on AXR. The plain radiographical appearances can be dramatic, showing a very dilated colon together with small-bowel distension. CT or colonoscopy is usually required to exclude mechanical

472

SECTION B  Abdominal Imaging

obstruction and to prevent unnecessary laparotomy. The caecum may exceed the critical diameter of 9 cm, when perforation is imminent, and surgeons should be alerted to this possibility.

Large-Bowel Obstruction Large-bowel obstruction is much less common than SBO. In the USA and Great Britain, carcinoma of the colon is the commonest cause of large-bowel obstruction, with approximately 60% of such carcinomas being situated in the sigmoid colon. Diverticulitis is the second most common cause of obstruction. In the USA, volvulus accounts for only 10% of cases of colonic obstruction, whereas for underdeveloped countries a figure of 85% is quoted. Adhesive obstruction of the large bowel is very unusual and obstruction is much more common on the left side of the colon than the right. The plain radiographic findings depend on the site of obstruction and whether the iliocaecal valve is competent. In a minority of patients, the iliocaecal valve remains competent and despite increasing intracolonic pressure and marked distension of the caecum, the small bowel is not distended; however, in most cases even with a closed iliocaecal valve, the small bowel is distended. In patients with an incompetent ileocaecal valve, the caecum and ascending colon are not unduly distended, but there is marked small-bowel distension. The obstructed colon almost invariably contains large amounts of air and can usually be identified by its haustral margin around the periphery of the abdomen (see Fig. 18.11B). With distal large obstruction, when both small- and large-bowel dilatation is present, the radiographical appearances may be difficult to distinguish from paralytic ileus. The plain radiographical appearances of large-bowel obstruction may be indistinguishable from pseudo-obstruction and any patient with suspected large-bowel obstruction therefore requires additional imaging to confirm the diagnosis. CT has replaced contrast enema in this setting (Fig. 18.18). There are some large-bowel obstruction syndromes leading to more specific radiological appearances on a plain radiograph. Sigmoid volvulus is the classical volvulus, occurring most frequently in old age or in patients with mental handicap or institutionalisation (Fig. 18.19). The usual mechanism is twisting of the sigmoid loop around its mesenteric axis. Although the classical radiographical findings may be present, in up to one-third of cases it is difficult to differentiate a twisted sigmoid from a distended but non-rotated sigmoid, or from more proximal colonic distension. When sigmoid volvulus occurs, the inverted U-shaped loop of sigmoid is usually extremely distended and is commonly devoid of haustra, an important diagnostic point. The anhaustral margin can often be identified overlapping, respectively, the lower border of the liver shadow (the liver overlap sign), the haustrated, dilated descending colon (the left flank overlap sign) and the left side of the bony pelvis (the pelvic overlap sign). The top of the sigmoid volvulus usually lies very high in the abdomen with its apex on the left side. As originally described on contrast enema, and now seen on CT, a smooth, curved tapering of the colon is observed at the torsion point, like a hooked beak (the bird of prey sign); a whirl sign is seen with twisting of the mesentery and mesenteric vessels. Caecal volvulus can only occur when the caecum and ascending colon are mobile on a mesentery. In comparison with sigmoid volvulus, it usually occurs in those aged 30–60 years. In about half the patients, the caecum twists and inverts so that the pole of the caecum and the appendix occupy the left upper quadrant (Fig. 18.20A). In other patients, the twist occurs in an axial plane without inversion and the caecum still occupies the right half of the abdomen. The distended caecum can frequently be identified as a large viscus, which may be situated almost anywhere in the abdomen. The attached gas-filled appendix may be seen. There is often marked gaseous or fluid distension of the small

A

B Fig. 18.18  Large-Bowel Obstruction. Computed tomography. (A) Multiple distended loops of colon. The dependent loops are fluid filled. These are not visible on a plain radiograph. (B) The cause of the obstruction is demonstrated to be a stricture in the sigmoid colon, subsequently proven to be carcinoma.

bowel, which may sometimes obscure the caecal volvulus itself. The left side of the colon is usually collapsed. Distinction between caecal and sigmoid volvulus on CT may be difficult and careful scrutiny of the large bowel is essential to determine the type of volvulus. Location of the obstructed loop and the ‘swirl sign’, if present, as well as location of anatomical appendages can certainly provide valuable clues. The ‘swirl’ with caecal volvulus is usually located in the right lower quadrant, whereas the ‘swirl’ of sigmoid volvulus is typically located at the pelvic brim (see Fig. 18.20B).

ABNORMAL BOWEL WALL PATTERN The gas within the bowel profiles the mucosa, allowing appreciation of the mucosal surface on AXR. Adjacent bowel loops and peritoneal fat can delineate the outer extent of the bowel wall and the wall thickness can, therefore, be judged.

Small-Bowel Ischaemia Please see the above section Gas in Bowel Wall. The small-bowel wall becomes thickened if there is acute ischaemia, as a result of haemorrhage

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract

A

B Fig. 18.19  Sigmoid Volvulus. (A) Markedly dilated distended gas-filled loop of sigmoid colon extending from the pelvis into the left upper quadrant giving ‘coffee-bean’ sign. (B) Contrast-enhanced CT (coronal reformatted) shows twisting of vessels in the sigmoid mesocolon with ‘whirl sign’ at level of pelvic brim.

A

B Fig. 18.20  Caecal Volvulus. (A) Supine abdominal radiograph shows an inverted distended caecum extending from the right side up to the left upper quadrant. (B) Coronal computed tomography reconstruction in the same patient shows a ‘whirl sign’ in the right lower quadrant (black arrow) with dilated caecum extending into the left upper quadrant.

473

474

SECTION B  Abdominal Imaging

and oedema. Thickened bowel wall outlined by gas and adjacent fat may be evident on plain radiographs, but CT is far more sensitive; furthermore, small amounts of gas in the bowel wall are better demonstrated using CT.

Large-Bowel Ischaemia Ischaemic colitis is characterised clinically by the sudden onset of abdominal pain, followed by bloody diarrhoea. The splenic flexure and proximal descending colon are most often involved, and an explanation for development of ischaemia in this distribution is that this section of colon is located at a ‘watershed’ between arterial supply from the superior and inferior mesenteric arteries. The wall, particularly the submucosa of the colon, is thickened as a result of haemorrhage and oedema. This can sometimes be detected on plain radiographs, but in most cases CT is necessary. The term ‘thumbprinting’ has been used to describe the plain radiographical appearance of the submucosal thickening (Fig. 18.21A and Fig. 18.22). The involved area of the colon acts as a functional obstruction so that the right side of the colon is frequently distended. In the long term, the affected area could fibrose with development of a stricture.

Inflammatory Bowel Disease The plain AXR can usually predict the extent of colonic involvement in acute inflammatory disease of the colon (see Fig. 18.21). An assessment of

A

the extent of colitis, the state of the mucosa and the presence or absence of toxic megacolon and/or perforation can be made. The extent of faecal residue is related to the extent of the colitis. The disease is likely to be inactive where there are formed faeces, while a complete absence of faecal residue suggests extensive colitis. Intraluminal gas tends to accumulate as the colitis becomes more severe. Severe mucosal changes can be missed, however, if there is no intracolonic air to delineate the mucosal outline. When the bowel becomes dilated and the haustra disappear, the ulceration has penetrated the muscle layer and the patient moves into a high-risk group where urgent surgery must be considered. This is known as toxic megacolon (Fig. 18.23). To identify the development of this situation, daily AXR may be justified in order to monitor progress. The reader is referred to Chapter 21 for information on small-bowel inflammatory disease.

Pseudomembranous Colitis Pseudomembranous colitis may follow the administration of antibiotics, particularly clindamycin, lincomycin or antibiotics with similar pharmacological characteristics. Clostridium difficile is frequently cultured in the stools. AXR is abnormal in approximately one-third of cases: colonic dilatation (32%), thumbprinting, thickened haustra and abnormal mucosa (18%) may be identified (see Fig. 18.22). The whole colon may be involved, but the transverse colon is the most frequently affected segment. The rectum is involved in most cases. Appearances may mimic

B Fig. 18.21  Inflammatory Bowel Disease. Supine abdominal radiograph. (A) A thick-walled oedematous segment of transverse colon with mucosal oedema and thumbprinting. (B) Computed tomography abdomen coronal reformats demonstrating the extent of transverse colon inflammation in this young lady with Crohn’s disease.

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract

475

Fig. 18.23  Toxic Megacolon. Abdominal radiograph showing gas-filled dilated ascending and transverse colon with a thickened wall. There is an absence of haustral markings, indicating full-thickness mural inflammation. Fig. 18.22  Pseudomembranous Colitis. There is gross mucosal oedema in the transverse colon with ‘thumbprinting’ (black arrow). Clostridium difficile was grown.

acute inflammatory bowel disease, and other forms of infectious colitis, with or without immunosuppression, can look similar. CT is non-specific in the diagnosis of pseudomembranous colitis, and has been shown to be normal in 39% of cases, but CT can demonstrate markedly thickened mucosa, with oedema seen in the submucosa, and sometimes nodular mucosal thickening and the characteristic accordion sign. There may also be mild pericolonic inflammatory change in the fat. Peritoneal free fluid (ascites) may be seen in up to 40% of cases.

ACUTE ABDOMINAL INFLAMMATORY CONDITIONS The AXR still has a role in the acute abdomen in the scenarios described above but it offers very little information in the differential diagnosis of many causes of acute abdominal pain. Patients with suspected appendicitis, diverticulitis and cholecystitis should not undergo plain radiography, since diagnostic signs are usually lacking and spurious findings may be misinterpreted. Acute appendicitis is the commonest acute surgical condition in the developed world and carries an overall mortality rate of approximately 1%. Diagnosis is frequently difficult by clinical examination as there is a long list of alternative diagnoses that may mimic acute appendicitis (Table 18.5). Clinical diagnosis alone results in a normal appendix being found at appendicectomy in 10%–15%, and in young women the negative appendicectomy rate is higher still. Many argue that this rate of preventable operative intervention is unacceptable and that this should lead to greater utilisation of US and CT.

TABLE 18.5  Diseases Mimicking Clinical

Appendicitis Diagnosed at Ultrasound Ectopic pregnancy Ovarian cyst ± torsion Salpingitis Endometriosis Diverticulitis Infectious ileocaecitis Crohn’s disease Malignancy Intussusception Meckel’s diverticulum Cholecystitis Urolithiasis Mesenteric adenitis

Plain radiographs of the abdomen are not indicated for suspected appendicitis. There are no specific plain radiographical signs of acute appendicitis but ileus can occur, and there may be obstruction as loops of small bowel become matted together or stuck to the inflamed appendix. There is a high positive correlation between the presence of an appendiceal faecolith and acute appendicitis (Fig. 18.24A).

Ultrasound in Appendicitis Graded compression US is a well-established technique for examining the appendix and is particularly well suited for children and thin patients. The US probe is applied with gradually increased pressure over the right iliac fossa to displace bowel loops and examine the appendix. US signs of acute appendicitis include visualisation of a blind-ending tubular structure, which is non-compressible, with a diameter of 7 mm or greater (Fig. 18.25). An appendicolith may be seen as a hyperechoic focus casting

476

SECTION B  Abdominal Imaging

Fig. 18.25  Appendicitis. Ultrasound (in transverse plane) of the Right Iliac Fossa (RIF) RIF shows a thickened hypoechoic tubular blind-ended structure in the right iliac fossa. The surrounding fat is hyperechoic.

TABLE 18.6  Acute Appendicitis: US Signs

A

B Fig. 18.24  Appendicitis. (A) Supine abdominal radiograph. Radio-opaque density (white arrow) is projected over the right iliac fossa in a 24-year-old who presented with acute abdominal pain. (B) Contrast-enhanced CT in the same patient shows an appendicolith (black arrow) within the base of the inflamed oedematous appendix (white arrows).

an acoustic shadow and the surrounding inflammatory mass, which consists mainly of fat, is hyperechoeic. An abscess or fluid around the appendix may be seen (Table 18.6). US of acute appendicitis has been reported to have a sensitivity of 78%–98% and specificity of 85%–98%. There are interpretative pitfalls: false-negative results can arise in focal appendicitis of the appendiceal tip, retrocaecal appendicitis, gangrenous or perforated appendicitis, a gas-filled appendix and a massively enlarged appendix, which is very unusual in the inflamed

Blind-ending tubular structure: Non-compressible Diameter 7 mm or greater No peristalsis Appendicolith casting acoustic shadow High echogenicity surrounding fat Surrounding fluid or abscess Oedema of caecal pole Maximal tenderness over appendix

appendix. Pitfalls leading to a false-positive examination include a dilated fallopian tube, peri-appendicitis, inflammatory bowel disease and inspissated stool mimicking an appendicolith. When the appendix has perforated it may be compressible at US. This phenomenon has been reported in 38% of paediatric perforations and 55% of adult perforations. The main drawback of US is that in most instances, and in most hands, a normal appendix is not visualised and, subsequently, a negative US result, where the appendix is not seen, is of little value. In a multicentre German study of 2280 patients with acute abdominal pain, US did not result in proven clinical benefit. Similar results have been reported elsewhere. US is more useful in patients considered clinically indeterminate for acute appendicitis rather than when the diagnosis is considered very likely or very unlikely. Nevertheless, US can diagnose a number of conditions that mimic appendicitis clinically (see Table 18.5). When an experienced radiologist is available, US is recommended in children where there is diagnostic doubt, in young women (due to the higher incidence of tubal disease) and in those patients who are pregnant.

Computed Tomography in Appendicitis A number of large prospective trials have demonstrated that CT is a highly accurate investigation for confirming or excluding appendicitis. CT signs of appendicitis include an appendix measuring greater than 6 mm in diameter, presence of appendicolith and enhancement of the wall of the appendix following intravenous contrast medium (Fig. 18.26; see also Fig. 18.24B). Surrounding inflammatory changes include increased fat attenuation, fluid, phlegmon, caecal thickening, abscess and extraluminal gas. Focal caecal thickening due to oedema at the origin of the appendix is referred to as a caecal bar.

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract

A

477

B Fig. 18.26  Appendicitis. Contrast-enhanced CT (axial (A) and sagittal (B) reformatted) in a patient with acute right iliac fossa pain. The appendix (white arrow) is thickened with enhancing wall and surrounding inflammatory changes in the peri-appendiceal fat.

CT techniques for diagnosing appendicitis vary at different centres. A focused technique examining the abdominopelvic junction exposes the patient to approximately one-third of the radiation dose of a full abdomen and pelvic examination (~3 vs 10 mSv). However, this technique may not reveal other relevant diagnoses. Although authors debate the figures, a normal appendix is seen more frequently at CT than US and thus CT carries a better true negative rate. Much may depend on body habitus; US is well suited to thin patients and, especially, children. Proponents of CT in acute appendicitis have reported sensitivities and specificities of 100% and 95%, respectively, while establishing an alternative diagnosis in 89%. Several studies have reported overall accuracy of 93%–98%. Excellent results such as these come from radiologists with expertise in CT of the acute abdomen, often examining patients where the diagnosis of acute appendicitis is deemed highly likely on clinical assessment. It does not follow that CT will perform as well in the hands of general radiologists with less experience in abdominal imaging, especially when investigating patients where the differential diagnosis is still broad. The high radiation dose from abdominal CT should always be considered. When performing CT in this setting, it must be remembered that performance of CT at some centres may result in delayed treatment of appendicitis, with potential impact on clinical outcome. Imaging should not be a substitute for good clinical assessment. Several studies have shown the negative appendicectomy rate to fall from over 20% to less than 9% with the liberal use of imaging. In many institutions in the USA, CT is performed on almost all patients with acute right iliac fossa pain. Other studies have found that liberal use of CT and US does not reduce the negative appendicectomy rate. A key strength of CT is its ability to make alternative diagnoses in right

iliac fossa pain including mesenteric adenitis, terminal ileitis, Meckel’s diverticulitis, typhlitis, epiploic appendagitis and omental infarction. The use of CT versus US in the setting of suspected acute appendicitis is influenced by local expertise with US, availability of CT, and patient factors such as age and sex. More recently, several studies have addressed the efficiency of low-dose CT and MRI in the assessment of suspected acute appendicitis, which is discussed later.

Other Inflammatory Conditions CT is the accepted first investigation for suspected acute diverticulitis, due to its ability to assess extracolonic complications: namely, inflammation, abscess, perforation and fistula formation and the degree of bowel obstruction, if present. CT also has the best chance of identifying alternative diagnoses in the clinical setting of left iliac fossa pain. Ultrasound remains the investigation of choice in suspected acute cholecystitis because of its superior ability to detect gallstones and to assess sonographically for tenderness over the gallbladder (sonographic Murphy’s sign). For this condition, the reader is referred to Chapter 24 on the biliary system.

IMAGING THE ABDOMEN WITH COMPUTED TOMOGRAPHY: RADIATION ISSUES When imaging acute abdominal conditions, the choice of imaging investigation and technique is guided firstly by the clinical question at hand, secondly by accessibility to imaging equipment and economic issues and thirdly by the local expertise of the radiographer and radiologist who acquire and interpret the study. The use of CT has increased exponentially in recent years; however, clinicians and their patients are

478

SECTION B  Abdominal Imaging

becoming increasingly conscious of the exposure to ionising radiation associated with CT. This concern results partially from recent coverage in the scientific literature and media, which highlights the potential for increased cancer risk to patients as a result of increasing use of CT. Carcinogenesis is the primary concern and it is a proven stochastic effect of exposure to ionising radiation. A stochastic effect occurs randomly and no lower threshold of ionising radiation exposure exists where cancer induction does not occur. Carcinogenesis usually occurs many years remote from the exposure and is, therefore, of particular concern in patients who are young and in those with chronic disease who are subjected to repeated CT examinations over the course of their illness. Young cohorts at particular risk for high cumulative effective dose include those with curable malignancies such as testicular cancer and Hodgkin’s lymphoma, those with inflammatory bowel disease, particularly Crohn’s disease, cystic fibrosis patients and patients who repeatedly present with renal colic. Currently, there is a considerable research and industry drive to reduce radiation exposure during CT while preserving image quality and diagnostic yield. One of the first technological steps towards CT dose reduction was to reduce inefficiencies of radiation delivery. Fixedtube kilovoltage and amperage settings were commonly used in oldergeneration CT systems. Using these fixed-tube settings for CT of the abdomen and pelvis resulted in wider areas such as the mid-abdomen receiving the same exposure as narrower regions such as the pelvis, frequently with no improvement in image quality. This inefficient method of dose delivery was the target of one of the most successful and now widely implemented dose-reduction technologies: namely, automatic exposure control or automatic tube current modulation (ATCM).

AUTOMATIC TUBE CURRENT MODULATION Automatic tube current modulation tailors the output of the CT tube to the patient’s size and shape based on the patient’s diameter and the x-ray attenuation of the tissues through which the radiation beam is passing. This automated process ensures that thicker regions of the body are imaged using higher tube currents than thinner, less attenuating areas. Initial trials examining the utility of ATCM found that image quality could be preserved while radiation exposure was significantly reduced. Early clinical trials demonstrated that dose reductions could be achieved in almost 90% of examinations and that the tube–current time product was reduced by an average of 32% while using ATCM. Further reductions in CT dose, beyond the elimination of inefficiencies in dose delivery, are inherently associated with an increase in image noise. Noise is defined as the statistical variation in attenuation values of CT images, which does not reflect anatomy and blurs image contrast. Increased image noise is particularly problematic when imaging pathological changes with a low lesion–background contrast. Accuracy in detecting small focal lesions of the liver, spleen and kidneys may be negatively influenced by even a slight increase in image noise, whereas when imaging the bowel or the renal tract (for potential urinary calculi), a higher amount of image noise is usually acceptable because of the higher lesion–background contrast. In patient groups with Crohn’s disease and with urinary calculi and also in patients with suspected appendicitis, clinical trials demonstrate preservation of diagnostic accuracy despite significant reductions in radiation dose and consequential increases in image noise. Emerging CT noise-reduction strategies have provided new strategies for reducing CT dose while maintaining image quality. These methods are already showing great promise in abdominopelvic CT and the widespread dissemination of technologies such as iterative reconstruction is likely to result in CT dose reductions on the order of 75% and greater in future years.

ITERATIVE RECONSTRUCTION ALGORITHMS Iterative image reconstruction (IR) algorithms were used to generate images for the first commercial clinical CT systems but the limited processing abilities of computers in the early days of CT forced manufacturers to use a more computationally efficient method of IR known as filtered back projection (FBP). The greater image noise associated with low-dose CT imaging is poorly handled by FBP alone and the use of IR has recently been revisited as a more appropriate method for low-dose imaging reconstruction. Modern computers with improved computational capability and speeds have allowed iterative IR to enter the clinical domain and iterative IR techniques currently represent the most exciting dose-optimising developments in CT. Multiple generations of IR are being developed and tested by CT manufacturers. IR algorithms may operate on the image data or, preferably, on the raw projection data from the CT system itself. Some IR algorithms operate on a combination of FBP and IR and are referred to as hybrid algorithms. IR images were initially described as being ‘waxy’, ‘plastic’ or ‘oversmoothened’ in appearance in the case of early algorithms but later generations of IR have yielded more acceptable images, which are mildly ‘mottled’ or ‘pixelated’. With time, radiologists have also become accustomed to images reconstructed with IR and are more accepting of these characteristics. Expert opinion in this area suggests that imagers tend to adapt to the new quality of these images in a relatively short period of time. Hybrid IR algorithms are typically noise efficient, computationally fast and studies have indicated that images have good low-contrast detail and preserved image quality even with radiation dose reductions of at least 30%. Early investigations of the utility of IR systems from different manufacturers have shown that a diagnostically acceptable CT of the abdomen and pelvis can be acquired at approximately 50% less dose than was previously possible with FBP. A recent prospective study investigating the diagnostic accuracy of low-dose CT, using hybrid IR, to detect active inflammation in Crohn’s disease with doses comparable to plain AXR (~1.4 mSv), were as effective as conventional-dose CT in detecting clinically significant observations despite reduction in image quality. In comparison to hybrid IR, pure iterative reconstruction algorithms result in diagnostically acceptable images de novo, which negates the requirement for blending with FBP data. Pure or model based iterative reconstruction (MBIR) is now commercially available and models the physical characteristics of the focal spot, the x-ray fan beam, the threedimensional interaction of the x-ray beam within the patient and the two-dimensional interaction of the x-ray beam within the detector. Pure IR is computationally demanding and despite the use of parallel processing technology, only three to four data sets can be reconstructed per hour, at present. Several clinical trials suggest that both low-dose and, more recently, conventional-dose abdominal CT reconstructed with pure IR are superior to hybrid IR and also outperform FBP in both subjective image quality indices and objective image noise scores, facilitating dose reductions of approximately 80% in selected clinical settings (Fig. 18.27). In recent years, most CT manufacturers have incorporated IR into the latest generations of commercially available CT systems. This raises an issue of widespread availability of this technology, as it is not feasible for all medical centres to replace their equipment with newer CT machines. Also, the widespread availability of radiographers and radiologists trained in the optimal use of modern CT technology is challenging at many centres. Current research is growing towards the development of ‘vendor-independent’ algorithms that may be applied to FBP-reconstructed images obtained at lower doses to eliminate propagated image noise levels. These are known as ‘adaptive non-local means’ (ANLM) algorithms and often require less computational time

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract

A

B

C Fig. 18.27  Low-Dose Computed Tomography With Iterative Reconstruction. A 36-year-old female patient (BMI 23.6) with a known diagnosis of Crohn’s disease. Axial contrast-enhanced images from low-dose computed tomography abdomen and pelvis (0.96 mSv) reconstructed with (A) filtered back projection, (B) hybrid iterative reconstruction and (C) pure iterative reconstruction show a progressive reduction in image noise and increase in image quality with more advanced iterative reconstruction.

in comparison to de novo IR techniques. Early clinical studies have validated this method and results are promising. A study comparing ANLM to IR and, FBP suggests that image quality was not significantly compromised and, when applied with FBP, performance was as effective as IR techniques.

ROLE OF MRI IN THE ACUTE ABDOMEN As previously discussed, current trends in CT imaging have led to increasing concern regarding patient radiation exposure, especially in patient subgroups with chronic illnesses who have the potential for high-lifetime cumulative exposures due to requirement for repeated CT imaging (e.g. Crohn’s disease and cystic fibrosis patients). Other groups for whom CT should be avoided include children and pregnant patients.

479

Avoidance of exposure to ionising radiation and the capability to produce excellent soft-tissue contrast in multiple imaging planes, without the absolute need for administration of intravenous contrast agents, make MRI an excellent alternative for an increasing number of indications. MRI has been shown to supersede other investigations in certain abdominal: for example, complications and relapse of Crohn’s disease (MR enterography), characterisation of solid visceral lesions (MRI liver and kidneys) and biliary abnormalities Magnetic resonance cholangiopancreatography (MRCP). Due to the aforementioned advantages, there is a growing demand for research for the potential expansion of the role of MRI in the acute abdomen. A meta-analysis incorporating 30 studies of 2665 patients investigating the diagnostic performance of MRI for the evaluation of acute appendicitis in the general population, as well as paediatric and gravid subgroups, demonstrated high sensitivity and specificity (95%–97%) in comparison with CT. However, replacement of CT with MRI has been limited in such cases due to its relatively high cost, longer acquisition times and limited access to MRI in comparison with CT. Issues such as patient tolerance, claustrophobia and relative increased difficulty in performing MRI on critically ill patients also hinder its use. Therefore, the current consensus by the American College of Radiology (ACR) ACR appropriateness criteria is that MRI utilisation in the acute setting should be reserved for selected patient groups as mentioned above and, in specific cases, usually as a second-line investigation or problem-solving tool following an inconclusive US or CT study. A retrospective study of MRI used in the investigation of acute abdominal pain during pregnancy demonstrated an accuracy of almost 95% in identifying those who needed emergency intervention (Fig. 18.28). Acute appendicitis is the commonest acute abdominal emergency encountered during pregnancy, with highest incidence in the second trimester. Other relatively common indications for MRI in pregnancy include acute cholecystitis, choledocholithiasis and ureterolithiasis. It also gives the advantage of excluding gynaecological disorders including ovarian torsion or pelvic abscess. MRI has also been used in the evaluation of acute appendicitis, as well as Crohn’s disease, in young patients presenting with acute pain. For suspected appendicitis, graded compression US is usually the first-line imaging investigation and, when equivocal, an MRI of the pelvis is considered. When choosing an MRI protocol for imaging of the acute abdomen, there is no fixed protocol or set of sequences. In many cases, MR protocols are tailored for the specific clinical question and to individual patients; in general, it is advised that a radiologist is on site to review studies and determine if further sequences are needed. Several institutes have proposed the following basic sequences in assessment of acute abdominal pain: T2 weighted single shot fast spin echo with and without fat saturation (T2 SSFSE +/− FS) for detecting inflammatory change and oedema, T2 weighted fast spin echo +/− short inversion recovery (T2 FSE +/− STIR) for superior anatomical delineation and T1 weighted gradient echo including in and out-of-phase sequences. The use of IV contrast enhancement is also considered in special circumstances, particularly when abscess is suspected (diffusion weighted imaging (DWI) may also be useful). Identifying the normal appendix on MRI may sometimes be difficult. The paucity of intra-abdominal fat in children and the presence of a gravid uterus can pose additional challenges. Nonetheless, imaging features in acute appendicitis are quite similar to those in CT: that is, the demonstration of a dilated appendix (>6 mm) with thickened hypointense walls. An appendicolith may be seen as a hypointense filling defect within the lumen. T2 fat-saturated or STIR sequences usually demonstrate high-signal intensity within the appendix and the surrounding region, representing oedema and inflammatory change. Fluid collections and abscesses may also be identified (Fig. 18.29).

480

SECTION B  Abdominal Imaging

A

A

B Fig. 18.29  Acute Appendicitis. T2 weighted fat saturation sequence coronal plane (A). The appendix (white arrow) demonstrates thickened walls with multiple intraluminal filling defects (B) representing appendicoliths. The surrounding hyperintensity represents free fluid and inflammatory oedema.

RADIATION DOSE REDUCTION IN CLINICAL PRACTICE B Fig. 18.28  (A) Bowel obstruction during pregnancy. Coronal T2 Weighted Magnetic Resonance showing significant large bowel dilatation and ascites in this gravid female patient during her second trimester, caused by an obstructing colonic mass. (B) Maximum intensity projection T2 weighted (MIP T2 weighted) short inversion recovery (different patient) demonstrating bilateral hydroureteronephrosis resulting from compression of distal ureters during the third trimester.

MRI findings in other acute and chronic abdominal conditions, including inflammatory bowel disease, choledocholithiasis and cholecystitis, are described in detail in their respective chapters. MRI has many advantages, including equivalent diagnostic capability to CT in the diagnosis of acute appendicitis and/or alternative diagnoses in the acute setting. Taking into account accessibility and other issues discussed earlier, MRI is currently not considered a first-line investigation for diagnosis of acute appendicitis but acts as a useful problem-solving tool in challenging or equivocal cases and in patients where exposure to ionising radiation should be avoided.

In practice, effective radiation dose reduction is best achieved by careful patient selection, rigorous justification of high-dose examinations and the application of suitable acquisition parameters particularly with CT. The monitoring and audit of one’s own practice is recommended and may serve to identify aberrant trends in patient management or imaging, which may result in an increase in patient cumulative effective dose. It is important to involve CT radiographers in dose-reduction strategies to increase awareness and to prioritise dose reduction as part of routine practice. Variations in CT dose between, as well as within, individual institutions have been documented. One study found a 32-fold variation in CT exposures among different centres, some even using the same CT machine. Furthermore, the utilisation of alternative imaging techniques, which do not use ionising radiation, should be exploited when answers to clinical questions can be equally and effectively obtained. Early recognition of patients that would require repeated imaging would often serve to guide future referrals and choice of examinations. For example, a recent study concluded that patients suffering from Crohn’s disease who are receiving immunomodulating treatment or have had surgery for complications are prone to receiving greater cumulative

CHAPTER 18  Current Status of Imaging of the Gastrointestinal Tract radiation doses due to repeated CT imaging. Patients such as these may, therefore, be considered ‘at-risk’ for high-lifetime, cumulative-effective doses and alternative imaging strategies should be prioritised. Radiation dose management software is a new tool being developed in the era of ‘big data’. It allows continuous online monitoring of institutional radiation exposures from CT, nuclear medicine, plain radiography and fluoroscopy. It also provides instant data regarding individual patient’s recent and lifetime cumulative radiation exposures from diagnostic imaging studies and interventional radiology procedures. In conclusion, radiation dose optimisation in abdominal imaging is a multidisciplinary process involving radiologists, radiographers, medical physicists and referring physicians. This multidisciplinary focus on dose reduction will result in a systematic reduction in cancer risk. Our responsibility is to first counsel patients accurately regarding the risks of ionising radiation exposure; second, to limit the use of those imaging investigations that involve ionising radiation to clinical situations where they are likely to change management; and third, to ensure that a diagnostic quality imaging examination is acquired with lowest possible radiation exposure when ever an imaging investigation that results in radiation exposure is deemed necessary.

SUMMARY BOX: Ways to Reduce Radiation Exposure in

Clinical Practice 1. Ensure proper indication of ordered radiological examinations 2. Consider alternative non- radiation modalities to reach diagnosis US or MRI 3. Use low dose CT algorithms and protocols: Automated Tube modulation, iterative reconstruction 4. Pay attention to accumulative exposure dose in patients with chronic conditions. 5. Adhere to updates in international bodies’ radiation exposure and monitoring guidelines.

FURTHER READING Ditkofsky, N.G., Singh, A., Avery, L., et al., 2014. The role of emergency MRI in the setting of acute abdominal pain. Emerg. Radiol. 21 (6), 615–624.

481

Ditkofsky, N.G., Singh, A., Avery, L., et al., 2014. Risk factors for radiation exposure in newly diagnosed IBD patients. Emerg. Radiol. 21, 615. Duke, E., et al. (2016). A systematic review and meta-analysis of diagnostic performance of MRI for evaluation of acute appendicitis. AJR Am. J. Roentgenol. 206 (3), 508–517. Gangadhar, K., Kielar, A., Dighe, M.K., et al., 2016. Multimodality approach for imaging of non-traumatic acute abdominal emergencies. Abdom. Radiol. (NY) 41, 136. Jha, P., Bentley, B., Behr, S., 2017. Imaging of flank pain: readdressing state-of-the-art. Emerg. Radiol. 24 (1), 81–86. Johnson, E., Megibow, A.J., Wehrli, N.E., et al., 2014. CT enterography at 100 kVp with iterative reconstruction compared to 120 kVp filtered back projection: evaluation of image quality and radiation dose in the same patients. Abdom. Imaging 39 (6), 1255–1260. Kambadakone, A.R., Prakash, P., Hahn, P.F., et al., 2010. Low-dose CT examinations in Crohn’s disease: impact on image quality, diagnostic performance, and radiation dose. AJR Am. J. Roentgenol. 195 (1), 78. McDonald, G.P., Pendarvis, D.P., Wilmoth, R., et al., 2001. Influence of preoperative computed tomography on patients undergoing appendectomy. Am. Surg. 67, 1017–1067. McLaughlin, P.D., O’Connor, O.J., O’Neill, S.B., et al., 2012. Minimization of radiation exposure due to computed tomography in inflammatory bowel disease. ISRN Gastroenterol. 2012, 790279. O’Connell, O.J., McWilliams, S., McGarrigle, A., et al., 2012. Radiologic imaging in cystic fibrosis: cumulative effective dose and changing trends over 2 decades. Chest 141 (6), 1575. Patino, M., Fuentes, J.M., Singh, S., et al., 2015. Iterative reconstruction techniques in abdominopelvic CT: technical concepts and clinical implementation. AJR Am. J. Roentgenol. 205 (1), W19–W31. Puylaert, J.B.C.M., 1986. Acute appendicitis: US evaluation using graded compression. Radiology 158, 355–360. Ramalingam, V., LeBedis, C., Kelly, J.R., et al., 2015. Evaluation of a sequential multi-modality imaging algorithm for the diagnosis of acute appendicitis in the pregnant female. Emerg. Radiol. 22, 125. Rao, P.M., Rhea, J.T., Novelline, R.A., et al., 1997. Helical CT technique for the diagnosis of appendicitis: prospective evaluation of a focused appendix CT examination. Radiology 202, 139–202. Singh, S., Kalra, M.K., Hsieh, J., et al., 2010. Abdominal CT: comparison of adaptive statistical iterative and filtered back projection reconstruction techniques. Radiology 257 (2), 373. Yu, H.S., Gupta, A., Soto, J.A., et al., 2016. Emergency abdominal MRI: current uses and trends. Br. J. Radiol. 89 (1061), 20150804.

19  The Oesophagus Edmund M. Godfrey, Sibu Varghese, Alan H. Freeman

CHAPTER OUTLINE Anatomy and Function, 482 Examination, 482

ANATOMY AND FUNCTION Anatomy (Table 19.1) The oesophagus is a fibromuscular tube that connects the pharynx in the neck to the stomach in the abdomen, traversing the thorax via the superior and posterior mediastinum. It begins below the cricopharyngeus muscle, at the lower edge of the cricoid cartilage and at the level of C6. In the neck, the oesophagus lies posterior to the trachea. As it descends through the mediastinum, it passes posterior to the aortic arch, the left main bronchus and the left atrium, each of which causes an impression (Fig. 19.1). At the diaphragm the oesophagus passes through the diaphragmatic hiatus at T10, accompanied by the vagus nerves; it ends at the gastric cardia at the level of T11. The abdominal oesophagus lies posterior to the left lobe of the liver. The oesophagus is therefore composed of a short cervical, a long thoracic and a short abdominal segment. In health, the oesophagus is lined by stratified non-keratinising squamous epithelium. At the gastro-oesophageal junction (GOJ) there is an abrupt transition to columnar epithelium, termed the ‘Z-line’ because of the irregular interdigitations between pale pink squamous and darker columnar epithelia (Fig. 19.2). The GOJ is usually found at a surprisingly constant 40 mm from the teeth. The wall of the oesophagus is made up of five layers: the mucosa, the muscularis mucosa, the submucosa, the muscularis propria and the adventitia.

Pathological Features, 485

branchial arches 4 and 6, whereas the smooth muscle of the lower oesophagus is derived from somite mesenchyme. The myenteric plexus is derived from neural crest cells.

Function The oesophagus actively moves ingested material from the pharynx to the stomach and thus prevents reflux of stomach contents. Passage of a food bolus is regulated by the upper and lower oesophageal sphincters. The upper oesophageal sphincter is a high-pressure zone at the pharyngo-oesophageal junction and comprises the cricopharyngeus, the thyropharyngeus and the superior part of the cervical oesophagus. The lower oesophageal sphincter is a 3-cm-high pressure zone at the GOJ and is composed of the lower oesophageal muscle fibres and the diaphragmatic hiatus. The GOJ is anchored by the phreno-oesophageal ligament, which allows the oesophagus to slide a short distance longitudinally through the diaphragmatic hiatus while acting as a seal between the thoracic and abdominal cavities. Unlike the upper and lower oesophageal sphincters, the oesophagus between these high-pressure zones is relaxed in the resting state. The swallowing reflex induces so-called primary peristaltic (or stripping) waves that travel at 3 to 4 cm/s. Secondary peristalsis occurs when oesophageal sensory receptors are activated by material persisting in the oesophagus after primary peristalsis. Tertiary contractions are nonpropulsive and are seen in a variety of motility disorders (Fig. 19.3).

EXAMINATION SUMMARY BOX: Oesophageal Anatomy • Begins below the cricopharyngeus at C6 • Passes through the diaphragmatic hiatus at T10 • Ends at the gastric cardia at T11 • Has five layers: mucosa, muscularis mucosa, submucosa, muscularis propria, adventitia/peritoneum

Embryology The stratified squamous epithelium of the oesophagus, together with its associated submucosal glands, is derived from the endoderm of the foregut. The striated muscle of the upper oesophagus is derived from

482

The oesophagus can be examined with any of the commonly used imaging techniques. The initial test of choice is usually endoscopy, with fluoroscopy reserved for frail patients or those who have had recent surgery. Computed tomography (CT) is often the first-line test in the context of trauma. Imaging is extensively used in the staging of oesophageal malignancy, particularly CT, positron-emission tomography–CT (PET-CT) and endoscopic ultrasound (EUS).

Plain Radiography In most circumstances, plain radiographs reveal little useful information regarding the oesophagus except in the context of foreign body ingestion.

CHAPTER 19  The Oesophagus

483

TABLE 19.1  Anatomy of the Oesophageal Segments Cervical and upper thoracic Mid-thoracic Lower thoracic and abdominal

Muscle

Length (cm)

Striated

8

Mixed Smooth

8 8

Innervation

Artery

Veins

Lymph

Sensory, motor and parasympathetic from vagus, some sensory from spinal nerves Vagus Vagus

Inferior thyroid

Inferior thyroid

Aortic branches Left gastric

Azygos system Left gastric (note the varices)

Lower deep cervical nodes Mediastinal nodes Left gastric nodes

Fig. 19.2  Endoscopic image demonstrating the normal Z-line.

Fig. 19.1  Double-Contrast Barium Swallow Demonstrating the Aortic, Left Main Bronchus and Left Atrial Impressions (Arrows).

Foreign bodies tend to lodge at one of the following oesophageal constriction points: • Cricopharyngeus • Aortic arch • Left main bronchus • Diaphragmatic hiatus Otherwise, a dilated, gas- or fluid-filled oesophagus (Fig. 19.4A and B) may be identified incidentally during chest radiography for other indications.

Ultrasound Most of the oesophagus is inaccessible to conventional ultrasound examination (but see ‘Endoscopic Ultrasound’, later). The short cervical and abdominal segments are amenable to imaging in this way, but this is rarely used in clinical practice.

Fluoroscopy Fluoroscopic examination of the oesophagus is performed for a wide variety of indications. Barium suspensions are preferred for most indications; a preparation of 100% w/v is often used to provide good mucosal coating and an appropriate density. If possible, double-contrast images should be obtained using an effervescent agent, usually with the patient in the erect position. These are complementary to prone, single-contrast images. Water-soluble contrast medium is used when a tear, perforation or anastomotic leak is suspected. Low osmolar agents such as iopamidol

Fig. 19.3  Single-contrast barium swallow showing tertiary contractions in the middle and lower oesophagus.

(Gastromiro) should always be used to prevent pulmonary oedema, which can occur following aspiration of high osmolar agents such as meglumine diatrizoate (Gastrografin). In some institutions, when a leak is suspected, water-soluble contrast medium is followed with a barium suspension. Using barium in this

484

SECTION B  Abdominal Imaging

A

Fig. 19.5  Single-Contrast Barium Swallow of a Malignant Stricture. Note the irregularity of the mucosa and shouldered, shelf-like margins (arrows indicate proximal and distal extent). Endoscopic biopsy confirmed a squamous cell carcinoma.

B Fig. 19.4  A dilated gas- and fluid-filled oesophagus is visible on the posteroanterior (A) and lateral (B) chest radiograph of this patient with achalasia (arrows indicate fluid level).

way has been shown to be more sensitive for contained perforations, although it adds nothing in the detection of free leakage into the neck or mediastinum. Fluoroscopic examination of the oesophagus is tailored to the indication, but a suggested technique is as follows: control images should be obtained if the patient has had oesophageal or gastric surgery. With the patient in the erect position, double-contrast images are obtained in the lateral and posteroanterior (PA) projections of the cervical and upper oesophagus at four images per second. Right anterior oblique images of the middle and lower oesophagus are obtained at two per second, also in double contrast. The patient is then moved to the prone position and images are obtained at one per second during three separate single-bolus swallows to assess oesophageal motility and fully distend the GOJ. This view is particularly important if wrap migration is suspected following fundoplication. The images obtained in the prone position are usually single contrast. Finally, a static image of the stomach to include the gastric fundus is obtained with the patient in the erect position. The standard fluoroscopic examination may be augmented with additional procedures. As an example, where a patient describes a clear

history of dysphagia, but the images obtained appear normal, a swallow of biscuit dipped in barium or a barium tablet may uncover an occult stricture. If there are pharyngeal symptoms, images of the pharynx obtained during phonation should be obtained. Although no longer the first-line test for dysphagia, fluoroscopy remains an important test. It is well suited to evaluate the oesophagus following trauma or surgery, in complex hiatal herniae and as a less invasive alternative to endoscopy in the frail patient (Fig. 19.5).

Endoscopy Oesophagogastroduodenoscopy (OGD/endoscopy) is the initial investigation of choice for most indications, particularly dysphagia. It permits the direct visualisation of the mucosa and, crucially, biopsies can be taken. In patients with high dysphagia, preliminary fluoroscopic assessment can be used to forewarn the endoscopist of a pharyngeal pouch, which if present, would potentially reduce the risk of perforation. An OGD is carried out with the patient in the left lateral position, under topical local anaesthesia or conscious sedation (usually with a benzodiazepine such as midazolam). In addition to a detailed diagnostic assessment of the mucosa, a wide variety of therapeutic manoeuvres may be carried out endoscopically. These include the treatment of upper gastrointestinal (GI) haemorrhage, balloon dilatation and/or stenting of strictures, radiofrequency ablation (RFA) of dysplastic epithelium and injection of botulinum toxin for motility disorders. Endoscopic mucosal resection (EMR) deserves special note, as it is both therapeutic and the preferred method for staging early oesophageal tumours.

Computed Tomography In the context of oesophageal disease, CT is most widely used in the staging of oesophageal cancer. A CT of the thorax, abdomen and pelvis should be acquired. Good oesophageal and gastric distension is important: the patient should be given 1–1.5 L of water to drink as well as effervescent granules and should be imaged in the prone position. Intravenous

CHAPTER 19  The Oesophagus

Fig. 19.6  Computed Tomography Image on Lung Window Settings. There is a fistula (arrow) between the trachea and the oesophagus in this patient with metastatic squamous cell carcinoma. A nasogastric tube is present in the oesophagus.

contrast medium should be used whenever possible, with the upper abdomen imaged in both the arterial and portal venous phases. For the investigation of patients with suspected oesophageal trauma (including Boerhaave syndrome) and in the postoperative setting, positive oral contrast medium is required. As for fluoroscopic examinations, this should always be carried out with a low osmolar agent. For suspected tracheo-oesophageal fistula, an initial acquisition without the use of oral contrast medium is usually diagnostic (Fig. 19.6).

Magnetic Resonance Imaging In current clinical practice, magnetic resonance imaging (MRI) is not used for imaging the oesophagus. Image quality is hampered by motion artefacts from cardiac motion, breathing and peristalsis (Fig. 19.7). Whole-body MRI is under evaluation as an alternative to PET-CT for the staging of metastatic disease in oesophageal cancer but has not yet entered clinical practice.

Endoscopic Ultrasound EUS is generally used to characterise abnormalities identified using other imaging techniques, in particular the staging of oesophageal cancer. Less frequently, EUS is used for the assessment of submucosal lesions of the oesophagus. The high frequency and close proximity of the ultrasound probe allow the delineation of five layers of the oesophageal wall: mucosa, muscularis mucosa, submucosa, muscularis propria and adventitia. The muscular layers are hypoechoic; hence, there is a fivelayered alternating pattern. Endoscopic ultrasound/fine-needle aspiration (EUS-FNA) enables the sampling of structures deep to the oesophageal mucosa, particularly thoracic and upper abdominal lymph nodes. This can be particularly useful in the staging of oesophageal and lung malignancy and in the diagnosis of tuberculosis. In addition to sampling, EUS can be used to place fiducial markers to guide radiotherapy.

Radionuclide Radiology Including Positron-Emission Tomography–Computed Tomography For patients with oesophageal cancer, 18F-fluorodeoxyglucose (FDG) PET-CT is now the standard of care if radical treatment is intended.

485

Fig. 19.7  Axial T2 weighted magnetic resonance image of a distal oesophageal adenocarcinoma (arrows).

The presence of FDG-avid lymph nodes on preoperative PET-CT is prognostically significant even within the group of patients with the same pathological stage. The most important reason that PET-CT is used in oesophageal cancer staging is the high proportion of patients who have unsuspected metastatic disease at presentation (Fig. 19.8) and the superiority of PET-CT over other techniques for identifying it. Technetium-based radionuclide imaging of the oesophagus can be used for the identification of oesophageal motility disorders and gastrooesophageal reflux disease (GORD). Patients can be imaged swallowing both liquid and solid material (usually 99mTc-labelled sulphur colloid and scrambled egg, respectively) (Fig. 19.9).

PATHOLOGICAL FEATURES Oesophageal Cancer Oesophageal cancer is the sixth most common cause of death from cancer in the United Kingdom. There are two major histological types: squamous cell carcinoma and adenocarcinoma. Although the squamous type has been more common historically (and still is worldwide), in the United Kingdom, due to the rise in obesity, it has been overtaken by adenocarcinoma. Accurate preoperative staging of oesophageal cancer is difficult. The mobility of the oesophagus and its proximity to other organs make the assessment of local invasion problematic. Malignant lymph nodes are usually not enlarged and may first arise some distance from the tumour. Furthermore, unsuspected metastases may be present in up to 30% of patients at diagnosis. It is not surprising, then, that a variety of different tests are required for accurate staging. The patient with oesophageal cancer can face a whirlwind of tests, including endoscopy, CT, EUS and PET-CT. This combination is crucial for determining appropriate therapy. Initial diagnosis is usually with endoscopy, as it permits histological confirmation with biopsy (Fig. 19.10). Despite this, a good-quality fluoroscopic examination can detect even early tumours (Fig. 19.11). If a stricture is identified using fluoroscopy and it appears unequivocally benign, with symmetrical, smooth narrowing and a gradual tapering to normal calibre, malignancy can be confidently excluded. The converse is also true: strictures with an ulcerated, irregular mucosa and

486

SECTION B  Abdominal Imaging

A

Fig. 19.8  Coronal positron-emission tomography–computed tomography image of a fluorodeoxyglucose-avid left supraclavicular lymph node (arrow) metastasis in a patient with a distal oesophageal adenocarcinoma.

shouldered, shelf-like margins can be considered malignant on imaging appearances alone. The vast majority of patients will go on to CT as their initial staging investigation. Although less sensitive than EUS and PET-CT, it is relatively specific for identifying locally advanced or metastatic disease. Patients with these findings on CT are therefore spared further staging investigations. In the case of early tumours that appear to be T1 endoscopically, EMR is the preferred initial staging technique. The introduction of the 8th edition of the TNM (tumour–node– metastasis) classification (Table 19.2) has resulted in a single change to oesophageal cancer staging: type 3 junctional tumours are now considered gastric rather than oesophageal. In other words, junctional tumours are classified as lower oesophageal tumours if their epicentre is within 2 cm of the junction and they extend to involve the oesophagus. Tumours with their epicentre more than 2 cm distal to the GOJ or cardial tumours that do not extend to involve the oesophagus regardless of site are classified as gastric tumours.

B Fig. 19.9  Radioisotope Imaging of the Oesophagus. A normal examination is shown with isotope in the stomach (A), whereas in achalasia the isotope is retained in the oesophagus (B). (Images courtesy Dr KK Balan, Addenbrooke’s Hospital Cambridge.)

Computed Tomography for Oesophageal Cancer The normal oesophagus when adequately distended should have a wall thickness of less than 5 mm on CT. Tumours are seen as regions of wall thickening, which may be circumferential or asymmetric (Fig. 19.12). CT is rather limited in the local staging of oesophageal tumours because it is unable to delineate the layers of the oesophageal wall and is therefore useful only for distinguishing between T1–3 and T4 (invasion of other structures). The sensitivity and specificity of CT for T4 disease in a study of 94 patients with oesophageal squamous cell carcinoma were

66% and 84%, respectively. Signs of T4 disease include tumour contact of more than 90 degrees with the aorta (Fig. 19.13); loss of the triangle of fat between the oesophagus, aorta and spinal column (see Fig. 19.13); and nodular protrusion into the airways (Fig. 19.14). For nodal staging of oesophageal cancer, CT is relatively insensitive, as the majority of involved nodes are not enlarged. In a meta-analysis, the sensitivity and specificity for regional nodal disease were 50% and 83%, respectively. In general, nodes with a short axis of greater than

CHAPTER 19  The Oesophagus

487

Fig. 19.10  Endoscopic Image of a Distal Oesophageal Adenocarcinoma. Note the gastric folds (arrows) visible adjacent to the tumour, indicating the level of the gastro-oesophageal junction.

Fig. 19.12  Computed tomography image of asymmetrical oesophageal thickening caused by adenocarcinoma.

TABLE 19.2  TNM, 8th Edition T

T1a T1b T2 T3 T4a T4b

N

N0 N1 N2 N3 M0 M1

M

Fig. 19.11  Double-contrast barium swallow image demonstrating a nodular mural lesion arising from the left side of the proximal oesophagus (arrows indicate proximal and distal extent). This was confirmed as a squamous cell carcinoma on endoscopic biopsy.

1 cm are considered involved on CT. Common sites of regional nodal disease include perioesophageal, subcarinal, left gastric (Fig. 19.15) and coeliac territories. A key distinction for node groups is that between the perioesophageal cervical nodes (Fig. 19.16), which are considered regional, and the supraclavicular lymph nodes (Fig. 19.17), which are metastatic. The anatomical landmark that distinguishes these sites is the vascular plane containing the common carotid artery. The most frequent sites for oesophageal cancer metastases are nonregional lymph nodes such as the supraclavicular (see Fig. 19.17) and

Invasion of the lamina propria or muscularis mucosa Invasion of the submucosa Invasion of the muscularis propria Invasion of the adventitia Resectable tumour invading the pleura, pericardium or diaphragm Unresectable tumour invading structures such as the aorta, airway, vertebrae, etc. No lymph nodes involved One to two regional lymph nodes Three to six regional lymph nodes Seven or more regional lymph nodes No distant metastasis Distant metastasis

retroperitoneal abdominal lymph nodes. As is the case with gastric cancer, the left supraclavicular node is more frequently involved than the right. Visceral metastases are seen in the liver, the lungs (Fig. 19.18), bones, muscles (Fig. 19.19) and the adrenal glands. As for tumours arising elsewhere in the body, squamous cell carcinoma lung metastases are more likely to cavitate than is adenocarcinoma. In keeping with its performance for regional nodal involvement, CT is insensitive but relatively specific for metastatic disease.

SUMMARY BOX: Computed Tomography for Oesophageal Cancer • Tumour contact greater than 90 degrees, loss of the periaortic fat, or nodular soft tissue extension into the major airways suggests an unresectable tumour (T4b disease) • Regional lymph nodes should be considered involved if they measure more than 1 cm in the short axis • Common sites for metastatic disease include non-regional nodes, liver, lungs, bones and adrenal glands

488

SECTION B  Abdominal Imaging

Fig. 19.13  Computed Tomography Image of a Mid Oesophageal Adenocarcinoma. There is infilling of the triangle between the oesophagus, aorta and vertebral column. The tumour contacts the aorta for greater than 90 degrees. The aortic contour is flattened adjacent to the tumour. These are all features of T4b invasion.

Fig. 19.14  Coronal Computed Tomography Image of a Patient With a Mid Oesophageal Adenocarcinoma. There is nodular protrusion of soft tissue from the tumour into the left main bronchus (arrow), in keeping with T4b invasion.

Endoscopic Ultrasound for Oesophageal Cancer EUS is superior to CT and PET-CT for T staging (Fig. 19.20). The sensitivity and specificity for identifying the various T stages of oesophageal cancer is high. In some patients with advanced tumours, the stricture is too tight to permit passage of the standard radial echoendoscope. An endobronchial ultrasound (EBUS) scope can be used in most of

Fig. 19.15  Computed Tomography Image of an enlarged left gastric lymph node (arrow) in a patient with adenocarcinoma of the oesophagus.

Fig. 19.16  Computed Tomography Image Demonstrating an Enlarged Left-Sided Perioesophageal Cervical Lymph Node. Many surgeons would consider this unresectable, given its position above the aortic arch. However, it is still considered a regional lymph node on the basis of TNM 8.

these cases if required. If the tumour is not traversable with the standard echoendoscope, the T stage is almost always T3 or T4. The number of abnormal lymph nodes seen on EUS correlates closely with patient survival. For nodal disease, EUS has a sensitivity higher than that of PET-CT or CT, but it is less specific. Although metastatic disease can be identified with EUS on occasion (for example, in the left adrenal gland or liver) it does not provide the

CHAPTER 19  The Oesophagus

Fig. 19.17  Computed tomography image demonstrating an enlarged right supraclavicular lymph node, which is lateral to the right common carotid artery. This lymph node is considered non-regional and therefore metastatic.

489

Fig. 19.19  Computed Tomography Image Demonstrating Metastases in the Left Pectoralis Major and Minor.

Fig. 19.20  Radial Endoscopic Ultrasound Image of a Distal Oesophageal Adenocarcinoma. The tumour is seen from 10 o’clock to 1 o’clock and does not breach the muscularis propria, in keeping with T2 disease.

Fig. 19.18  Computed tomography image demonstrating a cavitating right lung metastasis in a patient with an oesophageal mass (squamous cell carcinoma). Cavitation is seen more commonly in squamous than in adenocarcinoma metastases.

whole-body coverage necessary and so is always used in conjunction with PET-CT. Although previously a routine part of staging in some centres, the current National Institute for Health and Care Excellence (NICE) guidelines recommend against using EUS in this way and suggest that it should be reserved for cases where it will change management (for example, in patients with solitary FDG-avid lymph nodes outside the resection field).

Positron-Emission Tomography–Computed Tomography for Oesophageal Cancer In T1 tumours of the oesophagus, it is usually not possible to identify the tumour with PET, which should therefore be omitted if this stage is suspected endoscopically. If a tumour is not detectable by PET-CT, it will be T2 or less in 70% of cases. PET-CT otherwise suffers the same limitations as CT in terms of depth of mural invasion; EUS is therefore the preferred technique for T staging. Nodal disease can be assessed with reasonable sensitivity and specificity with PET-CT. The presence of FDG-avid lymph nodes is a negative prognostic marker and an indication for neoadjuvant therapy. PET-CT is the technique of choice for identifying metastases to nonregional lymph nodes (see Fig. 19.18) and other tissues such as the liver and skeletal muscle. The ability of PET-CT to correctly upstage up to 20%

490

SECTION B  Abdominal Imaging

A

B Fig. 19.21  Endoscopic Images of a Patient With an Oesophageal Adenocarcinoma. A self-expanding covered metallic stent immediately after insertion (A) and after full expansion (B).

of patients means that it should be used for all patients before radical treatment. For single metastases, cytological or histological confirmation is recommended in view of the 4% false-positive rate with PET-CT.

Treatment of Oesophageal Cancer Treatment for oesophageal cancer involves a broad range of interventions that are dependent on the stage and type of tumour as well as the fitness of the patient and local availability. For the earliest oesophageal tumours (T1a) that do not invade the submucosa, EMR is the preferred technique for removal. The EMR specimen is assessed histologically for deep invasion. If present, further treatment would be considered, including oesophagectomy. At the other end of the staging spectrum, patients with invasion of major structures (T4b) or metastatic disease are offered palliative treatment. This includes a variety of manoeuvres for maintaining oesophageal patency, most commonly stenting (Fig. 19.21A and B). Systemic treatment with palliative chemotherapy is used in patients with a good performance status. Radiotherapy has an important role, particularly for the more radiosensitive squamous cell carcinoma. For patients with resectable disease, most will have nodal involvement or a tumour extending through the muscularis propria (T3 disease). If a patient has disease staged clinically as cT2N0M0 or greater, the patient will likely benefit from neoadjuvant chemotherapy prior to surgery. Radical chemoradiotherapy is an alternative treatment for squamous cell carcinoma. A variety of surgical approaches are used. The oesophagus is almost always substituted with a gastric conduit. Both fluoroscopy (Fig. 19.22) and CT (Fig. 19.23) play an important role in the detection of postoperative complications. If there is necrosis of the gastric conduit, a colonic interposition (Fig. 19.24) may be used, usually after an interval of several months, to allow the patient to recover before further surgery.

Other Oesophageal Neoplasms Other than adenocarcinoma and squamous cell carcinoma, true neoplasms of the oesophagus are uncommon. They can be categorised as

Fig. 19.22  Water-soluble contrast swallow demonstrating an anastomotic stricture (arrows) at the oesophagogastric anastomosis 2 weeks after Ivor Lewis oesophagectomy. Note the jet of contrast medium as it passes through the narrowed lumen.

benign or malignant and according to whether they are mucosal or submucosal.

Benign Lesions Glycogenic acanthosis, though not a neoplasm, requires mention, as it is present in up to 30% of normal individuals. It manifests as mural nodules, usually measuring 2 to 5 mm, which are more easily seen as white/yellow plaques on endoscopy. These nodules/plaques are caused

CHAPTER 19  The Oesophagus

Fig. 19.23  Computed tomography image of an anastomotic leak 6 days after an Ivor Lewis oesophagectomy. Note the defect in the gastric conduit (arrow) and the associated right pneumothorax.

Fig. 19.24  A water-soluble contrast-swallow image demonstrating the normal appearance of a colonic interposition. Note the normal pattern of haustral folds in the colonic segment.

491

Fig. 19.25  Computed tomography image of a soft tissue attenuation perioesophageal lesion with internal calcification, confirmed with endoscopic ultrasound and fine-needle aspiration as a leiomyoma.

by the proliferation of glycogen-containing cells within the squamous epithelium. Glycogenic acanthosis is of no clinical consequence, although it may be associated with GORD, coeliac disease and, rarely, Cowden syndrome. The multiplicity of lesions is usually helpful in making the diagnosis at fluoroscopy. Papillomata are uncommon benign tumours of the oesophagus and comprise hyperplastic squamous epithelium. A papilloma usually appears as a solitary sessile polyp and will rarely measure more than 10 mm. In view of this non-specific appearance, biopsy is required to distinguish a papilloma from an early adenocarcinoma or squamous cell carcinoma. The most common benign submucosal tumour of the oesophagus is the leiomyoma. This is in contrast to the rest of the GI tract, where gastrointestinal stromal tumours (GISTs) predominate. Endoscopically, and on fluoroscopic studies, a leiomyoma appears as a smooth submucosal mass. On CT, a homogeneous well-defined soft tissue mass is the most common appearance, sometimes with a focus or two of punctate calcification (Fig. 19.25). These findings are rather non-specific, so EUS-FNA is needed to confirm the nature of the lesion. Although not neoplastic, a congenital foregut duplication cyst (Fig. 19.26) may be identified as a submucosal mass on fluoroscopic or endoscopic examination of the oesophagus. Such cysts are very straightforward to characterise with MRI or EUS; both techniques will demonstrate a simple cyst (Fig. 19.27). Fibrovascular polyps are very rare pedunculated submucosal lesions that usually arise from the upper oesophagus. As they tend to be rather soft, they can reach a considerable size before causing dysphagia. On CT they often have regions of both fat and soft tissue attenuation (Fig. 19.28). They are notable for their potentially unusual clinical presentation: regurgitation into the mouth, which can sometimes result in death by asphyxiation. Other submucosal lesions—such as schwannoma, neurofibromata and lipomata—are also rare.

492

SECTION B  Abdominal Imaging

Fig. 19.26  Computed tomography image of a perioesophageal lesion demonstrating intermediate attenuation (HU = 32).

Fig. 19.28  Computed tomography image demonstrating a grossly dilated oesophagus with a smooth intraluminal lesion (arrows) of mixed soft tissue and fat attenuation. This was confirmed as a fibrovascular polyp. (Image courtesy Dr DJM Tolan, Leeds Teaching Hospitals.)

GISTs are very rare in the oesophagus but much more common in the stomach and small bowel. They are of variable malignant potential. On EUS, oesophageal GISTs appear as submucosal lesions arising from the muscularis mucosa or the muscularis propria and are therefore difficult to distinguish from a leiomyoma without tissue sampling. Metastases to the oesophagus most commonly arise via direct extension, usually from carcinoma of the bronchus. Involved lymph nodes infiltrate the oesophagus, causing extrinsic compression and occasionally fistulae between the oesophageal lumen and the airways. Pancreatic cancer may also extend to direct involvement of the distal oesophagus and GOJ. Lymphatic or blood-borne metastases result in a submucosal mass or masses, as in breast cancer.

Hiatus Hernia

Fig. 19.27  T2 weighted axial magnetic resonance image demonstrating that the lesion is T2 hyperintense; this is diagnostic of a foregut duplication cyst.

Malignant Lesions Nearly all oesophageal malignant tumours are adenocarcinomas and squamous cell carcinomas. A number of rare malignancies are encountered, each representing around 1% of all oesophageal tumours. These include small cell carcinoma, GISTs, melanoma, lymphoma and metastases. Like small cell carcinoma of the lung, small cell carcinoma of the oesophagus is a highly aggressive primary malignancy. It tends to be advanced at diagnosis and to relapse quickly following treatment. Primary oesophageal malignant melanoma is similar to small cell carcinoma in that it confers a grave prognosis, is frequently advanced at presentation and behaves aggressively. It should be considered as a diagnosis when endoscopic findings are of a pigmented lesion, although non-pigmented, amelanotic forms are also encountered.

A hiatus hernia exists when abdominal organs pass through the oesophageal hiatus into the chest. Usually the herniated organ is the abdominal segment of the oesophagus with part of the stomach, although greater omentum, colon, spleen, pancreas and small intestine are sometimes involved. Hiatal hernias are divided into sliding and rolling varieties. The majority of hiatal hernias are of the sliding type (approximately 90%). Weakening of the phreno-oesophageal membrane allows superior displacement of the GOJ. The diaphragmatic hiatus is normally a muscular, slit-like opening in the right crus of the diaphragm passing anterior to posterior. In a well-established sliding hiatal hernia it becomes more circular, with atrophy of the surrounding muscular fibres. This change in configuration explains the main clinical consequence of sliding hiatal hernias: GORD. The diagnosis of a sliding hiatal hernia is made on fluoroscopy when gastric rugae are seen traversing the diaphragm, or when the oesophageal B ring (representing the squamo-columnar junction) is seen above the diaphragm. Assessment of small (95%) benign gastric ulcers heal in about 8 weeks when treated medically. Benign ulcers may heal completely without any radiographic residua. As benign ulcers heal, they may change shape from round or oval to linear crevices. There may be subtle retraction or stiffening of the affected wall (see Fig. 20.4C). An easily recognisable radiographic sign of healed gastric ulcer is the presence of folds converging to the site of the healed ulcer. There may be a residual central pit or depression (see Fig. 20.4D). The radiating folds should be uniform. Incomplete healing, irregularity of the folds, residual mass or loss of mucosal pattern all suggest the possibility of an underlying malignancy. The Carman meniscus sign is appreciated in the clinical setting of a large, flat ulcer with heaped-up edges. The edges of the ulcer trap a lenticular barium collection that is convex relative to the lumen when the edges are folded upon themselves during compression. These findings are indicative of a malignant gastric ulcer. Occasionally benign ulcers may heal with significant scarring. Severe retraction of the greater curvature from healed ulceration of the lesser curvature may cause narrowing of the mid-body of the stomach. Healing of antral ulcers may form prominent transverse folds or significant antral narrowing and deformity (see Fig. 20.4E). Such scarring may lead to significant obstruction, an important complication of PUD. The use of CT in the evaluation of gastric ulcers has been studied by several authors and found to be useful in the detection and characterisation of benign and malignant gastric ulcers. The differences between malignant and benign ulcers are important in treatment planning and follow-up of patients with gastric ulcers. Evaluation by VG and/or MPR has been studied and certain advantages and disadvantages have been elucidated. The benefits of VG include the ability to evaluate for mucosal changes, which include morphological changes. This is similar to conventional gastroscopy. Another benefit is that as optical endoscopic criteria have been well established for benign and malignant ulcers, then similar criteria can be applied to VG. Characteristics of malignant ulcers on VG include ‘irregular or angulated shape; uneven base; asymmetric edge; bulbous enlargement and fusion or disruption of gastric folds reaching the crater edge’. Benign ulcer criteria by VG include: ‘smooth, regular, round or oval shape; even base; sharply demarcated or round edges; converging gastric folds with smooth tapering and radiation’. The use of MPR images allows for better visualisation of ulcers in multiple planes, facilitating better detection and characterisation of the ulcers. MPR also provides mural and extramural information, including, but not limited to, enhancement patterns of the gastric wall, lymphadenopathy and evaluation of adjacent organs. Characteristics of malignant ulcers on MPR images include: ‘strong enhancement of the wall at the site of ulcer greater than adjacent wall; marked periulcer wall thickening with loss of normal wall stratification; perigastric fat infiltration, lymphadenopathy or metastatic disease’. Characteristics of benign ulcers include ‘no excessive enhancement when compared with the adjacent

504

SECTION B  Abdominal Imaging

A

B

C

D

E Fig. 20.4  Gastric Ulcer. (A) Fluoroscopic image demonstrating a benign ulcer along the lesser curvature (arrowhead). Note: the stomach fundus (F) and the antrum (A). (B) Hampton line, a thin line of radiolucency crossing the opening of an ulcer, a virtually infallible sign of a benign ulcer. (C) Focal retraction along the incisura angularis with small residual outpouching is present. Converging smooth folds no longer fill an ulcer crater. (D) Healing ulcer radiating folds converging to a linear scar. (E) Scarred antrum with constriction at site of previous ulcer causing narrowing and deformity.

CHAPTER 20  The Stomach

505

B

A

Fig. 20.5  Gastric Ulcer on Computed Tomography (CT). (A) Axial CT image showing benign ulcer (arrow) along the lesser curvature with a crater and surrounding smooth mound. (B) Axial CT shows perforated gastric ulcer in the posterior gastric antrum with leaking contrast (arrow), focal wall thickening and adjacent fat stranding.

gastric wall and mild periulcer wall thickening with preservation of wall stratification’ (Fig. 20.5A). But some of the features for benign and malignant lesions may overlap and differentiation of benign from malignant ulcers may not be determined by CT alone. MDCT evaluation remains complementary to optical endoscopy (gastroscopy). Surgical complications such as penetration and perforation are easily identified on MDCT using water-soluble contrast medium (see Fig. 20.5B).

Gastric Erosions Gastric erosions or aphthous ulcers are superficial ulcerations that do not penetrate the muscularis mucosa. They usually appear as small, shallow collections of barium 1 to 2 mm in diameter surrounded by a radiolucent rim of oedema. These are called ‘complete’ or ‘varioliform’ erosions (Fig. 20.6). When the halo of oedema is lacking, the erosions are called ‘incomplete’. These appear as short linear or serpentine lines or dots of barium. Gastric erosions are most easily detected radiographically when they are multiple and complete. They are most often detected on double-contrast barium studies but may also be seen with compression technique. Single or incomplete erosions are commonly detected endoscopically but are infrequently identified radiographically. Erosions heal without scarring. Gastric erosions are most often causally related to H. pylori infection. Other causes include alcohol, NSAID ingestion and Crohn disease. They may also be seen as a response to stress: for example, in patients with severe trauma.

Gastritis Gastritis is a descriptive term with sometimes conflicting pathological, endoscopic and radiographic definitions. It is now better understood that many causes of gastritis, including H. pylori, alcohol and NSAID gastritis, lead to similar morphological changes. The most common findings are thick (>5 mm) folds with or without nodularity (Fig. 20.7A). Erosions, while less commonly seen, are a frequent sign of H. pylori gastritis. Other signs of gastritis include antral narrowing, inflammatory polyps and prominent areae gastricae. The radiological findings are similar to endoscopic findings. There is considerable overlap of findings between patients who are biopsy-positive or -negative for H. pylori. Therefore it is often not possible

A Fig. 20.6  Gastric Erosions. Complete or ‘varioliform’ erosions in antrum seen on double contrast views.

at this time to distinguish between ulcers and gastritis caused by H. pylori or those caused by chemical irritants (alcohol, NSAIDs and other causes of gastritis). Owing to the prevalence of H. pylori, its association with many gastric diseases and effective treatment options, it is important for the radiologist to recognise findings that suggest the presence of the infection. It has been shown that thickened gastric folds, although nonspecific, are still the single most useful radiographic sign for the diagnosis of H. pylori gastritis and that the combination of thick folds and enlarged areae gastricae may be the most specific findings.

Atrophic Gastritis Atrophic gastritis is a combination of atrophy of the gastric glands with histological inflammatory changes. Atrophic gastritis is found in more

506

SECTION B  Abdominal Imaging of severe oesophageal disease. Prominent aphthous ulceration may be seen in such cases. In immunocompromised patients, cytomegalovirus, crytosporidiosis and toxoplasmosis may occur. Radiographic findings are non-specific but there are some suggestive signs. Deep ulceration, and even fistulisation to adjacent structures may be seen with cytomegalovirus. Cryptosporidium primarily affects the small bowel, causing severe diarrhoea and thick, small bowel folds. It rarely involves the stomach but has been shown to cause deep ulcers, antral narrowing and rigidity. Strongyloides is a parasitic infection with a worldwide distribution affecting the stomach, duodenum and proximal small bowel evidenced by thickened, effaced folds and narrowing. In advanced cases, the stomach may be narrowed and have thickened folds.

Crohn and Other Granulomatous Diseases

A

B Fig. 20.7  (A) Diffuse erosive gastritis with thick nodular folds. Erosions are scattered along the folds. (B) Atrophic gastritis. Diffuse atrophy of the mucosal folds in a narrowed featureless stomach.

than 90% of patients with pernicious anaemia and is characterised by loss of parietal and chief cells, leading to achlorhydria, and atrophy of the mucosa and mucosal glands. Atrophic gastritis causes a decrease in the production of intrinsic factor, which in turn causes malabsorption of vitamin B12. Radiographic findings of atrophic gastritis include loss of rugal folds and a tubular, featureless narrowed stomach (see Fig. 20.7B). Areae gastricae may be absent. The radiographic features are non-specific, but because of the important prognostic implications of atrophic gastritis, an appearance suggesting this diagnosis should trigger an appropriate clinical work-up. Atrophic gastritis may be associated with gastric polyps, carcinoma and ulcers, both benign and malignant. Intestinal metaplasia, which may be seen histologically in atrophic gastritis, is considered a premalignant condition. The diagnosis of intestinal metaplasia may be suggested by areas of focal enlargement of the areae gastricae.

Infectious Gastritis H. pylori gastritis is by far the most common infection affecting the stomach. Tuberculosis, histoplasmosis and syphilis are usually lumped together with other granulomatous processes causing gastritis. Ulceration, thick folds and mucosal nodularity are common features, with antral narrowing being a late feature of these diseases. Monilia (Candida) may involve the stomach. This almost always occurs in the presence

Granulomatous conditions of the stomach may be infectious or inflammatory, such as Crohn disease, sarcoidosis, tuberculosis, syphilis and fungal diseases. Although Crohn disease primarily affects the ileum and colon, gastroduodenal involvement occurs in up to 20% of cases. When the upper gastrointestinal tract is affected by Crohn disease, both the stomach and duodenum are commonly affected. However, involvement of the duodenum alone is more common than the stomach alone. A wide range of symptoms may be seen in patients with gastroduodenal Crohn disease; some are asymptomatic, others have symptoms more typical of PUD or with symptoms related to antral narrowing or gastric outlet obstruction. Diarrhoea caused by associated ileocolic disease is common. Gastrocolic fistula is an unusual complication. When it occurs, disease of the transverse colon extending to the stomach is most often the cause. Radiographic findings of gastric Crohn disease almost always demonstrate involvement of the antrum alone or antrum and body of the stomach. With early disease, findings include aphthous ulcers, larger discrete ulcers, thickened and distorted folds and sometimes a nodular (‘cobblestone’) mucosa (Fig. 20.8). These are indistinguishable from aphthous ulcers or erosions because of other causes. These findings represent the non-stenotic phase of Crohn disease. Stenotic disease caused by scarring and fibrosis can result in narrowing of the gastric antrum and pylorus into a funnel or ‘rams-horn’ shape, sometimes leading to foreshortening of the stomach which can be so severe that it simulates a partial gastrectomy. This scarred, funnel-shaped antroduodenal region may also be seen in other granulomatous diseases, including tuberculosis, syphilis, sarcoidosis and eosinophilic gastroenteritis. Antral narrowing may also mimic scirrhous gastric carcinoma.

Hypertrophic Gastritis Hypertrophic gastritis is characterised radiographically by thickened folds, often greater than 10 mm in width, predominantly in the fundus and body, which are the acid-producing regions of the stomach. While the term is often used descriptively, the entity of hypertrophic gastritis is associated with glandular hyperplasia and increased acid secretion. Histologically, inflammation is not a prominent feature; thus, ‘gastritis’ is somewhat a misnomer. The areae gastricae pattern will become more prominent (Fig. 20.9). There is a high prevalence of duodenal and gastric ulcers in these patients. Many cases that had been previously classified as hypertrophic gastritis may in fact have been due to H. pylori infection. The differential diagnosis is primarily Ménétrier disease and lymphoma.

Ménétrier Disease Ménétrier disease is a rare entity well known in the radiology literature because of its dramatic and characteristic appearance. This condition is characterised by hypertrophy of gastric glands, achlorhydria and hypoproteinaemia. Loss of protein from the hyperplastic mucosa into

CHAPTER 20  The Stomach

507

Fig. 20.10  Ménétrier Disease. Classic appearance with massively enlarged folds in the body without abnormality in the antrum.

Fig. 20.8  Crohn Disease. Multiple aphthous (superficial) erosions are present on the antrum. Duodenal folds are thick and nodular (cobblestone mucosa).

differentiate it from carcinoma, where the stomach becomes rigid and aperistaltic. While in the classic description of Ménétrier disease the antrum is spared, it has been found to be involved in up to 50% of cases causing diffuse involvement of the stomach. Another feature of Ménétrier disease is the finding of increased fluid in the small bowel, which in turn may prevent optimal mucosal coating with barium.

Zollinger-Ellison Syndrome Zollinger-Ellison syndrome is caused by gastrinomas, which are non-β islet cell tumours that secrete gastrin, stimulating acid secretion in the stomach. Seventy-five per cent of the tumours are found in the pancreas and 15% in the duodenum; the rest are extraintestinal. Patients with Zollinger-Ellison syndrome may have thickened gastric folds and increased gastric secretions. The tumours may be malignant and metastases, primarily to the liver, are present in up to half of the cases. Zollinger-Ellison is one of the manifestations of multiple endocrine neoplasia (MEN) type I (also called Wermer syndrome), which includes parathyroid, pituitary and adrenal tumours. In some cases MEN type I may be associated with carcinoid tumours.

Eosinophilic Gastroenteritis

Fig. 20.9  Hypertrophic Gastritis. Image from a patient with a recently healed lesser curvature ulcer. This characteristic enlargement and prominence of the areae gastricae can be correlated with an increased incidence of gastric hypersecretion and PUD.

the gastric lumen results in a protein-losing enteropathy, and may produce disabling symptoms. The disease is characterised by markedly enlarged, often bizarre gastric folds most prominent in the proximal stomach and along the greater curvature. Radiographically, an upper gastrointestinal barium study or CT shows massively thickened often lobular folds (Fig. 20.10). The folds remain pliable, which helps to

Eosinophilic gastroenteritis is characterised by focal or diffuse infiltration of the gastrointestinal tract by eosinophils. The clinical presentation includes crampy abdominal pain, diarrhoea, distension and vomiting, often in an atopic or asthmatic patient. Peripheral eosinophilia is a frequent accompaniment. Any segment of the gastrointestinal tract may be affected but it most often involves the stomach, especially the antrum and the proximal small bowel. The clinical and imaging features depend on which layers of the GI tract wall are involved. Involvement may be predominantly mucosal, muscular or subserosal. Many cases are panmural and eosinophilic ascites is often seen in such cases. Radiographically, eosinophilic gastritis is characterised by fold thickening of the stomach and small bowel. Antral narrowing and rigidity with mucosal nodularity are frequently seen (Fig. 20.11).

Corrosive Ingestion Ingestion of caustic chemicals may cause severe injury to the stomach, sometimes leading to gastric necrosis, perforation and death. Acids are more injurious to the stomach and duodenum. Because the stomach

508

SECTION B  Abdominal Imaging

Fig. 20.11  Eosinophilic Gastroenteritis. CT shows diffuse thickening of the gastric wall in a patient proven to have eosinophilic gastroenteritis. No ascites was present. Symptoms resolved with steroid therapy.

secretes hydrochloric acid, it has the ability to neutralise alkaline agents. However, the already acidic gastric contents have no ability to neutralise strong ingested acids such as sodium hypochlorite (household bleach). The consequences of ingesting a corrosive agent follow a distinct course. First, there is necrosis, with sloughing of the mucosal and submucosal layers. In severe cases, full-thickness necrosis of the gastric wall may lead to perforation. In less severe cases, the denuded gastric wall develops a granulating surface and the formation of collagen then leads to fibrosis and stricture. The final result is a deformed, contracted and occasionally obstructed stomach which often necessitates total gastrectomy. The radiographic findings in corrosive gastritis depend on the severity of the chemical insult and the time that has elapsed since injury. Initially, swelling and irregularity of the gastric mucosa are seen, occasionally with visible blebs. As the mucosa sloughs, barium flows beneath it and the mucosa may then be seen as a thin radiolucent line paralleling the outline of the stomach. After a week or two, fibrotic contraction of the stomach becomes evident (Fig. 20.12). In severe cases, the lumen of the stomach may be no larger than that of the duodenal bulb.

Amyloidosis Amyloidosis is a rare condition that may cause gastric fold and/or wall thickening and rigidity. Luminal narrowing may mimic infiltrative tumour (linitis plastica). The condition is caused by the deposition of amyloid, a protein–saccharide complex in the stomach. Findings are non-specific and the diagnosis is confirmed by biopsy.

NEOPLASTIC DISEASES Mucosal Polyps Gastric epithelial polyps are typically asymptomatic and found incidentally in upper endoscopic or radiological studies. One to 4% of patients in the Western world who undergo gastric biopsy have gastric polyps, with even higher rates reported from the East. The prevalence and the type of the polyps encountered depend on the population being studied: hyperplastic and adenomatous polyps are highest in an H. pylori-infected population and fundic gland polyps are related to increased proton pump inhibitor (PPI) usage. Larger polyps can be

Fig. 20.12  Corrosive Gastritis Following the Ingestion of Household Bleach. The distal stomach has undergone considerable scarring and contraction in a manner similar to syphilitic gastritis or linitis plastica.

symptomatic from bleeding or distal prolapse, producing gastric outlet obstruction. Polyps are important because of their variable intrinsic malignant potential and their occurrence in patients who have other risk factors for developing gastric malignancy. For these reasons most polyps are biopsied or resected if possible, along with multiple topographic biopsies of the rest of the gastric mucosa. Surveillance is indicated in patients with higher risk for gastric cancer. SUMMARY BOX: Polyps • Gastric polyps are significant not only because of their variable intrinsic malignant potential, but also their occurrence in patients who have risk factors for developing other gastric malignancy. • Hyperplastic polyps are the most common type seen in association with insulted gastric mucosa. Although they have low malignant potential, the risk increases with their size. • Fundic gland polyps appear to be on the rise because of increased use of PPI therapy and have virtually no malignant potential. • Adenomatous polyps or adenomas are similar to colonic adenoma and are premalignant. They may also be seen in polyposis syndromes. • Hamartomatous polyps are nonneoplastic and occur sporadically or in polyposis syndromes or they may coexist with adenomatous polyps.

Hyperplastic polyps are the most common type seen in association with insulted gastric mucosa (H. Pylori gastritis, atrophic gastritis, pernicious anemia, ulcers and erosions, gastric surgical sites and bile reflux gastritis). They are randomly distributed throughout the stomach, usually multiple, uniform and of an average size of 1 cm, but may reach much larger dimensions (Fig. 20.13A and B). Although they have low malignant potential, the risk increases in polyps more than 2 cm in size; therefore larger polyps are completely excised. Fundic gland polyps appear to be on the rise due to increased use of PPI therapy and are currently the commonest type found in upper endoscopy in the United States. They are traditionally considered a variant of hamartomatous polyps and distributed exclusively in the gastric fundus and proximal body (Fig. 20.14A and B). They are small (usually 5 mm, protruding into the lumen; type II superficial, which is further subtyped as IIa elevated

512

SECTION B  Abdominal Imaging

6–8 mm in short axis) considered to identify nodal metastases. Also the metastatic nodal margin, architecture and the enhancement pattern can be different. Nevertheless, CT (or MRI) evaluation of N staging is still challenging. In the preoperative evaluation, it is reasonable to use more sensitive criteria at the expense of specificity to allow detection of potentially pathological nodes. It is recommended to indicate all visible lymph nodes regardless of their size and leave it to histology for accurate N staging. The common nodal groups involved are perigastric, coeliac axis, para-aortic and porta hepatis nodes. Recent literature suggests that the accuracy of MDCT is comparable to EUS for both T and N staging. Both EUS and MDCT are now considered complimentary techniques for locoregional staging of gastric cancer. CT and MRI are the techniques of choice for M staging (liver, peritoneum, lung, pancreas, retroperitoneum, adrenal, ovary (Krukenberg tumour) and diaphragm) of the gastric cancer. With the higher sensitivity of CT and the high specificity of FDG-PET, fusion of these imaging techniques (PET-CT) will be more useful than either alone for peritoneal metastases. Diagnostic laparoscopy is highly sensitive for identifying peritoneal metastases. The detection of gastric carcinoma by FDG-PET has certain limitations, such as: (1) Some histological subtypes (mucinous, signet ring and poorly differentiated adenocarcinoma) are not FDG-avid, (2) highly variable and sometimes intense focal physiological uptake within the normal gastric wall, (3) spatial resolution limits the ability to distinguish primary mass and the immediate perigastric nodal metastases and (4) there is lack of unified criteria in how to interpret imaging findings. FDG-PET has a better positive predictive value for lymph nodal metastases when compared with CT, especially for N3 metastases, which is important in the management of gastric cancer as treatment strategy may change from curative surgery to palliative measures. Metabolic response in FDG-avid tumours helps to identify therapy responders from non-responders earlier than morphological imaging techniques (Fig. 20.21A and B).

A

B Fig. 20.21  Positron Emission Tomography-Computed Tomography (PET-CT) in Advanced Gastric Cancer. (A) Un-enhanced axial CT image shows a large polypoid mass (M) in the posterior body of the stomach. The perigastric fat planes are preserved. (B) Fluorodeoxyglucose-PET image at the same level shows intense uptake in this moderately differentiated adenocarcinoma.

Gastric Lymphoma The gastrointestinal tract is the commonest site of extra-nodal lymphoma and the stomach is the most frequent site of gastrointestinal lymphoma. Gastric involvement may be primary, that is confined to the stomach and the regional lymph nodes, or secondary, as part of generalised disease. Primary gastric lymphomas are almost exclusively of nonHodgkin’s type (NHL), mostly of large B-cell type. Lymphoma of mucosa-associated lymphoid tissue (MALToma) is an indolent subtype of marginal cell type NHL, which is closely linked with chronic H. pylori infection, in particular strains expressing the Cag-a protein. Normally, there is no lymphoid tissue in the gastric mucosa. It is postulated that H. pylori infection may trigger the acquisition of MALT and subsequent inflammatory response may be a prerequisite for the development of MALToma, which may transform to intermediate or high-grade B-cell NHL. However, high-grade NHL may also arise de novo. Inflammatory bowel disease, coeliac disease, HIV infection and immunosuppression after solid organ transplant are other risk factors for gastric lymphoma. Lymphoma may involve any portion of the stomach in a diffuse or focal pattern. Although early lymphoma is confined to the mucosa and submucosa, gastric lymphomas are usually advanced at presentation. Primary gastric lymphoma has no typical radiographic appearance, and may mimic any of the appearances of gastric carcinoma. Radiographic findings of lymphoma, in both double-contrast barium and CT studies, parallel their gross morphological subtypes. The most

common appearance is that of an infiltrating lesion extending over a large area of the stomach with diffuse mural and fold thickening (Fig. 20.22A). Ulcerations may or may not be present. Other presentations are those of a bulky, polypoid mass, multiple submucosal nodules or single or multiple ulcers. CT may be helpful in differentiating gastric lymphoma from carcinoma. Preservation of perigastric fat planes at CT is more likely to be seen in lymphoma than in carcinoma, particularly in the presence of bulky tumour. In addition, the stomach remains pliable even with extensive lymphomatous infiltration, and the lumen is preserved, making gastric outlet obstruction an uncommon feature. However, NHL should be recognised as another cause of linitis plastica, an appearance that results from dense infiltration of lymphomatous tissue in the gastric wall without associated fibrosis. Adenopathy is seen both with carcinoma and lymphoma, but if it extends below the renal hila or if the lymph nodes are bulky, lymphoma is more likely. While FDG-PET has a well-established role generally in lymphoma, evaluation of primary gastric lymphoma is challenging due to unpredictable physiological FDG uptake in the stomach and variability in the degree of uptake in different histological subtypes (see Fig. 20.22B). PET-CT helps detect extragastric involvement in diffuse large B-cell subtype. Also FDG-PET has proven to be helpful in evaluating the response to treatment.

CHAPTER 20  The Stomach

515

endocrine neoplasms that arise in the mucosa and/or submucosa. Three clinicopathological types are described. Type I is most commonly associated with enterochromaffin-like cell hyperplasia, hypergastrinaemia and chronic atrophic gastritis, with or without pernicious anemia. Type II is the least common, seen in the hypergastrinaemic state of Zollinger-Ellison syndrome as part of multiple endocrine neoplasia (MEN 1). Both types commonly present in MDCT as small multiple nodules with or without central ulceration in the setting of diffuse gastric wall thickening. Type III are sporadic tumours and not associated with hypergastrinaemic state. Unlike types I and II tumours, type III tumours are large solitary ulcerative lesions. They are aggressive and are more likely to be locally invasive with nodal and hepatic metastases. Carcinoid syndrome may be seen in patients with liver metastases.

Metastatic Disease A

Metastatic tumours to the stomach are uncommon and occur late in the course of the disease. The most common primary tumours that metastasise to the stomach are breast, malignant melanoma and lung. Blood-borne metastases to the stomach initially appear radiographically as small intramural masses, usually multiple, that are indistinguishable from benign disease. As the disease progresses, they may assume similar morphology to primary gastric tumours. These may contain central ulcerations, having a bull’s-eye appearance. This is most frequently seen in metastatic melanoma, lymphoma and Kaposi sarcoma. Breast carcinoma may produce a linitis plastica-type appearance, indistinguishable from primary gastric carcinoma.

MISCELLANEOUS CONDITIONS Positional Abnormalities

B Fig. 20.22  Primary Gastric Lymphoma, Large B-Cell NHL. (A) CT shows diffuse wall thickening in the body of the stomach with giant folds (F). (B) The corresponding FDG-PET image shows intense tracer uptake in the involved area. Also seen is a hypermetabolic lymphomatous perigastric node (arrow).

SUMMARY BOX • The GI tract is the predominant site of extranodal lymphoma involvement. Primary lymphomas of the GI tract are rare, while secondary GI involvement is relatively common. • Stomach is the most common extranodal site of lymphoma (NHL type). The vast majority (>90%) of these lesions are approximately equally divided into two histological subtypes; either diffuse large B cell lymphoma or marginal zone B cell lymphoma of mucosa associated lymphoid tissue (MALToma). • H. pylori infection is highly associated with the development of MALToma of the stomach. • The diagnosis of gastric lymphoma is usually established during upper endoscopy with biopsy; PET-CT used to detect extra-gastric involvement and in evaluating treatment response.

Carcinoid Though the stomach is the least common site of gastrointestinal carcinoids, they are clinically significant because of the associated endocrinopathies that involve the stomach itself. Gastric carcinoids are well-differentiated

The stomach is attached to several peritoneal reflections, permitting its relative mobility. These include the gastrohepatic ligament (lesser omentum), the gastrosplenic ligament and the gastrocolic ligament, which is part of the greater omentum. The oesophagogastric junction passing through the oesophageal hiatus of the diaphragm normally fixes the proximal stomach while the distal stomach is anchored at the pyloroduodenal junction.

Hiatus Hernia Hiatal hernias are the most common positional abnormality in which the stomach herniates into the chest through the diaphragmatic hiatus when there is widening of the opening between the diaphragmatic crura. The prevalence of hiatal hernia increases with age and is present in over 50% of the aged population. Most hiatal hernias are small, involving a protrusion of a part of the gastric fundus at least 1.5 to 2 cm above the diaphragm. At the opposite extreme the entire stomach may be intrathoracic. Hiatal hernias may be divided into four types (Fig. 20.23A). Sliding hiatal hernias are the most common type of hiatal hernia (type 1) (see Fig. 20.23B). In this type of hernia the gastrooesophageal junction slides proximally through the diaphragmatic hiatus to assume an intrathoracic location. Small or moderate-sized hiatal hernias are often reducible, changing in size and configuration during barium evaluation and are best demonstrated with the patient recumbent in the right anterior oblique position. Sliding hiatal hernias are often accompanied by gastro-oesophageal reflux and reflux oesophagitis. In comparison to sliding hiatal hernias, in paraoesophageal hernias (type 2) the gastro-oesophageal junction is in its normal position below the diaphragm. The proximal stomach herniates through the oesophageal hiatus usually to the left of the distal oesophagus in the posterior mediastinum. This type of hernia is important because it is more prone to incarceration and obstruction than a sliding hernia.

516

SECTION B  Abdominal Imaging

A

C

B

Fig. 20.23  Hiatal Hernia. (A) Types of hiatal hernias: type I, sliding hernia; type II, paraoesophageal; type III, mixed type hiatal hernia; type IV, intrathoracic stomach. (B) Fluoroscopic image demonstrates type 1 hiatal hernia with gastric folds in the herniating portion of the stomach cardia (arrow) above the diaphragm with an incidental Schatzki ring (arrowhead). (C) Multidetector computed tomography shows barium-filled stomach in the chest (type 3).

Mixed hiatal hernia (type 3) is a combination of both types 1 and 2 together, with the gastro-oesophageal junction above the hiatus (see Fig. 20.23C). Type 4 hiatal hernia is composed of an intrathoracic stomach, which may demonstrate organoaxial rotation. Traumatic diaphragmatic hernias result from a tear in the diaphragm either from a direct penetrating injury or from a sudden increase in intra-abdominal pressure during blunt trauma. These hernias are almost always on the left side. Herniation may occur immediately after trauma or may be delayed by many years. Diagnosis is often difficult both due to lack of specificity of symptoms and it is often confused with simple elevation of the hemidiaphragm. On barium studies the recognition of the gastric hernia lateral to the normal oesophageal hiatus is crucial. CT is helpful in confirming the diagnosis.

Gastric Volvulus Gastric volvulus occurs when the stomach twists on itself between the points of its normal anatomical fixation. It is clinically important as it may cause gastric outlet obstruction or vascular compromise resulting in a surgical emergency. Classically, it presents with violent retching

with minimal amount of vomitus, severe epigastric pain and difficulty in passing a nasogastric tube. Gastric volvulus is most common in the elderly but may occur at any age. Most of the time gastric volvulus involves a stomach that is partially or totally intrathoracic and that rotates between the normally positioned gastric ligaments. Other predisposing factors include phrenic nerve palsy, eventration of the diaphragm, traumatic diaphragmatic hernia, gastric distension and abnormalities of the spleen. Gastric volvulus is often divided into two types depending on the plane of torsion (Fig. 20.24A). In organoaxial volvulus the stomach rotates along its long axis, which is a line drawn between the cardia and the pylorus. Rotation may be to the right or left. The configuration of the torsed stomach depends on the original shape and position of the stomach (horizontal or vertical). If the normal stomach was in a horizontal position, volvulus flips the stomach upwards so that the greater curvature is superior to the lesser curvature (see Fig. 20.24B). If the stomach was originally vertically orientated, volvulus causes a right–left twist. Mesenteroaxial volvulus is less common but more likely to have significant clinical consequences. In this type of volvulus the stomach

CHAPTER 20  The Stomach

517

A

A

B Fig. 20.25  Gastric Emphysema. (A) Abdominal radiograph in a patient with ischaemic gastritis after extensive abdominal surgery showing curvilinear lucency outlining the stomach wall. (B) Computed tomography of a patient with infectious emphysematous gastritis.

B Fig. 20.24  (A) Gastric volvulus: Organoaxial rotation about the long axis of the stomach (axis AA). Mesenteroaxial rotation on axis perpendicular to long axis of stomach and line joining the midpoints of the lesser and greater curvatures (B) Gastric volvulus, organoaxial type. Multidetector computed tomography coronal image shows an inverted stomach with the greater curvature above the lesser curvature (arrow).

rotates on an axis perpendicular to the long axis of the stomach along a line joining the middle of the lesser curvature to the greater curvature. This corresponds to the axis of the mesenteric attachments of the greater and lesser omentum. The characteristic appearance is an ‘upside-down stomach’. This type of volvulus is often associated with traumatic diaphragmatic ruptures. Radiographic signs of gastric volvulus include an air-fluid level of the stomach in the mediastinum and upper abdomen on upright plain radiography. On barium studies and on CT, the stomach will be vertically flipped (the greater curvature above, the lesser curvature below) in the

case of organo-axial rotation and horizontally flipped (with the distal antrum and pylorus assuming a position cranial to the fundus and proximal stomach) in the case of mesoenteroaxial rotation. The torsed area can usually be identified as the source of the obstruction.

Gastric Pneumatosis Disruption of the gastric mucosa permits air to enter the gastric wall. When this occurs without an underlying infection it is called gastric emphysema (Fig. 20.25A). Causes include corrosive ingestion, gastric ulcer, gastric outlet obstruction, chronic obstructive pulmonary disease (COPD), ischaemia and trauma. Air in the gastric wall caused by an acute infection with a gas-forming organism is called emphysematous gastritis (see Fig. 20.25B). Escherichia coli and Clostridium welchii are the usual causative agents. Acute panmural infectious gastritis by non-gas-forming organisms may also occur and is referred to as phlegmonous gastritis. Causative infectious agents include α-haemolytic streptococcus, Staphylococcus aureus, E. coli, C. welchii and Streptococcus pneumoniae. Infectious gastritis with any organism is unusual but is always a fulminant process associated with a high morbidity.

518

SECTION B  Abdominal Imaging

The radiological findings in both emphysematous gastritis and gastric emphysema include thin, curvilinear lines of radiolucent gas paralleling the gastric wall. CT is highly sensitive in detecting gas within the gastric wall.

Prepyloric Web (Antral Mucosal Diaphragm) Prepyloric web, believed to be a congenital lesion, occurs as a diaphragmlike, exaggerated fold of gastric mucosa orientated perpendicular to the long axis of the stomach. The thin web demarcates the distal antrum into a third small chamber between the proximal antrum and duodenal bulb. It appears as a thin persistent circumferential smooth band within 3 to 4 cm of the pylorus. The antral chamber produced by the web may mimic a second duodenal bulb. While usually asymptomatic, it may present with symptoms of gastric outlet obstruction.

Diverticula Gastric diverticula are most common in the posterior aspect of the fundus, below the oesophagogastric junction and near the lesser curvature, rarely in the antrum (Fig. 20.26A and B). These are true diverticula, that is containing muscularis propria, and thus are capable of peristalsis. They may be several centimetres in size and readily fill with barium. They rarely present a diagnostic dilemma but may mimic a submucosal mass if they fail to fill with barium. They may be mistaken for a gastric ulcer. In certain cases the CT appearance may be less specific if there is no luminal contrast medium or air. Intramural or partial gastric diverticula describes a rare anomaly in which there is invagination of gastric mucosa into the gastric wall. These diverticula are usually smaller than 1 cm in size and have a lenticular shape in profile with a small opening into the gastric lumen. They are typically located on the greater curvature of the distal antrum and are generally considered to be asymptomatic. Radiologically they may present a problem as they can be mistaken for ulcers, or for aberrant (ectopic) pancreatic rests, which typically occur in the same region.

A

Hypertrophic Pyloric Stenosis Hypertrophic pyloric stenosis (HPS) is a relatively frequent congenital disorder diagnosed in infancy. Presentation in adults occasionally occurs. The morphological features are due to hypertrophy and hyperplasia of the circular muscle with some contribution by the longitudinal muscle. The hypertrophied muscle lengthens and narrows the pyloric channel. Radiographically, there is lengthening of the pyloric channel (2–4 cm long) with smooth symmetrical narrowing. The hypertrophied muscle bulges retrogradely into the antrum, creating a ‘shoulder’. In infants, ultrasound (US) is the technique of choice in the diagnosis of HPS. A similar appearance may be seen with acquired hypertrophy of the distal antrum and pylorus. This is usually a sequela of peptic or other inflammatory disease. This form of pyloric stenosis typically lacks the retrograde bulge of muscle.

Varices Gastric varices are seen in most patients who have portal hypertension and oesophageal varices. Gastric veins provide one of the collateral pathways when there is obstruction of the portal vein. The presence of gastric varices in the absence of oesophageal varices is a sign of splenic vein thrombosis most often associated with pancreatitis or pancreatic carcinoma. Varices are most often seen in the fundus around the oesophagogastric junction sometimes involving the proximal body. They appear as widened, effaceable polypoid folds. They may be nodular-appearing, ‘grape-like’ or appear mass-like, in which case they may mimic gastric cancer. Rarely

B Fig. 20.26  Gastric Diverticulum. (A) The patient is upright and a barium/ air level is present in the fundal diverticulum. (B) Multidetector computed tomography in a different patient with fundal diverticulum (arrow).

they occur in the antrum without fundal involvement. Transabdominal US and EUS are important techniques for definitive diagnosis of gastric varices. Varices are seen as multiple nodular and serpiginous submucosal masses in the fundus in double contrast studies (Fig. 20.27A). They are also well demonstrated on contrast-enhanced MDCT and MRI as submucosal enhancing vessels in venous phase imaging (see Fig. 20.27B).

Gastric Distention Gastric distension may be obstructive or non-obstructive. PUD and gastric cancer account for more than 90% of cases of gastric outlet obstruction. PUD used to be the most common cause of gastric outlet

CHAPTER 20  The Stomach

519

TABLE 20.4  Normal Values for

Standardised Solid Gastric Emptying Study Time (h)

% Gastric Emptying

1 2 4

>10 but 40 >90

abnormal gastric wall when gastric malignancy is the cause of obstruction (see Fig. 20.19). Non-obstructive gastric dilatation may be acute or chronic. Rapid sudden gastric distension or gastric ileus occurs most often as a complication of abdominal surgery or acute trauma. Chronic gastric dilatation is due to hypomotility and is further discussed in the next section.

Gastroparesis

A

B Fig. 20.27  (A) Gastric varices. Multiple nodular and serpigenous submucosal masses in the fundus suggesting gastric varices. (B) Corresponding contrast enhanced axial computed tomography (arrow) showing submucosal varices during the venous phase.

obstruction in adults; however, in recent decades with improved treatment of PUD, malignancy is now the leading cause of obstruction. Duodenal and pyloric channel and distal antral ulcers are the usual culprits. Obstruction is caused by one or more factors including spasm, scarring, acute inflammation and muscle hypertrophy. Other less common causes are pancreatitis or pancreatic cancer, Crohn disease, sarcoidosis, tuberculosis, and syphilis. Abdominal radiographs show the outline of the dilated air-filled stomach or downwards displacement of the transverse colon by a fluid or an air-filled stomach. Barium studies are less than ideal for the delineation of the obstruction. CT is superior in depicting a mass or

Gastric dysmotility is characterised by delayed gastric emptying in the absence of a mechanical obstruction with symptoms of gastric stasis such as early satiety, nausea, vomiting, bloating, and abdominal pain. Although most cases are idiopathic, a variety of metabolic and infiltrative diseases, neuropathies and medications have been implicated in its pathogenesis. In patients with suspected gastroparesis, imaging plays a dual role of (a) excluding any unsuspected mechanical obstruction, usually with CT or upper endoscopy and (b) objectively documenting gastric stasis using scintigraphic gastric emptying study. Technique: Patient preparation for scintigraphic gastric emptying study involves withholding medications that affect gastric emptying for at least 48 hours before the study. In patients with diabetes, hyperglycemia should be treated and the test performed only when blood glucose level is 10% at 4 hours and/or >60% at 2 hours when using the standard radioactive meal described above (Table 20.4). The 4-hour time point is considered more sensitive for detection of delayed gastric emptying compared with the other time point. Although the severity of symptoms do not always correlate with the rate of gastric emptying, delayed gastric emptying has been classified based on the extent of gastric retention on scintigraphy at 4 hours as mild (10–15%), moderate (15–35%) and severe (>35%). Gastric hypermotility is suggested if there is less than 30% retention (more than 70% emptying) at 1 hour (see Fig. 20.28C).

520

SECTION B  Abdominal Imaging

A

B

C Fig. 20.28  (A) Gastric emptying study time activity curve:Emptying in a monoexponential pattern soon after fluid enters the stomach; this pattern is typical for fluids. (B) Initial lag period followed by linear emptying, typical of solids. (C) Stomach activity emptied consistent with moderate degree of gastroparesis.

CHAPTER 20  The Stomach

Ectopic Pancreas Ectopic pancreas (pancreatic rest, aberrant pancreas) is the presence of pancreatic tissue in the submucosa of the gastrointestinal tract. On imaging studies it is most commonly visible along the greater curvature of the antrum. The ectopic pancreatic tissues are usually solitary deposits that appear as sharply defined submucosal nodules usually 3 N2 Distant metastasis in one organ M1a Distant metastasis in >1 organ M1b Metastasis to the peritoneum with or without distant organ involvement

Dukes

5-Year Survival

A

85%–95% Fig. 22.21  2D axial CT colonogram demonstrating a flat cancer manifesting as lobulated fold thickening (arrows).

B

60%–80%

C

30%–60%

D D D

10 cm) and is associated with vascular engorgement and mesenteric fluid. However, there is clear overlap in imaging findings

Fig. 22.40  Endoluminal view projection of computed tomography colonography (CTC). Digital cleansing has been employed to remove high intensity faecal tagging solution and a three-dimensional reconstruction performed to render the endoluminal surface. The technique can be useful to provide an overview of the colon akin to BCBE but with the superior diagnostic performance of CTC.

586

SECTION B  Abdominal Imaging

between the two (Fig. 22.42). Perfusion CT parameters may have a role in distinguishing between cancer and diverticulitis. US is useful in episodes of mild diverticulitis with localised pain. Graded compression over the area of tenderness reveals the pericolic abscess as a low reflective collection related to the bowel wall and surrounded by reflective inflamed fat. MRI may also have a role. Fistula formation most commonly involves the bladder. Conventionally a contrast enema is performed in cases of suspected colovesical fistula, although CT or MRI is increasingly the first-line investigation. The fistulous track may not always be seen, but the presence of gas in the bladder (in the absence of recent instrumentation), abscess related to the bladder wall, bladder wall thickening and/or adjacent adherent colon are highly suggestive (Fig. 22.43). Giant cyst formation is a rare complication and represents a pseudocyst with no epithelial lining

formed from expansion of a walled-off subserosal perforation. Its association with diverticular disease is apparent on CT or DCBE. The differential diagnosis includes a duplication cyst. Perforation or volvulus are rare complications.

Fig. 22.41  Axial computed tomography shows acute diverticulitis manifest by a thickened sigmoid with diverticula and adjacent inflammatory fat stranding (arrow). A pelvic abscess (arrowhead) has formed next to the inflamed sigmoid.

Fig. 22.43  Coronal oblique computed tomography showing thickened sigmoid colon with a gas-filled track (arrow) extending towards the inflamed, thickened bladder dome (arrowhead). This was confirmed as a colovesical fistula at surgery.

EPIPLOIC APPENDAGITIS Infarction of an epiploic appendage is most common either in the sigmoid or caecum where the appendages are most prominent, and causes acute pain and tenderness similar to diverticulitis or appendicitis. Most resolve spontaneously in about 2 weeks. The typical appearance on US is of a non-compressible pericolic hyperechoic ovoid mass immediately under the abdominal wall, and on CT focal hyperattenuation with a central area of fat density (Fig. 22.44).

Fig. 22.42  (Left) Abdominopelvic contrast enhanced computed tomography (CT) and (Right) T2 weighted magnetic resonance imaging (MRI) at corresponding levels through a segment of circumferential mural thickening on a background of extensive sigmoid diverticulosis. While there is minimal shouldering or ulceration and hence limited evidence of malignancy on CT, MRI shows intermediate signal intensity mural tissue extending into the adjacent pericolonic fat; endoscopy confirmed locally advanced tumour.

CHAPTER 22  The Large Bowel

Fig. 22.44  Axial computed tomography of epiploic appendagitis showing the typical, central ovoid fat-density mass (arrow) with peripheral enhancement (arrowheads).

COLITIS Endoscopic techniques remain the primary diagnostic modality for colonic inflammation (colitis), facilitating direct mucosal inspection and histological sampling. However, radiological imaging plays a large role, both in diagnosis and particularly in follow-up and detection of complications. Contrast enema (particularly using barium) can produce exquisite mucosal detail, but cross-sectional techniques are now much more widely used. There are numerous causes of colitis, and imaging features are often non-specific.

Imaging Features of Colitis The hallmark of colitis on cross-sectional imaging is wall thickening (4 mm or more). Contrast enhancement is seen best in the enteric phase (typically 45 to 50 s post-IV administration of contrast agent). The degree and pattern of contrast enhancement may give an indication of disease activity in IBD, although may not particularly reduce the differential diagnosis in unknown disease. For example, a layered or striated pattern (central and peripheral enhancement with a central layer of relative reduced enhancement) may be seen in IBD, infection and ischaemia, among others (see Table 22.1). The ability to visualise extramural tissue is a significant advantage of cross-sectional techniques over luminal radiology; mesenteric oedema, abscesses, fistulation and lymphadenopathy are well seen. A phlegmon is an ill-defined inflammatory mass without overt abscess formation. These present as poorly defined focal masses of increased attenuation in adjacent omentum or mesentery. Abscesses are of low density (10 to 30 HU) and often contain gas bubbles either from gas-forming bacteria or a direct communication to bowel. On US, the bowel wall is typically stratified in ulcerative colitis (UC) with differentiation between the submucosal and muscularis propria (Fig. 22.45), although in chronic Crohn disease (CD) this may be lost. The surrounding fat is more reflective with acute inflammation. High-resolution isotropic cross-sectional images allow some appreciation of mucosal disease in a well-distended bowel. However, high-definition thin-layer contrast studies, such as the DCBE, are required to show superficial changes, although with the ubiquitous nature of endoscopy, the need for such detailed mucosal assessment is

587

Fig. 22.45  High-resolution transverse ultrasound image of the descending colon shows marked mural thickening and an exaggerated mural stratification pattern in keeping with colitis. The arrow shows the echogenic submucosa.

reducing in clinical practice. Aphthoid ulceration is typical of CD, does not occur in UC, but may be seen in amoebiasis, tuberculosis (TB), Behçet’s disease, and human immunodeficiency virus (HIV)-related infections. Reflux ileitis with a patulous ileocaecal valve and granular distal 10–15 cm of ileum is a typical feature of UC and disappears rapidly following colectomy. Extensive ulceration in acute colitis may virtually denude the mucosa, leaving oedematous remnants as pseudopolypoid elevations. With ulcer healing, mucosal tags may form sessile, filiform, adherent or bridging polyps, and are best called post-inflammatory polyps.

Inflammatory Bowel Disease IBD usually refers to UC and CD, although indeterminate colitis is also often included. UC is a relapsing and remitting disease characterised by bloody diarrhoea. The rectum is always involved, and the colitis is in continuity to its proximal extent. The patient with CD typically presents with diarrhoea, abdominal pain and weight loss, with peak ages of onset at 20 and 50 years. Several classifications have been proposed, for example, the Montreal Classification, which considers age at diagnosis, disease location (ileal, colonic, ileocolonic or upper disease), stricturing, fistulation and anal involvement. Anal disease is associated with colonic involvement and affects 5% of individuals. The rectum is involved in only about half of cases and the disease is characteristically patchy and asymmetric. Stool culture and laboratory tests are needed to exclude infective colitides and multiple biopsies are required to define the presence and nature of the colitis.

Differential Features UC and Crohn colitis are different diseases, though they have some features in common. UC typically presents with a granular mucosa, rectal involvement and symmetrical disease that is in continuity to its proximal extent, with a shortened and narrowed bowel. Inflammatory changes are limited to the mucosa. In comparison, CD is asymmetrical with aphthoid ulceration. Inflammation is transmural, and deep fissuring ulceration may lead to fistula formation. The transmural nature of CD

588

SECTION B  Abdominal Imaging

is reflected in the bowel wall thickness, which can be measured on CT, MRI, US or plain radiography. The normal thickness is less than 3 mm; it may be 5 to 8 mm in UC, but can be grossly thickened in CD (on average 11 ± 5.1 mm). This tends to be greater in CD than UC (7.8 ± 1.9 mm). Extraluminal manifestations of IBD such as phlegmon, abscess and fistula formation suggest CD. Small-bowel involvement in UC is limited to reflux ileitis.

Disease Activity in Inflammatory Bowel Disease Disease activity assessment in IBD is important with the widespread use of powerful anti-inflammatory medication such as anti tumour necrosis factor alpha (TNF-α) agents. Endoscopic scoring systems exist, mainly based on the presence and extent of ulceration, and the presence of stenosis. Histopathological features of inflammatory activity include ulcer formation, neutrophilic infiltration and crypt abscess formation. Cross-sectional imaging can help differentiate acute from chronic disease and assess treatment response. Increasing wall thickness and contrast enhancement, particularly in a layered pattern, are correlated with disease activity on CT, MRI and US, as is mural T2 signal hyperintensity on MRI (Fig. 22.46). Mesenteric vascular encouragement (comb sign), mesenteric oedema/fluid and lymphadenopathy are also associated with active disease. Fibrotic disease tends to exhibit moderate heterogeneous contrast enhancement, low signal on T2W MRI and no adjacent mesenteric oedema or fluid. An increase in submucosal fat is characteristic of chronic UC and mesenteric fibrofatty proliferation accompanies chronic changes of CD. Fatty proliferation is responsible for widening the presacral space in rectal disease. Recent research suggests that MRI quantification of body composition metrics such as visceral adiposity may be a useful predictive biomarker of disease severity or progression. Differentiation of fibrostenotic, chronic and active inflammatory strictures continues to challenge established cross-sectional imaging

A

techniques. However, US strain elastography, shear wave elastography and magnetisation transfer MRI all show promise for evaluating mural fibrosis both in vitro and in human trials.

Carcinoma in Colitis There is an increased incidence of CRC in UC. Patients at risk are those with an extensive colitis of more than 10 years’ duration. The carcinomas arise from dysplastic changes within the diseased epithelium and not from adenomas as in the general population. The tumours are consequently frequently multiple and infiltrative. Colonoscopic surveillance with multiple biopsies is mandatory for patients with chronic UC, and imaging plays a limited role. Dysplasia is essentially a histological diagnosis, as it may be found in a flat mucosa and be unrecognisable radiologically. Dysplasia-associated lesions or masses (DALMs), similar to villous adenomas, represent severe dysplasia and are occasionally visible radiologically.

OTHER COMMON CAUSES OF COLITIS In general, the imaging appearances of colitis are non-specific, but there are features which may narrow the differential diagnosis. Location is important (Table 22.6), but can only be used as a guide.

Ischaemic Colitis Ischaemia is a common cause of colitis in the elderly (over 90% of those affected are over 60 years old), typically presenting with abdominal pain and rectal bleeding of sudden onset. There are a variety of causes: (1) mesenteric occlusion, arterial and venous; (2) mechanical, from strangulation or raised intracolonic pressure, such as proximal to an obstruction; or (3) low-flow states. In younger patients, hypercoagulable states, vasculitis, long-distance running and use of cocaine are also

B Fig. 22.46  (A) Post-gadolinium T1 high-resolution interpolated volumetric examination fat-suppressed magnetic resonance imaging (MRI) at 3T shows a thickened sigmoid colon secondary to Crohn colitis. The layered enhancement pattern suggests active disease (arrow). (B) T2 weighted MRI with fat suppression shows a thickened hepatic flexure in a different patient with Crohn disease. The increased mural T2 signal (arrow) represents oedema and is also indicative of active inflammation.

CHAPTER 22  The Large Bowel

589

TABLE 22.6  Common Location of Colitis According to Aetiology Diffuse

Mainly Right-Sided

Mainly Left-Sided

Ulcerative colitis CMV Escherichia coli Pseudomembranous colitis

Crohn disease Salmonella TB Yersinia Amoebiasis Neutropenic enterocolitis (typhlitis) Immunosuppressive states (including AIDS) Ischaemic colitis Hypovolaemic states in young patients Cocaine users

Ulcerative colitis Shigellosis Lymphogranuloma venereum Gonorrhoea Ischaemic colitis Radiation Diverticulitis

CMV, Cytomegalovirus; TB, tuberculosis.

or significantly thickened from bacterial superinfection. Mesenteric fat stranding and free fluid may be seen with transmural necrosis, venous occlusion or superadded infection. Pneumatosis or portomesenteric gas indicates transmural necrosis. Large vessel occlusion or aneurysm formation should also be looked for. Sacculation is common and is a feature of strictures due to either ischaemia or CD.

Radiation Colitis Radiation-induced bowel damage is a late complication, often presenting years after otherwise successful therapy where the total dose has exceeded 45 Gy (4500 rad). The rectum is directly involved in the treatment of rectal cancer and may be indirectly affected following therapy for gynaecological or prostatic malignancy. The pathogenesis of radiation enteritis is occlusive endarteritis with thrombosis and fibrosis. In the early phase there is mucosal injury with an acute colitis, while the chronic changes involve a proctitis with possible ulceration, rectal stricturing, or fistula formation, usually to the vagina or bladder. Strictures may be smooth and symmetrical, but if there has been ulceration, these may be deformed with thick irregular folds. Perforation is rare. A barium enema will demonstrate the deformity or fistula. CT and MRI show generalised changes of rectal wall thickening, an increase in the mesorectal fat density, thickening of the mesorectal fascia and widening of the presacral space (Fig. 22.48). Specific changes of complications such as fistula formation also may be seen. Fig. 22.47  Coronal computed tomography shows mural thickening (arrowheads) due to acute ischaemic colitis extending from around the splenic flexure distally. The entire descending colon was affected (not shown). The mucosa retains brisk enhancement.

causes. The region of the splenic flexure is most commonly affected (due to the watershed in vascular supply), but anywhere in the colon can be involved. Indeed, in low-flow states, the right colon is more commonly affected. The mucosa is most susceptible to vascular compromise but can repair, whereas necrosis of the submucosal and muscle layer creates more fibrosis, leading to stricture formation. Transmural necrosis or ‘bowel infarction’ is life threatening, requiring immediate surgical intervention. Plain radiographs may reveal narrowing and thumbprinting, and help exclude toxic megacolon, free perforation and intramural or portal venous gas. US may show a thickened wall (mean 7.6 mm) with stratification. On CT (Fig. 22.47) wall thickening is more marked with venous occlusion. A low-attenuation target sign is due to submucosal oedema. Wall thickness does not correspond to the extent of necrosis. The bowel may be relatively thin with transmural necrosis,

Behçet’s Syndrome This is a chronic multisystem vasculitis that may affect the colon, usually the ileocaecal region in about 30% of cases. Deep discrete ulcers may lead to haemorrhage or perforation. CT shows a polypoid mucosa with wall thickening and marked enhancement. There is relatively little nodal enlargement or fibrofatty change, which helps distinguish the condition from CD. Also, pericolonic inflammatory changes are minimal, unless there has been a perforation.

Infectious Colitis Salmonella, Shigella and Campylobacter may all present as a localised or diffuse colitis, with a granular or ulcerated mucosa. There may be marked ileus in the acute stages of salmonellosis, and toxic megacolon has been reported. Cytomegalovirus (CMV) causes vasculitis with a thick wall, lymphadenopathy and large ulcers that may bleed, and is typically ileocolic in distribution. CT demonstrates wall thickening, serosal enhancement, mesenteric lymphadenopathy and often ascites. Chlamydia trachomatis causes lymphogranuloma venereum. A chronic proctitis is complicated by fistula formation, extensive fibrosis and eventual stricture formation.

590

SECTION B  Abdominal Imaging

Pseudomembranous Colitis

Neutropenic Colitis

This colitis results from the effects of cytoplasmic endotoxins produced by overgrowth of Clostridium difficile, usually as a result of broadspectrum antibiotic therapy. It is characterised endoscopically by yellowish plaques formed by sloughed mucosal cells that create the pseudomembrane. The presentation is with diarrhoea, pyrexia and leucocytosis, but may be fulminant with perforation from necrosis; hence, this is a potentially life-threatening condition. Plain radiographs may show a generalised ileus and nodular haustral thickening. CT, US and MRI may show gross wall thickening, with marked mucosal enhancement and extensive low attenuation from submucosal oedema. This produces a very prominent target sign, often described as the ‘accordion sign’ (Fig. 22.49). This is typical of pseudomembranous colitis but may be seen in acquired immunodeficiency syndrome (AIDS)-related or ischaemic colitis, and in severe oedema from cirrhosis. Pericolic stranding is minimal and ascites may be present.

Neutropenia may be due to a number of causes, commonly chemotherapy with bone marrow transplantation. CT shows less bowel wall thickening than in pseudomembranous colitis, but pneumatosis is common (21%). Changes are typically right sided and may be limited to the caecum, which is why the term ‘typhlitis’ is often used (Fig. 22.50). Mesenteric stranding and small-bowel involvement are common.

Fig. 22.48  Axial T2 weighted abdominopelvic magnetic resonance imaging showing thickened, featureless sigmoid colonic segment (arrows) in contrast to the normal mural thickness of the adjacent caecum (arrowheads). While the features are relatively non-specific, in the context of prior radiotherapy, these features are compatible with radiation enteritis.

A

Parasitic Colitis In trichuriasis, the small, coiled worms may be seen on the mucosal surface on DCBE. Strongyloides stercoralis may simulate UC. In Chagas

Fig. 22.50  Coronal contrast enhanced computed tomography showing terminal ileal and proximal colonic thickening, submucosal oedema and mural stratification in this neutropenic patient undergoing chemotherapy for peripheral sarcoma. Interpretation is challenging due to the paucity of intraabdominal fat and low dose acquisition in this young patient but in this context, morphology and distribution are typical of neutropenic enterocolitis, commonly referred to as typhlitis.

B Fig. 22.49  (A) Ultrasound showing florid bowel wall and haustral fold thickening (arrows) secondary to pseudomembranous colitis. (B) The corresponding computed tomography confirms marked colonic thickening and mucosal hyperenhancement (arrow).

CHAPTER 22  The Large Bowel disease, a megacolon results from the neurotoxic effect of the protozoon Trypanosoma cruzi. In schistosomiasis, ova are deposited in the submucosa of the large bowel. The inflammatory response results in the formation of numerous polyps. Fibrosis may later cause stricture formation, and calcification may be visible in the bowel wall.

Tuberculosis Most intestinal TB used to be secondary to pulmonary disease, but it is now more likely to be of primary bovine origin from drinking unpasteurised milk. Ulcerative, hypertrophic or fibrotic forms are described. The ulcers tend to be large and circumferential with a shaggy edge, while the hypertrophic form presents with an inflammatory mass and stenosis of the bowel lumen. TB is commonest in the ileocaecal region but may be seen in any part of the gastrointestinal tract and may be indistinguishable from CD. It must always be considered in a patient from an endemic area, whatever the appearance of the colitis. Certain changes are suggestive of TB: a conical contracted caecum with a patulous ileocaecal valve and a dilated terminal ileum and transverse ulceration with a short hourglass stricture sharply demarcated from normal bowel. The hypertrophic form with a large exophytic mass may be difficult to distinguish from a lymphoma. Muscle involvement suggests TB, with actinomycosis as a differential. Wall stratification, increased vascularity and fibrofatty proliferation are not typical of TB and all favour CD, whereas ascites, peritoneal involvement, and lymphadenopathy favour TB (Fig. 22.51). Central caseous necrosis in lymph nodes creating a hypoechoic centre on US and peripheral enhancement on CT or MRI is highly suggestive of TB.

Amoebiasis In endemic areas, approximately 20% of the population harbour the cystic form of the protozoan Entamoeba histolytica. The radiological features of invasive amoebiasis include a segmental or diffuse colitis, with a granular or ulcerated mucosa. Aphthoid ulceration may be seen,

Fig. 22.51  Coronal computed tomography of ileocaecal tuberculosis with caecal and terminal ileal thickening (arrows) and ascites (arrowheads).

591

and amoeboma formation occurs in about 10% of cases. These inflammatory masses comprising granulation cause an irregular stricture that may simulate a carcinoma; they are often multiple and are usually found at the flexures and the caecum. Embolic spread to the liver is seen in about 15% of cases. It is essential to examine fresh stools for trophozoites in all patients with colitis to exclude amoebiasis.

Acquired Immunodeficiency Syndrome HIV infection and immunosuppression lead to a complex of intestinal infection and neoplasia, often superimposed on venereally acquired infections such as gonorrhoea, chlamydia and herpes simplex. CMV infection is common. Cryptosporidium and Mycobacterium avium-intracellulare cause nonspecific changes on barium enema. Kaposi sarcoma most frequently involves the rectum with diffuse submucosal nodules that may coalesce into a mass. The lymph nodes are hyperaemic and show increased attenuation. A specific diagnosis from imaging is difficult, even with typical changes, as multiple infections are common.

Defunctioned Colon The defunctioned colon always has a low-grade bacterial colitis causing narrowing and loss of haustration (Fig. 22.52). Barium may be retained for years and hence a water-soluble contrast agent is recommended for all fluoroscopic examinations of defunctioned bowel.

Acute Fulminant Colitis Any colitis including CD, ischaemic colitis, amoebiasis, antibioticassociated colitis and salmonellosis can become fulminant where the inflammation becomes transmural and ulceration extends deeply into the muscle layer with neuromuscular degeneration, potentially leading to toxic dilatation and perforation. This complication is most commonly seen in UC, and accounts for most UC-related deaths. The major signs of toxic megacolon are dilatation, loss of normal haustral contours and mucosal islands (Fig. 22.53). Plain radiography remains the mainstay in the diagnosis and monitoring of the condition, although cross-sectional techniques, particularly MRI, US and CT, have an increasing role. Dilatation of greater than 5 cm is associated with ulceration deep into the muscle layer, and represents an initial stage of

Fig. 22.52  Axial T1 weighted fat-suppressed T1 weighted fast spin echo showing typical features of diversion colitis. Despite the unenhanced sequence, the defunctioned descending colon (arrow) is thickened with stratification of mural layers and subtle perienteric fat stranding.

592

SECTION B  Abdominal Imaging

TABLE 22.7  Causes of Large-Bowel

Strictures Physiological Surgical Malignant

Diverticular disease Ischaemia Radiation colitis Inflammatory bowel disease

Miscellaneous

Distended bladder, spasm Anastomosis, site of colostomy Annular, scirrhous, metastatic carcinoma, lymphoma Pericolic abscess Sacculation common as with Crohn’s strictures In radiation field so usually rectosigmoid Ulcerative colitis, Crohn disease, tuberculosis, lymphogranuloma venereum, amoebiasis Extrinsic mass, endometriosis, pelvic lipomatosis, trauma

distension of the small bowel is a poor prognostic sign for successful medical treatment. Fig. 22.53  Coronal T2 Weighted Magnetic Resonance Imaging Showing Gross Dilatation of the Transverse Colon (Arrows). In the correct clinical setting, a haustral appearance and florid dilatation allows the diagnosis of toxic megacolon.

Perforation Perforation is most likely to occur during an acute attack of UC, within the first year of onset of the disease. Perforation is the result of deep ulceration, which may be due to severe localised disease, or as part of toxic megacolon. Free perforation is recognised by the presence of intraperitoneal gas, but sealed perforations cannot be reliably detected from plain radiographs. CT has high sensitivity for extraluminal gas in suspected perforation. Free perforation is rare in CD as, unlike in UC, the chronic transmural inflammatory nature of the disease causes adherence to adjacent structures. Cross-sectional imaging is required to demonstrate a localised pericolic abscess.

MISCELLANEOUS CONDITIONS Large-Bowel Strictures

Fig. 22.54  Axial T2 weighted magnetic resonance imaging shows rectal and sigmoid mural thickening (arrowheads) and pelvic free fluid (arrow) in a patient with ulcerative colitis.

the process. In established cases the dilatation may be greater than 8.5 cm. Haustration is always absent, and toxic megacolon should not be diagnosed if it is preserved. Changes are observed mostly in the transverse colon in the supine position, as this is the least dependent part of the colon where intraluminal gas will collect. Mucosal islands are oedematous remnants of mucosa and indicate the very extensive nature of the ulceration. The colon has the consistency of wet blotting paper and perforation is frequent. The distinction between severe colitis and early toxic megacolon may be difficult, and serial radiographs are helpful to monitor progress. MRI may have a role in staging and monitoring acute colitis—unprepared T2W axial and coronal images can efficiently image the colon and dilation, mural thickening, perimural mesenteric oedema and fluid are all easily appreciated as hallmarks of acute disease (Fig. 22.54). Gaseous

It is important to distinguish functional colonic narrowing from pathological causes (Table 22.7). The incidence of localised spasm is reduced by using a smooth-muscle relaxant, but it may still occur at one of the ‘physiological sphincters’ in the colon. There are seven such sites, of which Cannon point in the mid-transverse colon is the best known. Spasm is easily abolished by further IV relaxants and gas insufflation. The DCBE gives a purely luminal view with a positive predictive value of 96% for malignant and 86% for benign strictures. Classically, a fibrotic stricture has a smooth lumen with tapering ends, whereas a malignant one has an irregular lumen with shouldered ends. Scirrhous carcinoma is a rare exception, which may look more benign than malignant. Narrowing in diverticular disease is common. The retention of mucosal folds and the spiculated necks of compressed diverticula are important features in distinguishing a benign from a malignant stricture. In chronic UC there is considerable hypertrophy of the muscularis mucosae and submucosal thickening with fat. The smooth muscle changes are probably responsible for the generalised shortening of the colon, and may produce localised strictures in the left colon in 10% to 20% of patients with extensive long-standing UC strictures. Strictures in CD (Fig. 22.55) are usually asymmetrical with sacculation and secondary to ulceration on the antimesenteric border. As noted earlier, differentiation from malignancy can be problematic, particularly with cross-sectional techniques; any irregular raised area, shouldering or asymmetry suggests malignancy. Cross-sectional imaging allows evaluation of the mural thickness, enhancement pattern and extracolonic findings, which may help limit

CHAPTER 22  The Large Bowel

Fig. 22.55  T1 high-resolution interpolated volumetric examination fatsuppressed magnetic resonance imaging Post-gadolinium contrast enhancement shows dilated colon upstream of a distal transverse colonic stricture (arrowheads) in this patient with Crohn’s disease. Mesenteric induration around the vasa recta gives rise to the characteristic ‘comb sign’ (arrows).

the differential diagnosis. The site of the stricture is also significant: radiation strictures are related to the field of therapy and so invariably affect the rectosigmoid colon; endometriosis usually involves the anterior wall of the rectosigmoid; ischaemic strictures are most common in the region of the splenic flexure.

Pseudodiverticula Sacculation of the bowel wall is frequent in CD, secondary to fibrosis in healing eccentric ulceration. Pseudodiverticula may be seen in ischaemic strictures, but are rare in other forms of colitis and never occur in UC. Wide ‘square’-shaped diverticula in focal areas of bowel wall weakness are seen in scleroderma.

Appendicitis Appendicitis is the commonest abdominal emergency in the UK. Around 9% of men and 7% of women will have appendicitis at some point, typically adolescents or young adults. Imaging is only needed if the clinical diagnosis is uncertain. CT is probably the most sensitive and specific test (Fig. 22.56), but a strategy based on initial US (with CT only if this is equivocal) has good positive and negative predictive value and reduces radiation exposure in this (often young) group. MRI, particularly with DWI, also shows substantial promise particularly during pregnancy or for paediatric presentations (Fig. 22.57).

Lipomatous Disorders of the Large Bowel Lipomatous disorders include lipomatous infiltration of the ileocaecal valve, solitary lipomas or pelvic lipomatosis. Lipomatous infiltration of the ileocaecal valve causes diffuse enlargement of the valve, the surface of which may be smooth or lobulated. Two-thirds of all gastrointestinal lipomas are in the colon, with most being solitary lesions in the right

593

Fig. 22.56  Coronal oblique computed tomography showing a dilated, blind-ending appendix with surrounding inflammatory fat stranding (arrow) and an appendicolith (arrowhead) in the orifice.

colon. Those greater than 4 cm may cause pain, bleeding, or intussusception. Lipomas are submucosal, so that the luminal surface is smooth, with no mucosal line at the edge of the lesion. The fat content, with a Hounsfield unit reading from −80 to −120 (around −100 HU), is apparent on CT, which is the optimum method for diagnosis. Pelvic lipomatosis is a rare condition of unknown aetiology in which there is proliferation of adipose tissue in the pelvis. The bladder and rectum are compressed. On plain radiographs, there is increased radiolucency of the pelvis and exceptionally good delineation of the sacrum. The presence on CT of a diffuse increase in pelvic fat is diagnostic.

Pneumatosis Coli The origin of the gas cysts seen in this condition in the submucosal and subserosal layers of the bowel wall is uncertain. Small mucosal tears probably allow gas or gas-forming bacteria to enter the wall. Once a pocket is established in this way, diffusion into and out of the cyst may balance so that the lesion becomes self-perpetuating. Prolonged oxygen therapy alters these diffusion gradients, collapsing the cysts. Pneumatosis coli may be asymptomatic or present with diarrhoea or constipation, and rectal bleeding may result from superficial erosions. Pneumoperitoneum from the rupture of a cyst is rare. The cysts are well-defined, closely packed, gas-filled lesions about 1 to 2 cm in diameter. A segment of the left colon is usually involved and the intramural location of the cysts is confirmed on imaging (Fig. 22.58). The plain radiographic changes are typical and should not be confused with those of other causes of gas in the bowel wall, such as necrotising enterocolitis, where there may be numerous minute foamy pockets of gas within a necrotic segment of bowel, or crescentic linear gas shadows running parallel to the bowel wall and portal venous gas in gross cases.

Volvulus For volvulus to occur, the colon must be on a mesentery; thus, the sigmoid is the commonest site. However, the caecum, transverse colon and splenic flexure are potential areas (Fig. 22.59), as is any part of the colon on a

594

SECTION B  Abdominal Imaging

Fig. 22.57  Axial T2 weighted magnetic resonance imaging (left) with corresponding high B-value (B600) diffusion weighted imaging sequence (right) showing focal inflammation involving the appendix tip in this paediatric patient. The axial T2 weighted sequence shows subtle mural thickening and stratification at the appendix tip (arrow) while the diffusion-weighted imaging confirms restricted diffusion (arrow) due to active inflammation.

Fig. 22.58  Coronal computed tomography demonstrates extensive bubbly mural (arrows) and mesenteric gas (arrowheads) in this asymptomatic patient (who was being imaged for an unrelated complaint).

persistent dorsal suspending mesentery. Water-soluble contrast studies are useful for confirmation to show the ‘bird’s beak’-type twist. The ‘whirl sign’ on CT reflects the twisted bowel and mesentery and is proportional to the degree of rotation. Complicating features are bowel ischaemia and perforation. A formal DCBE or CTC is indicated with intermittent suspected volvulus when the patient is asymptomatic, to confirm abnormal redundancy of part of the colon and to rule out any obstructing lesion.

Intussusception Colonic intussusception in adults is almost always secondary to a tumour and is perhaps more commonly seen on CT (Fig. 22.60) than the barium enema of old.

Endometriosis Gastrointestinal involvement occurs in 12% to 37% of cases, mainly involving sigmoid and small-bowel loops in the pelvis, though the caecum also may be affected. Serosal implants invade the muscularis

Fig. 22.59  Volvulus. Coronal oblique computed tomography shows a dilated caecum in the left upper quadrant (arrowhead shows part of the ileocaecal valve). At the site of the twist there are two overlapping transition points (arrows), sometimes called the ‘X-marks-the-spot’ sign.

propria, causing fibrosis with contraction of the wall and a mass effect. The mucosal surface remains intact, though rectal bleeding is a symptom. Contrast studies show a localised mass effect with characteristic contracted mucosal folds. MRI is increasingly used and can demonstrate the low signal fibrotic plaques, which can obliterate the pouch of Douglas and involve the rectosigmoid (Fig. 22.61).

CHAPTER 22  The Large Bowel

595

TABLE 22.8  Retrorectal Mass Developmental cysts

Sacral lesions

Anorectal lesions

Epidermoid Dermoid Enteric (cystic hamartoma or rectal duplication) Teratoma Anterior sacral meningocele Chordoma Lymphangioma Lipoma GIST Anal gland cyst

GIST, Gastrointestinal stromal tumour.

Fig. 22.60  Coronal oblique computed tomography with oral contrast agent showing the typical ‘target’ appearance of intussusception. Dense oral contrast material (arrow) is present in the lumen of the intussusceptum which is surrounded by the indrawn mesenteric fat. This was colo-colic intussusception due to an adenocarcinoma serving as the lead-point.

Fig. 22.62  Sagittal magnetic resonance imaging of a large tailgut cyst (arrows) with cystic and solid (C) components, the latter due to the development of a carcinoid tumour within the cyst.

Retrorectal Lesions Tailgut cysts present as a mass that may be complicated by infection, bleeding, or malignant change. Most are developmental in origin (Table 22.8). MRI provides the most complete examination to show the nature of the mass, signal characteristics of any cyst, any mural mass that might indicate malignancy, local infiltration and exclude any meningocele or sacral lesion (Fig. 22.62).

FUNCTIONAL DISORDERS OF THE ANORECTUM

Fig. 22.61  Sagittal T2 weighted pelvic magnetic resonance imaging shows low signal intensity nodularity and obliteration of the pouch of Douglas (arrow) with tethering and angulation of the adjacent rectal wall to a deep endometriotic rectovaginal nodule (arrowhead). The absence of submucosal anterior rectal wall oedema suggests full thickness infiltration is not currently present.

Hirschsprung disease usually presents in infancy, and megarectum is mainly a childhood problem, but both may present in early adult life with a history of chronic intractable constipation. Plain radiographs demonstrate extensive faecal build-up, often outlining an enormously dilated rectosigmoid. A water-soluble contrast enema, without bowel preparation, will distinguish short-segment Hirschsprung disease from megarectum. The lateral view of the pelvis in megarectum will show a dilated distal bowel (>6.5 cm in diameter) extending right down to the pelvic floor and often filling the entire pelvic cavity. There is a sudden transition proximally into colon of normal calibre. In Hirschsprung disease there is a short abnormal segment, which is narrowed and may

596

SECTION B  Abdominal Imaging

Fig. 22.64  Endoanal ultrasound shows an old scar (between arrows) at the site of an obstetric injury involving the external anal sphincter.

Fig. 22.63  Sagittal true fast imaging with steady-state precession (TrueFISP) magnetic resonance defecography shows a large anterior rectocele (arrow) into which the rectally administered contrast material has entered and become ‘trapped’. The rectum (arrowheads) itself is almost empty. This patient needed to digitate per vaginam to empty her rectum.

contract abnormally, leading into the funnel of the transition zone and the dilated normal proximal bowel. Constipation is a very common symptom and is frequently due to slow colonic transit, which may be shown using radio-opaque markers with a plain film taken 5 days after the ingestion of 20 different geometric markers on days 1, 2 and 3. Retention of greater than 4 of day 1, greater than 5 of day 2 and greater than 12 of day 3 markers is abnormal. Difficult defecation is also a component of constipation and may be due to anismus, which is the failure to relax the pelvic floor during attempted defecation. Evacuating proctography (using either barium or MRI) shows delayed and incomplete evacuation (30 s). An anterior bulge of the rectum is common during evacuation, particularly in women, but rectoceles are probably only functionally significant if there is retention within the rectocele at the end of evacuation (Fig. 22.63). Clinically, this is associated with perineal digitation to achieve complete emptying. Rectal prolapse starts with an infolding of the distal rectal wall entering the anal canal, an intra-anal intussusception. This is a typical finding in the solitary rectal ulcer syndrome. If the intussusception extends through the anal canal, it becomes an external prolapse. Incontinence is a complex symptom, and, as with constipation, depends very much on the inter-relationship of the anorectum with the colon. Sphincter damage is best shown on endoanal US (Fig. 22.64) and striated muscle thinning or atrophy on MRI.

ANAL FISTULA Anal fistulae are generally considered to be secondary to cryptogenic anal gland infection. Discharge of an abscess creates a track through

Fig. 22.65  Coronal short Tau inversion recovery (STIR) magnetic resonance imaging shows an extrasphincteric fistula (arrow) running in the ischioanal fossa to the skin surface.

part of the sphincter to the peri-anal skin. Discharge through the skin then completes the fistulous track from its internal opening in the anal canal out onto the skin. Fistulae are commonly classified according to the Parks classification based on their relationship to the muscles of the anal sphincter complex (trans-sphincteric, intersphincteric, suprasphincteric or extrasphincteric). MRI is the imaging modality of choice, although US has a role. On fat-suppressed T2W sequences, fistulae appear as high signal against the lower signal sphincter complex and adjacent fat (Fig. 22.65). Tracks may also link into abscesses, and into supralevator collections, which are more difficult to detect clinically, but very well

CHAPTER 22  The Large Bowel demonstrated on MRI. Preoperative MRI has been shown to reduce recurrence rates in complex fistula disease. Tracks within the sphincter and the internal opening are well shown on endosonography.

FURTHER READING Atkin, W., Dadswell, E., Wooldrage, K., et al. SIGGAR investigators, 2013. Computed tomographic colonography versus colonoscopy for investigation of patients with symptoms suggestive of colorectal cancer (SIGGAR): a multicentre randomised trial. Lancet 381 (9873), 1194–1202. Battersby, N.J., How, P., Moran, B., et al. MERCURY II Study Group, 2016. Prospective validation of a low rectal cancer magnetic resonance imaging staging system and development of a local recurrence risk stratification model: the MERCURY II study. Ann. Surg. 263 (4), 751–760. Bruining, D.H., Zimmermann, E.M., Loftus, E.V., Jr., et al. Society of Abdominal Radiology Crohn’s Disease-Focused Panel, 2018. Consensus recommendations for evaluation, interpretation, and utilization of computed tomography and magnetic resonance enterography in patients with small bowel Crohn’s disease. Gastroenterology 154 (4), 1172–1194. de Haan, M.C., van Gelder, R.E., Graser, A., et al., 2011. Diagnostic value of CT-colonography as compared to colonoscopy in an asymptomatic screening population: a meta-analysis. Eur. Radiol. 21 (8), 1747–1763. Halligan, S., Wooldrage, K., Dadswell, E., et al. SIGGAR investigators, 2013. Computed tomographic colonography versus barium enema for diagnosis of colorectal cancer or large polyps in symptomatic patients (SIGGAR): a multicentre randomised trial. Lancet 381 (9873), 1185–1193. Fini, L., Laghi, L., Hassan, C., et al., 2014. Noncathartic CT colonography to screen for colorectal neoplasia in subjects with a family history of colorectal cancer. Radiology 270 (3), 784–790. Kim, D.H., Hinshaw, J.L., Lubner, M.G., et al., 2014. Contrast coating of the surface of flat polyps at CT colonography: a marker of detection. Eur. Radiol. 24 (4), 940–946. doi:10.1007/s00330-014-3095-z. Neri, E., Halligan, S., Hellström, M., et al. ESGAR CT Colonography Working Group, 2013. The second ESGAR consensus statement on CT colonography. Eur. Radiol. 23 (3), 720–729.

597

Obaro, A., Plumb, A., Fanshawe, T., et al., 2018. Post-imaging colorectal cancer or interval cancer rates after CT colonography: a systematic review and meta-analysis. Lancet Gastroenterol. Hepatol. 3 (5), 326–336. Panés, J., Bouzas, R., Chaparro, M., et al., 2011. Systematic review: the use of ultrasonography, computed tomography and magnetic resonance imaging for the diagnosis, assessment of activity and abdominal complications of Crohn’s disease. Aliment. Pharmacol. Ther. 34 (2), 125–145. Panes, J., Bouhnik, Y., Reinisch, W., et al., 2013. Imaging techniques for assessment of inflammatory bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J. Crohns Colitis 7 (7), 556–585. Patel, U.B., Taylor, F., Blomqvist, L., et al., 2011. Magnetic resonance imaging– detected tumor response for locally advanced rectal cancer predicts survival outcomes: MERCURY experience. J. Clin. Oncol. 29, 3753–3760. Pickhardt, P.J., Hassan, C., Halligan, S., et al., 2011. Colorectal cancer: CT colonography and colonoscopy for detection–systematic review and meta-analysis. Radiology 259 (2), 393–405. Pickhardt, P.J., Kim, D.H., 2013. CT colonography: pitfalls in interpretation. Radiol. Clin. North Am. 51 (1), 69–88. Regge, D., Iussich, G., Segnan, N., et al., 2017. Comparing CT colonography and flexible sigmoidoscopy: a randomised trial within a population-based screening programme. Gut 66 (8), 1434–1440. Stoop, E.M., de Haan, M.C., de Wijkerslooth, T.R., et al., 2012. Participation and yield of colonoscopy versus non-cathartic CT colonography in population-based screening for colorectal cancer: a randomised controlled trial. Lancet Oncol. 13, 55–64. Taylor, F.G., Quirke, P., Heald, R.J., et al. Magnetic Resonance Imaging in Rectal Cancer European Equivalence Study Study Group, 2014. Preoperative magnetic resonance imaging assessment of circumferential resection margin predicts disease-free survival and local recurrence: 5-year follow-up results of the MERCURY study. J. Clin. Oncol. 32 (1), 34–43. Taylor, S.A., et al. on behalf of the METRIC study investigators, 2018. Diagnostic accuracy of magnetic resonance enterography and small bowel ultrasound for the extent and activity of newly diagnosed and relapsed Crohn’s disease (METRIC): a multicentre trial. Lancet Gastroenterol. Hepatol. 3 (8), 548–558.

24  The Biliary System Robert N. Gibson, Tom R. Sutherland

CHAPTER OUTLINE Biliary Anatomy, 656 Methods of Investigation, 657 Disorders of the Gallbladder, 659 Role of Radiology in Investigation of Jaundice, 665

Benign Bile Duct Pathology, 668 Neoplastic Bile Duct Pathology, 673 Interventional Techniques, 676

BILIARY ANATOMY

Gallbladder Anatomical Variants

The intrahepatic pattern of bile duct branching is best described according to the system of Healey and Schroy, to which can be applied the Couinaud system for numbering segments. The typical pattern and its variations are shown in Figs 24.1 and 24.2. The confluence of the bile ducts is a bifurcation in about 60% of individuals and a trifurcation in about 12% (Fig. 24.3). A right sectoral duct crosses to the left to join the left hepatic duct in 28% of cases (22% right posterior sectoral, 6% right anterior sectoral) (Fig. 24.4). Occasionally, a right posterior sectoral or segmental duct (more often posterior than anterior) courses inferiorly and either enters the common hepatic duct directly or cystic duct (Fig. 24.5). Other uncommon left-branching variations are shown in Fig. 24.2E and F. The cystic duct typically joins the common hepatic duct in the middle third of the extrahepatic bile duct—often referred to as the ‘common duct’ on ultrasound (US) for convenience—which then continues as the common bile duct (CBD). The cystic duct usually joins the right side of the common duct but can pass behind or in front of the common duct to join it from the left. The cystic duct can join the common duct at a very low level, in which case it may be mistaken for the common duct on imaging. Uncommonly it may join a right-sided duct, which is usually a low, aberrant right sectoral or segmental duct (see Fig. 24.5). Some of these variations predispose patients to duct injury at cholecystectomy. Other variations include ducts of Luschka or subvesical ducts and cystohepatic ducts. There is some confusion over nomenclature but it seems that the terms ‘subvesical duct’ and ‘duct of Luschka’ both describe an intrahepatic duct running adjacent to the gallbladder fossa, unaccompanied by a portal vein branch, and emptying into either the right hepatic or common hepatic duct. The term ‘cystohepatic duct’ is probably best reserved for small ducts that drain directly into the gallbladder or cystic duct. The significance of these variants is their proximity to the gallbladder and the potential for injury at cholecystectomy, resulting in a bile leak.

Agenesis of the gallbladder is extremely rare, with a prevalence of 0.03%–0.07%. A double gallbladder occurs in about 0.03%, usually with a shared cystic duct, and the accessory gallbladder is often diseased. True gallbladder septa are uncommon and, when occurring at the fundus, form a Phrygian cap. Frequently, an apparent septum is merely gallbladder wall folding, which can vary with patient position. The gallbladder can be abnormal in position, being retrohepatic, suprahepatic, left-sided or intrahepatic, the latter potentially presenting as a liver abscess if complicated by acute cholecystitis. A number of forms of left-sided gallbladder exist: 1. The gallbladder lies under the left hepatic lobe to the left of the falciform ligament 2. Independent development of a second gallbladder from the left hepatic duct with regression or failure of development of a right gallbladder 3. Herniation of the gallbladder through the foramen of Winslow 4. Transposition of the viscera.

656

SUMMARY BOX:  Anatomy Related • Knowledge of biliary segmental anatomy and its variants, especially in the perihilar area is important diagnostically and in planning surgical or other intervention. • Key investigations for biliary imaging are ultrasound (including contrastenhanced ultrasound and endoscopic ultrasound), CT, CT intravenous cholangiography (where available), MRI and MRCP. Supplementary techniques include biliary scintigraphy and PET. • PTC and ERCP are now uncommon as purely diagnostic tools, and their role is mainly in intervention.

CHAPTER 24  The Biliary System

657

VIII

VII

II

RPSD I

VI

IV

V RASD RHD CHD

Fig. 24.1  Typical Pattern of Intrahepatic Biliary Branching. Segments are numbered according to the system of Couinaud. CHD, Common hepatic duct; RHD, right hepatic duct; RPSD, right posterior sectoral duct; RASD, right anterior sectoral duct. (From Blumgart LH, Fong Y (eds). Surgery of the Liver and Biliary Tract, 3rd edn. London: WB Saunders; 2000:365, with permission.) RPSD

RHD

LHD

RASD

D

LHD

CHD

II

RHD

RPSD

III

B

E

RPSD

LHD

IV

RHD

RASD

C

Fig. 24.3  Biliary Duct Anatomy. CT intravenous cholangiography (CT-IVC) (surface-rendered maximum intensity reformat) shows trifurcation at the biliary confluence and segments numbered according to Couinaud. Arrowhead shows right anterior sectoral duct; arrow shows right posterior sectoral duct.

Sectoral duct

A

RASD

LHD

II

IV

III

F

Fig. 24.2  Variations of Biliary Branching Patterns. The more common patterns are A–C. Segments are numbered according to the system of Couinaud. CHD, Common hepatic duct; RHD, right hepatic duct; LHD, left hepatic duct; RPSD, right posterior sectoral duct; RASD, right anterior sectoral duct. (From Blumgart LH, Fong Y (eds). Surgery of the Liver and Biliary Tract, 3rd edn. London: WB Saunders; 2000:365, with permission.)

METHODS OF INVESTIGATION Ultrasound Transabdominal ultrasound (US) is frequently the first imaging technique employed for patients presenting with hepatobiliary-type symptoms as it is more accurate than Computed tomography (CT) for diagnosing acute biliary, especially gallbladder, disease. Imaging is usually performed following a 4-hour fast, allowing the gallbladder to fill and reducing obscuring upper abdominal gas. The wall of a normal non-contracted gallbladder is less than 3 mm thick and is smooth. US allows a dynamic

Fig. 24.4  Biliary Duct Anatomy. CT intravenous cholangiography (CT-IVC) (maximum intensity reformat). Right posterior sectoral duct (arrow) passes to the left to drain into left hepatic duct.

assessment and by moving the patient helps differentiate stones, sludge and polyps. Doppler US allows assessment of vascularity, while focal gallbladder tenderness can be determined using probe pressure. The normal cystic duct may not be visible; however, the extrahepatic bile duct can be seen as a tubular structure anterior to the portal vein and lacking blood flow on Doppler. Contrast-enhanced US (CEUS) using second-generation microbubble agents can be performed at transabdominal, endoscopic and intraoperative US.

658

SECTION B  Abdominal Imaging Sodium iotroxate is safer than older IV biliary agents, with reported complications in 3.5% of patients (3.0% minor, 0.3% moderate and 0.2% severe) and an estimated mortality rate of 0.005%. Adequate excretion of contrast agent relies on near-normal hepatocyte function, so the technique is of no value in the investigation of jaundice, and usually fails if bilirubin levels are more than about two times normal.

Magnetic Resonance Cholangiopancreatography

Fig. 24.5  Biliary Duct Anatomy. Coronal oblique magnetic resonance cholangiopancreatography showing the cystic duct (arrowhead) running with a low right posterior sectoral duct (arrow) and the two joining before the common duct. This ductal configuration can predispose to inadvertent duct injury.

CEUS is achieved by the use of intravenous (IV) injection of very small volumes of microbubbles comprising an inert gas contained by a stabilising shell. The small bubble size allows passage through the pulmonary to the systemic circulation. The bubbles resonate on lowpower US insonation, generating harmonic frequencies, the display of which can be separated from the fundamental image in a way analogous to digital subtraction angiography. Most agents are purely intravascular and are therefore blood pool agents. The bubbles gradually break down safely with insonation, and one injection allows diagnostic enhancement for approximately 5 to 8 minutes. The microbubbles are extremely strong signal enhancers compared with CT and magnetic resonance imaging (MRI) contrast agents, and the use with real-time US allows very high spatial and temporal resolution. They have a very high safety profile and can be used in the presence of renal impairment and cardiac pacemakers, both significant advantages over CT and MRI. CEUS is useful in selected patients: for example, in differentiating sludge from tumour, identifying perforation in cholecystitis and better demonstrating hilar cholangiocarcinoma.

Computed Tomographic Cholangiography Computed tomographic intravenous cholangiography (CT-IVC) relies on an infusion of iodinated contrast agent, such as sodium ipodate, which undergoes biliary excretion. CT is performed around 30 minutes after the infusion, allowing high-resolution reformatted images to be obtained. Prone imaging may be performed after supine images if intraductal gas is present or contrast is layering. As a functional imaging technique, the presence of contrast agent within the duodenum proves that the biliary tree does not have a complete obstruction. This functional information also allows the direct demonstration of bile leaks, biliary communication with cysts and segmental obstruction, which is not obtained with routine non-contrast magnetic resonance cholangiopancreatography (MRCP). Unfortunately access to CT-IVC is reliant on access to the contrast agent that is becoming more limited. CT-IVC and MRCP are complementary investigations, with both offering casedependent advantages.

MRCP has substantially replaced diagnostic percutaneous transhepatic cholangiography (PTC) and endoscopic retrograde cholangiopancreatography (ERCP). It relies on heavily T2 weighted sequences that display stationary water as high signal. Multiplanar thin and thick section acquisitions are obtained using fast spin-echo techniques. As conventional MRCP is not reliant on excretion of contrast material, it is suitable for jaundiced patients, a clear advantage compared with CT-IVC. More recently, MR has been combined with hepatobiliary contrast agents. These agents, which include gadobenate dimeglumine and gadoxetic acid disodium, shorten T1 relaxation, providing positive contrast images on T1 weighted sequences. Biliary excretion can occur as early as 10 minutes after injection, depending upon liver function, although often images are obtained after 30 minutes. The contrast agent provides functional as well as anatomical information but, as with CT-IVC, depends on near-normal excretory hepatocyte function. As T1 weighted MR sequences are used, it is possible to use near-isotropic threedimensional gradient-echo acquisitions. Contrast-enhanced MR cholangiography using hepatobiliary contrast agents has similar applications to CT-IVC, except that it seldom produces the high spatial and contrast resolution achieved with CT-IVC. Diagnostic pitfalls with MRCP include localised signal voids caused by surgical clips and intraductal gas or blood. Bile flow voids may mimic small stones but the former are centrally placed and have less well-defined margins than stones. Acquisition times are longer for MRCP than CT-IVC and, therefore, more prone to motion and respiratory artefacts.

Endoscopic Retrograde Cholangiopancreatography ERCP provides direct opacification of bile ducts and pancreatic ducts, with success rates of 92%–97%, and provides dynamic information during contrast medium introduction and drainage. It allows visual assessment of the duodenum and ampulla of Vater and enables biopsy and brushings, as well as interventional procedures such as sphincterotomy and stone extraction and biliary stenting and biliary stricture dilatation. Complication rates vary depending on the indication for the procedure, the presence of coexisting disease and the experience of the endoscopist, with severe complication rates of 0.9% to 2.3% and total complication rates of 8.4%–11.1%, the most common significant complication being acute pancreatitis. The main diagnostic pitfall with ERCP is the underfilling of ducts above a stricture.

Percutaneous Transhepatic Cholangiography PTC has been substantially replaced by ERCP and MRCP. Its role now is mostly as part of transhepatic biliary intervention. A 22-G Chiba needle is used to puncture and opacify the intrahepatic ducts. Any coagulation disorder should be reversed before the procedure, which is performed with broad-spectrum IV antibiotic cover and conscious sedation or, occasionally, general anaesthesia.

Intraoperative Cholangiography Intraoperative cholangiography (IOC) is performed routinely or selectively during cholecystectomy to detect choledocholithiasis, confirm duct stone clearance and delineate anatomy to minimise risk of bile duct injury.

CHAPTER 24  The Biliary System

659

T-Tube Cholangiography If the CBD has been explored at cholecystectomy, a T-tube is usually left in place and cholangiography performed via this tube after about 7 days, before its removal. Cholangiography should confirm stone clearance and the free passage of contrast medium into the duodenum. Care must be taken to avoid the injection of air bubbles.

Hepatobiliary Scintigraphy Hepatobiliary iminodiacetic acid (HIDA) scintigraphy uses a derivative of iminodiacetic acid, a bilirubin analogue, labelled with technetium-99m (99mTc). It is injected intravenously and serial gamma camera images are obtained over 2–4 hours. It relies on near-normal bilirubin levels, although some agents can be excreted with moderate elevations of bilirubin. Serial image acquisitions show accumulation of the isotope in the liver, bile ducts, duodenum, small bowel and gallbladder (providing it is present and the cystic duct is patent).

Endoscopic Ultrasound

Fig. 24.6  Gallstone. Ultrasound of the gallbladder shows a small stone in the cystic duct (callipers) with multiple small echogenic stones in the dependent gallbladder producing posterior acoustic shadowing.

Biliary endoscopic US (EUS) provides high-frequency grey-scale imaging, colour Doppler and CEUS for the evaluation of the extrahepatic biliary tree and pancreas. EUS has similar sensitivity and specificity to MRCP in diagnosing causes of biliary obstruction, being especially well suited to anatomically lower or more distal causes. Though relatively invasive, advantages are that it allows direct visualisation of the duodenum, fine-needle aspiration cytology and potentially biliary drainage. More sophisticated and expensive systems of ‘mother-daughter’ probes allow intraductal examination of the CBD, but are not routinely available.

DISORDERS OF THE GALLBLADDER Gallbladder Stones The prevalence of gallbladder stones in adults in Western communities is approximately 15%. They are asymptomatic in about 80% but in this group about 15% will develop symptoms over 15 years and they confer a small lifetime risk of gallbladder carcinoma. About 70% of gallbladder stones are solely or predominantly cholesterol in type, with up to 30% being black pigment stones composed mainly of calcium bilirubinate. Less than 10% of stones are opaque on plain radiographs, the larger stones showing laminated or peripheral calcification. On CT, a minority of gallbladder stones are visible, being hyperdense, hypodense or of mixed density. US is the most accurate investigation for the diagnosis of gallbladder stones, which appear as echogenic foci producing acoustic shadows, and stone mobility is frequently identifiable (Fig. 24.6), although is not essential for diagnosis. The sensitivity of US is greater than 95%. Falsenegative diagnoses are more common than false-positive ones and are usually because of small stones in patients in whom there is poor acoustic access to the gallbladder because of obesity or other unfavourable anatomy. False-negative diagnoses are reduced by careful US technique. Small stones are differentiated from small polyps by the demonstration of mobility or the presence of an acoustic shadow. Non-visualisation of the gallbladder on US can be due to a previous cholecystectomy, non-fasting state, an abnormal gallbladder position, emphysematous cholecystitis or because the gallbladder is filled with stones. The latter can be recognised by identifying the so-called ‘double-arc shadow’ sign in the gallbladder fossa, consisting of two parallel curved echogenic lines separated by a thin anechoic space with dense acoustic shadowing distal to the deeper echogenic line (Fig. 24.7).

Fig. 24.7  Gallbadder Stones. Gallbladder filled with stones producing the “double-arc” sign, a hypoechoic line (arrow) between two echogenic lines, with a distal acoustic shadow.

Low Phospholipid-Associated Cholelithiasis Low phospholipid-associated cholelithiasis (LPAC) is the more common of the two main clinical manifestations of ABCB4 gene mutations. The other is progressive familial intrahepatic cholestasis type 3 (PFIC3), which is a disease characterised by the development of biliary cirrhosis, portal hypertension and liver failure, generally in childhood. The gene defect results in reduced phospholipid excretion into the bile, resulting in reduced solubilisation of cholesterol (and hence cholesterol stones or microlithiasis) and reduced cytoprotection against the damaging effect on biliary epithelium of certain bile salts. The exact prevalence of LPAC is unclear but the condition is being recognised more frequently. The main clinical feature is symptomatic and recurring cholelithiasis (and its complications) and notably intrahepatic cholelithiasis. Female-to-male ratio is approximately 3 : 1, age is typically less than 40 and presentation is often recurrence

660

SECTION B  Abdominal Imaging

of biliary symptoms post cholecystectomy, and there may be a history of cholestasis of pregnancy or a family history of LPAC or cholecystectomy at a young age. The key imaging features, apart from those common to gallstones, are intrahepatic stones that are often tiny and associated with either shadowing or ring-down artefact. Bile duct dilatation and larger stones may occur, and this may be in a segmental or subsegmental distribution, sometimes quite peripheral, and is generally best demonstrated by MR. US and MR have a complementary role when the diagnosis is suspected, as US can show the tiny intrahepatic stones missed by MR; CT is generally less useful. Less common manifestations of this disease spectrum include cholangiopathy similar to primary sclerosing cholangitis (PSC), biliary cirrhosis and intrahepatic cholangiocarcinoma.

Sludge Sludge is commonly seen on US and appears as fine, non-shadowing dependent echoes. It is composed of calcium bilirubinate granules, cholesterol crystals and glycoproteins. It is more commonly seen in chronic fasting states, critically ill patients, those receiving total parenteral nutrition and in pregnancy. Sludge resolves spontaneously in 50% of patients and gallstones will develop in 5%–15%. Small stones are difficult to locate within sludge, so careful imaging through sludge is important (Fig. 24.8). Usually, sludge layers in a dependent fashion, but occasionally it mimics a tumour mas: that is, ‘tumefactive sludge’. Sludge can usually be differentiated from tumour by its mobility, lack of internal blood flow on Doppler examination, lack of focal gallbladder wall abnormality or lack of enhancement on CEUS. Blood (haemobilia) and pus (empyema) may have a similar appearance to sludge, and the clinical setting aids in their differentiation. Sludge, blood and pus can also occur in the bile ducts.

Milk of Calcium Bile Milk of calcium bile, or limy bile, is an uncommon condition in which the gallbladder bile becomes viscous, probably as a result of stasis, and contains a high concentration of calcium bilirubinate. On US, it causes diffuse echoes, similar to sludge, but is more echogenic with a tendency to layer out and produce an acoustic shadow. On CT and, occasionally, on plain radiographs, it is visible as layering high-density material.

Fig. 24.8  Biliary Sludge. Ultrasound shows mobile echogenic sludge within the gallbladder.

Cholecystitis

Acute Calculous Cholecystitis US is the best initial imaging investigation in patients with suspected acute cholecystitis, which, in 90%–95% of cases, is due to gallstones (acute calculous cholecystitis). The positive predictive values of stones combined with either tenderness localised to the gallbladder (positive sonographic Murphy sign), or the presence of a gallbladder wall thickness of >3 mm, are 92% and 95%, respectively (Fig. 24.9). The negative predictive value of the absence of gallbladder stones and a negative sonographic Murphy sign is 95%. US can be definitive in about 80% of cases. Gallstone(s) may be impacted in the gallbladder neck, and this region must be carefully examined. Other US signs are gallbladder distension (diameter >5 cm), pericholecystic fluid, gallbladder wall striations and, occasionally, wall hyperaemia on Doppler examination. Fine echoes within the gallbladder may be due to sludge or pus (gallbladder empyema). If liver function tests suggest duct obstruction, a careful evaluation of the CBD should be made for choledocholithiasis. CT is less accurate than US for acute cholecystitis, but is widely used to evaluate patients with acute abdominal pain. The CT findings in acute cholecystitis include gallbladder wall thickening, subserosal oedema, gallbladder distension, high-density bile, pericholecystic fluid and inflammatory stranding in pericholecystic fat (Fig. 24.10). Gallstones are identifiable in the minority. Gallbladder wall enhancement is variable and not a reliable predictor of cholecystitis because normal gallbladders can show wall enhancement. Transient pericholecystic liver rim enhancement may be seen. Functional studies, such as MRI with hepatobiliary contrast agents and hepatobiliary scintigraphy, assess cystic duct patency, with a positive result being absence of contrast/tracer within the gallbladder. Gallbladder wall thickening may result from many causes other than cholecystitis. These include non-fasting state, generalised oedematous states, hepatitis, pancreatitis, gallbladder wall varices, adenomyomatosis and carcinoma, although the latter two usually cause focal rather than diffuse thickening.

Gangrenous Cholecystitis This condition is suggested on US by pronounced irregularity or asymmetrical thickening of the gallbladder wall, internal membranous echoes resulting from sloughed mucosa and pericholecystic fluid. The clinical findings, paradoxically, may diminish with progression to gangrenous

Fig. 24.9  Acute Cholecystitis. The gallbladder has an oedematous wall (short arrow) with an impacted shadowing stone in the neck (long arrow).

CHAPTER 24  The Biliary System

661

gallbladder distension, gallbladder wall thickening, echogenic contents and, occasionally, sloughed membranes/mucosa and pericholecystic fluid. Positive diagnosis is often difficult, as sludge and gallbladder distension may occur without cholecystitis in this group. All investigations—US, CT and biliary scintigraphy—are less accurate than in acute calculous cholecystitis. Biliary scintigraphy is possibly the most accurate technique. In suspected cases, percutaneous cholecystostomy may be diagnostic and is therapeutic, with the vast majority clinically improving and not requiring subsequent cholecystectomy. Chronic acalculous cholecystitis is a controversial entity as there are no clear clinical, pathological or imaging criteria for its diagnosis. The clinical setting is usually unexplained biliary-type pain, and patients have often previously undergone numerous other negative investigations. US may show gallbladder wall thickening and, by definition, no stones. Cholescintigraphy followed by the IV infusion of cholecystokinin (CCK), or one of its analogues, can be used to assess gallbladder contractibility. An ejection fraction less than 35% on CCK-cholescintigraphy is generally considered to be an indicator of gallbladder dysfunction and helps select patients who may benefit from cholecystectomy.

Xanthogranulomatous Cholecystitis

Fig. 24.10  Acute Cholecystitis on Computed Tomography. The gallbladder wall is oedematous with pericholecystic oedema. Stones and sludge layer in the gallbladder neck.

change. CT and MRI signs suggesting gangrenous cholecystitis are gas in the wall or lumen, discontinuous and/or irregular mucosal enhancement (Fig. 24.11A), which may also be seen on CEUS (Fig. 24.11B) along with internal membranes (Fig. 24.11C) and pericholecystic abscess. Of these, interrupted wall enhancement is the most sensitive sign (70.6%) and is highly specific (100%). Gallbladder perforation occurs in 5%–10% of patients with acute cholecystitis and may be present as free spillage into the peritoneal cavity, a localised pericholecystic abscess, or the development of a fistula. It is suggested by pericholecystic fluid and the features of gangrenous cholecystitis, often with relative collapse of the gallbladder lumen. Localised disruption of the gallbladder wall is seen on US in 40% and on CT in 80% (Fig. 24.12). CEUS can help identify perforation by showing local absence of gallbladder wall enhancement.

Emphysematous Cholecystitis

Xanthogranulomatous cholecystitis is an unusual form of chronic cholecystitis that may mimic malignancy. It usually presents with the clinical features of cholecystitis or biliary obstruction (a variant of Mirizzi syndrome), is typically diffuse and usually associated with stones. To differentiate xanthogranulomatous cholecystitis from carcinoma, five features are useful to search for, which are continuous mucosal enhancement, intramural hypoattenuating nodules, diffuse mural thickening, absent hepatic invasion and absent intrahepatic duct distension.

Gallbladder Mucocele Gallbladder mucocele, also known as gallbladder hydrops, is the result of chronic gallbladder obstruction, without superimposed infection, allowing accumulation of large volumes of sterile mucinous fluid. It is usually caused by impacted stones and less often by polyps, tumours or adjacent adenopathy. The gallbladder is markedly distended, fluid-filled and may present as a mass.

Gallbladder Fistulae Gallbladder fistulae are rare. The great majority are due to chronic stone disease rather than neoplasm. Most communicate with the duodenum and most of the remainder to the colon. Cholecystoduodenal fistulae may result in bowel obstruction due to the impaction of larger stones in the distal small bowel, so-called gallstone ileus, a condition associated in a minority of patients with a visible gallstone on plain radiographs or CT and gas in the biliary tract.

Emphysematous cholecystitis accounts for only 1% of acute cholecystitis but has a relatively high mortality rate. It is more common in men (the reverse of the usual female predominance in cholecystitis). About 50% of patients with this condition are diabetics and stones are present in less than 50%. The diagnosis may be evident on plain radiographs and is readily made on CT (Fig. 24.13A), which shows intramural and/or intraluminal gas caused by gas-forming organisms, while on MRI the gas appears as signal void (see Fig. 24.11C). On US, intramural gas appears as focal or diffuse bright echogenic lines. Intraluminal gas, in the non-dependent portion of the gallbladder, causes a curvilinear, brightly echogenic band with shadowing (Fig. 24.13B), which can make recognition of the gallbladder difficult and lead to a false-negative US result. Small foci of intramural gas may cause ring down artefact and mimic adenomyomatosis.

Porcelain gallbladder is an uncommon condition occurring in 0.2% of cholecystectomy specimens and consisting of complete or scattered mural calcification. There is an association with gallbladder carcinoma, although the incidence of coexisting carcinoma is less than previously thought (10 mm. Further investigation should be correlated with the pretest probability of obstruction. The authors’ practice is to attempt to visualise the entire duct and measure the largest internal diameter, which tends to be in the suprapancreatic portion. If only the very upper end of the common duct is seen and is not dilated, distal dilatation is unlikely. Conversely, if there is mild dilatation of the suprapancreatic portion but the duct tapers to a normal size in its pancreatic portion, further imaging is not mandatory and should be guided by the clinical picture. Hilar biliary obstruction will produce only intrahepatic duct dilatation, whereas more distal obstruction will result in extrahepatic and intrahepatic dilatation. Approximately 95% of patients with bile duct obstruction have biliary dilatation. In the remaining 5%, there are usually sufficient clinical/biochemical indicators of duct obstruction to suggest that cholangiography is warranted. Most cases of biliary obstruction without

Fig. 24.20  Low Biliary Obstruction. Longitudinal US shows a very dilated bile duct and a large pancreatic head carcinoma (arrows).

duct dilatation are due to choledocholithiasis, PSC or postoperative stricturing. If duct obstruction (i.e. duct dilatation) is present, the anatomical level should be determined: namely, whether it is hilar/perihilar or low/ mid common duct (Fig. 24.20). This aids the differential diagnosis (Table 24.1) and selection of further imaging tests. At US, if choledocholithiasis is demonstrated, patients can proceed to endoscopic sphincterotomy or cholecystectomy. If choledocholithiasis is not demonstrated but stones are highly likely on clinical grounds (e.g. pain with jaundice), most patients should proceed to ERCP, particularly in the presence of sepsis. In patients with comorbidities that contraindicate ERCP or surgery, MRCP is helpful in providing confirmatory evidence of stones or suggesting another cause of obstruction. US detects the level of obstruction in up to 95% and cause in up to 88%. If the cause is not evident and stones are not considered the most

CHAPTER 24  The Biliary System

667

TABLE 24.1  Causes of Major Bile Duct

Obstruction Anatomical Location Hilar Low/mid duct Either

Malignant Gallbladder carcinoma*** Hepatocellular carcinoma** Pancreatic carcinoma**** Ampullary carcinoma** Cholangiocarcinoma*** Metastases*** Lymphoma* Benign biliary tumours*

Benign

Pancreatitis (acute or chronic)** Stones**** Mirizzi syndrome** Postoperative strictures*** Primary sclerosing cholangitis*** Other cholangiopathy* Haemobilia* Parasites*

Type I

Type II

Type IIIa

Type IIIb

Asterisks indicate approximate relative incidence (**** = most common). Low/mid-duct obstruction is more common than hilar obstruction.

Type IV Fig. 24.22  Modified Bismuth classification of malignant hilar biliary obstruction based on proximal extent of tumour.

Fig. 24.21  Low Biliary Obstruction. Magnetic resonance cholangiopancreatography shows adjacent strictures (arrows) of the common bile duct and pancreatic duct (“double-duct sign”) in a patient with pancreatic carcinoma.

likely diagnosis, CT is usually the next most useful test, although MRCP and MRI may be substituted, depending on local access and expertise. MRCP identifies the presence of obstruction in up to 99%, the level of obstruction in 96% and detects tumour in 88% of patients with a malignant cause. CT is highly accurate for identifying the level and cause of obstruction, having similar accuracy to MRI and MRCP (Fig. 24.21), the exception being that MRCP has a higher accuracy for detection of choledocholithiasis. The next questions with presumed malignant obstruction relate to tumour resectability and biliary decompression options. In malignant hilar obstruction, evaluation should assess the proximal extent of stricturing into the right and left hepatic ducts, the presence of lobar atrophy, the patency of the portal veins (main, right and left branches) and the presence of any intrahepatic or local extrahepatic metastases. The proximal extent of stricturing is classified according to the modified Bismuth classification (Fig. 24.22). In malignant low obstruction, usually

due to pancreatic carcinoma, the main factors to assess are tumour size, vascular involvement (portal vein, superior mesenteric vein and superior mesenteric artery, coeliac trunk and common hepatic artery), lymph node metastases and hepatic metastases. Even with modern imaging techniques, it can be difficult to accurately differentiate benign from malignant strictures. Serum bilirubin levels have an association with malignancy, with levels over 100 µmol/L being 71.9% sensitive and 86.9% specific, while levels over 250 µmol/L are 97.1% specific for malignancy. Further information concerning pancreatic lesions giving rise to biliary obstruction may be found in Chapter 25. US (including Doppler), CT, MRI (including MRCP and magnetic resonance angiography (MRA)) and EUS can all provide information about tumour resectability. Positron emission tomography (PET) imaging is more helpful in identifying metastases than in identifying primary biliary tumours. Resectability assessment should, ideally, identify signs of nonresectability without excluding appropriate patients from potentially curative surgery. CT, including computed tomography angiography (CTA), has a good overall accuracy for assessment of resectability. For hilar tumours, MRCP allows assessment of the proximal extent of the lesion and determination of its Bismuth classification. EUS helps in the assessment of vascular involvement by pancreatic head or periampullary tumours.

668

SECTION B  Abdominal Imaging

Core biopsy or fine-needle aspiration of suspected malignant obstructing lesions can be guided by US (transabdominal or endoscopic) or CT. If surgical resection is considered, a preoperative biopsy is not usually appropriate. If palliative stenting is being performed, biopsy is preferable after decompression to reduce the risk of a bile leak.

BENIGN BILE DUCT PATHOLOGY Choledocholithiasis At least 90% of bile duct stones originate in the gallbladder, so-called secondary stones. Primary stones are those that arise in the bile duct and these are pigment stones. For patients younger than 60 years undergoing cholecystectomy, 8%–15% have duct stones, with the figure increasing substantially in older patients. Current guidelines utilising clinical and biochemical data have a poor sensitivity and specificity for predicting choledocholithiasis, and so accurate non-invasive imaging is essential to prevent unnecessary intervention.

Ultrasound US sensitivity varies greatly, with the upper range being 50%–80%, and it is better in jaundiced patients. The specificity is, however, about 95%. Duct dilatation and acoustic shadowing are each absent in about 30% of cases. Positive stone diagnosis depends on the demonstration of an intraductal echogenic focus in both the longitudinal and transverse planes (Fig. 24.23A). Conditions that may mimic stones on US are: 1. Intraductal gas—usually recognisable by its linear nature and movement.

2. Haemobilia and sludge—they produce more diffuse echoes than stones. 3. Surgical clips, hepatic artery calcification and duodenal diverticula— these do not lie within the duct lumen. 4. Parasites. As bile duct diameter increases, the likelihood of duct stones also increases, but a duct diameter of 90% and an even higher specificity.

Unenhanced CT The reported sensitivity of unenhanced CT in detection of choledocholithiasis varies. Two studies report sensitivities of 88% and 80% and specificities of 97% and 100%, respectively. Another study, however, reported a sensitivity of only 60%, which is probably closer to what is achieved in routine practice. Stones usually appear as a ring density or soft-tissue density surrounded by lower-density bile, or sometimes as uniformly high density (Fig. 24.24).

Cholangiography ERCP has a high accuracy for the diagnosis of choledocholithiasis but it does not detect all stones; in one study, its sensitivity was 89% in comparison with EUS. The sensitivity and specificity of CT-IVC for detection of CBD stones is 85%–96% and 88%–98%, respectively. A strength of this technique is the diagnosis of stones 90 mL/min/1.73 m2).

733

Stage 2: mild impairment (glomerular filtration rate 60–89 mL/min/1.73 m2). Stage 3: moderate (glomerular filtration rate 31–59 mL/min/1.73 m2). Stages 4 and 5: severe (glomerular filtration rate 15–29 and 0.3 mg/dL (26.4 µmol/L), (2) >50% increase in serum creatinine and (3) urine output decreased to 38.5°C) and costovertebral angle tenderness. Symptoms include flank or suprapubic pain, dysuria, urinary frequency/urgency, rigors or nausea/vomiting. Tubulointerstitial inflammation causes parenchymal swelling and renal enlargement. Inflammation can affect the perinephric fat, Gerota’s fascia and, very rarely, the extrarenal space. Extrarenal extension usually occurs in untreated cases, uncontrolled diabetics or immunocompromised patients. Microabscesses can coalesce to form a renal abscess, which may rupture into the perirenal space or extend to adjacent organs such as the liver. Usually infection ascends through the collecting system; occasionally it seeds haematogenously or rarely enters through lymphatics. Ascending infection from the urethra to bladder causes cystitis; this extends to the upper tract in half of cases. Major risk factors include long-term catheterisation, faecal soiling in chronically ill patients or vesicoureteric reflux (VUR). Ureteric peristalsis is hindered by gram-negative endotoxins and anatomic obstruction to urinary flow; therefore, VUR is not essential for infection ascent. Bacteria enter the renal parenchyma through collecting ducts at renal papillae, causing an infectious-inflammatory response extending from the collecting tubules into the renal interstitium. Leucocytes and debris tend to occlude renal tubules, and patchy or lobar vasoconstriction of intrarenal arteries and arterioles also occurs. Haematogenous seeding tends to occur among intravenous drug abusers, those with extrarenal infection (dental, cutaneous, or coronary valves) and immunocompromised or paediatric patients. Urinary tract obstruction may precipitate haematogenous seeding through reduced urinary clearance of the responsible organisms. In contradistinction to ascending infection, haematogenous seeding begins in the cortex and involves the medulla 24–48 hours after bacterial inoculation. Initially, haematogenous infection causes multiple, small, peripheral round lesions without a lobar distribution. Once the medulla is involved, differentiation between ascending and haematogenous infection is not easily made with imaging. Lymphatic spread of infection to the kidney is rare, occurring with severe retroperitoneal infection or abscess, bowel perforation or infection. Lymphatics may permit lower UTIs to ascend to the kidney.

Imaging of Acute Pyelonephritis In general, APN is diagnosed clinically and does not require imaging. Most imaging techniques add little to management-related decisions

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection in the uncomplicated patient with APN if the patient responds to therapy within 72 hours. The indications for imaging of suspected APN are the same as for any UTI. The type of imaging is determined by the indication, suspected abnormality, patient background, availability and personal preference. The Society of Uroradiology recommends that imaging should document whether APN is unilateral or bilateral and diffuse or focal and if renal enlargement is present. Ultrasound.  US is often the first imaging investigation of the urinary tract for suspected APN. US has a low sensitivity, is operator dependent and has limited diagnostic accuracy in obese patients. US does not demonstrate renal abnormalities in 75–80% of patients with uncomplicated infection and often underestimates the infection severity, particularly with multifocal disease or perinephric extension. Nevertheless, US is the technique of choice for APN assessment during pregnancy due to lack of ionising radiation. Features of APN on US include renal enlargement caused by generalised renal oedema from inflammation and congestion. Renal length exceeding 15 cm or increase in length of greater than 1.5 cm compared with the unaffected kidney can be attributed to APN on US. Oedema can also efface renal sinus fat. Renal parenchymal echogenicity may be normal, hyperechoic from haemorrhage or hypoechoic due to oedema. Loss of corticomedullary differentiation may be observed. This is suggestive of but not diagnostic for APN on US. Hydronephrosis, either as a cause for infection or secondary to pyelonephritis, can be seen on US and internal echoes in a dilated pelvicalyceal system should raise concern for pyonephrosis. In children where infection can easily extend to the pelvis, there may be relatively increased renal sinus echogenicity. US can diagnose underlying congenital anomalies that predispose to infection such as pelviureteral junction obstruction or duplicated pelvicalyceal system. Doppler US increases the sensitivity for detection of renal parenchymal abnormalities due to APN. Power Doppler US can help demonstrate areas of parenchymal hypoperfusion due to arteriolar vasoconstriction and interstitial oedema. It is suggested that contrastenhanced ultrasonography (CEUS) better demonstrates hypoperfusion from APN and has comparable sensitivity and specificity with intravenous contrast medium–enhanced CT. CEUS does not demonstrate a striated nephrogram or delayed persistent enhancement from tubular obstruction since it is a purely intravascular contrast agent. The main advantage of CEUS for APN assessment is for abscess detection. Careful analysis of all phases of enhancement is important to differentiate pure APN from abscess. CEUS can differentiate infarction from APN; although both may return wedge-shaped abnormalities, APN usually shows some degree of enhancement, whereas infarction does not enhance. Infarction usually does not produce cortical deformity, whereas pyelonephritis usually causes deformity when it reaches sufficient size. Infarction may show penumbral enhancement due to collateral blood supply. US of the urinary bladder should be performed in cases of suspected APN. Measurement of bladder wall thickness and calculation of postmicturition residual urine volume can support a diagnosis of bladder outflow obstruction, as can an enlarged prostate in male patients. Computed tomography.  The sensitivity and specificity of CT for APN is 86.8% and 87.5%, compared with 74.3% and 56.7% for US, respectively. CT is considered the imaging technique of choice for diagnosis or follow-up of APN in complicated patients despite radiation exposure, use of iodinated contrast media and the relatively high cost. CT is readily available and sensitive and can also assess for complications or extrarenal involvement. CT for pyelonephritis should include an unenhanced phase, an early nephrographic phase and a pyelographic phase. Occasionally, CT 3 hours after intravenous contrast medium administration may demonstrate retention of contrast medium and help to differentiate APN from tumour or infarction. Unenhanced CT evaluates renal size, areas of increased attenuation due to haemorrhage,

735

calcification or perinephric stranding and areas of hypoattenuation due to gas, mass or fluid. Unenhanced CT detects causes of renal obstruction such as stones or soft-tissue masses. APN is typically hypoattenuating due to oedema or necrosis; rarely, hyperattenuation is observed due to haemorrhagic bacterial nephritis. A normal unenhanced CT does not exclude APN, and intravenous contrast medium– enhanced imaging is required. Nephrographic-phase imaging allows accurate detection and assessment of APN. Loss of corticomedullary differentiation with one or more wedge-shaped zones or streaky areas of reduced enhancement extending from the papilla through the medulla to the renal capsule are observed (Fig. 26.16). APN causes homogeneous or heterogeneous reduced parenchymal enhancement due to tubular obstruction by inflammatory debris, interstitial oedema and vasospasm (Fig. 26.17). Increased time-to-peak enhancement results in delayed persistent

A

B Fig. 26.16  Acute Focal Nephritis. Computed tomography following intravenous contrast agent enhancement demonstrates a hypoattenuating wedge-shaped area in the right kidney with perinephric fat stranding (arrows in images A and B).

736

SECTION B  Abdominal Imaging

Fig. 26.17  Pyelonephritis. Computed tomography: Heterogeneous enhancement of an enlarged right kidney in a patient with flank pain and fever.

enhancement 3 to 6 hours after contrast medium administration. A striated nephrogram may be observed (Fig. 26.18); however, this may occur in normal subjects when low-osmolar iodinated contrast is used in a dehydrated patient. Peripherally located rounded hypoattenuating renal lesions with clinical features supportive of APN favour haematogenous aetiology, since ascending infection usually creates wedgeshaped abnormalities. When the entire kidney is involved, differentiation between haematogenous and ascending infection is difficult. Early enhanced CT is highly sensitive for pelvicalyceal involvement. Thickened enhancing walls may be the only sign of ascending infection. Later, extrarenal extension or involvement of Gerota’s fascia may also occur. The presence of non-enhancing fluid locules within an area of APN suggests fluid liquefaction and abscess formation (Fig. 26.19). Diffuse renal involvement with APN causes impaired contrast medium excretion proportional to the severity of infection. Pelvicalyceal filling defects at the excretory-phase CT can be due to sloughed papillae, blood clots or fungus balls. CT is important for differentiating pyelonephritis from tumour or infarction. Imaging features of APN should show improvement after medical treatment but tumour or infarction will not. Occasionally, CT features of APN lag behind clinical resolution. This can create clinical uncertainty and precipitate biopsy for histopathological reassurance. Extrarenal manifestations of APN that are observed on CT include gall bladder wall thickening, periportal oedema and renal vein or inferior vena cava thrombosis. Magnetic resonance imaging.  MRI has gained wide acceptance in the diagnosis of APN. MRI is recommended when CT cannot be performed (pregnancy, previous reactions to contrast medium, children, etc.). MRI features of APN resemble those observed at CT: renal enlargement, alterations of signal intensity due to haemorrhage or oedema, striated nephrogram or parenchymal or perinephric fluid (Fig. 26.20). Signal void in the urinary tract on MRI may be due to either stone or gas. With MRI, APN demonstrates heterogeneous low signal on T1 weighted sequences and high-signal intensity on T2 weighted sequences. MRI has excellent soft-tissue contrast resolution which provides accurate information regarding APN, helping to differentiate APN from renal scar, and detection of the extrarenal extension at least as well as CT. MRI with intravenous gadolinium helps to depict areas of renal parenchymal involvement. Gadolinium-enhanced inversion recovery

A

B Fig. 26.18  Acute Pyelonephritis. Nephrographic-phase computed tomography (images A and B) shows a striated pattern of enhancement in the left kidney and perinephric fat stranding.

imaging can help to detect intrarenal APN. This method accentuates contrast differences as normally perfused renal tissue enhances normally APN remains hypointense after the administration of intravenous contrast media. MRI using a gadolinium-enhanced short-tau inversion recovery (STIR) sequence can diagnose APN accurately in children compared with dimercaptosuccinic acid (DMSA) scintigraphy. MR urography using heavily T2 weighted sequences such as half-Fourier acquisition single-shot turbo spin echo (HASTE) can detect urinary tract dilatation in cases of obstruction and perinephric inflammatory extension, especially when fat suppression is used. Diffusion-weighted imaging (DWI) helps characterise tissues without ionising radiation or contrast medium injection. On DWI, APN can be more confidently diagnosed when areas of restricted diffusion are demonstrated, these areas are hyperintense

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection

737

A

A

B Fig. 26.19  Abscess in a Transplanted Kidney. Contrast-enhanced computed tomography demonstrates non-enhancing fluid locules in a transplanted kidney consistent with early abscess formation (arrows in images A and B).

on high b value MRI > 600 (Fig. 26.21), and hypointense on ADC mapping. Quantitative assessment of affected areas shows low ADC values in comparison with the cortex and medulla, but this is not specific and can be also seen in tumours and ischaemia. Accurate clinical information is therefore essential to tailor imaging findings. Intravenous urography.  IVU was once commonly used to diagnose APN but is currently only used in very limited circumstances, often due to restricted access to CT where its purpose is mainly to detect anatomical abnormalities that predispose to APN. IVU depicts abnormalities in approximately 25% of patient with APN. This is determined by the degree and extent to which parenchymal involvement disturbs renal excretion. IVU findings include diffuse renal enlargement, delayed contrast excretion with dense persistent striated nephrogram, delayed faint filling of the calyces and effacement or dilation of the collecting system. Caliectasis may be due to either urinary obstruction or impaired peristaltic activity secondary to released endotoxins. The sensitivity of IVU is insufficient to reliably diagnose APN, characterise the type of parenchymal involvement or demonstrate complications. Intravenous

B Fig. 26.20  Acute Pyelonephritis on Magnetic Resonance Imaging. T2 weighted magnetic resonance imaging (A and B) demonstrating a striated signal in the right kidney and perinephric fluid.

contrast medium administration for IVU is not possible in patients with severe renal impairment and is not recommended in the routine assessment of APN. Renal scintigraphy. Renal scintigraphy using technetium 99m (99mTc)-DMSA is commonly used to assess for renal scarring as part of APN evaluation. Approximately 60% of the administered dose of 99mTcDMSA is taken up by the proximal tubular cells, mainly through peritubular arterioles and some through filtration and tubular reabsorption. The remaining DMSA is filtered and excreted at low concentrations in the urine. Tracer accumulates in the cortical tubules within 1 hour and remains for 24 hours. Homogeneous distribution of tracer throughout the cortex on anterior and posterior high-spatial resolution imaging 2–3 hours after DMSA administration can confirm that the cortex is normal. In the early phase of APN, there is reduced tracer uptake due to ischaemia. In later stages, there is associated tubular obstruction with impairment of the kidney function and isotope accumulation. Renal

738

SECTION B  Abdominal Imaging

A

B

C

Fig. 26.21  Acute Pyelonephritis in Transplanted Kidney. (A) Ultrasound of the kidney shows a right renal upper pole hyperechoic lesion (arrow). (B) The lesion is hypoenhancing on a postcontrast T1 weighted axial magnetic resonance image. (C) There is diffusion restriction in the lesion on a coronal diffusion-weighted image (b = 800).

scarring, on the other hand, appears photopenic. DMSA should ideally be performed at least 6 months after APN to allow the reversible changes to resolve. Photopenia can occur due to renal abscess, hydronephrosis, cysts or duplex morphology. Scintigraphy should be performed with US for diagnostic purposes. Assessment of left and right renal DMSA uptake relative to one another is generally performed on the posterior view. The lower limit of normal for renal uptake is approximately 45%. Less than 45% uptake during APN is predictive of permanent renal injury. DMSA scintigraphy is 92% sensitive for the diagnosis of experimentally induced APN in an animal model. DMSA is superior to IVU for renal scar detection and more or less equal to US, but MRI is superior for diagnosis of APN. In children, it is arguable if cortical imaging should be performed to determine parenchymal involvement before voiding cystourethrography (VCUG). Some authors have suggested that the diagnosis of parenchymal scar at DMSA scintigraphy is an indication for VCUG.

Renal and Perirenal Abscess Abscess formation affects the renal parenchyma more often than the perinephric space, usually as a complication of APN but occasionally due to haematogenous or direct spread from acute diverticulitis or pancreatitis. Inflammation from APN causes vasospasm, which can cause liquefactive necrosis and microabscess formation. Coalescence of small microabscesses leads to abscess formation. Absent clinical response to medical management should raise concern for abscess formation. Patients receiving renal replacement therapy, intravenous drug abusers and diabetic patients (75% of cases) are particularly at risk for renal abscess formation. Most abscesses are unilateral and solitary; multifocal lesions suggest haematogenous dissemination. A perirenal abscess may rupture through Gerota’s fascia to become a paranephric abscess. Urine culture is negative in 20%–25%, usually when an abscess is isolated from the pelvicalyceal system.

Imaging of Renal and Perirenal Abscess CT is the imaging study of choice for the diagnosis of renal and perirenal abscess. At contrast medium–enhanced CT, the contents of an evolving abscess may be dense, but a mature abscess will contain hypoattenuating non-enhancing fluid. An enhancing rim, due to a pseudocapsule may be observed, and a halo of hypoenhancement may occur at nephrographicphase CT. Renal fascial thickening, fat stranding, septal thickening and perinephric fat obliteration are commonly found (Fig. 26.22). Gas is occasionally seen on CT, as can extension into the psoas muscle, perirenal, anterior or posterior pararenal spaces and the pelvis. US or CT can guide diagnostic aspiration or drainage.

Contrast medium–enhanced and DW-MRI have important roles in renal disease assessment, particularly during pregnancy, in children and if CT is contraindicated. MRI can help differentiate an abscess from a renal mass and, when serum creatinine is raised, conventional MRI with DWI can help accurately diagnose renal and perirenal abscesses. An intrarenal abscess generally appears as an ovoid or rounded thickwalled lesion with low signal intensity on T1 weighted images and increased signal intensity on T2 weighted images, depending on the presence of protein, fluid and cellular debris (Fig. 26.23). Fluid–fluid levels, irregular septations or perinephric oedema can be observed. Gadolinium-enhanced MRI can help detect extrarenal extension. Cytotoxic oedema and increased fluid viscosity restrict diffusion and help differentiate collections from parenchyma (Fig. 26.24). Ultrasound can be used for initial evaluation of suspected renal or perirenal abscess, and it has an important role in follow-up to assess treatment response. At US, a renal abscess appears as a hypoechoic thick-walled lesion containing internal echoes, which tend to disappear as the abscess matures and the thickened walls become more distinct. The kidney may be displaced or rotated and is usually enlarged with distorted contour. A perirenal abscess appears as a cystic mass of variable echogenicity adjacent to the kidney. The presence of gas inside the abscess causes ‘dirty’ acoustic shadowing. There is usually no colour Doppler flow inside an abscess on power Doppler US and the mobile debris inside the abscess should not be misinterpreted as colour flow. CEUS will show no enhancement within an abscess, but 50% of mature abscesses will have mural enhancement, and diagnosis of a cystic neoplasm should also be considered. IVU is not recommended for renal abscess detection.

Emphysematous Pyelonephritis Emphysematous pyelonephritis is a life-threatening, fulminant, necrotising upper UTI associated with gas within or surrounding the kidney, acute flank pain, rapid-onset fever and chills and occasionally shock. Gas confined to the renal pelvis is called emphysematous pyelitis, but in the perinephric space, the term ‘perinephric emphysema’ is more appropriate. Uncontrolled diabetes mellitus is a major predisposing condition (90%) and, to a lesser extent, immunocompromise or urinary obstruction due to tumour, stone or sloughed papillae from papillary necrosis. Escherichia coli, Klebsiella pneumoniae and Proteus mirabilis is typically responsible.

Imaging of Emphysematous Pyelonephritis CT is the most reliable and sensitive technique for evaluating and characterising emphysematous pyelonephritis. Findings include parenchymal

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection

A

C

enlargement and destruction, small bubbly or linear streaks of gas radiating from the papillae, gas–fluid levels and focal tissue necrosis with or without renal and perirenal abscess (Fig. 26.25). CT differentiates type 1 emphysematous pyelonephritis (parenchymal destruction with loculated gas without perinephric fluid or collection) from type 2 (parenchymal loculated gas with renal or perirenal fluid or collection). Type 2 emphysematous pyelonephritis has a better prognosis, likely due to better immune response. CT-guided drainage in addition to antibiotics can be used for management of emphysematous pyelonephritis, particularly in patients with a solitary kidney. Emphysematous pyelitis without parenchymal involvement can be depicted easily at CT, but MRI is reserved for cases where CT is contraindicated since signal void due to gas on T1 and T2 weighted images simulates calculi or fluid movement. Perinephric and intraparenchymal fluid collections are accurately delineated on MRI. Emphysematous pyelonephritis should be differentiated from other causes of air inside the kidney such as

739

B

Fig. 26.22  Perirenal and Psoas Abscess. Computed tomography demonstrates a large perinephric fluid collection (images A and B) communicating with the renal collecting system and extending into the left psoas muscle (arrow in image C).

air reflux from the bladder, air within a renal abscess, renal fistulae or retroperitoneal air from perforated viscus. Abnormal gas collections can be detected on plain abdominal radiography in 70% of patients with emphysematous pyelonephritis (Fig. 26.26). At US, multiple non-dependent echogenic foci with low-level reverberations are seen in an enlarged kidney. Adjacent bowel gas or renal stones create diagnostic difficulty. US is inferior to CT for this indication.

Xanthogranulomatous Pyelonephritis Xanthogranulomatous pyelonephritis (XGP) is a chronic granulomatous inflammatory process caused by abnormal host response to bacterial infection often associated with renal pelvic stones, which results in parenchymal destruction, replacement with lipid-laden macrophages and renal impairment. The renal pelvis is initially affected; later the corticomedullary territories become involved and eventually extension

A

B

C

D Fig. 26.23  Renal Abscess. (A) Ultrasound shows a hypoechoic right renal lesion with internal echoes. (B) Axial T1 weighted magnetic resonance imaging (MRI) shows a hypointense parenchymal lesion. (C) Axial T2 weighted MRI demonstrates hyperintensity relative to the surrounding renal parenchyma with a thick margin. (D) There is marginal enhancement on gadolinium contrast-enhanced T1 weighted axial image.

A

B

Fig. 26.24  Magnetic Resonance Imaging of Renal Abscess. Magnetic resonance imaging (MRI) was carried out to evaluate left acute pyelonephritis with possible abscess formation on ultrasound of diabetic patient who was allergic to iodinated contrast medium. (A) There is a hyperintense left renal midpole lesion on coronal T2 weighted imaging. (B) Coronal diffusion-weighted MRI demonstrates diffusion restriction in the lesion.

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection

741

into the perinephric space or retroperitoneum occurs. It is commonest in middle-aged women and diabetic patients (approximately 10%); P. mirabilis and E. coli are most often cultured and, occasionally, transitional cell carcinoma, haematuria, retroperitoneal haemorrhage or renal vein thrombosis occurs. Patients have non-specific symptoms such as loin pain, low-grade fever, chills, dysuria and weight loss.

Imaging of Xanthogranulomatous Pyelonephritis CT is indicated and essential for preoperative assessment, in particular for documenting extrarenal disease in XGP. A non-functioning enlarged kidney, containing a central calculus within a contracted renal pelvis, expansion of the calyces and inflammatory changes in the perinephric fat are strongly suggestive of XGP (Fig. 26.27). Although hypoattenuating

Fig. 26.25  Emphysematous Pyelitis. Axial unenhanced computed tomography image shows an air locule inside a dilated left renal pelvis with peripelvic and perinephric fat stranding and thickening of perinephric fascial planes.

A

B Fig. 26.26  Emphysematous Pyelonephritis. A 38-year-old female patient with uncontrolled diabetes. Plain radiograph shows abnormal soft-tissue opacity and air lucencies projected over the left hypochondrium, lumbar and iliac regions.

Fig. 26.27  Xanthogranulomatous Pyelonephritis. Unenhanced axial (image A) and coronal (image B) computed tomography shows an enlarged right kidney with a staghorn calculus, multiple renal parenchymal collections and perinephric fat stranding.

742

SECTION B  Abdominal Imaging

branching patterns (due to its lipid content) extending from the contracted renal pelvis may suggest hydronephrosis, the low attenuation corresponds to an extensive inflammatory infiltrate rather than fluid in almost all cases. Xanthomatous material does not enhance and approximately 10% of cases are acalculous. Psoas abscess and fistula formation (which is often cutaneous or colonic) are among the more clinically significant patterns of disease progression (Figs 26.28 and 26.29). The disease may be focal in 10% of cases, presenting as a mass in an otherwise functioning kidney; the mass is usually hypoattenuating, and only rim enhancement may be seen. Focal disease may occur in one moiety of a duplex kidney or the kidney may be atrophic. MRI accurately depicts extrarenal extension of XGP and can differentiate focal XGP from renal tumour. On T1 weighted MR images, the solid component of the lesion may be isointense or hyperintense, due to the xanthine and fatty component. On T2 weighted images, the signal intensity of the solid component is isointense to slightly hypointense relative to the normal kidney. The parenchymal cavities filled with fluid and pus

show high signal intensity on T2 weighted images and low signal intensity on T1 weighted images, varying according to the protein concentration in the cavity. Restricted water movement may be observed due to increased viscosity at DW-MRI. Staghorn stones are often seen on plain abdominal radiography, and affected kidneys are usually enlarged and non-functioning in the setting of XGP, or partially enhance and excrete on IVU. Inflammatory extension to the perirenal space will obscure the kidney margins. US typically demonstrates renal enlargement and a staghorn calculus. Loss of corticomedullary differentiation, stenosis of the renal pelvis and replacement of the renal parenchyma by hypoechoic masses with internal echoes may be observed.

Pyonephrosis Pyonephrosis refers to an obstructed infected kidney. Pyonephrosis represents an emergency; the kidney is usually enlarged, and if not drained immediately, permanent parenchymal damage and septicaemia

A

B

C

D Fig. 26.28  Xanthogranulomatous Pyelonephritis. Contrast-enhanced computed tomography reveals multiple parenchymal hypodense areas with a staghorn calculus (arrow in images B and D). There is a large multiloculated perinephric collection involving the right psoas and extending through the abdominal wall into the subcutaneous tissues (asterisk in images A, B and C).

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection

743

A

A

B Fig. 26.29  Xanthogranulomatous Pyelonephritis. Contrast-enhanced computed tomography demonstrates air within the right kidney (image A) and a fistula between the ascending colon and right kidney (arrow in image B).

will ensue. In the adult, pyonephrosis occurs from acute or chronic obstruction secondary to calculus, tumour, stricture or congenital anomaly with superimposed UTI and extensive parenchymal involvement. Patients normally present with fever, chills, loin pain or septicaemia, although a small proportion (15%) may be asymptomatic.

Imaging of Pyonephrosis CT is the investigation of choice for imaging pyonephrosis; it delineates the hydronephrotic component, often its cause, severity and information regarding renal function. Typical findings include hydronephrosis with high-density fluid, perirenal inflammation and thickening of the renal pelvis (>2 mm) with enhancement of the urothelial lining. Excretoryphase CT demonstrates contrast-debris layering. MRI provides little additional information over CT, but MRI may be preferred to avoid intravenous contrast medium administration; on T1 weighted images the signal intensity of the infected fluid will be hypointense to slightly

B Fig. 26.30  Pyonephrosis on Magnetic Resonance Imaging. (A) Axial T2 weighted magnetic resonance imaging (MRI) shows fluid–fluid levels inside dilated right renal calyces. (B) Axial contrast-enhanced T1 weighted MRI demonstrates thickened renal pelvis with mural enhancement.

hyperintense, and on T2 weighted images it will be intermediate to low intensity. MRI may demonstrate fluid–fluid layering and renal pelvic thickening (Fig. 26.30). DW-MRI can help to differentiate between hydronephrosis and pyonephrosis since thick infected fluid in pyonephrosis may yield restricted diffusion. US can detect pelvicalyceal dilatation with echogenic debris or fluid–fluid levels. Echogenic debris in a dilated pelvicalyceal system is the most reliable sign of pyonephrosis, with a sensitivity of 90%, specificity of 97% and accuracy of 96%. US is important for diagnostic aspiration and drainage catheter insertion. IVU is not recommended for assessment of suspected pyonephrosis. Plain abdominal radiography can show renal enlargement or an opaque stone.

744

SECTION B  Abdominal Imaging

Chronic Pyelonephritis Chronic pyelonephritis is characterised by chronic renal inflammatory and fibrotic change induced by recurrent or continuous infection, VUR or urinary tract obstruction. Renal damage and scarring from childhood VUR is termed ‘reflux nephropathy’. Progressive renal scarring can lead to end-stage renal disease, although scarring may occur without VUR.

Imaging of Chronic Pyelonephritis Chronic pyelonephritis (reflux nephropathy) is characterised by renal scarring, atrophy and cortical thinning, hypertrophy of residual normal tissue (which may mimic a lesion with mass effect), calyceal clubbing secondary to retraction of the papilla from overlying scar, thickening and dilatation of the calyceal system and overall renal asymmetry. On US, chronically pyelonephritic kidneys are usually small with focal contour irregularities and areas of fibrosis may also be seen. DMSA demonstrates photopenia in scarred areas of kidney (Fig. 26.31). CT with intravenous contrast medium differentiates non-enhancing areas of infarction from the scar tissue; it also can differentiate pseudotumours of hypertrophied parenchyma from neoplasia. MRI provides information equivalent to CT without intravenous contrast medium. A rare chronic severe form of renal and bladder infection and inflammation is called alkaline-encrusted pyelitis and cystitis. It is characterised by encrustations and calcifications of the urothelium due to gram-positive urea-splitting organisms. On unenhanced CT, linear hyperdense calcifications occur along the thickened urothelium (Fig. 26.32). IVU is less sensitive than CT and US and requires a functioning kidney to demonstrate calyceal and contour irregularities. IVU in many cases cannot differentiate between renal scar secondary to pyelonephritis and focal contour changes in cases of renal infarction, papillary necrosis and fetal lobulation.

A

Renal Tuberculosis The urinary system is the most common extrapulmonary site of TB, normally seeded haematogenously in periglomerular or peritubular regions from active or quiescent pulmonary TB. It has an incidence of 4–8% in patients with pulmonary TB, but only 50% of patients with renal TB have concomitant pulmonary manifestations. Renal TB has an insidious onset characterised by non-specific symptoms including flank pain, dysuria, low-grade fever, malaise or weakness due to parenchymal destruction, impaired function and renal calcification (autonephrectomy). Presentation is frequently late: activated infection has spread from the cortex into the medulla, papillae and the collecting system with haematuria and culture-negative pyuria on urinalysis. Extrarenal spread can involve perinephric and retroperitoneal tissues and adjacent organs including the gastrointestinal tract or skin.

Imaging of Renal Tuberculosis Imaging appearances of renal TB are non-specific and rely on detection of features of papillary necrosis and parenchymal destruction. The presence of three or more of the following features is highly suggestive of TB: pelvicalyceal thickening; ulceration; and fibrosis with or without stricture. Infundibular strictures obstruct renal segments, creating a phantom calyx against a background of normal renal tissue. Strictures distort the collecting system and create cavities and contour deformities. Later calcification can create a thin rim surrounding a necrotic area or completely replace renal parenchyma (autonephrectomy). Calcification occurs in 40%–70% of renal TB cases. Reactivated disease causes inflammation and vasoconstriction, which causes hypoperfusion and a striated nephrogram at contrast medium– enhanced CT. CT can also detect papillary necrosis, which gives calyces a moth-eaten appearance, and CT examines for extrarenal spread. CT can detect parenchymal thinning and scarring and is optimal for detection

B Fig. 26.31  Chronic Pyelonephritis. (A) Ultrasound shows hydronephrotic changes with generalised atrophy of the left renal parenchyma. (B) Technetium 99m dimercaptosuccinic acid (DMSA) demonstrates reduced tracer uptake with photopenic area at the left renal lower pole due to scarring.

of calcification (Fig. 26.33). Fibrotic strictures of infundibula, renal pelvis and ureters are highly suggestive of TB. TB granulomas contain caseous material or calcification on CT. CT is not as sensitive as excretory urography for detecting early urothelial changes. MRI is good at depicting TB cavities (Fig. 26.34), sinus tracts, fistulous communications and extrarenal spread. MRI features of renal macronodular tuberculoma include hypointensity on T1 weighted images and a thick, irregular, hypointense peripheral wall with intralesional fluid–debris level on

CHAPTER 26  Common Uroradiological Referrals: Haematuria, Loin Pain, Renal Failure and Infection T2 weighted images. TB granulomas appear as mildly enhancing soft-tissue masses on CT or MRI (Fig. 26.35). Putty kidney represents end-stage renal TB; dystrophic calcifications involve the entire non-functioning kidney (Fig. 26.36). Focal hyperechoic or hypoechoic renal masses, diffuse parenchymal hyperechogenicity from calcification or renal abscess formation can be observed on US. IVU can detect parenchymal calcification, cavitary lesions, infundibular stenosis with amputated calyces or pelviceal stenosis with hydronephrosis.

Cystitis Cystitis refers to bladder inflammation. Symptoms include urgency, frequency, dysuria, haematuria, or suprapubic pain. Sources include infection (bacterial, TB, schistosomiasis), medications (cyclophosphamide) or radiation. Women younger than 50 years are especially predisposed due to short urethral length and urethral proximity to the anus. Imaging and urine cultures are unnecessary in uncomplicated cystitis, but recurrent cystitis merits urine culture.

745

Imaging of Cystitis Imaging is indicated for complicated cystitis and to differentiate pyelonephritis from cystitis, especially in children. US can assess mucosal thickness (normal mucosa 1 cm), rotation (>45 degrees) and the number of fragments exceeding these parameters. Closed manipulation is usually adequate to reduce most two-part fractures. Based on the above, a

Fig. 45.24  Reversed Hill-Sachs Lesion. Axial fat-suppressed image demonstrating a large antero-medial defect (arrowed) in the humeral head indicative of previous anterior dislocation and a reversed Hill-Sachs lesion. Displaced antero-medial labrum (Bankart lesion) indicates additional anterior instability. Combined appearances indicate multidirectional instability.

Fig. 45.23  Posterior Dislocation. Anteroposterior view demonstrating incongruent articular surfaces. Axial view confirms posterior dislocation of the humeral head.

1150

SECTION C  The Musculoskeletal System

multiple-part fracture of the humeral head, which is undisplaced, is classified as a Neer ‘1 part’ fracture. This pattern accounts for the great majority (85%) of such fractures, which, if minimally displaced, are not considered as separate fragments. Two-, three- and four-part fractures are seen with fractures involving, respectively, the greater and lesser tuberosities in addition to the surgical neck. Axillary artery and vein and brachial plexus injury can occur with surgical neck fractures. Transverse fractures of the humerus involve either the surgical neck or, less commonly, the shaft inferior to this. Shaft fractures are more often spiral and can include segmental fractures. Medial displacement of the humeral shaft proximal to the fracture occurs due to the pull of pectoralis major muscle.

ACROMIO-CLAVICULAR JOINT Clavicle fractures are very common; as the bone is just under the skin surface, they are easy to diagnose clinically. Fractures most often involve the middle third (80%), outer-third fractures occur in 15% of cases and medial third fractures are uncommon, accounting for only 5% of cases. The AC joint is held in place by the AC ligament and the two coraco-clavicular (CC) components known as conoid and trapezoid ligaments. Minor disruptions are best detected by looking at the inferior cortex of both the clavicle and the acromion. These should align. If they do not, then there is an AC joint subluxation. The normal distance between the superior surface of the coracoid and the inferior clavicle is no more than 13 mm. Disruptions of the AC joint are graded from type 1 to type 6 as described by Rockwood. Grade 1 is a sprain of the ligaments only and there is only minor separation, if any, at the AC joint. In grade 2 injuries the AC joint ligaments are torn but the CC ligaments are intact, and the clavicle is elevated but not above the superior border of the acromion (Fig. 45.25). In a grade 3 injury the CC ligaments are also disrupted. As a result, the coraco-clavicular distance may exceed 13 mm and the clavicle is elevated above the superior border of the acromion (Fig. 45.26). Surgical stabilisation will be required. Detecting the difference between a grade 2 and grade 3 injury may be facilitated by weight-bearing radiography using weights strapped to the wrist to distract the AC joint.

Fig. 45.25  Acromio-Clavicular Joint Subluxation (Grade 2). Elevated lateral end of the clavicle but no widening of the coraco-clavicular distance to signify disruption of the intervening ligament.

THE ELBOW There are no variations in the standard views of the elbow, which comprise the AP and lateral projections. A useful soft-tissue sign in the elbow is the fat pad sign. The normal elbow has small pads of fat closely applied to the distal humerus both anteriorly and posteriorly. These are not normally visible as they lie within the bony fossae of the distal humerus; however, if there is a joint effusion, such as may occur following trauma, the fluid displaces the fat pads away from the humerus and out of their respective fossae. These fat pads can then be identified on a lateral radiograph as lucent areas. The anterior fat pad may just be visible normally but it should not be displaced away from the humerus. A visible posterior fat pad is always abnormal. The presence of displaced fat pads indicates an effusion and, even if a fracture is not readily identified, further careful inspection of the radiograph is required to exclude a subtle fracture that may have been overlooked (Fig. 45.27). In children, a supracondylar fracture is the commonest lesion to be overlooked; in adults, it is a radial head fracture. Note, however, that the absence of the fat pad sign does not exclude a fracture and that effusions are not always caused by trauma.

Children

Supracondylar Fractures Supracondylar fractures are the commonest injury around the elbow in children. Some fractures are grossly displaced with the distal bone fragment displaced posteriorly. The anteriorly displaced humeral shaft often has sharp bone edges that abut the adjacent brachial artery as well as jeopardising the median and ulnar nerves. Subtle Greenstick supracondylar fractures account for 50% of injuries in this region (Fig. 45.28) and detection of subtle injuries can be facilitated by using the anterior humeral line. On a lateral radiograph a line drawn down the anterior humeral cortex will pass through the capitellum so that at least one-third of the capitellum lies anterior to the line (Fig. 45.29). If this rule is broken then it is likely that there is a subtle Greenstick or Salter–Harris growth plate injury of the distal humerus, allowing the distal humerus and capitellum to be displaced posteriorly. Lateral epicondyle fractures.  The second commonest elbow injury in children is of the lateral condyle. Recognition of these injuries is made difficult by the presence of developing ossification centres in the immature skeleton; there are four in the distal humerus and one each in the radius and ulna. The order of appearance with few exceptions

Fig. 45.26  Acromioclavicular Joint Subluxation (Grade 3). Distance from coracoid to clavicle is in excess of 13 mm, indicative of coracoclavicular ligament rupture; surgical repair is usually indicated.

CHAPTER 45  Appendicular and Pelvic Trauma

Fig. 45.27  Displaced Fat Pads Are Visible Both Anterior and Posterior to the Humerus. This indicates that there is a joint effusion. Careful scrutiny of the radiograph is now essential to look for a subtle injury.

Fig. 45.28  A Subtle Supracondylar Fracture. This is best identified using the anterior humeral line. A line along the anterior humeral cortex does not pass through the capitellum, which is displaced posterior to the line.

1151

Fig. 45.29  The anterior humeral line (solid line) should pass through the capitellum with at least one third of the capitellum lying anterior to this line. This is especially useful in demonstrating subtle supracondylar fractures, as the line will then pass through the anterior third or miss the capitellum. The radiocapitellar line (hatched) should pass through the middle of the radial shaft intersecting the anterior humeral line within the capitellum.

is capitellum, radius, internal epicondyle, trochlea, olecranon and lateral epicondyle—remembered by the mnemonic CRITOL from the first letter of each epiphysis. Knowledge of this order makes it possible to determine whether a bone fragment adjacent to the humerus is a normal ossification centre or represents a fracture. As the lateral epicondyle is the last ossification centre to appear, then if there is a bone fragment adjacent to the lateral aspect of the distal humerus look for the other ossification centres. If it is not possible to identify the five other centres, then the area in question must represent a fracture (Fig. 45.30). It should be noted that although there may only be a small bony component to this injury, there will be a large fracture through the growing cartilaginous distal humerus. Medial epicondyle fractures.  Medial epicondylar avulsion fractures in children often occur as a result of an avulsion force by the attachment of the common flexors and, less commonly, from direct trauma to the elbow. Soft tissue swelling around the medial epicondyle should alert the reporter to an avulsion injury. Widening of the growth plate should be looked for. The apophysis may be displaced and if not visible when expected to be present in accordance to the CRITOL order of appearance of ossification centres, then intra-articular dislocation should be looked for (Fig. 45.31). Dislocation of the radial head occurs in both children and adults. It may be difficult to appreciate on the AP radiograph but easier on the lateral. A helpful method of detecting these injuries is the use of the radio-capitellar line (see Fig. 45.29), which is drawn along the mid shaft of the radius proximal to the tuberosity. It should pass through the capitellum on every view; if not, the radial head is dislocated (Fig. 45.32). This rule is also true in children, although care is needed in the

1152

SECTION C  The Musculoskeletal System

A

Fig. 45.30  Anteroposterior View of Child’s Elbow. A linear lucency is visible at the lateral epicondyle. This represents a fracture and should not be confused with a normal unfused epiphysis. The lateral epiphysis is the last to appear (CRITOL) and owing to the fact that only the capitellum is visible on the radiograph, this cannot be the lateral epiphysis and must be a fracture.

very young when the capitellum is unossified. The bones of the forearm are bound strongly at either end by strong ligaments. It is difficult to fracture one of these bones without the other, so fractures of the radius and ulna commonly occur together; however, if only one bone appears fractured and, especially if the fracture is angled or overlapped, a covert injury of the other bone is very likely. This typically takes the form of a dislocation of the non-fractured bone. Accordingly, a fracture of the ulnar shaft is often associated with a dislocation of the radial head at the elbow, known as a Monteggia fracture dislocation (Fig. 45.33). Conversely, a radial shaft fracture is associated with a dislocation of the ulna at the wrist and is known as a Galeazzi fracture dislocation.

B Fig. 45.31  Anteroposterior (AP) (A) and lateral (B) views of the elbow. On the AP view there is no visible medial epicondyle ossification centre when expected in accordance to CRITOL. The medial epicondyle ossification centre has been displaced and is seen lying within the joint, seen projected over the trochlea on the AP view (arrow) and projected between the olecranon and humerus on the lateral view (arrow).

Adults Fractures of the distal humerus in adults rarely cause diagnostic problems. Most fractures are readily apparent and typically involve the distal humeral condyles often with intra-articular extension. Capitellar fractures are uncommon and may be difficult to see on the AP radiograph. The lateral view, however, provides the diagnosis. The displaced capitellum has an appearance like a half moon and is seen lying antero-superior to the forearm bones. Radial head fractures may involve the radial neck only or may be intra-articular, with disruption of the radial articular surface. The fat pad sign described above is useful in drawing attention to the radial head as a possible site for subtle radial head fracture. The commonest ulnar fracture is of the olecranon and involves the articular surface with the proximal fracture fragment often displaced owing to the unopposed pull of the triceps muscle. Coronoid fractures of the ulna are uncommon but are often associated with elbow dislocation. Large coronoid fractures often predispose to persistent elbow instability (Fig. 45.34).

Fig. 45.32  Lateral Adolescent Elbow. Fat pads are evident. A line drawn along the proximal radius mid shaft does not pass through the capitellum. The radial head must be dislocated.

CHAPTER 45  Appendicular and Pelvic Trauma

1153

III II I

Fig. 45.33  Radiograph of the Forearm. There is an angulated fracture of the midshaft of the ulna. When this is present there must be an abnormality of the radius. If no fracture is evident, a dislocation of the radial head is almost invariably present, as is the case here. The combination is known as a Monteggia fracture.

Fig. 45.34  Lateral view of the elbow demonstrating a fracture through the coronoid process (arrow).

Fig. 45.35  Carpal Arcs.

Fig. 45.36  Buckle Fracture. Buckling of the posterior radial cortex is demonstrated on the lateral projection. The volar cortex is intact. A subtle increased linear density is the only indicator of injury on the posteroanterior view.

SUMMARY BOX: Upper Limb • MRI allows early diagnosis of stress fractures, which are easily missed on plain radiographs. • A bony Bankart lesion greater than 5 mm in depth and width is associated with a high probability of re-dislocation. • A posterior elbow fat pad is always abnormal. • Subtle supracondylar fracture detection can be facilitated by using the anterior humeral line. • Any missing ossification centre of the elbow should raise suspicion of injury and a displaced apophysis looked for.

THE WRIST Standard views for evaluation of the wrist include a PA and lateral radiographs, with additional views required when there is clinical suspicion of a scaphoid fracture. The AP radiographs demonstrate uniform spacing (1 to 2 mm) around each carpal bone. Variation in this spacing usually signifies ligamentous injury with associated subluxation or dislocation. The carpal bones are organised into three smooth arcs: the first delineates the proximal surface of the proximal row of carpal bones (scaphoid, lunate, triquetrum), the second is formed by the distal articular surface of the same bones and the third is along the proximal curvature of the capitate and hamate. These arcs should normally be smooth with no steps or interruption (Fig. 45.35).

The distal radius has a volar tilt of approximately 10 degree on the lateral projection, which represents the angle between a line drawn along the long axis of the radius and a line drawn from the dorsal to the volar rim of the radius, and has a normal range of 2 to 20 degree. Alteration of this angle occurs when there is a fracture of the distal radius. Normal relationship of the distal radius and carpal bones is best assessed on the lateral view; this is discussed further in the section on carpal injuries. Displacement of the distal radial epiphysis and disruption of carpal alignment are often only evident on the lateral view. Fractures of the triquetrum are usually only seen on the lateral radiograph and this, along with evaluation of normal longitudinal alignment, emphasise the critical importance of this projection.

Radius and Ulna Most injuries result from a fall on the outstretched wrist. The type of injury sustained is age related, a consequence of the reduced elasticity and strength of bones with increasing age and the existence of relatively weak growth plates in children.

Children In young children, fractures of the radius proximal to the epiphyseal plate predominate in the age group 6 to 10 years, including Buckle (Torus) (Fig. 45.36) and Greenstick fractures (Fig. 45.37). Before epiphyseal fusion, injuries also commonly involve the epiphyseal plate and

1154

SECTION C  The Musculoskeletal System are encompassed by the Salter–Harris classification. Dorsal displacement of the epiphysis with or without an associated fracture of the adjacent metaphysis is common, resulting in either a Salter–Harris type 1 or 2 injury respectively (Fig. 45.38).

Adults Distal radial fractures are the commonest wrist injury in patients over the age of 40, with the increasing frequency relating to progressive osteoporosis in the older age groups.

Colles Fracture The Colles fracture is the commonest wrist injury sustained in adults and includes fracture of the radius within 2 cm of distal radial articular surface, dorsal angulation or displacement of distal fragment and associated fracture of ulnar styloid process. The AO classification includes varying degrees of intra-articular extension but there is no associated radio-carpal subluxation (Fig. 45.39).

Smith Fracture The Smiths fracture describes a ‘reversed Colles fracture’ of the distal radius. The distal radial fragment is displaced and angled volarly and medially, which is sustained through a fall on a volar flexed wrist (Fig. 45.40).

Barton Fracture Barton fracture differs from the Colles and Smith fractures as there is subluxation of the carpus and a rim fracture at the radiocarpal interface. This is a shearing injury through the articular surface and can involve volar or dorsal rim fractures and corresponding displacement. Fig. 45.37  Greenstick Fracture. Lateral radiograph following a fall on outstretched hand, demonstrating an oblique cortical fracture on the volar aspect of the radius and buckling of the dorsal cortex.

Chauffeur Fracture (Hutchinson Fracture) The Chauffeur fracture refers to an intra-articular fracture of the radial styloid and was initially sustained due to the backfire of the starting

Fig. 45.38  Salter-Harris Type 2 Fracture. The posteroanterior view demonstrates metaphyseal deformity but the lateral view illustrates epiphyseal and linked dorsal metaphyseal displacement uncovering the volar 2/3rds of the radial growth plate and metaphysis.

CHAPTER 45  Appendicular and Pelvic Trauma

A

1155

B

Fig. 45.39  Colles Fracture. Lateral view demonstrating fracture within 2 cm of the radio-carpal interface with substantial dorsal tilt. The posteroanterior view demonstrates an ulnar styloid fracture.

A

B

Fig. 45.40  Smith Fracture. Volar displacement of the distal radius. The posteroanterior view is indistinguishable from the Colles fracture.

handle of a car; however, it is now most commonly sustained due to axial compression transmitted through the scaphoid. Many fractures do not conform exactly with the aforementioned descriptions; accordingly, rather than using the eponymous names, it is often better to give a full description of the fracture lines, fragment displacement, angulation and extent of articular involvement.

Carpal Injuries Scaphoid

Scaphoid injury accounts for approximately 60% of all carpal fractures. Typically, when scaphoid injury is suspected, PA, lateral and oblique radiographic projections are employed. The intention of the oblique views is to elongate the scaphoid, projecting it clear of other carpal bones,

allowing the x-ray beam to pass through the scaphoid perpendicular to its long axis and parallel to the line of fracture (Fig. 45.41). Additional projections, including a second PA view in ulnar deviation with or without tube angulation, can also be employed to extend and elongate the scaphoid and improve fracture identification. The blood supply of the scaphoid is highly relevant to the risk of complications (non-union and avascular necrosis). Arteries enter distally then pass proximally through the scaphoid to supply its proximal pole and the probability of avascular necrosis (AVN) and non-union increases the more proximal the fracture (Fig. 45.42). AVN and non-union may require additional imaging, particularly MRI, to establish their existence if symptoms persist months after the injury and there is uncertainty regarding union (Fig. 45.43).

1156

SECTION C  The Musculoskeletal System

Most scaphoid fractures are visible on initial radiographs; nevertheless, there are a small number that are not identified initially but assuming a potential risk of AVN and non-union, the patient is managed as if a fracture is present and the wrist is immobilised. These patients are reviewed at approximately 10 days when further radiographs are obtained if symptoms persist, in the belief that either resorption or displacement may occur around the fracture site, rendering it more visible on delayed radiographs. Alternative imaging techniques are now increasingly utilised in the acute and subacute phases to allow definitive treatment to be instituted earlier. MRI is considered to be the best imaging investigation in detection of occult scaphoid fractures (Fig. 45.44). MRI is more

Fig. 45.41  Scaphoid Series and Distal Pole Fracture. Lateral, Posteroanterior and Both Obliques. The forehand oblique (bottom right) elongates the scaphoid and demonstrates the distal pole fracture.

Dorsal

sensitive than CT and scintigraphy, although CT can identify most significant fractures (Fig. 45.45).

Triquetrum Triquetral fractures are the next most common injured carpal bone, accounting for 20% of injuries. These fractures are not evident on the PA view but can be seen on the lateral radiograph as a bone fragment on the dorsal aspect of the wrist (Fig. 45.46). The small fragment is most commonly an avulsion by the radio-lunato-triquetral ligament. Other more complex injuries can be identified by CT, but this does not, in the authors’ experience, change management. Other carpal bone

Fig. 45.43  Non-Union and AVN Scaphoid. Persistent fracture line in the waist of scaphoid and low signal marrow space especially in the proximal pole, typical of avascular necrosis.

Palmar

Union rate %

100%

90% 84% 57% 27%

Fig. 45.42  Diagrammatic representation of the main arterial supply to the scaphoid and the relative probability of non-union with fracture lines at different levels in the scaphoid.

CHAPTER 45  Appendicular and Pelvic Trauma

Fig. 45.44  MR of Acute Scaphoid Fracture. Radiographically normal, acute magnetic resonance image using short tau inversion recovery images in the coronal plane demonstrates occult fracture of the scaphoid waist (low signal line) with adjacent marrow oedema.

1157

Fig. 45.46  Triquetral Fracture. Small fragment demonstrated along the dorsum of the mid carpus. Cross-sectional imaging will invariably confirm this as triquetral fracture.

Lunate Dislocation On a PA radiograph, the lunate normally has a trapezoid appearance but, when dislocated, it assumes a triangular shape. The normal carpal arcs and the intercarpal spaces are lost if the lunate displaces. The diagnosis is usually established on the lateral projection. A vertical line drawn through the centre of the normal lunate on the lateral view should, if extended, pass through the radius and capitate (Fig. 45.47); this alignment is lost if the lunate dislocates (Fig. 45.48). Lunate dislocation is most commonly an isolated finding: additional fractures are rare but interosseous ligamentous disruption in the proximal carpal row is invariably present.

Perilunate Dislocation Perilunate dislocation (PLD) is more common and is often associated with other carpal injuries, most commonly a fracture of the scaphoid (transcapho-PLD). In PLD, the lunate remains aligned with the radius but the capitate and the rest of the distal carpal row are displaced dorsally (Fig. 45.49).

Mid-Carpal Dislocation Fig. 45.45  CT Reconstruction of Scaphoid Fracture. Coronal CT reconstruction demonstrating a radiographically occult scaphoid waist fracture.

injuries are all relatively rare, each accounting for approximately 2% to 3% of carpal injuries. CT or MRI has a useful role to play in these injuries, which are often difficult to detect on plain radiography.

Carpal Dislocation and Subluxation Most carpal fractures and dislocations are caused by falling on the outstretched hand with the wrist forced into hyperextension. Carpal dislocations most commonly involve the lunate.

The lunate is subluxed rather than dislocated in a volar direction and the carpus subluxed posteriorly. It is assumed that following posterior dislocation of the carpus (PLD), partial spontaneous realignment displaces the lunate in a volar direction.

Scapholunate Disassociation Scapholunate disassociation describes disruption of the interosseous scapholunate ligaments, which leads to rotatory subluxation (flexion) of the scaphoid. The ligaments are injured as a result forced wrist extension. On plain radiographs the injury can be identified by a widening of the intercarpal distance between scaphoid and lunate. The normal gap is 40% of posterior wall, indicating a requirement for surgical intervention.

Fig. 45.65  Anterior Column Fracture. Typical features, on sequential CT images (A to C), demonstrating coronal fracture plane dividing the acetabulum into anterior and posterior halves (arrowed) and extension across obturator ring (these two features are common to Anterior and Posterior column fractures). The exit point anteriorly above acetabulum identifies this as an Anterior rather than a Posterior column injury.

1166

SECTION C  The Musculoskeletal System

Fig. 45.67  Ischial Tuberosity Avulsion. Hamstring avulsion of the left ischial tuberosity (arrowed).

A

B Fig. 45.66  Transverse Fracture. Plain radiograph (A) and CT surface reconstruction (B) demonstrating transverse fracture pattern dividing acetabulum into upper and lower halves in contra-distinction to columntype injuries.

Pelvic Avulsion Injuries The commonest sites are the anterosuperior iliac spine (sartorius origin), anteroinferior iliac spine (rectus femoris origin) and the ischial tuberosity (hamstring origin) (Fig. 45.67). Predisposition of these sites relates to all of these muscles units crossing two joints with associated risk of isometric contractions generating excessive forces at the tendon origins. These injuries are more common in athletic children and adolescents due to the immaturity and relative weakness of the apophyseal growth plate. This may be identifiable on plain radiographs if a thin sliver of bone is avulsed, with the apophysis through the line of provisional calcification. Alternatively, ultrasound (US) or MRI may be required if the plain radiographs are negative and functional disability is significant. Surgical repair will be considered for hamstring avulsions.

Pelvic Insufficiency and Stress Fractures Stress fractures are sustained through excessive loading, usually but not exclusively during athletic training. The vogue for distance running, in particular, has precipitated an increase in lower limb stress injuries, including the pelvis. Plain radiographs are usually normal and, in

Fig. 45.68  Pelvic Stress Fracture. Incomplete stress fracture line above the right acetabulum on T1-spin echo imaging in a 20-year-old soldier.

these circumstances, MRI is the preferred investigation. Stress injuries are usually associated with marrow oedema; this may have a linear orientation and occur anywhere in the pelvic ring (Fig. 45.68). The posterior column of the acetabulum and the sacrum are favoured sites due to their relatively high loading. Sacral fractures or stress reactions often parallel the sacroiliac joint. The fracture line may be identifiable but is often obscured on STIR or fat-saturated sequences and is best demonstrated on T1-SE. CT is capable of identifying overt fractures but is less capable of identifying stress reactions without a fracture line and does not adequately assess other potential differential diagnoses relating to soft-tissue injury. Nuclear medicine is sensitive at identifying stress injuries but unless the findings are bilateral and symmetric, its findings are usually non-specific.

CHAPTER 45  Appendicular and Pelvic Trauma Insufficiency fractures occur when normal loads are applied to a structurally deficient pelvic ring. These fractures are frequently multiple and, characteristically, are bilateral and relatively symmetric on MRI. This most commonly occurs due to osteoporosis, but other metabolic disorders also predispose fracture, including osteomalacia and parathyroid-related bone disease. Insufficiency fractures in the posterior pelvis and sacrum are common following pelvic radiation therapy for gynaecological malignancy. MRI is the investigation of choice in elderly patients with unexplained pelvic pain, a where insufficiency fractures are suspected.

Pathological Fractures Pathological fractures in the pelvis are not uncommon and are most frequently associated with metastatic tumour and myeloma but are also encountered with other disorders, including osteopetrosis and Paget disease (Fig. 45.69). The diagnosis in metastatic malignancy can be supported by the identification of other areas of increased isotope uptake with the pelvis on skeletal scintigraphy. Other processes are also recognised to cause pathological fractures, including osteopetrosis and Paget disease.

1167

and 4 fractures. As a general rule, intracapsular fractures are treated with either a hemiarthroplasty or in patients who were independently mobile before fracturing with a total hip replacement. In young patients under the age of 60 with undisplaced (Garden type 1 and 2) intracapsular fractures, most surgeons favour internal fixation with dynamic hip screw or multiple screws. Most fractures are readily apparent but approximately 15% are difficult to detect. Looking carefully at the trabecular pattern will assist in identifying these: interruption of the trabecular lines indicates a subtle fracture. Impacted undisplaced fractures may be identified by the presence of a sclerotic line and/or interruption of the normal trabecular pattern. Use of the lateral view may be extremely helpful in identifying fractures when there is uncertainty on the AP view (Fig. 45.73). Nevertheless, CT and/or MRI may be required to confirm/exclude subtle fractures.

Intertrochanteric and Subtrochanteric Fracture Inter- and subtrochanteric fractures are extracapsular and the blood supply is therefore at less risk. These fractures can be treated by internal

The Hip and Proximal Femur Fracture of the proximal femur is one of the commonest injuries, particularly in elderly patients with osteoporosis. It is the commonest reason for urgent admission to an orthopaedic ward. The standard view is an AP of the pelvis and a lateral of the painful hip. The pelvic radiograph allows comparison of the injured hip with the uninjured side. As these patients will generally be in pain, the lateral radiograph is obtained using the so-called groin lateral projection. This is obtained with a horizontal beam with 20 degree cephalic angulation centred on the greater trochanter. The opposite thigh is flexed. This does not require any movement by the patient of the painful injured side.

Intracapsular Subcapital Transcervical Basicervical Intertrochanteric

Extracapsular

Femoral Neck Fracture Fractures are classified by location from subcapital to intertrochanteric (Fig. 45.70). Subcapital, transcervical and basicervical fractures are intracapsular. The blood supply to the femoral head is derived from recurrent arteries closely applied to the femoral neck. The risk of disrupting the blood supply resulting in avascular necrosis of the femoral head is greatest with intracapsular fractures, particularly if the fracture is displaced (Fig. 45.71), resulting in avascular necrosis in 15% to 35% of patients. The Garden classification is a system of categorising intracapsular fractures (Fig. 45.72). The blood supply to the femoral head is more likely to be compromised in Garden type 3

Fig. 45.69  Paget Disease. Pathological fracture through Pagetoid bone involving obturator ring and ischium (arrowed).

Subtrochanteric

Fig. 45.70  The Sites of Proximal Femoral Fractures.

Fig. 45.71  Displaced Subcapital Fracture of the Femoral Neck. The blood supply will almost inevitably be interrupted.

1168

SECTION C  The Musculoskeletal System Normal

Garden classification:

1

Undisplaced incomplete, including impaction in valgus

3

Complete fracture varus angulation

2 Complete fracture, no displacement

fixation with plate and dynamic screw fixation or intramedullary nailing. Intertrochanteric fractures are easily identified when severely comminuted (Fig. 45.74), but some fractures may be overlooked when only minimally displaced. Care should be taken when air is trapped in an overlying skin crease. This may mimic a fracture or, conversely, mask an underlying fracture (Fig. 45.75). Despite very careful observation, approximately 1% of fractures are initially occult and cannot be identified on radiographs. If these are not identified and the patient is encouraged to mobilise and weight-bear, then the fracture may become displaced. Because of the risk of subsequent avascular necrosis, great effort must be made to detect these fractures. In the absence of an identifiable fracture on the radiographs, and if weight bearing is limited at 24 hours, further investigations is indicated. MRI is specific and sensitive and is

4 Completely displaced

Fig. 45.72  Garden Classification of Hip Fractures. Stage 1: Undisplaced incomplete, including impaction in valgus. Stage 2: Complete fracture no displacement. Stage 3: Complete fracture varus angulation. Stage 4: Completely displaced.

Fig. 45.73  Subtle Interruption of the Trabecular Lines of the Femoral Neck. There is ill-defined sclerosis extending across the neck. These signs suggest an impacted fracture.

Fig. 45.74  A Comminuted Intertrochanteric Fracture on the Left.

Fig. 45.75  Air Is Trapped in the Skin Crease of the Groin, Producing a Linear Lucency Traversing the Underlying Femur. On close inspection, however, the lucency crossing the bone is separate from the soft-tissue lucency. Therefore, this patient has a fracture as well as air in the groin crease.

CHAPTER 45  Appendicular and Pelvic Trauma

1169

Fig. 45.77  Fracture through the left superior pubic ramus and a subtle fracture through the left inferior pubic ramus.

Fig. 45.76  The Anteroposterior Radiograph of the Hip Was Normal. Magnetic resonance image coronal short tau inversion recovery shows a low signal transcervical fracture line. There is surrounding high signal marrow oedema.

now the investigation of choice; if MRI is not available or is contraindicated, then CT should be considered (Fig. 45.76). Hip fractures are often caused by underlying osteoporosis so, for this reason, patients under the age of 75 who sustain a hip fracture should be assessed for underlying osteoporosis (usually by DEXA) and treated appropriately. There are two fractures that may mimic the signs and symptoms of a hip fracture, namely, fractures of the pubic rami and greater trochanter. These should be looked for in every patient but especially when no femoral neck fracture is apparent (Fig. 45.77). A significant number of patients with isolated greater trochanteric fractures on plain radiographs will have intertrochanteric fracture extension. MRI is therefore recommended in patients with greater trochanteric fractures (Fig. 45.78).

Lesser Trochanter Fracture Isolated fractures of the lesser trochanter are uncommon. They can be seen in athletic teenagers caused by sudden forceful iliopsoas contraction. In adults, especially if non-traumatic, there should be a very high suspicion of underlying, usually metastatic, tumour infiltration (Fig. 45.79).

Atypical Femoral Fractures There is a causative relationship between the long-term use of bisphosphonates used in the treatment of osteoporosis and atypical femoral fractures that are typically subtrochanteric. These fractures are associated with minimal or no trauma, originating at the lateral cortex. The fractures may be complete or incomplete. Complete fractures are substantially transverse in orientation with a characteristic medial unicortical spike. Both complete and incomplete are associated with thickening (beaking) of the lateral cortex. The fractures are often bilateral (Figs 45.80 and 45.81).

Fig. 45.78  Magnetic resonance image coronal T1 shows a fracture through the greater trochanter with intertrochanteric extension (arrow).

SUMMARY BOX: Pelvis • The risk of femoral head avascular necrosis is greatest with intracapsular Garden type 3 and 4 fractures. • In the absence of an identifiable fracture on plain radiograph, MRI is the investigation of choice if ongoing clinical suspicion. • A significant number of patients with isolated greater trochanteric fractures on plain radiographs will have intertrochanteric extension on MRI. • There should be a high suspicion of underlying pathology in isolated, lesser trochanteric fractures in adults. • Long-term bisphosphonate treatment is associated with subtrochanteric femoral fractures.

1170

SECTION C  The Musculoskeletal System

Fig. 45.79  Anteroposterior radiograph of the pelvis showing a displaced left lesser trochanter avulsion fracture secondary to metastatic infiltration (arrow).

Fig. 45.81  Same Patient as in Fig. H.11. Coronal short tau inversion recovery magnetic resonance image showing incomplete fracture of the left femur. There is cortical breaking of the lateral cortex (long arrow) and underlying incomplete fracture with surrounding marrow oedema (short arrows).

Fig. 45.82  Horizontal Beam Lateral Radiograph. There is a fluid level evident within the suprapatellar pouch. The dark layer on top represents fluid fat, which has layered on top of the much brighter fluid blood. The fluid fat has escaped from the bone marrow, which means there must be a fracture present.

Fig. 45.80  Thickening of the left lateral cortex owing to incomplete fracture in a patient on bisphosphonates (arrow).

THE KNEE On the AP radiograph, the fibula denotes the lateral side of the knee. The medial and lateral joint spaces are usually equal and the cortices of the tibial plateau, both medial and lateral, are sharply defined and sclerotic. A line dropped perpendicularly from the lateral femoral condyle should meet the margin of the adjacent tibial plateau, with (at most) 5 mm of the plateau lying lateral to the line. If this rule is broken, suspect a lateral tibial plateau fracture. The lateral view following trauma should be obtained using a horizontal beam: that is with the patient lying flat and the x-ray beam parallel to the floor. This enables a useful soft-tissue sign to become evident, which is a fluid–fluid level superior

to the patella lying within the suprapatellar pouch. It consists of fat layered on blood, known as a lipohaemarthrosis. The only fat that is fluid at body temperature is marrow fat. Therefore to be evident on the radiograph there must be a break in a bony cortex, allowing fat to escape into the knee joint; this is usually accompanied by blood. If this sign is present, there must be an intra-articular fracture and this is almost always visible on the radiograph with careful inspection (Fig. 45.82). Supracondylar and condylar fractures of the femur are caused by severe trauma and usually present little diagnostic difficulty.

Tibial Plateau Fractures Fractures of the tibial plateau typically occur when the lateral side of the knee is struck, forcing the knee into valgus. The lateral femoral condyle is driven down into the lateral tibial plateau, resulting in fracturing and depression of the plateau. Medial fractures are much less common, accounting for only 10% of such injuries. Most injuries are

CHAPTER 45  Appendicular and Pelvic Trauma

1171

Fig. 45.85  A lateral plateau fracture with displacement of bone laterally such that a line drawn down the lateral cortex of the femur does not run smoothly onto the lateral margin of the tibia. Fig. 45.83  An Obvious Fracture of the Lateral Tibial Plateau Which Has Been Depressed Inferiorly. There is an associated fracture of the fibula.

A

B

Fig. 45.84  (A) Area of sclerosis inferior to the lateral plateau indicative of trabecular condensation as a result of a plateau fracture. (B) Computed tomography demonstrates the fracture lines and also shows the degree of depression of the fracture. The sclerotic impacted trabeculae correlate with the plain radiographic findings.

easily detected (Fig. 45.83). Many patients now undergo CT, not for diagnosis but to aid surgical planning. The number of fracture fragments and the degree of plateau depression will determine the need for surgery and the type. Depression of the articular surface by more than 10 mm is a key observation in this regard. Tibial plateau depression is measured as the vertical distance between the lowest point on the intact medial plateau and the lowest depressed lateral plateau fracture fragment. More subtle injuries may be detected by use of three observations: loss of clarity of the normal lateral plateau cortex impaction of the subcortical trabeculae resulting in sclerosis (Fig. 45.84), and lateral displacement of the tibial margin beyond a vertical line drawn inferiorly from the lateral femoral condyle. Small bone fragments around the knee usually signify significant injury. They should not be dismissed as flakes or chips as many are avulsion injuries by tendons or ligaments. Severe force and distraction of the knee is required to avulse the bony attachments seen on

Fig. 45.86  There are small bone fragments seen centrally in the joint space projected over the tibial spines. Bone fragments in this position are caused by avulsion by the anterior cruciate ligament.

the radiographs. There are three findings of particular significance (Fig. 45.85).

Anterior Cruciate Ligament Avulsion Fracture The anterior cruciate ligament (ACL) inserts into the anterior tibial eminence. In adolescents, in particular, this bony attachment is weaker than the ligament. In some instances the ligament avulses its bony attachment, resulting in a bone fragment that is seen within the joint centrally on the AP view and in the anterior half of the joint on the lateral (Fig. 45.86). The Meyers and McKeever classification system

1172

SECTION C  The Musculoskeletal System

describes four subtypes dependent on displacement. Type I describes a minimally displaced fragment, Type II describes anterior elevation of the fracture fragment. In Type III and IV there is displacement of the fragment with additional rotation or comminution of the fracture fragment in type IV.

Segond Fracture

distal femur. If a patellar injury is suspected clinically and none is seen on conventional radiographs, then an additional radiograph, the skyline (or sunrise) view, should be taken with the knee in flexion. This radiographically projects the patella off the femur and may allow a vertical fracture to be seen (Fig. 45.90). Stellate fractures are usually caused by a direct blow to the patella and may only be seen on the AP view with difficulty.

The presence of a small bone fragment adjacent to the upper lateral tibia, known as a Segond fracture, is an avulsion injury by the lateral capsular ligament (Fig. 45.87). Its importance is that the mechanism of a forced varus injury that produces this finding results in a very high incidence of rupture of the ACL and tears of the medial meniscus, both said to be as high as 75% to 100% of cases.

Fibula Avulsion Fracture An avulsion fracture of the fibular head is referred to as the arcuate sign (Fig. 45.88). Its presence indicates an injury to at least one of the posterolateral corner structures that insert onto the fibular head (lateral collateral ligament, popliteo-fibular ligament, arcuate ligament, biceps femoris muscle tendon). This is usually associated with cruciate ligament injury in approximately 90% of cases, most commonly the PCL. Its recognition is important because the patient may develop chronic posterolateral instability.

Patella Fractures Patellar fractures are either transverse vertical or stellate (comminuted). Patellar injuries occur either from a direct blow or as a result of contraction of the powerful quadriceps muscles, resulting in a transverse fracture that accounts for 60% of cases. These injuries are readily visible on a lateral view (Fig. 45.89). Vertical fractures are much less common (15%) and are visible on the AP view with difficulty because of the overlying

Fig. 45.88  Anteroposterior knee radiograph shows an avulsion fracture of the fibular head (arrow).

Fig. 45.87  A Very Small Bone Fragment Is Seen Adjacent to the Lateral Tibia Also Overlapping the Fibula Head. This should not be misinterpreted as a minor flake injury. It indicates a high likelihood of injuries of the anterior cruciate ligament and medial meniscus.

Fig. 45.89  Lateral Knee. A horizontal patellar fracture is readily identified.

CHAPTER 45  Appendicular and Pelvic Trauma

1173

A

B

Fig. 45.91  Bipartite Patella. The bone fragment seen arising from the superolateral quadrant of the patella is a normal variant, a bipartite patella. It should not be mistaken for a fracture.

Fig. 45.90  (A) A vertical patellar fracture is seen with some difficulty. (B) If index of suspicion is high and no fracture is seen, then an additional skyline view will identify such fractures more readily.

Bipartite patella is a normal variant. The bone fragment always occurs in the superior lateral quadrant and has well-defined sclerotic margins (Fig. 45.91). The fragment is usually larger than the defect. These can be symptomatic as they can become unstable and are associated with cartilage irregularity.

Patellar Dislocation Patellar dislocation is rarely evident as such on plain radiographs. Dislocation is usually transient and has relocated long before the patient attends the emergency department. The only occasional evidence of the dislocation is when an osteochondral fragment is sheared from the lateral femoral condyle or medial patella facet, or avulsed from the patella by the medial patellofemoral ligament. These bone fragments may be visible on the radiograph. Again, the skyline view may be helpful in demonstrating these small bone fragments. This is another example of a small bone fragment indicating a significant injury. It should be distinguished from an ACL avulsion injury described earlier. In the latter, the bone fragment lies centrally in the joint on both views. The former is seen often on one view only and lies away from the joint centre (Fig. 45.92).

THE ANKLE Ankle injuries are common and the interpretation of these injuries based on plain radiographs requires an appreciation of the mechanism of injury and the probability of associated ligamentous injury. The ankle joint should be considered as a ring of bone and ligaments. A single breach is invariably stable but any combination of two or more bone and ligamentous injury will result in an unstable joint potentially requiring operative reduction and fixation. The five common sites of injury are the lateral ligament, lateral malleolus, syndesmosis, medial malleolus and deltoid ligament.

Fig. 45.92  There Is a Curvilinear Bone Fragment Seen in the Anterior Aspect of the Knee. This is not visible on an anteroposterior view. Bone fragments like this are typically found as a result of patellar dislocation, although in most cases there are no radiographic abnormalities.

Standard views include the mortice and lateral views The mortice view is an AP with approximately 10 degrees of internal rotation of the foot. The medial and lateral malleoli are equidistant to the table top, which projects the fibula clear of the talus and lateral gutter, facilitating identification of fractures of the lateral malleolus and widening of the lateral gutter, which is indicative of instability. The lateral radiograph

1174

SECTION C  The Musculoskeletal System

of the ankle should superimpose the medial and lateral malleoli and must also include the calcaneus and base of the fifth metatarsal, both of which can present as potential ankle fractures. Oblique fractures of the distal fibula are often visible only on the lateral view. A simple classification of ankle fractures is that of Danis-Weber (Fig. 45.93). This considers only the level of the fibular fracture in relation to the talar dome but under-represents the important soft tissue structures of the syndesmosis and associated ligaments of the distal tibia and fibula, which are also commonly injured. There are three types of injury in this classification (Types A, B and C). Type C occurs above the talar dome and is the pattern most likely to be associated with syndesmotic injury and requirement for operative reduction. Anatomical descriptions of ankle fractures affect the malleoli: medial and/or lateral either singly (unimalleolar) or in combination (bimalleolar), although this is even more limited than the Danis-Weber system in appreciating the severity and mechanism of injury. There are also associated fractures of the posterior tubercle of the distal tibia, sometimes referred to as the posterior malleolus, resulting in a trimalleolar fracture. The Lauge-Hansen classification addresses both the mechanism of injury and defines for each mechanism a predictable order of structural failure. This is based on the position of the foot (supination or pronation) and the direction of the force applied (Adduction, Abduction and Lateral rotation). Recognition of one of the four unique patterns of fracture allows for recognition of the deforming force, informing how to reduce the injury. A more comprehensive understanding of associated soft injuries is required as each fracture pattern follows a predictable order of bone and/or soft-tissue failure, entitled the Lauge-Hansen Ankle Classification Supination-Adduction (SAD): Stage 1: Lateral ligament rupture or low transverse lateral malleolar fracture below ankle line (Danis-Weber Type A)

Inversion

A

Inversion Rotation

Eversion

Eversion Rotation

Stage 2: Oblique fracture base of medial malleolus Supination-Lateral Rotation (SLR) Stage 1: Anterior tibiofibular ligament rupture Stage 2: Spiral/oblique fibular fracture crossing the joint line (DanisWeber Type B) Stage 3: Posterior tibiofibular ligament rupture or avulsion of its tibial insertion Stage 4: Transverse medial malleolar fracture Pronation-Abduction (PAB) Stage 1: Deltoid ligament rupture or transverse medial malleolar fracture Stage 2: Anterior and posterior tibiofibular ligament rupture Stage 3: Oblique comminuted fibular fracture at the joint line Pronation-Lateral Rotation (PLR) Stage 1: Deltoid ligament rupture or transverse medial malleolar fracture Stage 2: Anterior tibiofibular ligament and interosseous membrane rupture Stage 3: High spiral fracture of fibula 6 cm or more above the joint line Stage 4: Posterior tibiofibular ligament rupture or avulsion of its tibial insertion The pattern of injury according to this classification can usually be defined by the site and appearance of the fibular fracture. Other associated fractures or features of instability allow the interpreting radiologist to appreciate associated ligamentous injuries according to the stages within each of the patterns (see Fig. 45.93A)

Supination-Abduction (Fig. 45.94) Simply put, this represents the simple inversion injury. Low transverse fibular fracture (Danis-Weber Type A) defines its type (Fig. 45.95), which is a single column fracture that is stable. An oblique medial malleolar fracture (stage 2) (Fig. 45.96), by definition will have been preceded by the fibular fracture described above or, if not present, a lateral ligament rupture can be assumed. Stage 2 injuries are unstable.

Supination-Lateral Rotation (Fig. 45.97) The fibular fracture is oblique and is frequently only seen on the lateral projection (Fig. 45.98). The fracture line is high posteriorly and low anteriorly and crosses the joint line. Posterior and medial malleolar fractures define stages 3 and 4 respectively and are associated with ankle instability (Trimalleolar fracture) (Fig. 45.99).

Lauge-Hansen system defining fibular fracture patterns 2 1

A B

B

C

Danis-Weber classification system of fracture patterns

Fig. 45.93  (A) Diagrammatic representation of the patterns of fracture in the Danis-Weber Classification. (B) Patterns of fibular fracture defined the mechanism of injury according to the Lauge-Hansen system.

Fig. 45.94  Supination-adduction (simple inversion) diagrammatic representation of the patterns of injury (numerics represent the order of structural failure in this and subsequent patterns).

CHAPTER 45  Appendicular and Pelvic Trauma

1175

Fig. 45.95  Supination-Adduction Injury Stage 1. Typical low transverse fibula fracture below the joint line (commensurate with Danis-Weber type 1 injury). Intact medial pillar.

Pronation-Abduction (Fig. 45.100) Often referred to as a simple eversion injury. The medial pillar is injured initially (Deltoid ligament or Medial Malleolus), and the fibular is then fractured and usually angles laterally distal to the fracture line under the abducting force (Fig. 45.101). The lateral cortex is often comminuted. This fibula fracture is stage 2 of this pattern and in the absence of a transverse medial malleolar fracture, rupture of the deltoid ligament and tibiofibular ligaments can be assumed (Fig. 45.102).

Pronation-Lateral Rotation (Fig. 45.103) Several points distinguish this from the SLR injury. No oblique fibular fracture crossing the joint line is seen, with the posterior malleolar fracture in a PLR injury. A posterior malleolar fracture in isolation would immediately define this as a stage 4 PLR injury and is highly unstable (Fig. 45.104). Proximal leg views may be required to demonstrate the high fibular fracture. Proximal fibular fractures of this type are often referred to as Maisonnueuve fracture (Fig. 45.105).

Fig. 45.96  Supination-Adduction Injury Stage 2. The oblique medial malleolar injury is characteristic of the stage 2 injury. The lateral malleolus is either fractured as per A 3 or the lateral ligaments are ruptured.

Pilon Fracture

Triplane Fracture

Fractures that disrupt the weight-bearing articular surface of the distal tibia caused by a high-energy axial loading. These fractures are generally readily apparent and are often associated with comminuted fractures of the fibula as well as the tibia. The distal tibio-fibula syndesmosis usually remains intact and forms the focus onto which other fragments are operatively reduced and stabilised (Fig. 45.106A and B).

Triplane fractures represent a S-H type 4 injury, characterised by a sagittal fracture through the epiphysis, a horizontal fracture through the growth plate and a coronal fracture plane through the posterior tibial metaphysis (Fig. 45.108). There are two types of Triplane fracture with either two or three bone fragments. The three-fragment Triplane injury is common in younger patients with less mature growth plates and is the more likely pattern to require operative reduction

Paediatric Ankle Fractures

fracture fragment can vary considerably in size and displacement great than 5 mm usually requires operative reduction.

Tillaux Fracture

Talar Dome Fractures

A Tillaux fracture describes a fracture of the distal tibia resulting from external rotation. It is an avulsion injury by the anterior inferior tibiofibular ligament of the anterolateral aspect of the tibial physis, an area of relative immaturity and delayed fusion of the growth plate. This is the equivalent of a Salter-Harris (S-H) type 3 injury (Fig. 45.107). The

Talar dome osteochondral injury more commonly occurs following inversion and is more frequently associated with lateral ligament injury. Lateral dome injuries are more common in men and are usually thin and linear, whereas the medial talar dome injuries are deeper and hemispherical in shape (Fig. 45.109). The mortice view facilitates

1176

SECTION C  The Musculoskeletal System

1

2

4

3

2

2 1

4

3

Fig. 45.97  Supination-Lateral Rotation (SLR). Diagrammatic representation of the order of structural failure with added rotation.

Fig. 45.99  Supination-Lateral Rotation (SLR) Injury Stage 3. Oblique fibular fracture defines this as an SLR injury. The posterior malleolar fragment indicates a minimum stage 3 injury.

visualisation of these fractures and displacement of fragments into the gutters. Occult injuries and the stability of fragments is best evaluated by MRI, whereas CT can locate displaced fragments pre-operatively. Failure to identify these injuries can lead to early osteodegenerative change.

Calcaneal Fractures

Fig. 45.98  Supination-Lateral Rotation (SLR) Injury Stage 2. Oblique fibular fracture, high posterior to low anterior, frequently only visible on the lateral and typical of supination-lateral rotation injury.

The calcaneus is the most commonly fractured tarsal bone, caused by axial loading, usually resulting from a fall from a height on to the heels. Fractures are bilateral in 10% of cases; 25% are extra-articular but the majority have articular involvement. Fracture of the lumbar spine is encountered in approximately 10% of cases. There are two types of fracture. The tongue type (Fig. 45.110) involves a fracture from the posterior calcaneal apophyseal cortex extending forwards to involve the posterior facet of the subtalar joint. The second variety is a vertical oblique fracture from the superior to the inferior calcaneal margin, resulting from a cleaving action of the lateral process of the talus under compressive force (Fig. 45.111). Both fracture types may occur in the same patient. It is important to determine whether the fracture extends to the articular surface of the subtalar joint owing to the fact that articular involvement indicates a more severe injury and the degree of displacement or compression defines the need for reconstruction. The lateral view of the ankle provides a satisfactory view of the calcaneus but an AP view of the ankle does not provide an optimal

2

2

3

3

2

1 Fig. 45.100  Pronation-Abduction (PA) injury (Simple Eversion): Diagrammatic representation of the order of structural failure in PA injuries.

2 3

1 4

3 3 Fig. 45.101  Pronation-Abduction Injury Stage 3. Transverse (tractional) fracture medial malleolus, widened syndesmosis (ruptured tibio-fibular ligaments) and distal fibular fracture above joint line with lateral angulation. 2

4

1

Fig. 45.103  Pronation-Lateral Rotation Injury. Diagrammatic representation of the order of structural failure.

Fig. 45.102  Pronation-Abduction Injury Stage 3. Fibular fracture above the joint line with lateral displacement but intact medial malleolus. The deltoid ligament should be injured according to the order of failure. This is supported by the widened medial joint space.

1178

SECTION C  The Musculoskeletal System

Fig. 45.104  Pronation-Lateral Rotation (PLR) Injury Stage 4. Isolated posterior malleolar fragment. In the absence of a low fibular fracture, consequently this is a PLR 4 injury and medial pillar, tibio-fibular ligament and associated high fibular fracture will all be present on other views.

A

Fig. 45.105  Pronation-Lateral rotation injury (Stage 3)/Maissoneuve injury: High spiral fibula fracture.

B

Fig. 45.106  (A) Pilon fracture. Comminuted distal tibia fracture, consequence of axial loading. (B) Pilon fracture (post-operative). All fragments are fixed to the lateral tibial metaphysis and plafond, which usually remains bound to the fibula.

CHAPTER 45  Appendicular and Pelvic Trauma

A

1179

B

Fig. 45.107  (A) Tillaux fracture. Fracture of the anterolateral aspect of the tibial epiphysis (arrowed). (B) Tillaux fracture. CT demonstrating typical positioning of the Tillaux fracture fragment.

A

B

C

Fig. 45.108  (A to C) Triplane fracture. Anteroposterior and lateral views demonstrating the typical sagittal and coronal fracture plains respectively (arrowed). The MRI scan demonstrates the connection of the sagittal and coronal fracture plains through the disrupted growth plate (arrowed).

second view. If calcaneal injury is suspected, a further axial projection radiograph taken at right angles to the lateral is obtained. CT is now invariably used to assess the severity of injury and aid surgical planning; intra-articular fractures are commonly complex. Several radiographic signs are of great diagnostic importance. Bohler angle is constructed by drawing two lines through the three most superior points on the calcaneus. The first is from the superior aspect of the posterior calcaneus to the highest midpoint, which is the posterior aspect of the articular facet, and the second line from this point passes anteriorly and is drawn to the superior tip of the anterior process. These two lines should normally subtend an angle of 30 degrees (Fig. 45. 112).

Intra-articular fractures associated with a reduced angle are more likely to require surgical reconstruction of vertical height. An abnormal area of sclerosis caused by trabecular condensation signifies an impacted fracture, whilst loss of the normal congruity of the articulation between the talus and calcaneus at the posterior subtalar facets suggests fracturesubluxation. Finally, a fracture of the anterior spine of the Os calcis is a subtle observation on the lateral projection but, if missed, can progress to non-union and degenerative change at Chopart articulation (Fig. 45.113). This fragment is angular and often fragmented and should be distinguished from an ununited secondary ossification centre, the Os calcaneus secondarius, which is more rounded and smooth.

1180

SECTION C  The Musculoskeletal System

Fig. 45.109  Osteochondral Defect (OCD) Talar Dome. Typical appearance of an OCD in the medial aspect of the talar dome. Fig. 45.111  Vertical fracture line is seen running from superior to inferior, posterior to the subarticular joint. This is therefore an extra-articular injury.

30°

A

Fig. 45.110  Tongue-type fracture of the calcaneus with V-shaped fracture line extending from the posterior aspect to the body, centrally. There is loss of articular congruity of the subtalar joint. B

Talar Fractures Talar injuries comprise fractures (50%), fracture-dislocations (25%), dislocation without fracture (20%) and compression fractures (5%). Fractures of the talus most often involve the talar neck, but they are uncommon and result from severe force. Fractures of the neck are caused by a force driving up into the sole of the foot. The degree of displacement directly relates to the risk of the retrograde blood supply to the talar body being disrupted, leading to avascular necrosis. A more common injury is an avulsion by the joint capsule of a small flake of bone from the dorsum of the talar head, which is usually of little clinical significance. Lateral process fractures result from shearing forces in eversion, which, importantly, extends into the posterior subtalar articulation, can be overlooked and often fails to unite, and leads to subtalar osteoarthritis (Fig. 45.114).

Fig. 45.112  (A) Line drawing showing how Bohler angle is constructed and measured. (B) The calcaneus is abnormal with flattening of Bohler angle as the highest point in the middle is depressed so that joining the three points results in an almost straight line. A posterior fracture is evident.

FOOT INJURIES Inversion injuries of the ankle may also result in a fracture of the forefoot. The peroneus brevis muscle inserts onto the base of the fifth metatarsal and is prone to avulse its insertion during inversion or forced muscle contraction. The fracture is well demonstrated on radiographs of the foot and is the reason why all lateral ankle x-rays should encompass the base of the fifth metatarsal. This injury should be distinguished from

CHAPTER 45  Appendicular and Pelvic Trauma

1181

Fig. 45.113  Fracture Anterior Spine Os Calcis. Persistent midfoot pain post injury. Sagittal CT reconstruction demonstrates angular fracture fragment (arrowed).

Fig. 45.115  Fifth Metatarsal Base Fracture. Fracture line traversing the metatarsal base following an inversion injury (arrowed).

Fig. 45.114  Lateral Talar Process Fracture. Comminuted fracture of the lateral talar process (arrowed), likely to extend into the subtalar joint. CT recommended to assess articular involvement and displacement.

a more distal fracture of the fourth and fifth metatarsals, referred to as a Jones fracture, which has a propensity to heal poorly. This injury is also caused by inversion of the plantar flexed foot (Fig. 45.115). An unfused apophysis at the base of the fifth metatarsal in the skeletally immature should not be mistaken for a fracture. The apophysis is separated from the metatarsal by a lucent growth plate, which has a longitudinal orientation; conversely, fractures orientate transversely (Fig. 45.116).

Lisfranc Injury A Lisfranc injury refers to fractures and/or dislocations involving the tarsometatarsal articulation. When there is an associated fracture or dislocation is severe, the abnormality is readily identified. Most

Fig. 45.116  Fifth Metatarsal Base Unfused Apophysis (Normal). Lucency running longitudinal to the expected line of metatarsal base fracture.

1182

SECTION C  The Musculoskeletal System

Fig. 45.117  Lisfranc Injury. Failure of longitudinal alignment of the medial aspects of the 2nd metatarsal base and intermediate cuneiform, indicative of lateral displacement of the metatarsal.

Fig. 45.118  Lisfranc Injury. CT in a patient with persistent mid-foot symptoms and normal radiographs, demonstrating an irregular fragment at the tarsometatarsal articulation indicative of a Lisfranc injury without displacement.

commonly, there is a fracture of the base of the second metatarsal with displacement of the second to fifth metatarsals laterally. Subtle injuries may be identified by observing loss of the normal alignment of the second and third metatarsal bases with their respective cuneiform bones (Fig. 45.117). Normally, a line drawn down the medial cortex of the second metatarsal will align with the medial cortical line of the intermediate cuneiform on the AP view. On the oblique view, a line along the medial cortex of the third metatarsal should align with the medial cortex of the lateral cuneiform bones. Some injuries are occult on all radiographs and when there is strong index of suspicion clinically, standing weight-bearing views can accentuate displacement or CT may reveal subtle misalignments or small avulsion fractures (Fig. 45.118).

Stress Fracture Stress fractures are caused by repeated low-impact trauma. They are common in the foot, with a predilection for the shaft of the second metatarsal. Stress fractures are often seen in military recruits from

SUMMARY BOX: Lower Limb • A line dropped perpendicular from the lateral femoral condyle on an anteroposterior view should meet the margin of the adjacent tibial plateau with, at most, 5 mm of the plateau lying lateral to this. • Small bone fragments around the knee usually signify a significant injury (small flakes great shakes!) • There are two types of calcaneal fractures; the tongue type involves a fracture from the posterior calcaneal apophyseal cortex extending forwards to involve the posterior facet of the subtalar joint and a vertical oblique fracture from the superior to the inferior calcaneal margin. • Subtle tarsometatarsal dislocations may be identified by observing loss of normal alignment of the second and third metatarsals with their respective cuneiform bones.

Fig. 45.119  Stress Fracture Third Metatarsal. Periosteal new bone around the middle third of the third metatarsal.

prolonged marching, hence the name of March fracture; however, these are now increasingly recognised in endurance athletes of all levels. Radiographically, they are infrequently identified as a typical lucent line but, more often, as a periosteal reaction caused by increased osteoblastic activity (Fig. 45.119). Persistent pain should prompt a repeat examination after 10 to 14 days to establish a definitive diagnosis. MRI should be reserved at an early stage for professional athletes where a definitive early diagnosis is essential and at later stages for symptoms unresponsive to conservative measures.

FURTHER READING Badillo, K., Pacheco, J.A., Padua, S.O., et al., 2011. Multidetector CT evaluation of calcaneal fractures. Radiographics 31 (1), 81–92.

CHAPTER 45  Appendicular and Pelvic Trauma Berquist, T.H., 1991. Imaging of Orthopaedic Trauma, second ed. Raven Press. Davies, A.M., Whitehouse, R.W., Jenkins, J.P.R., 2003. Imaging of the Foot and Ankle. Springer. Feldman, Frieda, Staron, Ronald B., 2004. MRI of seemingly isolated greater trochanteric fractures. AJR Am. J. Roentgenol. 183, 323– 329. Gottsegen, C.J., Eyer, B.A., White, E.A., et al., 2008. Avulsion fractures of the knee: imaging findings and clinical significance. Radiographics 28 (6), 1755–1770. Greenspan, A., 2011. Orthopaedic Imaging: A Practical Approach, fifth ed. Wolters Kluwer/Lippincott Williams & Wilkins. Iyer, R.S., Thapa, M.M., Khanna, P.C., et al., 2012. Paediatric bone imaging: imaging elbow trauma in children? A review of acute and chronic injuries. AJR Am. J. Roentgenol. 2012, 1053–1068. Khurana, B., Sheehan, S.E., Sodickson, A.D., et al., 2014. Pelvic ring fractures: what the orthopedic surgeon wants to know. Radiographics 34 (5), 1317–1333. Macmahon, P.J., Dheer, S., Raikin, S.M., et al., 2009. MRI of injuries to the first interosseous cuneometatarsal (Lisfranc) ligament. Skeletal Radiol. 38 (3), 255–260.

1183

Marshall, R.A., Mandell, J.C., Weaver, M.J., et al., 2018. Imaging features and management of stress, atypical, and pathologic fractures. Radiographics 38 (7), 2173–2192. Melenevsky, Y., Mackey, R.A., Abrahams, R.B., et al., 2015. Talar fractures and dislocations: a radiologist’s guide to timely diagnosis and classification. Radiographics 35 (3), 765–779. Porrino, J.A., Kohl, C.A., Taljanovic, M., et al., 2010. Diagnosis of proximal femoral insufficiency fractures in patients receiving bisphosphonate therapy. AJR Am. J. Roentgenol. 194 (4), 1061–1064. Raby, N., Berman, L., De Lacey, G., 2014. Accident and Emergency Radiology: A Survival Guide, third ed. Elsevier. Sarwark, J., 2010. Essentials of Musculoskeletal Care, fourth ed. American Academy of Orthopaedic Surgeons. Scheinfeld, et al., 2015. Acetabular fractures: what radiologists should know and how 3D CT can aid classification. Radiographics 35 (2). Solomon, S., Warwick, D., Nayagam, S., 2010. Apley’s System of Orthopaedics and Fractures, ninth ed. Hodder and Arnold.

46  Bone, Joint and Spinal Infections Jaspreet Singh, Radhesh Lalam

CHAPTER OUTLINE Introduction, 1184 Epidemiology, 1184 Classification, 1185 Paediatric Musculoskeletal Infections, 1185 Adult Musculoskeletal Infections, 1189 Diabetic Foot, 1201

INTRODUCTION Infection of the musculoskeletal system is encountered in everyday clinical practice and the scope of this chapter includes both osteomyelitis and soft-tissue infection. Osteomyelitis is defined as infection of the bone marrow and adjacent osseous structures, with or without extension into the soft tissues. It is generally categorised as acute, subacute or chronic based on the clinical course and histopathological findings. Acute osteomyelitis presents with symptoms typically within 2 weeks of infection and is associated with acute inflammatory changes in the bone marrow. In chronic osteomyelitis, there is presence of necrotic bone and symptoms may be delayed until 6 weeks after onset of infection. The route of infection is usually haematogenous spread but can also occur due to direct inoculation or spread from contiguous soft tissue. Infection in the soft tissues can involve differing anatomical planes such as skin and subcutaneous tissues (cellulitis), fascia (fascitis), muscles (myositis or pyomyositis) or joints (septic arthritis). Abscesses can result from infection either in the bone or the soft tissues. One of the greatest challenges of osteomyelitis is to be able to diagnosis early stages of infection in order to provide prompt treatment and prevent long-term debilitating sequelae. Imaging plays a key role in the early diagnosis of musculoskeletal infection and a variety of techniques, including radiography, computed tomography (CT), scintigraphy and magnetic resonance imaging (MRI), are often used to aid in diagnosis. Conventional radiography is usually the initial investigation in the setting of suspected musculoskeletal infection. Radiographs can be normal at the outset and it can take up to 2 weeks for changes to become apparent as periosteal reaction takes time to become radiographically apparent and approximately 50% of the trabecular bone needs to be destroyed for lysis to be visible on the plain radiograph. Ultrasound (US) is useful in soft-tissue infection and helps in detecting joint effusions and fluid collections and also demonstrating subperiosteal abscesses in paediatric patients. It is also useful in guiding aspiration or biopsy in order to establish diagnosis. CT is useful in chronic osteomyelitis to identify the extent of bone destruction and also to demonstrate any sequestra. CT can also identify gas in the soft tissues in certain aggressive

1184

Septic Arthritis, 1203 Musculoskeletal Tuberculosis, 1205 Unusual Musculoskeletal Infections, 1207 Differential Diagnosis, 1209 Management, 1210 Spinal Infection, 1212

infections. Scintigraphy is extremely sensitive for the diagnosis of infection (although not perfectly specific) but anatomical detail is limited. It remains useful in a diagnosis of infection in the presence of an implant. MRI has a number of advantages and is the best investigation to detect early osteomyelitis compared with other imaging techniques. It is extremely sensitive and can detect osteomyelitis within 3 to 5 days of disease onset. Although not specific for infection, it has a high negative predictive value in excluding infection. Osteomyelitis has a variable imaging appearance and often mimics other bone diseases. It is important that the radiologist is familiar with the clinical spectrum of presentations, appropriate diagnostic tests and the need to perform these without delay to enable prompt management.

EPIDEMIOLOGY Staphylococcus aureus is the commonest organism causing bone and joint infections in any age group, accounting for up to 80% of cases of osteomyelitis. Gram-negative organisms, including Pseudomonas and Enterobacter, are responsible for most of the remaining 20% of cases. Acute infections of prosthetic implants are usually caused by S. aureus. Coagulase-negative staphylococci such as Staphylococcus epidermidis account for most chronic osteomyelitis associated with orthopaedic implants and account for approximately 90% of pin tract infections. Bacteria adhere to bone matrix and orthopaedic implants via receptors to fibronectin and other receptor proteins. They form a slimy coat and elude the host defence mechanisms by hiding intracellularly and by developing a slow metabolic rate. Polymicrobial infection occurs in most cases of osteomyelitis affecting the diabetic foot, comprising of mixed gram-positive and gram-negative bacteria. Wound swabs are often inaccurate and dominated by contaminants. Cultures from bone biopsy or operative specimens are more reliable in planning appropriate antibiotic treatment. Brodie abscess, a true small, focal intraosseous collection of pus, is an uncommon manifestation of bone infection, except in East Africa, where reportedly it is a common occurrence. S. aureus is again the predominant organism cultured in these patients.

CHAPTER 46  Bone, Joint and Spinal Infections Vertebral osteomyelitis usually results from haematogenous seeding or by direct inoculation at the time of spinal surgery. S. aureus is the commonest microorganism amongst others that include methicillinresistant S. aureus (MRSA), streptococci and Escherichia coli. As with other orthopaedic implants, coagulase-negative staphylococcal infections are seen after usage of fixation devices. In children, other than staphylococci, Haemophilus influenzae type b, Streptococcus pneumoniae and Streptococcus pyogenes also occur. One of the significant changes in the epidemiology of bone and joint infections is the increasing incidence of MRSA and the emergence of multidrugresistant (MDR) organisms. Osteomyelitis complicating war injuries are predominantly caused by MDR organisms and include MRSA, Enterobacteriaceae and Acinetobacter. There is commonly a history of previous surgical procedures and these conditions usually need more aggressive management. Fungal infections of the bone and joints are uncommon and occur predominantly in the immunosuppressed patient. Such infection may resemble tuberculous disease and present as chronic multifocal osteomyelitis or polyarthritis.

CLASSIFICATION Osteomyelitis can be classified by the type of infection or, more commonly, by the duration since onset of the illness. Acute osteomyelitis is defined as infection diagnosed within 2 weeks of onset of symptoms whilst subacute osteomyelitis is diagnosed if symptoms exceed 2 weeks’ duration. If the infection is diagnosed months after the onset of symptoms, then it is defined as chronic osteomyelitis. There are, however, other classifications in the literature, depending on the anatomical areas of bony involvement and on pathogenesis.

PAEDIATRIC MUSCULOSKELETAL INFECTIONS Paediatric, unlike adult, bone infections more commonly occur in healthy bones, without pre-existing trauma. The usual route of infection is haematogenous, via the arterial blood supply. Acute haematogenous osteomyelitis (AHO) is the most common form of bone infection in children. About 50% of cases occur in children less than 5 years of age and it is twice as common in boys than in girls. Immunocompromised children and children with underlying haematological disorders such as sickle cell disease are more prone to these infections. Infections are commonest in long bones such as the femur, tibia or humerus, with most cases limited to single bone involvement. Less than 10% of cases involve two or more locations, the corollary of which is that multifocal involvement does not exclude acute osteomyelitis (the same is true for septic arthritis).

Pathophysiology Haematogenous spread is the most common route to bones and joints, though implantation at trauma or surgery and spread from contiguous infection may also occur. Infection usually starts in the metaphysis due to its rich blood supply. From here the infection may spread to the bone cortex and penetrate the loosely attached periosteum, eliciting a periosteal reaction. It may perforate the periosteum and spread to adjacent muscles and soft tissues, forming abscesses. Less commonly, the infection may spread to the epiphyseal plates and joints, causing septic arthritis. In neonates, the metaphyseal capillaries form a connection with the epiphyseal plate, thus increasing the chance of joint infection. In later infancy, these vessels atrophy and there is thickening of the cortex, thus reducing the risk of growth plate involvement and septic arthritis. S. aureus accounts for approximately 60%–90% of childhood AHO, followed by group A β-haemolytic streptococci (10%). Other causes

1185

include H. influenzae and S. pneumoniae. Salmonella infections were historically common in patients with sickle cell disease, but this seemed to reflect the organism prevalence in the population. Pseudomonas infections can be seen after puncture wounds of the feet. In neonates, again, S. aureus infections predominate, but group B streptococcus and E. coli infections are also common.

Clinical Features A few well-recognised clinical presentations of bone and joint infections are seen in children. Acute osteomyelitis, more common in boys, usually occurs between ages 1 and 10 years. The lower limbs are usually involved. Children usually present with pain and reluctance to use the affected limb. The cardinal signs of acute inflammation, including pain, redness and swelling, may not be present initially, may develop later and may be associated with systemic illness. Often pain is localised and prompt diagnosis and treatment offer a good prognosis. Chronic osteomyelitis has a more insidious onset over a few weeks and is more difficult to diagnose. There is usually minimal loss of function. Systemic signs are usually absent, but local tenderness may be present. Radiographs are usually helpful as bone changes are usually present by the time medical attention is sought. Chronic osteomyelitis is further discussed under adult infections. Septic arthritis is also twice as common in boys, with a peak incidence below 3 years of age. Symptoms include pain and swelling around the joint with reluctance to move the limb. Pseudo-paralysis and painful passive movements may be present. There are systemic symptoms and there is rapid progression of symptoms. Diagnosis is more difficult in neonates. The joint is warm and ultrasound can demonstrate a joint effusion. The lower limbs are usually involved, particularly the hip or knee. Neonatal osteomyelitis and septic arthritis are probably different manifestations of the same condition: 75% of the children are not severely ill and present as failure to thrive. Prognosis is not as good as there is usually a delay in diagnosis. Disseminated staphylococcal disease presents as a rapidly progressive severe life-threatening illness with virulent bacteraemia and multiorgan involvement, but fortunately is rare.

Investigations and Management Thorough clinical examination and laboratory investigations are extremely important in diagnosis, before imaging evaluation is performed. Conventional radiography is still one of the most important investigations for diagnosis. The other imaging investigations, ultrasound, CT, MRI and scintigraphic studies are also used. The risk/benefits and costs of these tests, their radiation exposure, requirements for sedation or anaesthesia need to be assessed carefully as investigations are planned.

Plain Radiographs In acute osteomyelitis, plain radiographs are extremely useful to exclude other lesions such as fractures, Perthes disease and slipped femoral epiphysis. In septic arthritis the joint space is initially expanded with fluid and may give rise to some asymmetry in the position of the epiphysis, but ultrasound is more reliable to demonstrate fluid in the joint. In subacute osteomyelitis, periosteal reaction, new bone formation and occasionally lucent lesions (Brodie abscess) in the metaphyseal region can be present. Chronic osteomyelitis demonstrates bone sclerosis, bone destruction and periosteal new bone formation.

Ultrasound Ultrasound (US) has an important role in demonstrating increased joint fluid early in acute septic arthritis and is also used to guide

1186

SECTION C  The Musculoskeletal System

aspiration of the joint effusion for diagnostic and therapeutic purposes (Fig. 46.1). In septic arthritis the cellular debris-laden effusion may be echoic, rendering it less conspicuous. Demonstrating movement of this debris by varying the ultrasound probe pressure during the examination may reduce false-negative interpretation. Subperiosteal abscess can be demonstrated as hypoechoic fluid along the bone surface in acute and subacute disease. Soft-tissue abscesses can also be seen as hypoechoic fluid with thick walls that may show increased colour Doppler flow. Direct aspiration under ultrasound guidance and culture and sensitivity of the aspirate is extremely helpful to confirm diagnosis and for further treatment with the appropriate antibiotics. Soft-tissue abnormalities can also be demonstrated with ultrasound, though MRI is more reliable.

A

B

Computed Tomography This investigation is less useful in a paediatric population due to its radiation dose and lack of specific advantages, and MRI is preferred.

Magnetic Resonance Imaging MRI is the investigation of choice for the diagnosis of acute, subacute and chronic osteomyelitis with high sensitivity and specificity. The characteristic signs of acute osteomyelitis in children include bone marrow oedema, which is seen as low T1 signal in the bone marrow, along with high signal on T2 and short tau inversion recovery (STIR) images (Fig. 46.2). Abscesses are seen both early and late in disease, as well-defined low-signal collections on T1, high signal similar to fluid on T2 images, with enhancing walls following intravenous enhancement. This may be seen in the bone, adjacent soft tissues or subperiosteal location due to the loose attachment of the periosteum to the underlying bone in children. Septic arthritis is diagnosed by the presence of a joint effusion, abnormal bone marrow signal localised to either side of the joint or synovial thickening which also shows post-contrast enhancement. Physeal involvement is characterised by low T1 and hyperintense T2 signal along the growth plate associated with widening of the growth plate and enhancement on post-contrast imaging (Fig. 46.3). Later in the course of the infection, chondrolysis occurs. Fat-saturated T1 weighted images with intravenous contrast medium can increase confidence in the diagnosis of osteomyelitis and also help diagnose complications such as septic arthritis, physeal involvement, intraosseous, subperiosteal and soft-tissue abscesses. Intravenous gadolinium is most useful to identify non-enhancing abscess collections within a background of soft-tissue and bone marrow oedema and to aid differentiation between granulation tissue and true abscess. It is generally not useful if unenhanced images show no evidence of soft-tissue or bone marrow oedema. The merits of administering intravenous contrast agents should be weighed against its injudicious use in children. The rim sign is seen in chronic osteomyelitis as a low-signal area of fibrosis surrounding an area of active infection. Active infection has high signal on T2, whereas the rim has low signal on all sequences. There may also be thickening and remodelling of cortex due to chronic infection. Periosteal reaction may not be present at this stage. Bone abscess is seen as fluid signal surrounded by a thick wall that has low signal on T1 and significant enhancement. Abscesses may extend through the cortex into the surrounding soft tissues with the formation of sinuses. There is high T2 signal returned from surrounding soft tissues, with loss of normal fat signal on T1 and T2 images (Fig. 46.4). Post-surgical or interventional procedures cause bone signal changes that can persist for up to 12 months. Hence, whenever possible, MRI should be obtained before any interventions are performed, to avoid the need for additional procedures. Whole-body MRI may be useful to demonstrate multiple sites of involvement.

Nuclear Medicine

C Fig. 46.1  Septic Arthritis in a Child. (A) Ultrasound of the right wrist demonstrates presence of echogenic fluid and synovial thickening with increased colour Doppler flow. (B and C) Sagittal T1 weighted pre- and post-gadolinium-enhanced images demonstrate diffuse thickening and enhancement of the synovium at the dorsal and volar aspect (arrows) of the radiocarpal articulation in keeping with septic arthritis.

In early stages, three-phase technetium (99mTc)-methylene diphosphonate (MDP) skeletal scintigraphy may be useful for demonstrating increased radioactivity in the affected bone and also for demonstrating multifocal disease. Increased activity may be seen in dynamic perfusion, early blood pool and delayed images. However, such uptake is non-specific and may be seen in other conditions, including trauma and tumours.

Chronic Recurrent Multifocal Osteomyelitis This uncommon form of non-bacterial inflammatory osteomyelitis occurs in children with a female predominance. The aetiology is not known, but may be related to autoimmune disorders due to its association

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

1187

Fig. 46.2  Osteomyelitis of the Tibia With Abscess Formation. (A) AP radiograph of the left knee shows subtle lucency in the proximal tibial metaphysis. Plain radiographs can be normal in early osteomyelitis. (B and C) Coronal T1 and coronal short tau inversion recovery images demonstrate diffuse bone marrow oedema of the proximal tibial metaphysis with extension into the epiphysis and associated periosteal and soft-tissue oedema. (D) Axial T1 weighted image after gadolinium administration demonstrates non-enhancing intraosseous abscess formation and also abscess in the posterior soft tissues (arrow).

with inflammatory bowel disease and psoriasis-like skin conditions. A genetic aetiology is also postulated due to an association with mutation of the LPIN2 gene. It is characterised by multifocal non-pyogenic inflammatory bone lesions and a clinical course of exacerbations and remissions. It is more common in the lower limbs and may be symmetrical. Metaphyseal lesions occur in approximately 75% of cases. Symptoms are vague and it may present as monoarthritis or polyarthritis. Other less frequently involved sites include the spine, clavicle, pelvis and mandible. When a single site of disease is present, the term chronic non-bacterial osteomyelitis is used. There may be mild elevation of inflammatory markers but laboratory diagnosis is generally unhelpful. Microbiology cultures are negative. Plain radiographs may reveal osteolytic lesions with surrounding sclerosis. Whole-body MRI can be helpful by identifying multiple sites of the disease, some of which may be asymptomatic. Typically, MRI reveals periostitis, bone marrow oedema and signs of transphysitis. Many lesions may heal spontaneously, though symptoms may persist for several years. Diagnosis is made by exclusion and from clinical history and typical radiological findings affecting multiple sites (Fig. 46.5) and recurrent negative cultures and bone biopsy. Majeed syndrome is an autosomal recessive disorder comprising a triad of chronic recurrent multifocal osteomyelitis (CRMO), congenital dyserythropoietic anaemia and inflammatory dermatosis (Summary Box 46.1).

Synovitis, Acne, Pustulosis, Hyperostosis and Osteitis Syndrome Like CRMO, synovitis, acne, pustulosis, hyperostosis and osteitis (SAPHO) is a complex inflammatory disorder with multifocal osteitis the predominant feature. It tends to affect 30 to 50 year olds with a female predominance. It is characterised by the presence of SAPHO. Skin manifestations and osteoarticular involvement commonly occur. Skin manifestation includes palmoplantar pustulosis and acne. There is bilateral symmetric pain and swelling of the bones and joints. The most common site involved is the upper anterior chest wall involved in 60%–95% of cases. This includes sternoclavicular joints, medial ends of the clavicle, costosternal junctions and manubriosternal joint. Spinal involvement is also common and is seen in 50% of the cases. It is often centred in the thoracic spine with contiguous vertebral involvement. The other sites of involvement include sacroiliac joints, long bones and mandible. The imaging features of SAPHO relate to synovitis, hyperostosis and osteitis. Whole-body MRI is useful to assess for asymptomatic sites of involvement and demonstrate the multifocal pattern of involvement. Biopsy demonstrates acute and chronic inflammatory cells. Relationship to Propionibacterium acnes has been postulated but is debatable. Treatment is by non-steroidal anti-inflammatory drugs (NSAIDs) in the first instance, but severe cases may need immunosuppressive treatment.

1188

SECTION C  The Musculoskeletal System

A

B

C

D

Fig. 46.3  Osteomyelitis With Epiphyseal Abscess Formation in a Child. (A) AP radiograph of the knee demonstrates a lucency in the medial aspect of the femoral epiphysis. (B) Sagittal T1 weighted image showing low T1 signal within the abscess surrounded by a rim of high T1 signal intensity (penumbra sign) (arrow). There is surrounding low T1 signal change within the epiphysis in keeping with bone marrow oedema. (C) Sagittal proton density fat saturated (PDFS) and (D) coronal PDFS images demonstrating intense high signal change within the abscess cavity with surrounding bone marrow oedema.

SUMMARY BOX 46.1: Synovitis, Acne, Pustulosis, Hyperostosis and Osteitis Syndrome/Chronic Recurrent Multifocal Osteomyelitis • Both are characterised by multifocal non-pyogenic inflammatory bone lesions and a clinical course of exacerbations and remissions. • Chronic recurrent multifocal osteomyelitis (CRMO) occurs predominantly in children and adolescents. • In CRMO, the most common sites of disease are the metaphyses of long bones, accounting for approximately 75%. • In synovitis, acne, pustulosis, hyperostosis and osteitis (SAPHO), the most common site involved is the upper anterior chest wall, which includes sternoclavicular joints, medial ends of the clavicle, costosternal junctions and manubriosternal joint. Spinal involvement is also common, often centred in the thoracic spine with contiguous vertebral involvement. • Identification of additional sites of asymptomatic disease involvement can aid in the diagnosis, especially when disease is present at typical locations. Therefore whole-body imaging should be performed when the diagnosis is considered.

Sclerosing Osteomyelitis of Garré This is a chronic form of osteomyelitis, usually occurring in children, and commonly affects the mandible. This is considered as a form of CRMO affecting the mandible. Patients present with pain and hard swelling of the mandible. Initial findings include lytic lesions of the mandible associated with sclerosis. With disease progression there is non-suppurative ossifying periostitis with subperiosteal new bone formation and sclerosis. The diagnosis can be made presumptively by radiology, and biopsy reveals features of chronic osteomyelitis, with cultures usually negative.

Necrotising Fasciitis This is a rapidly progressive life-threatening infection occurring in young children and more common in boys. This condition also occurs in adults. Risk factors include immunocompromise, diabetes, intravenous drugs or alcohol abuse and patients with peripheral vascular disease. The condition is polymicrobial in origin, and includes gram-positive aerobes like group A β-haemolytic streptococcus, S. aureus, gram-negative organisms like E. coli, Pseudomonas aeruginosa and anaerobes like Bacteroides. Clostridium infection leads to a gas-producing necrotic infection, gas gangrene, which is rapidly progressive and leads to systemic toxicity and shock. Characteristically, the gas is intramuscular in gas gangrene, giving rise to gas loculi that are elongated and aligned with

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

1189

Fig. 46.4  Osteomyelitis of the Distal Femur in a Child. (A) Plain radiograph demonstrating abnormal texture of the medullary bone with periosteal reaction. (B) Coronal T1 and (C) coronal short tau inversion recovery (STIR) images show low T1 signal intensity within the bone marrow which is hyperintense on the STIR sequences. There is associated periosteal and soft-tissue oedema. Presence of intramedullary fat globules is seen as areas of high signal intensity on T1 weighted images (arrow), which is a pathognomonic sign for acute osteomyelitis. (D) Sagittal T1 weighted image post-gadolinium administration demonstrates an abscess formation at the posterior aspect of the distal femur with peripheral rim enhancement (arrow).

the muscle fascicles, whilst more globular gas loculi are seen in fasciitis, when a gas-forming organism is present. The infection involves deep subcutaneous tissues and is associated with a high mortality. Infection causes thrombosis of small blood vessels, leading to necrosis and rapidly involving several fascial planes. The condition has been associated with trauma, burns, eczema and varicella infections. The extensive soft-tissue damage leads to multiorgan failure and shock. Ultrasound can be useful in diagnosing the abnormal muscle echo texture, but MRI is more reliable and is needed to assess the extent of tissue damage and demonstrate the spread of infection along the facial planes. CT may also be useful, but MRI is preferred due to better softtissue contrast and lack of radiation. Emergency surgical debridement is needed to stop the progress of this fulminating condition; relevant imaging must therefore be performed immediately.

ADULT MUSCULOSKELETAL INFECTIONS Spontaneous musculoskeletal infections in adults are less common than in children and are usually due to trauma, previous surgery or underlying immunodeficiency disorders. Trauma and open wounds can result in seeding of bone with microorganisms and to development of osteomyelitis. Whilst haematogenous infection is common in childhood

SUMMARY BOX 46.2: Osteomyelitis • Ultrasound is the preferred technique to investigate children with suspicion of septic arthritis. Ultrasound-guided aspiration is safe and easy to perform. • Magnetic resonance imaging is the preferred technique for early detection of osteomyelitis. The fat globule sign on T1 weighted images is pathognomonic for acute osteomyelitis, whilst the penumbra sign is pathognomonic for a Brodie abscess in subacute osteomyelitis. • Computed tomography is useful in the evaluation of chronic osteomyelitis and provides information regarding intraosseous cavities which may contain a sequestrum with the surrounding reactive involucrum formation. • Diabetic foot osteomyelitis primarily arises from the contiguous spread of an overlying ulcer. On T1 weighted images, low signal intensity in marrow adjacent to an ulcer or a sinus tract is the primary sign of osteomyelitis.

osteomyelitis, in adults, it is mostly responsible for vertebral osteomyelitis. Apart from trauma, infection can also develop in prosthetic implants after surgery. Underlying conditions like diabetes, vascular insufficiency, decubitus ulcers and sinuses can predispose to development of osteomyelitis in adults. Osteomyelitis is classified by the time since onset, as described earlier (Summary Box 46.2).

1190

SECTION C  The Musculoskeletal System

A

B

C

D

E

Fig. 46.5  Chronic Recurrent Multifocal Osteomyelitis of the Left Clavicle. (A) Conventional radiograph demonstrates features of osteitis and hyperostosis of the right clavicle. Involvement extends laterally from the medial aspect, sparing the lateral end. (B) Coronal oblique reformatted computed tomography demonstrates hyperostosis with sclerosis and areas of intervening lucency. (C and D) Coronal T1 and fat-saturated T2 weighted sequences demonstrate clavicular hypertrophy with intraosseous and periosteal oedema. (E) Coronal short tau inversion recovery weighted image from a whole-body magnetic resonance imaging also demonstrates changes in the right clavicle in keeping with multifocal involvement.

Pathogenesis Healthy adult bone is usually resistant to infection, but when affected can be difficult to treat. The presence of dead bone and implants make it difficult to treat by antimicrobial agents and removal of the debris and the prosthetic implants is necessary to eradicate the infection. The bacteria attach to the bone matrix and orthopaedic implant devices by developing receptors to fibronectin and other structural proteins. They develop a slimy coat and a very slow metabolic rate, hide in intracellular locations and are thus able to elude host defences and antibiotics. The presence of implants also causes cell dysfunction, which decreases the ability of polymorphonuclear cells to phagocytise bacteria. Reactions between the bacteria and the host defences cause release of cytokines and consequent osteolysis. Patients with sickle cell disease are prone to developing enteric bacterial osteomyelitis. In sickle cell disease, there is impaired gut defence due to sickling in the vasculature of gut. This enables the entry of organisms into the bloodstream, and haematogenous spread of infection into the bones. Typically this is Salmonella infection, but this probably reflects the prevalence of Salmonella in countries where sickle cell disease is common.

Clinical Features Clinical features are variable, but acute infections generally present with pain, swelling and redness of the affected area associated with systemic illness. Joint swelling, reduced mobility and features of overlying cellulitis may be present. In chronic infections, discharging sinuses may be present. Treatment may render chronic osteomyelitis inactive, but reactivation may occur, with recurrence of symptoms (pain, swelling, erythema,

fever) and new radiographic features (bone destruction, periosteal reaction). Characteristic radiographic features of chronic osteomyelitis include intraosseous cavities that may contain separated fragments of necrotic bone (a sequestrum), with the surrounding bone becoming thickened and sclerotic (involucrum). The cavity may communicate with the surrounding soft tissue through cloacae in the involucrum, with sinus tracks to soft-tissue abscesses or cutaneous ulcers. Extrusion of sequestra may occur through these sinuses. The chronic ulcer associated with these sinuses can undergo malignant transformation into a squamous cell carcinoma (Fig. 46.6).

Investigations and Management Laboratory investigations may demonstrate an increased white cell count, elevated C-reactive protein (CRP) and ESR (erythrocyte sedimentation rate). In acute infections, blood cultures may be positive. Culture of the pus from discharging sinuses is also useful, but generally has a low yield rate for microorganisms. Table 46.1 summarises the main imaging findings in acute, subacute and chronic osteomyelitis.

Plain Radiographs In acute infections, plain radiographs are useful to exclude other lesions such as fractures or malignancy. Focal abnormalities occur in acute osteomyelitis, usually in the metaphyseal region. These are commonly lytic lesions with a narrow zone of transition, but bone sclerosis can also occur. There may be associated soft-tissue abnormalities. In subacute osteomyelitis, periosteal reaction and cortical thickening also occur. As the disease progresses, bone sclerosis, thickening, resorption and destruction resulting in deformities may occur (Fig. 46.7). Septic arthritis can destroy the joint, resulting in joint fusion and deformities.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

1191

Fig. 46.6  Malignant Transformation in a Chronic Ulcer Related to Chronic Osteomyelitis. (A) Plain radiograph demonstrates erosion and destruction of the posterior aspect of the calcaneum with overlying skin ulceration. (B and C) Sagittal T1 and axial proton density fat-suppressed sequence demonstrating a diffuse area of low T1 and high T2 signal within the calcaneum with ulceration of the overlying skin (arrow) in keeping with acute on chronic osteomyelitis. (D and E) Sagittal T1 pre- and post-gadolinium administration sequence demonstrates an enhancing mass-like abnormality within the soft tissues (arrow) adjacent to the site of chronic ulceration. Biopsy confirmed malignant transformation to a squamous cell carcinoma.

TABLE 46.1  Imaging Findings in Osteomyelitis Acute

Subacute

Chronic

Plain Radiograph

Computed Tomography

Magnetic Resonance Imaging

NM

Minimal findings Soft-tissue swelling may be seen Lucent or sclerotic lesion, periosteal reaction, soft-tissue swelling

Not useful

May show increased uptake, but takes a few days

Bone sclerosis, cortical thickening, sequestrum and cloaca, bone destruction, resorption and deformities

Much better than plain radiographs to demonstrate cloaca and sequestrum, periosteal new bone formation and abscess

Bone marrow oedema can occur as early as 24–48 h, seen as low T1, and high T2 signal Bone marrow changes, cortical abnormalities seen as thickening, bone abscess, periosteal reaction, increased T2 signal in soft tissues, abscess formation. Post-gadolinium T1 weighted sequences outline abscess cavities clearly Better soft-tissue and bone marrow resolution to demonstrate medullary and cortical changes; sequestra and cloaca well demonstrated; useful to outline soft-tissue abscess and sinus tracts

Cortical and marrow abnormalities, including abscess, periosteal reaction, soft-tissue oedema and abscess

PET/CT, Positron emission tomography/computed tomography; WBC, white blood count.

Three-phase bone scintigram, 111In WBC scintigraphy and combined studies are useful, especially to assess multifocal involvement. PET/ CT generally not used in this context, but may be useful in exceptional circumstances Generally useful if there is a problem with diagnosis. Combined WBC and bone marrow scintigram is useful. May highlight multiple sites of involvement

1192

SECTION C  The Musculoskeletal System

A

B

Fig. 46.7  Chronic Osteomyelitis of the Tibia. There is cortical thickening and chronic periosteal new bone formation, forming an involucrum (long black arrows) around an indistinct medullary cavity. There is a cloaca (short black arrow).

Fig. 46.8  Ultrasound Extremities. (A and B) Ultrasound of the wrist demonstrates non-compressible synovial thickening with increased Doppler flow involving the tendon sheaths of the extensor tendons at the dorsal aspect of the wrist. The Doppler flow suggests active inflammation. The signs are non-specific and associated with inflammatory conditions, but are also an important early finding in infective tenosynovitis.

Ultrasound Ultrasound is a simple, non-expensive bedside investigation that can be extremely useful in the acute setting with ease of performing interventions in the same setting. Cellulitis is easily demonstrated as oedema and thickening of the subcutaneous tissues. This creates a cobblestone pattern due to anechoic strands randomly traversing the subcutaneous tissues. Infective bursitis is demonstrated by the presence of excess fluid in the bursa, wall thickening with increased colour Doppler flow due to inflammatory changes in the affected bursa. Prepatellar and olecranon bursae are the most commonly affected. Fluid aspiration for diagnosis by microscopy and culture under ultrasound guidance also offers therapeutic benefit. Tenosynovitis shows thickening of the tendon sheath associated with fluid surrounding the tendon itself. There may be non-compressible thickening of the tendon sheath, which also demonstrates increased colour Doppler flow due to hyperaemia. It is difficult to differentiate inflammatory tenosynovitis from infection, and aspiration and cytology will confirm the diagnosis (Fig. 46.8). In the appropriate clinical setting, septic arthritis is confirmed by the presence of fluid in the affected joint and can usually be readily

demonstrated by ultrasound, though turbid fluid may be echogenic and more difficult to see. Ultrasound is particularly useful in the hip, where guided diagnostic aspiration can be performed. Early diagnosis can avoid serious consequences, especially in children, where an effusion can be the only sign localising infection to that joint. In the hands, wrists and feet joints, diagnostic aspiration under ultrasound reduces the risk of contamination of other compartments. Thickening of the synovium is seen in septic arthritis associated with lack of compressibility and increased colour Doppler flow due to the presence of inflammation, but can also be seen in other inflammatory and non-inflammatory arthritis. Abscesses are usually well defined and show hypoechoic or anechoic fluid within, usually with a thick capsule which shows increased Doppler flow. Subperiosteal abscesses can also be demonstrated by ultrasound before a periosteal reaction is evident radiographically, but osteomyelitis generally needs further cross-sectional imaging. Pyomyositis shows abnormal echogenicity in early stages, but in later stages abscess formation is seen (Fig. 46.9). With prosthesis-related infection, ultrasound can be extremely useful to demonstrate fluid collections and diagnostic aspiration can be performed at the same time.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

1193

Fig. 46.9  Pyomyositis. (A) Ultrasound demonstrates a deep hypoechoic soft-tissue collection consistent with abscess formation. (B and C) Coronal T1 with gadolinium enhancement and coronal short tau inversion recovery sequence demonstrates extensive intramuscular fluid collection with high T2 signal and irregular enhancing walls. (D and E) Axial T1 pre- and post-gadolinium sequence demonstrates intramuscular abscess formation with irregularly enhancing peripheral rim.

Computed Tomography High-resolution, multiplanar reconstruction and wide availability result in CT being commonly used in the diagnosis and assessment of osteomyelitis. The main disadvantages of CT are exposure to ionising radiation and limited soft-tissue contrast. The advantages of CT are that it can demonstrate periosteal reaction, subtle bone erosion, cortical destruction, abscess formation and soft-tissue swelling. CT may also demonstrate thickening of trabeculae and medullary abnormalities. In chronic osteomyelitis, CT is better than MRI for the demonstration of cortical destruction and demonstrating the presence of gas. CT is also superior to MRI in the demonstration of sequestra (Fig. 46.10), involucra and cloacae and can guide therapeutic options. Soft-tissue abnormalities can be seen with CT, but MRI is superior for demonstration. In the spine, CT is much more sensitive in demonstrating trabecular destruction and endplate erosions than conventional radiography. Paravertebral abscesses may be demonstrated clearly with CT. Vertebral disc space narrowing is not reliably detected on axial images and requires sagittal reconstructions. Spinal canal stenosis and associated fractures of bones can also be clearly demonstrated. In the absence of trauma, the presence of fat/fluid levels in the soft tissues around the bone, especially when associated with spongy bone destruction, is an important and specific sign of underlying osteomyelitis. Periprosthetic infection may be demonstrated by CT. Artefacts due to beam hardening, from high-density metallic prosthetic components,

Fig. 46.10  Chronic Osteomyelitis With Sequestrum. Axial computed tomography of the tibia demonstrates a cloaca within the posterior cortex (black arrows) with a small sequestrum within it (white arrow).

1194

SECTION C  The Musculoskeletal System

can be minimised by using an extended CT number scale, a high kVp, acquiring data in thin sections then generating thicker-section maximum intensity projection reformations and the use of iterative image reconstruction software designed to suppress metal artefacts.

Magnetic Resonance Imaging MRI has high accuracy and can be positive as early as 3 to 5 days after the onset of infection. The good soft-tissue contrast, high diagnostic accuracy and wide availability of MRI makes this the investigation of choice in suspected osteomyelitis. T1 and T2 weighted spin-echo (SE) sequences should be obtained in at least two planes: axial and coronal or sagittal. A STIR sequence or fat-suppressed T2 sequence is useful to identify the bone marrow oedema and soft-tissue oedema easily. Usual slice thickness is 3 to 4 mm (Fig. 46.11). Fat-suppressed images can identify ulceration, abscess formation and sinus tracts due to the accentuated fluid signal against a background of suppressed soft-tissue and marrow signals. Gadoliniumenhanced T1 weighted SE images with fat saturation help improve diagnostic confidence, although the overall sensitivity and specificity do not change significantly (Fig. 46.12). Bone marrow oedema is one of the earliest signs of osteomyelitis. In acute osteomyelitis there is increase in intramedullary water due to

A

B

D

oedema, inflammation and ischaemia, resulting in areas returning low T1 signal and increased T2 signal; this is even better appreciated on fat-saturated or STIR images. The marrow oedema is usually ill defined in its early stages. Later on with more localised bone destruction the oedema and signal changes appear well defined. Bone marrow oedema is more extensive in osteomyelitis than in degenerative or infective arthritis. Cortical disruption may lead to the development of periosteal reaction. Cortical disruption is seen as a break in the normal low signal of the cortical bone. Periostitis shows as a thin linear pattern of oedema with enhancement, surrounding the outer cortical margin. Chronic periostitis and periosteal reaction are seen as thickening of low signal of cortical bone in both T1 and T2 weighted images. In subacute osteomyelitis, an intramedullary abscess (Brodie abscess) may be seen. The central fluid component has low-to-intermediate T1 signal and hyperintense T2 signal. The rim of the abscess cavity is often higher signal intensity than the main abscess on unenhanced T1 weighted images. This is referred to as the ‘penumbra sign’, and is useful in discriminating subacute osteomyelitis from other bone lesions (Fig. 46.13). The magnetic resonance (MR) characteristics of a sequestrum are similar to the bone it is derived from. If it is from cortical bone, it has low signal, with a higher signal if derived from cancellous bone. The

C

E

Fig. 46.11  Osteomyelitis Originating in a Bone Infarct. (A–C) Coronal T1, coronal short tau inversion recovery and coronal T1 weighted image with gadolinium enhancement demonstrate a typical bone infarct in the left distal femur. At the cranial aspect of the infarct, the margin is ill defined with a diffuse area of low T1 and high T2 signal intensity, which demonstrates irregular contrast enhancement consistent with an area of osteomyelitis (arrows). (D) Axial proton density fat saturated sequence demonstrates high T2 signal within the medullary cavity with a defect in the posterior cortex (arrow) and extension of the high T2 signal into the posterior soft tissues. (E) Sagittal reconstructed computed tomography image demonstrates the cortical defect posteriorly (arrow).

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

1195

Fig. 46.12  Osteomyelitis of the Zygoma as a Complication of Middle Ear Infection. (A) Axial computed tomography image on bone and soft-tissue windows shows soft-tissue filling the right middle ear associated with bone destruction around the temporomandibular joint, antibiotic-impregnated beads within the bone cavity and an abscess around the arch of the zygoma. (B) Axial T1 weighted image, (C) axial fat-suppressed post-enhanced T1 weighted image, (D) axial T2 weighted image and (E) coronal fat-suppressed enhanced T1 weighted image all demonstrate the perizygomatic abscess, evidence of infection in the right middle ear and artefact from the metal wire in the antibiotic beads.

exudate in the presence of active infection tends to show a low T1 signal and high T2 signal, which may show enhancement, which is seen to surround the central sequestrum. The involucrum has the signal of normal living bone, but is commonly thickened and sclerotic and may show oedema. A cloaca is seen as a high signal defect in cortical bone at the edge of a cavity. Collections of pus may be seen extending from the cloaca to the subcutaneous tissue (Fig. 46.14). Soft-tissue oedema is demonstrated as low T1 signal on the background of high signal of the subcutaneous fat and has hyperintense signal compared with normal fat on T2 images. Fat-saturated proton density (PD)-weighted images and STIR images can demonstrate the soft-tissue changes more clearly. MRI can be useful in differentiating between acute osteomyelitis and acute bone infarction, especially after intravenous gadolinium contrast agents. Those with osteomyelitis showed a thick, irregular peripheral enhancement around a non-enhancing centre. Medullary infarctions show thin, linear rim enhancement or a long segment of serpiginous central medullary enhancement. This is particularly important in acute bone crises in sickle cell disease. Metal artefact suppression MRI techniques are useful for imaging close to prostheses. These include avoidance of gradient-echo sequences, use of STIR rather than fat suppression, acquiring the images on a high image matrix with a wide bandwidth and repeating sequences with swapped phase and frequency.

Nuclear Medicine Three-phase skeletal scintigraphy is generally useful over conventional positron emission tomography (PET) in the diagnosis of osteomyelitis. However, in the presence of previous trauma, metal devices, neuropathic joints and pre-existing bone conditions, this becomes less reliable and indium and gallium studies play a role. The limiting factor in these investigations is the need for a second investigation such as MRI to confirm the diagnosis due to poor specificity and spatial resolution (Fig. 46.15). Positron emission tomography/computed tomography (PET/CT) has some clear advantages over the conventional. Normal bone cortex has only low fluoro­deoxyglucose (FDG) uptake and the normal medulla shows slightly increased uptake compared with cortical bone. Hence, in osteomyelitis, increased uptake is a sensitive sign, but it is also positive in trauma, inflammatory diseases, and normal healing processes up to 4 months after surgery. Prosthetic joint infection: differentiation between infection and loosening is paramount in the management of these cases, as infection may need removal of the metal prosthesis. PET may not be accurate here, as hypercellular marrow around a prosthesis secondary to inflammation may show increased uptake both in loosening and infection. Combined 111In-labelled leucocytes and 99mTc-sulphur colloids have an accuracy of more than 90% in diagnosing prosthetic infection (Figs 46.16 and 46.17).

1196

SECTION C  The Musculoskeletal System

A

B

C

D

Fig. 46.13  Brodie Abscess of the Distal Radius. (A) Conventional radiograph demonstrates a lucency in the distal radial metaphysis. (B) Coronal T1 weighted unenhanced image demonstrates the rim of abscess cavity which is higher signal intensity (penumbra sign) (arrows). (C) Coronal short tau inversion recovery weighted image demonstrates intense high T2 signal within the abscess cavity in keeping with fluid content diffuse adjacent high T2 signal due to bone marrow oedema. (D) Coronal fat-suppressed T1 weighted sequence with gadolinium enhancement. This demonstrates a non-enhancing abscess cavity centrally with intense peripheral rim enhancement.

PET/CT is extremely useful in the diagnostic work-up of osteomyelitis and is the radionuclide test of choice for spinal infection. PET/CT has advantages and can be particularly useful to diagnose vertebral osteomyelitis and in localisation of abscesses (Fig. 46.18). Gallium-67 scintigraphy is used primarily for spinal infections when 18F-FDG imaging cannot be performed. Gallium single-photon emission computed tomography (SPECT) can also be equally useful. However, the procedure requires several patient visits to the department. The Nuclear Medicine investigation of choice for diagnosing diabetic foot infections is a labelled leucocyte study with an accuracy of 80%. Combined bone–leucocyte scintigraphy is the investigation of choice in detecting osteomyelitis on the background of neuropathic joints. PET/CT does not appear to have a definitive role in this setting.

Osteomyelitis Secondary to Prosthetic Devices Infections seen within 3 months of implant surgery are called ‘early’ and occur due to contamination during surgery or in the early postoperative days. The usual causative organism is S. aureus. Subacute infections occur between 3 and 24 months after the implant surgery and are usually caused by virulent coagulase-negative staphylococci or S. epidermidis. Chronic infections occur after 24 months and are usually due to haematogenous spread of infections from other sources.

Diagnosis of prosthetic infections can be a clinical challenge and may warrant removal of the prosthetic device. Because of the similarities in the imaging appearances of aseptic loosening and prosthetic infection, no single investigation is conclusive. Plain radiography may not be useful in diagnosis but may demonstrate soft-tissue swelling or evidence of loosening (Fig. 46.19). It is important to evaluate serial radiographs to allow detection of minimal changes which may be progressive. MRI has its limitations due to the presence of artefacts from the metalwork, but sometimes is useful especially after intravenous gadolinium administration to demonstrate bone marrow oedema and fluid collections. Ultrasound is useful to diagnose fluid collections around the prosthetic device and also to perform diagnostic aspirations for culture and sensitivity. Contrast-enhanced CT can also demonstrate the presence of a deeper collection that may not be clearly seen on ultrasound. CT-guided aspiration can also be performed in such deep collections (Fig. 46.20). Skeletal scintigraphy is widely available and shows increased uptake of tracer around prostheses in the presence of infection. The appearance is generally due to osteolysis and thus can be similar in aseptic loosening. Gallium imaging, the uptake of which is related to inflammation in general, increases the overall accuracy to about 70%–80%, but this is still less than satisfactory to separate infection from aseptic loosening. Text continued on p. 1201

CHAPTER 46  Bone, Joint and Spinal Infections

B

A

C Fig. 46.14  Osteomyelitis of the Olecranon With a Cloaca. (A) Lateral radiograph of the right elbow shows abnormal bone texture of the olecranon with thinning of the cortex and a focal lytic area due to chronic osteomyelitis. (B) Axial T1 weighted and (C) T1 fat-saturated enhanced images showing abnormal bone marrow signal with low T1 signal which shows enhancement after intravenous contrast medium. There is a well-defined defect in the cortex which represents the cloaca (arrow), with formation of an abscess in the overlying subcutaneous tissue.

1197

1198

A

SECTION C  The Musculoskeletal System

B

C

D Fig. 46.15  Osteomyelitis of the Left Femur: Nuclear Medicine. (A) Blood pool and (B) delayed 99mTc bone scintigraphy showing increased uptake in the left femoral condyle. (C) First circulation (selected image at 50 seconds) and (D) 3 days delayed 111In-labelled white cell scintigram showing increased uptake in the same area confirming the presence of osteomyelitis.

CHAPTER 46  Bone, Joint and Spinal Infections

1199

A

C

A

B

Fig. 46.16  Osteomyelitis of Left Femur. Combined 111In and 99mTc-sulphur colloid bone scintigram. This is currently the gold standard for prosthesisrelated bone infection. (A) Sulphur colloid is taken by normal bone marrow of left femur. (B) 111In-labelled white cells accumulate around the site of infection only.

A

B

B

D

Fig. 46.17  Prosthetic Infection. (A) Anterior (left) and posterior (right) views of a 99mTc-MDP bone scintigram showing increased uptake around a hip prosthesis. This was proved to be due to infection, though loosening or any other problem around the hip can produce increased uptake. Confirmation with magnetic resonance imaging or ultrasound would be necessary. (B) Coronal reformatted fused positron emission tomography/ computed tomography (PET/CT), (C) axial fused PET/CT and (D) axial PET images showing increased uptake around the prosthesis. Again this is non-specific and further evaluation would be warranted to confirm the diagnosis.

C

Fig. 46.18  Positron Emission Tomography/Computed Tomography (PET/CT) in Osteomyelitis. (A) Axial CT through the mandible shows a lytic area with thickening of the adjacent mandible. (B) Axial PET image and (C) axial PET-CT. Continued

1200

SECTION C  The Musculoskeletal System

D

E

Fig. 46.18, cont’d (D) coronal and (E) sagittal PET/CT fusion images show abnormal uptake of tracer in the left mandible, suggesting active infection.

A

B

Fig. 46.19  Prosthetic Infection. (A and B) Plain radiographs at immediate postoperative and 3-month interval demonstrate asymmetric periprosthetic lucency, which is localised to the proximal aspect of the femoral component more towards the lateral aspect (arrows). Development of asymmetric periprosthetic lucency in the first few months after arthroplasty should raise suspicion of infection.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

1201

E

Fig. 46.20  Postoperative Infection After Femoral Intramedullary Nail. (A) Conventional radiograph demonstrates abnormal bone texture involving the lateral cortex and the medullary cavity. There is loosening of the distal interlocking screws. (B and C) Coronal T1 and coronal short tau inversion recovery weighted images demonstrate areas of low T1 and high T2 signal intensity around the distal femur consistent with soft-tissue abscess formation. There is also soft-tissue oedema. (D and E) Coronal and axial T1 weighted images with gadolinium enhancement demonstrate multiple areas of abscess formation in the soft tissues circumferentially around the distal femur with intense peripheral and septal enhancement.

Combined labelled leucocyte and 99mTc-sulphur colloid imaging is currently the gold standard for imaging of prosthetic infections. Sulphur colloid is taken up in normal marrow, whilst labelled leucocytes accumulate at sites of infection as well as bone marrow. Hence if there is activity in the labelled leucocyte images in the suspected area without corresponding activity on sulphur colloid images, then prosthetic infection is confirmed with an accuracy of 95%. There are, however, several limitations including costs, the labour-intensive in vitro labelling process, availability and inconvenience for elderly and unwell patients. 111In-labelled polyclonal immunoglobulin lacks specificity. 99m Tc-sulphur colloid does not consistently differentiate infection from aseptic inflammation. Anti-granulocyte scintigraphy (AGS) with monoclonal antibodies or antibody fragments labelled with 99mTc has a reasonably high discriminating ability to identify prosthetic infection with a sensitivity of 83% and specificity of 80%. PET/CT may also play a role in diagnosis of prosthetic infection, but is not widely used for this purpose at the current time and its role is still debatable.

DIABETIC FOOT Radiology plays an important role in the multidisciplinary approach to management of diabetic foot infections and osteomyelitis. The high risk of amputation in these patients if diagnosis is delayed necessitates the

need for prompt and accurate investigation and treatment. The lifetime risk of foot ulceration in diabetic patients is as high as 25% and more than 50% of these become infected, which may need hospitalisation. Osseous involvement occurs in 20%–50% of cases. Microangiopathy is the major initiating event in the cascade of foot infection and ulceration in these patients. Microangiopathy reduces the end-organ perfusion reserve contributing to development of bone and soft-tissue infections. A combination of motor, sensory and autonomic neuropathy contributes to development of foot ulcers, impaired healing and superadded infections. The cause of foot infection is usually polymicrobial, commonly S. aureus, followed by S. epidermidis. Extension of soft-tissue infection into the bone causes osteomyelitis. The clinical presentation of patients with diabetes and a foot complication is extremely important in the diagnostic algorithm. A warm, swollen foot with intact skin is probably an acute Charcot neuroarthropathy, whilst the presence of a cutaneous ulcer that can be probed down to the underlying bone surface is almost 100% diagnostic of underlying osteomyelitis. Imaging tests need to be over 95% sensitive and specific to alter these pre-test likelihoods. The American College of Radiology introduced appropriateness criteria to select the appropriate imaging investigation(s) which will provide the most help in the management of diabetic osteomyelitis.

1202

SECTION C  The Musculoskeletal System

Fig. 46.21  Diabetic Foot Osteomyelitis. Cortical bone destruction is evident along the lateral edges of the fifth metatarsal head and base of the adjacent proximal phalanx, with overlying soft-tissue abnormality due to cutaneous ulceration.

Plain Radiography Plain radiography of the toes, forefoot, foot, ankle or heel (tailored to the suspected site of infection) have appropriateness criteria score of 9 (9, most appropriate, to 1, least appropriate) in the diagnosis of osteomyelitis in diabetic foot infections in all forms of clinical presentation (Figs 46.21–46.23). Radiographs have a sensitivity of 60% and specificity of 80% in the diagnosis of acute osteomyelitis in diabetic foot infection. The earliest changes include soft-tissue swelling with loss of fat planes, though this is a rather non-specific finding. The classic findings include the triad of osteolysis, periosteal reaction and bone destruction. These changes, however, take 10–20 days to be apparent on radiography. The changes may progress to destruction of the cortex, increased bone sclerosis due to sequestrum formation or loss of blood supply with bone necrosis. Bone resorption and auto-amputation occur in chronic cases. The appearances are often difficult to assess in the presence of associated neuropathic arthropathy. Serial radiography is usually necessary to assess progressive changes. When the diagnosis of osteomyelitis is uncertain, use of MRI with or without contrast enhancement or threephase skeletal scintigram is suggested.

Magnetic Resonance Imaging After initial radiography, MRI is the investigation of choice in the evaluation of pedal osteomyelitis. Bone marrow and soft-tissue

abnormalities are usually demonstrated much earlier compared with plain radiographs. MRI has a sensitivity of 90% and specificity of 82.5% in the diagnosis of osteomyelitis and is superior to 99mTc bone scintigraphy, plain radiography and white blood cell studies. High-resolution MRI examines the affected digit or foot with an extremity coil using thin slices (3–4 mm) and a small field of view (8–10 cm). Markers may be helpful for localisation of skin ulceration and clinical sites of swelling, giving increased accuracy of reporting. The patient is usually supine, or the forefoot may be imaged in the prone position with toes in an extremity coil. Imaging in at least two planes is needed for diagnosis and cross-referencing. Axial views are good for the anatomy of tendons and compartments. Sagittal and coronal views help to demonstrate ulcers and sinus tracts, especially when used with fat saturation (Fat Sat). A combination of T1 SE, T2 SE or Proton Density Fat Saturated and STIR images may be used in different planes. Intravenous contrast medium is often useful. T1 is good for demonstrating the anatomy of bones and tendons. T2/STIR sequences are good for the demonstration of fluid and oedema. Intravenous contrast medium may help to define abscesses and sinuses clearly, but does not have proven value. Intravenous enhanced MRI is advised when it will affect management and there is diagnostic doubt after unenhanced MR images have been reviewed. MR may demonstrate bone marrow oedema, periosteal reaction, cellulitis, joint effusion, sinus tracts, foot ulcerations and callus formation and evidence of gangrene (see Fig. 46.22). Bone marrow oedema is seen as hyperintense signal on T2 imaging, accompanied by corresponding low-signal intensity on T1 images. Absence of corresponding low-signal changes on T1 images is more likely to represent osteitis than osteomyelitis, even if bone marrow enhancement is present. STIR sequences are useful to assess bone marrow oedema, but may overestimate disease by exaggerating high signal against the suppressed signal from normal soft tissues. Periosteal reaction is seen as low signal separated from the cortical bone by highsignal fluid collection or oedema; it is usually seen in the metatarsals. Periosteal reaction other than in metatarsals should raise the possibility of infection. An intraosseous fluid collection is highly indicative of abscess formation. The penumbra sign is also a sign indicative of osteomyelitis; this is an area of intermediate signal on T1 image, surrounding a central low-signal area representing the fluid collection. Joint effusion is evidenced by presence of fluid signal in the joint space associated with thickened (and enhancing) synovium. Skin ulcers are represented as breaches in the skin signal intensity, usually low on T1 and T2 images, and have T2 a hyperintense signal around them caused by oedema. There may be hypertrophy of the edge of the ulcer, which is indurated and has low T2 signal. Sinus tracts are demonstrated as linear fluid-containing tracts, hyperintense on T2 images, better seen on fat-saturated T2 or STIR sequences. If the tract is healing or healed, fluid may not be seen and the tract may yield low T2 signal; tract walls will enhance, producing a tramline appearance. Cellulitis is seen as soft-tissue oedema caused by infection and inflammation of the subcutaneous tissue. The normal subcutaneous fat yields high T1 and T2 signal; in cellulitis, low T1 signal and hyperintense T2 signal on fat-saturated sequences are present. In the diabetic foot it is often difficult to diagnose infections on the background of changes associated with neuropathic joints (see Fig. 46.23). This remains the prime diagnostic challenge in imaging of the diabetic foot. Both osteomyelitis and neuropathic osteoarthropathy can demonstrate bone marrow oedema, soft-tissue oedema, joint effusion and enhancement; the conditions often coexist and correlation with clinical features is always recommended. There are, however, some features which may help to differentiate between the two and enable one to come to a conclusion of osteomyelitis, in the background of neuropathic joints. Some of these features will be briefly discussed below.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

1203

Fig. 46.22  Diabetic Foot Osteomyelitis. (A) Conventional radiograph demonstrates erosive destructive change at the lateral aspect of the fifth tarsometatarsal (TMT) joint consistent with osteomyelitis. (B–D) Sagittal T1, sagittal short tau inversion recovery, sagittal post-enhanced T1 weighted images demonstrate an overlying skin ulcer in the region of the fifth TMT joint (arrows). A tract is seen extending to the underlying fifth metatarsal and the TMT joint. Low T1 and high T2 signal seen within the base of the fourth and fifth metatarsals and also the cuboid. There is a joint effusion in the fifth TMT joint with peripheral synovial enhancement in keeping with associated septic arthritis.

Neuropathic changes include multiple bone involvement, joint deformity, subluxation or dislocation, cortical fragmentation or low signal changes in subchondral bone on both T1 and T2 images, correlating with osteosclerosis on plain radiography. Periosteal new bone is seen exclusively in the metatarsals and phalanges. Neuropathic disease tends to affect intertarsal and tarsometatarsal joints in 60% cases followed by metatarsophalangeal (MTP) joints in 30%. Single bone involvement usually favours osteomyelitis. Neuropathy is primarily an articular disease; thus, bone marrow oedema may be juxta-articular, centred on subchondral bone, whereas oedema from osteomyelitis may be more diffuse and generally on one side of the joint. However, inflammatory or infective arthritis also tends to produce subchondral distribution of bone oedema. Associated soft-tissue changes, cellulitis, abscess formation and sinus tract suggest infection. Joint effusion may be present in neuropathic joints with or without infection. However, superadded infection tends to produce thick rim enhancement, whereas neuropathic joint effusion without infection has thin rim enhancement. Bone fragmentation and proliferation, subluxation and dislocations can occur in neuropathic joints with or without infection. Increased erosion occurs after infections, though it is also seen in neuropathic joints without infection. Intra-articular bodies are more common in neuropathic joints with or without infection. Soft-tissue-related fluid collections are seen more

often and are larger near infected joints. Soft-tissue signal abnormalities due to oedema are seen in both entities, but loss of subcutaneous fat signal is seen more with infection. Though ulceration is seen near both infected and non-infected joints, sinus formation is seen only in the presence of underlying osteomyelitis or septic arthritis. Bone marrow signal abnormalities are seen in both, but intensity and extent of signal abnormalities are greater in the presence of infection. Bone marrow oedema associated with osteomyelitis has low T1, high T2 or STIR signal and shows post-contrast enhancement. Subchondral bone cysts are seen more in neuropathic joints without infection. Imaging of post-amputation diabetic patients for infection at the amputation margin is also a clinical challenge. The criteria for diagnosis of infection are the same as above after surgery, though it should be borne in mind that signal changes secondary to oedema from surgery should not be mistaken for osteomyelitis. Surprisingly little postoperative oedema is seen after amputation in these patients.

SEPTIC ARTHRITIS Most septic arthritis results from haematogenous seeding of the synovial membrane. Because synovium lacks a basement membrane. Infection easily spreads into the joint. Infection can also spread from other sources

1204

SECTION C  The Musculoskeletal System

A

B

D

E

C

Fig. 46.23  Diabetic Neuropathic Osteoarthropathy. (A and B) Conventional radiographs demonstrate involvement of the tarsometatarsal joints, a typical location for neuropathic involvement. There is involvement of multiple joints with bone fragmentation and joint deformity. (C–E) Coronal, axial proton density fat-saturated and sagittal short tau inversion recovery sequence demonstrate areas of bone marrow oedema, destruction of the articular surfaces and surrounding soft-tissue oedema. Note that there is no abscess formation or sinus tract associated with skin ulceration.

of infections, including endocarditis, sepsis, intravenous drug abuse and inoculation of foreign bodies. Patients usually present with sudden onset of monoarticular arthritis, associated with systemic symptoms and clinical signs of a joint effusion. Joint effusion may be difficult to detect in shoulders, hips and sacroiliac joints. The knee joint is most commonly affected. Hips, shoulders or ankles are also commonly affected. Sternoclavicular joint infections occur in intravenous drug abusers. In one study, the MTP joint was most commonly affected, followed by small joints of the foot, knee, sacroiliac joints and joints of upper limbs. Early diagnosis is critical in septic arthritis to avoid disabling outcomes. Delay in diagnosis may lead to development of cartilage and bone destruction due to the release of enzymes by the action of neutrophils, synovial cells and bacteria, leading to permanent disability. Before the era of MRI, early imaging findings in septic arthritis were considered non-specific and it was essentially a clinical diagnosis. MRI has greatly improved the diagnostic confidence of septic arthritis and the clinical outcome. Plain radiographs are not diagnostic in early septic arthritis but may reveal signs suggestive of a joint effusion. Subsequent cartilage destruction will result in joint space narrowing, provided the joint is not held open by an effusion. Lysis of the subchondral bone plate, erosions and adjacent bone destruction then occur (Fig. 46.24). When these latter features

are evident, the prognosis is poor and diagnosis and urgent management should be achieved before these features have developed. Joint effusion, synovial thickening and increased vascularity can be demonstrated with ultrasound and this is more useful in small and superficial joints. CT may be useful if MRI is contraindicated. CT will reveal joint effusions, and may show bone erosions, bone destruction and synovial enhancement. CT or fluoroscopy may be used for guiding diagnostic aspiration, if ultrasound is difficult. MRI findings in septic arthritis have now been well established and can occur as early as 24 hours after the onset of infection. MRI is especially useful in deep joints like shoulders and hips where clinical examination is difficult. Gadolinium-enhanced MRI with fat suppression has a sensitivity of 100% and specificity of 77% (Fig. 46.25). There is usually a joint effusion, particularly in the larger joints, which may also be seen in other forms of arthritis. The degree of effusion is not reliable in differentiating between infective and non-infective arthritis. Enhancement of the joint effusion may be present. There may be decrease in the perfusion of the bones. Synovial thickening may be present, which is seen as intermediate signal on T2 images. Thickened synovium usually shows significantly more enhancement, compared with normal synovium. Synovial enhancement and joint effusion have the highest correlation with clinical diagnosis of septic arthritis. High-resolution ultrasound may be more useful to demonstrate synovial thickening, especially in

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

F

1205

Fig. 46.24  Septic Arthritis Left Hip. (A) Plain radiograph demonstrates joint space loss with destruction of the articular surfaces. (B) Sagittal reconstructed computed tomography image also demonstrates joint space loss, areas of osteolysis and destruction of the articular surfaces. (C–E) Coronal T 1, coronal short tau inversion recovery and enhanced coronal T1 weighted images demonstrate joint effusion with thickened enhancing synovium, bone marrow oedema and soft-tissue abscess formation within the inferomedial soft tissues and also in a subiliacus location. There is destruction of the femoral head. (F) Axial T1 weighted fat-suppressed sequence with gadolinium enhancement. This also demonstrates bone marrow oedema, joint effusion, enhancement of the synovium and adjacent soft tissues with abscess formation.

smaller joints. Synovial thickening is due to inflammation and vascular proliferation. Synovial outpouching may be present due to increased fluid pressure within the joint. Septic arthritis progresses to destruction of articular cartilage and then the subchondral bone plate, which can be seen as irregularity of articular cartilage with high signal changes on T2 images with associated bony irregularities and erosions. Bone marrow signal abnormalities may occur due to oedema, and are usually not very extensive. The presence of extensive bone marrow changes, especially low signal on T1 images, should alert one to the possibility of underlying osteomyelitis. Soft-tissue signal abnormalities may also be seen around the affected joints due to inflammation. Cellulitis may develop in muscles, leading to abscess formation in extreme cases. The diagnosis of septic arthritis is made on the basis of a combination of clinical symptoms and radiology findings. The final diagnosis is made on the presence of positive culture on arthrocentesis, which can be performed under ultrasound, CT or fluoroscopy. Ultrasound is useful for small joints, hips, knees, ankles, paediatric patients and pregnant women. The shoulder, sternoclavicular and sacroiliac joints are more difficult with ultrasound and may require fluoroscopic or CT guidance. If pus is aspirated, as much fluid as possible should be aspirated to achieve decompression. If there is only minimal fluid, saline irrigation

of the joint may be performed and the aspirate sent for culture. If culture is negative, a white blood cell (WBC) count of more than 50,000/ mL is useful to make a diagnosis of septic arthritis. Material for culture should be obtained urgently and antibiotic therapy instigated as soon as possible. Relevant imaging should be performed as soon as possible and should not delay treatment.

MUSCULOSKELETAL TUBERCULOSIS Tuberculous infections of the bones and joints are common in the developing countries, but are being seen with increasing frequency in developed countries due to increases in immigrant populations and the incidence of human immunodeficiency virus (HIV), acquired immune deficiency syndrome (AIDS) and other immunosuppressive conditions. The causative organism is Mycobacterium tuberculosis. Haematogenous spread of the bacillus from a primary or reactivated focus in the body, to the bones or vertebrae (see later), is the most common mechanism of bone infection. Several forms of musculoskeletal tuberculosis (TB) are seen, which include tuberculous spondylitis, osteomyelitis, septic arthritis (Fig. 46.26), dactylitis, multifocal bone tuberculosis and soft-tissue infections. Spinal

1206

SECTION C  The Musculoskeletal System

A

B

C

Fig. 46.25  Septic Arthritis. (A) Axial positron emission tomography (PET) and (B) axial and (C) coronal fused PET/computed tomography (PET/CT) images showing uptake of 18F-fluorodeoxyglucose along the wall of an abscess cavity surrounding the dislocated proximal left femur after removal of an infected prosthesis. (D) PET maximum intensity projection (MIP) image outlines the abscess cavity and sinus tracks.

TB is the most common form of tuberculous bone infection. It is discussed separately in the section on spinal infection.

Pathogenesis Pathologically, chronic granulomas develop and are characterised by multinucleated giant cells, lymphocytes and macrophages, with central caseating necrosis. In the spine this causes rarefaction and destruction of the vertebral endplates and infection then spreads to adjacent discs. There is late preservation of the intervertebral disc due to a lack of proteolytic enzymes which is somewhat dissimilar to pyogenic spinal infections where disc destruction is an early feature. In long bones, tuberculous arthritis usually starts as a bone infection in the metaphysis, which spreads to the epiphysis and then to the joint. Haematogenous spread is a less common mode of spread in tuberculous arthritis. In the bones there is rarefaction, trabecular destruction, progressive demineralisation and bone and cartilage destruction. The consequent lytic lesions are well defined, with little surrounding bone regeneration and periosteal reaction. Paraosseous abscess formation, called cold abscesses (as they are not warm or tender), may occur. Further extension into the soft tissues leads to sinus formation and skin ulceration.

D

Investigations (Table 46.2) Plain Radiography

Plain radiographs are useful in established infection, but in acute infections abnormality is absent or subtle. Small bones may ultimately demonstrate abnormal marrow changes, which include lytic lesions with rather thick and well-defined borders. Bony destruction may also be seen in established cases with joint deformities.

Computed Tomography CT is more sensitive than plain radiography in demonstrating cortical and trabecular bone destruction, and periosteal reaction. CT is generally useful for planning guided interventions—drainage of abscesses or planning/guiding bone biopsies.

Magnetic Resonance Imaging MRI can be useful in distinguishing between tuberculous and pyogenic arthritis, but the features show considerable overlap. Bony erosions are more commonly seen in tuberculous arthritis compared with pyogenic arthritis. Subchondral bone marrow oedema is more prominent in

CHAPTER 46  Bone, Joint and Spinal Infections

1207

TABLE 46.2  Features Which Aid in Distinguishing Pyogenic From Tuberculous Infection Pyogenic Infection

Tuberculous Infection

Clinical symptoms are more acute and severe, with systemic toxicity and raised acute inflammatory markers Subchondral bone marrow oedema prominent Bone erosions less common Destruction of articular cartilage more common Irregular synovial thickening that shows avid post-contrast enhancement Surrounding soft-tissue changes are prominent, ill-defined and more extensive Usually a single site

Insidious onset, less toxic, present as chronic infections Less prominent More common Cartilage destruction occurs, but less common Synovial thickening is smooth, and shows enhancement Well-defined inflammation and abscess with little surrounding signal changes, cold abscess Multiple sites of involvement common

A

UNUSUAL MUSCULOSKELETAL INFECTIONS Atypical Mycobacterial Musculoskeletal Infections

B

C

Fig. 46.26  Tuberculous Arthritis. (A) Plain radiograph demonstrates extensive bone destruction in the glenoid and humeral head. This appearance in the humeral head has been termed ‘caries sicca’ (dry rot). (B) Magnetic resonance coronal T1 weighted and (C) coronal enhanced fat-suppressed T1 weighted sequences demonstrate a large joint effusion, extending into the adjacent soft tissues with surrounding enhancing walls and the bone destruction also evident on plain radiograph.

pyogenic arthritis, best seen on T2 images with fat suppression. This may be better seen after gadolinium enhancement. Articular cartilage destruction may be seen in both, but is earlier and more prominent in pyogenic arthritis (Fig. 46.27). Synovial thickening demonstrated as intermediate T2 signal is seen in both conditions. After gadolinium enhancement, synovial changes are more easily analysed and show smooth thickening in tuberculous arthritis compared with irregular thickening in pyogenic arthritis. Surrounding soft-tissue changes are irregular and ill-defined in pyogenic arthritis, while in tuberculous infection tend to be better defined. Abscess cavities in tuberculous infection tend to have smooth and thin walls, with less prominent surrounding inflammation as these are cold abscesses (Fig. 46.28), whereas pyogenic abscesses tend to be more thick walled with pronounced surrounding inflammation.

Atypical mycobacterial infections are also on the rise and are more resistant to treatment than tuberculosis. Musculoskeletal infections occur in approximately 5%–10% of atypical mycobacterial infections. Most osseous infections are caused by Mycobacterium kansasii and M. scrofulaceum, followed by M. avium-intracellulare and M. fortuitum. The mode of infection may be haematogenous or by direct inoculation from surgical implants or trauma. Patients present with similar symptoms as mycobacterial infections but may be milder and more protracted. Fevers, chills, malaise and weight loss may occur. Early radiographic manifestation includes soft-tissue swelling due to inflammatory changes and regional hyperaemia. General radiographic observations include a tendency for metaphyseal or diaphyseal involvement. Bone resorption, osteolysis with periosteal reaction and bone marrow oedema occur, but take several weeks to be evident radiographically (Fig. 46.29). Multiple sites of involvement, well-defined lytic lesions with marginal sclerosis and osteopenia are less striking than that seen in tuberculosis. In subacute and chronic forms bone abscess, sequestrum, involucrum and cloaca formation occur with the formation of sinus tracts. Spinal involvement, monoarticular arthritis and soft-tissue infections of tenosynovitis, septic bursitis or carpal tunnel syndrome occur and have radiographic changes similar to tuberculous disease.

Hydatid Disease Infection of bone by the parasitic tapeworm Echinococcus is rare, accounting for less than 5% of cases of hydatidosis. Spinal and pelvic sites are commonest. Uni- or multiloculate intraosseous cysts which may expand the bone and extend into adjacent soft tissues are seen but are not specific. Serological tests are positive.

Bone Infections in Sickle Cell Disease Sickle cell disease is an autosomal recessive condition that occurs as a result of a defect in haemoglobin S, affecting the β-chain. The mild, heterozygous form of the disease is called sickle cell trait, which results in a carrier status, without symptomatic disease. The effect of sickle cell disease is abnormal sickling of red blood cells within the capillaries

1208

SECTION C  The Musculoskeletal System

A

B

C

D

Fig. 46.27  Septic Arthritis of the Shoulder. (A) Plain radiograph demonstrates reduction in bone density with an erosion at the craniolateral aspect of the humeral head. (B–D) Coronal T1, coronal T2 fat-suppressed and axial proton density fat-suppressed sequences demonstrate widespread chondrolysis, bone marrow oedema and joint effusion with surrounding soft-tissue oedema in keeping with septic arthritis.

and blood vessels, which results in vascular occlusion and infarction affecting various organs, called a ‘sickling crisis’. Bone infarction can occur, resulting in severe pain. The effect of autosplenectomy due to recurrent splenic infarcts and the presence of bone infarcts predispose these patients to bone infections. Infection is usually caused by Salmonella organisms, the source of infection being the gut, followed by staphylococci infections. Long bones and small bones are commonly affected. Radiographic appearances are similar to other forms of osteomyelitis, but a diagnostic challenge can occur due to pre-existing replacement of fatty marrow by red marrow secondary to intramedullary marrow hyperplasia, causing abnormal bone marrow signal in the absence of infection. Also, aseptic infarction can cause severe bone pain and can cause abnormal signal changes on MRI and be confused with infection. Familiarity with the patterns of MRI appearances of bone infarcts and normal red bone changes is necessary to avoid overdiagnosis of osteomyelitis. Bone infarcts typically affect small tubular bones of the hands and feet in children and long bone in adults. Lucency and periosteal reaction

occurs in the early stages, which may proceed to sclerosis and bone infarcts and destruction. Avascular necrosis of the epiphyses of the femoral and humeral head is common and can be bilateral. Blood cultures can be positive in up to 50% of cases and are essential for diagnosing osteomyelitis. Radiographic changes of bone infection are usually subtle in the early stages and may not be evident for up to 10 days after the onset of infection. Osteolysis and periosteal reaction are the initial findings. MRI is the optimal investigation for diagnosing bone infection as it can demonstrate bone marrow changes, abscesses and periosteal reaction, and helps to differentiate from bone infarction and avascular necrosis. Bone infarcts tend to have a serpiginous appearance on MRI and avascular necrosis affects typical sites. Septic arthritis is less common than osteomyelitis. Combined leucocyte-labelled and 99mTc-sulphur colloid scintigraphy may also help to differentiate between infection and infarction. Early diagnosis and treatment is important in preventing long-term complications in these patients.

CHAPTER 46  Bone, Joint and Spinal Infections

A

1209

B Fig. 46.28  Tuberculous ‘Cold’ Abscess. (A) Magnetic resonance sagittal and (B) axial T2 weighted sequences demonstrate a large abscess extending over the surface of the psoas muscle in the pelvis, arising from tuberculous discitis at L4/L5.

Musculoskeletal Fungal Infections High-risk patients for fungal infections include immunosuppressed HIV-positive patients, organ transplant recipients and patients on chemotherapy and long-term corticosteroids. Travel to endemic areas also increases the risk of infection. Infection occurs through skin inoculation or by the haematogenous route, usually via the lungs. They cause granuloma formation and may show soft-tissue nodules, discharging sinuses, chronic multifocal osteomyelitis and joint infection. Mixed sclerotic and lytic lesions can occur. Synovial thickening and features of chronic granulomatous arthritis are present and resemble osteoarticular tuberculosis. Aspergillus fumigatus infection occurs via haematogenous spread from invasive pulmonary aspergillosis. The spine is most commonly affected, but infection can occur in other joints. Multifocal osteomyelitis, septic arthritis and discitis can occur. Diagnosis is made by synovial or bone biopsy. Blastomycosis is endemic in the western United States, and bone infection affects the spine and lower limbs. Radiological features include those of chronic arthritis. Osteolytic lesions resembling bone tumours may occur and need bone biopsy for diagnosis. Candida osteomyelitis presents as a lytic lesion, without significant periosteal reaction, in immunosuppressed patients. Arthritis affects the larger joints. Cryptococcosis occurs in immunosuppressed patients and may proceed to disseminated infection affecting larger joints, osteolytic lesions in flat bones and avascular necrosis. Mycetoma tend to cause granulomatous infection of plantar subcutaneous tissue and proceed to cause chronic osteomyelitis, with multiple discharging sinuses. Mixed sclerotic and lytic lesions of bone are seen on radiographs. Chronic infection of the bones and soft tissues of the foot due to fungi or actinomycetes implanted by penetrating injury

from thorns was described as ‘Madura foot’. The soft-tissue abscesses in this condition characteristically contain multiple small cavities, each with a low signal centre on T2 weighted MRI (the dot in a ring sign) (Fig. 46.30).

Musculoskeletal Infections in Human Immunodeficiency Virus Patients Bone and joint infections were originally considered relatively rare in patients with HIV infections. Infection is usually caused by opportunistic organisms like Candida, Clostridium and Mycobacterium avium complex. However, more recent studies show musculoskeletal infections as a relatively common manifestation in HIV patients. The incidence also increases in IV drug-using HIV-infected patients. Apart from musculoskeletal manifestations of painful articular syndrome and non-infectious arthritis, musculoskeletal infections also occur in HIV patients. These include infectious myositis, septic arthritis, osteomyelitis and tuberculous arthritis. Patients present with fever and arthritic symptoms, commonly affecting knees or ankles. S. aureus, Salmonella and Penicillium infections were the commonest seen in these groups. Radiographic findings include periarticular osteopenia, osteolysis and soft-tissue swellings and share similarities to other forms of osteomyelitis. Outcomes were generally poor in patients with bone infections.

DIFFERENTIAL DIAGNOSIS Several other disorders may mimic osteomyelitis. Inflammatory arthritis may mimic septic arthritis in its early stages and a good clinical history is often the clue to diagnosis. As discussed earlier, both conditions may give rise to joint effusion, synovial thickening and post-contrast enhancement, though symptoms are usually severe with septic arthritis. Bone infarcts may mimic osteomyelitis in the acute stages and clinical features and laboratory tests often help in diagnosis. Bone infarcts have

1210

SECTION C  The Musculoskeletal System

A

B

C

D

Fig. 46.29  Mycobacterium Avium-Intracellulare Infection of the Knee. (A) Plain radiograph demonstrates abnormal bony architecture in the proximal tibia with intervening lucencies and areas of bone destruction involving the articular surface. (B) Sagittal proton density fat-suppressed sequence demonstrates large joint effusion, bone marrow oedema and bony destruction of the proximal tibia. (C) Axial T1 weighted image demonstrates areas of low T1 signal in the distal femur, joint effusion and an area of bone destruction in the lateral aspect of the trochlea. (D) Coronal short tau inversion recovery image demonstrates bony involvement both at the tibial and femoral aspects with marked surrounding soft-tissue oedema.

serpiginous appearances on MRI, whilst osteomyelitis demonstrates more significant bone marrow oedema and post-contrast enhancement. Tumours are an important differential diagnosis in osteomyelitis and share several imaging features, including bone marrow oedema and contrast enhancement. Both lytic and sclerotic lesions with wide zones of transition and cortical destruction can occur with tumours and infection. Biopsy is sometimes needed to confirm the diagnosis and needs liaison with referring clinicians and pathologists. It is axiomatic that biopsy material should always be sent for both histological and microbiological assessment. Acute diabetic neuropathic joints are also difficult to differentiate from infection, as discussed earlier in detail. Degenerative arthritis can also cause diagnostic difficulties on imaging and clinical symptoms and history is important in management. Granulomatous

diseases like giant cell tumour or Langerhans cell histiocytosis can also mimic infections of bone. Deposition diseases such as amyloidosis may also cause focal erosive bone lesions similar to infection.

MANAGEMENT A good clinical history and thorough clinical examination is the most important step in early diagnosis of these infections. Laboratory tests are extremely useful in many instances to highlight the presence of acute infections. Blood cultures should be obtained whenever possible, before the start of treatment with antibiotics, as it helps in choosing the appropriate antibiotics. Culture of pus from sinus tracts, abscess cavities and effusions should also be performed whenever possible to

CHAPTER 46  Bone, Joint and Spinal Infections

A

C

1211

B

D Fig. 46.30  Mycetoma of Ankle. (A) Sagittal T1 weighted, (B) short tau inversion recovery, (C) axial T2 weighted and (D) sagittal fat-saturated enhanced T1 weighted images of the ankle. A subcutaneous lesion demonstrating multiple low signal rings with central low signal dots on the T2 weighted image and multiple avid ring enhancements with contrast is seen. The appearances on imaging mimicked a haemangioma, but the ‘dot in a ring’ appearance is typical of a mycetoma, which was confirmed on biopsy. (Courtesy of Dr R Mehan, Bolton Royal Hospital, UK.)

obtain a microbiological diagnosis. It is important to have a multidisciplinary approach to choose appropriate investigations. Percutaneous biopsy is an alternative to surgery in appropriate circumstances, being cost-effective and less invasive. Percutaneous bone or soft-tissue biopsy may be performed under ultrasonic, fluoroscopic or CT guidance to obtain a definitive diagnosis. This is a relatively safe procedure and extremely useful to confirm the diagnosis as the histological yield from the procedure is high. Abscess cavities and soft-tissue aspirates also require culture and sensitivity, along with cytology and

histology, to analyse typical features of certain infections, like caseating necrosis in tuberculosis. Discussion with the referring team is necessary to plan the appropriate route for biopsy, in case the lesion in question turns out to be a malignancy. Departmental protocols should be present to ensure rapid transfer of specimens to the microbiology department in the appropriate transport medium. Histological examination may reveal changes that are compatible with acute or chronic osteomyelitis. In acute osteomyelitis, there are acute inflammatory cells, neutrophils, congestion and thrombosis of

1212

SECTION C  The Musculoskeletal System

Fig. 46.31  Computed Tomography-Guided Bone Biopsy. Diagnosis of bone infection is confirmed by bone biopsy. Samples should be sent both for histological and microbiological evaluation. Computed tomography remains the most useful tool for planning and guiding biopsy.

medullary and periosteal blood vessels and necrotic bone. Chronic osteomyelitis shows lack of neutrophils, areas of fibrosis in bones, macrophages, lymphocytes and histiocytes. Bone biopsy can be usually performed as an outpatient procedure under CT or fluoroscopic guidance (Fig. 46.31). The procedure is performed under local anaesthesia, augmented with nitrous oxide analgesia or conscious sedation if required. The coaxial bone biopsy technique is the one that is commonly used. Once the penetrating cannula is placed on the surface of the lesion, a biopsy needle is inserted and the specimen is aspirated. Positive culture rates are usually low and are about 35% for long bones. Cultures from the spine may reveal a higher yield rate, perhaps by sampling the infected disc. Surgical biopsies produce similar yields and thus carry no advantage over percutaneous biopsy. Positive culture rates are very low once antibiotic therapy has been instigated. Treatment of osteomyelitis and soft-tissue infections needs aggressive antibiotic therapy, surgical debridement and removal of prostheses, depending on the severity of infection. Follow-up is necessary to assess response to treatment and modify treatment accordingly. Septic arthritis requires immediate arthrotomy and joint lavage and optimal antibiotic therapy, adjusted as required by discussion with the microbiologist.

SPINAL INFECTION Summary Box 46.3.

SUMMARY BOX 46.3: Spinal Infection • Lack of increased T2 signal intensity of an associated disc, intradiscal vacuum phenomenon and lack of soft-tissue involvement are characteristic of Modic type 1 degeneration. • Primary pyogenic infection can occur in facet joints or the epidural space without associated spondylodiscitis. • Multilevel involvement is not rare and magnetic resonance imaging of the whole spine should be performed to identify the complete extent of epidural involvement. • A well-defined paraspinal abscess with thin enhancing rim and subligamentous spread to three or more vertebral levels are findings more suggestive of tuberculous spondylitis than of pyogenic spondylitis. • Vacuum phenomenon, facet involvement and bone sclerosis are suggestive of neuropathic spinal arthropathy.

Vertebral Osteomyelitis The incidence of vertebral osteomyelitis is 2.4/100,000 population per annum and tends to increase with age. Vertebral osteomyelitis may be acute, subacute or chronic. Vertebral osteomyelitis may be pyogenic, tuberculous or rarely fungal, such as Candida. Several studies have shown pyogenic infections to be the most common cause, followed by tuberculosis even in TB endemic areas.

Pyogenic Vertebral Osteomyelitis Acute pyogenic vertebral osteomyelitis is usually from haematogenous seeding. Direct extension from adjacent soft-tissue infections and direct inoculation of infection during surgery are also common. The usual causative pathogen is S. aureus, but in the presence of spinal implants, coagulase-negative staphylococcal infection also occurs. Low-virulence coagulase-negative staphylococcal infection occurs in prolonged bacteraemia after pacemaker infections. The primary site of infection in acute vertebral osteomyelitis may be urinary tract infections, skin or soft-tissue infections, vascular access sites, septic arthritis or bursitis and endocarditis. There may be underlying disease, including diabetes, coronary artery disease, immunosuppression, cancer or renal failure on dialysis (Fig. 46.32). Pyogenic infection is now thought to begin as osteomyelitis of vertebral endplates because of the presence of a dense end-arterial network at that site. From here, the infection spreads to the adjacent disc space. Through the disc space, infection spreads to contiguous vertebrae. Contiguous endplate destruction occurs with extension of the abscess into the paravertebral soft tissues and the epidural space.

Symptoms Back pain is the commonest clinical presentation and is present in 86% of patients. The lumbar spine is most commonly affected (58%), followed by thoracic (30%) and cervical spine (11%). Vertebral osteomyelitis is complicated by direct seeding in different compartments, including epidural, paravertebral and disc space abscess, with the paravertebral location being the commonest. Systemic signs such as fever may not always be present. Severe sharp pain should alert one to the possibility of epidural abscess. Motor and sensory involvement may occur depending on the degree of spinal cord compression. Neurological complications are particularly common in cervical vertebral infections.

CHAPTER 46  Bone, Joint and Spinal Infections

A

1213

B

C

D Fig. 46.32  Pyogenic Vertebral Osteomyelitis After Pelvic Sepsis. Patient with a presacral abscess following radiotherapy and AP resection for rectal cancer. (A) Contrast enema showing a sinus track to the prevertebral region at L5/S1. (B) Sagittal fused positron emission tomography/computed tomography (PET/CT) image shows increased uptake of fluorodeoxyglucose (FDG) tracer in the presacral region, but not the disc. (C and D) Sagittal T1 and T2 images show the presacral abnormality consistent with granulation tissue extending to the L5/S1 disc, with typical appearances of infective discitis.

Investigations Usual inflammatory markers: raised white cell count with neutrophilia. Elevation of the CRP and ESR is invariably present. Blood cultures may be positive in up to 58% of patients. Higher yield rates up to 77% can be obtained from bone biopsy cultures. Plain radiographs.  Plain radiographs are not very sensitive in acute vertebral osteomyelitis, but are useful in subacute or chronic conditions.

Early changes may be seen as minor endplate irregularities but MRI is much more sensitive. Endplate destruction, paravertebral soft-tissue swelling and kyphoscoliosis are present in established osteomyelitis. In the acute setting, plain radiographs are mandatory to exclude other lesions, such as metastases or fractures. Computed tomography.  CT is generally only used when there are contraindications to MRI. Enhanced images with multiplanar

1214

SECTION C  The Musculoskeletal System

reformatting help to delineate the lesion and demonstrate the extent of any abscess. The main role for CT in vertebral osteomyelitis is to plan and direct the biopsy, thereby confirming the diagnosis and obtaining material for culture. Magnetic resonance imaging.  MRI is the investigation of choice for vertebral osteomyelitis and should be promptly arranged, especially in the presence of neurological symptoms and signs, as it also excludes other causes such as intervertebral disc prolapse and malignancy. Motor and sensory symptoms depend on the degree and level of spinal cord or cauda equina involvement. MRI is extremely sensitive and has 90% accuracy in the diagnosis of acute vertebral osteomyelitis. The imaging features are typical for infective discitis and osteomyelitis, and in many instances these entities can be differentiated from tuberculous osteomyelitis. Typically, one disc space and the two adjacent vertebral bodies are involved. The disc usually yields low signal on T1 and increased signal on T2 weighted images associated with loss of the intranuclear cleft. There is destruction of

the endplates of the adjacent vertebrae above and below the disc space. There is adjacent bone marrow oedema and inflammatory tissue. The bone marrow oedema shows as low signal on T1 and high signal on T2 and STIR images, involving the vertebral bodies and endplates. As the infection proceeds untreated, this may lead to vertebral destruction and collapse. Epidural and/or paraspinal abscess may be noted as high signal fluid collections on T2 images. Intraspinal collections or granulation tissue may cause cord compression. Gadoliniumenhanced T1 images with fat saturation are useful to define the extent and effects of epidural or paravertebral abscess collections (Fig. 46.33). STIR images of the whole spine may be useful to assess multilevel involvement. In pyogenic osteomyelitis there is usually homogeneous and diffuse enhancement of the vertebral bodies as opposed to the heterogeneous and localised enhancement seen in tuberculous vertebral osteomyelitis. Disc abscess with rim enhancement is more common in pyogenic infections, whilst vertebral intraosseous abscess with rim enhancement

A

B

C

D Fig. 46.33  Pyogenic Infective Spondylodiscitis. (A–C) Sagittal T2, sagittal T1 pre- and post-contrast enhanced MR images demonstrate changes centred around the L5/S1 intervertebral disc with oedema of the vertebral bodies of L5 and S1 and paravertebral abscess formation anteriorly (arrow). (D) Sagittal reconstructed computed tomography image demonstrates destruction of the endplates adjacent to the L5/S1 intervertebral disc.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

F

1215

Fig. 46.34  Septic Facet Joint Arthritis With Epidural Abscess Formation. (A) Sagittal short tau inversion recovery image demonstrates bone marrow and soft-tissue oedema centred around the right-sided facet joint at L3/L4 level. (B–D) Axial T1 pre- and post-gadolinium-enhanced images and axial T2 weighted images demonstrate diffuse enhancement of the facet joint and the surrounding soft tissues with an abscess formation in the posterior paraspinal muscles (arrow). (E) Sagittal T1 weighted gadolinium-enhanced image demonstrates extensive epidural abscess formation from L1 down to the sacrum (arrows). (F) Coronal reconstructed computed tomography image demonstrates destruction of the articular surfaces of the involved facet in keeping with septic arthritis (arrow).

occurs more commonly in TB. Paraspinal abscesses are more commonly associated with TB than with pyogenic infections. Although pyogenic spinal infection usually begins as osteomyelitis of the endplates, in rare instances, primary infection can develop in the facet joints or the epidural space without associated spondylodiscitis. Facet joint infections may be complicated by abscess formation in the epidural space or in the paraspinal muscles (Fig. 46.34). Primary epidural involvement is rare, but can result in inflammatory phlegmon or abscess formation. It is usually secondary to spondylodiscitis and is important to recognise as it can lead to rapid loss of neurological function. Multilevel involvement is not rare and MRI of the whole spine should be performed to identify the complete extent of epidural involvement (Fig. 46.35). Isolated sacroiliac joint infection is rare but can be a cause of low back pain. It is usually unilateral, in contrast to predominantly bilateral involvement in spondyloarthropathies. However, unilateral involvement of the sacroiliac joints has been noted in psoriatic arthritis, reactive arthritis and the early stages of ankylosing spondylitis. MRI findings of extensive soft-tissue abnormalities, such as muscle oedema, capsular bulging/distension and fluid collections, are suggestive of infectious sacroiliitis (Fig. 46.36). Postoperative spinal infection is one of the most serious complications of spine surgery and the incidence ranges from 0.7% to 12%. Surgical site infections are commonly classified as superficial or deep. Superficial

infections are limited to the skin and subcutaneous layer without fascial involvement. Deep infections occur below the fascia and may also result in spondylodiscitis, osteomyelitis and paraspinal or epidural abscess. Infection can also occur at the site of bone graft harvest, usually from the posterior aspect of the iliac bone (Fig. 46.37). Certain disorders may mimic spinal infection such as Modic type 1 degenerative disc disease and neuropathic spine. Lack of increased T2 signal intensity of an associated disc and lack of soft-tissue involvement are characteristic of Modic type 1 degeneration. Occasionally, severely degenerate discs can demonstrate fluid signal on MRI and can be difficult to differentiate from infection. In these instances, the paravertebral and epidural soft tissues should be assessed carefully. In degeneration, the soft-tissue fat planes around the disc will appear normal whilst in infection these fat planes will be obliterated and there is often an associated soft-tissue mass including fluid collections in infections. Severely degenerate discs often demonstrate vacuum phenomenon, sometimes best seen on CT (Fig. 46.38), which is another useful sign to differentiate from infection. Neuropathic spine is a destructive process in response to repeated trauma in the background of sensory loss. Vacuum phenomenon, facet involvement, bone sclerosis and large osteophytes with pseudarthrosis formation are the common findings (Fig. 46.39). Nuclear medicine.  Skeletal scintigraphy is less important than it was, as most centres rely on MRI. Three-phase 99mTc bone scintigraphy

1216

SECTION C  The Musculoskeletal System

A

B

C

D Fig. 46.35  Epidural Abscess Formation in Pyogenic Infective Spondylodiscitis. (A and B) Sagittal short tau inversion recovery and post-gadolinium-enhanced T1 weighted images demonstrate bone marrow oedema and enhancement of the T1 and T2 vertebral bodies with abscess formation in the anterior soft tissues and also extensive epidural abscess formation which extends inferiorly into the lower thoracic spine (arrow), a significant distance away from the original source of infection in the T1/T2 disc space. For this reason, magnetic resonance imaging of the whole spine should be performed in all cases of suspected spinal infection. (C and D) Axial T1 weighted images post-gadolinium enhancement demonstrate the thoracic epidural abscess causing displacement and compression of the thoracic cord (C) and a paravertebral soft-tissue abscess extending into the left lung apex (D).

has about 67% accuracy, but positive results are seen only after a few days. It is also less sensitive to the detection of epidural abscess. 111Indium leucocyte scintigraphy and antileucocyte scintigraphy are more specific, but have sensitivities of around 20%, and hence are not generally useful. Gallium imaging is also a useful adjunct to MRI in the evaluation of spinal infections (Fig. 46.40). Because of limited availability, PET/CT is generally not used in the diagnosis of vertebral osteomyelitis but has been found to be useful in some studies.

Treatment Acute vertebral osteomyelitis is usually treated with intravenous antibiotics. Image-guided intervention under CT may be necessary for diagnostic purposes and CT or ultrasound is used for guiding the drainage of

abscesses. Surgery is usually required in the presence of implants; in chronic infections due to treatment failure, removal of the implant is generally recommended.

Tuberculous Vertebral Osteomyelitis Tuberculosis of the spine is still prevalent in developing countries and in TB endemic areas. Poverty, malnutrition and overcrowding predispose to the development of primary tuberculous infection. It is increasingly seen in developed countries in the immigrant population and also increases with the increase in incidence of pulmonary tuberculosis due to immunosuppressive conditions, including HIV infection and AIDS. Tuberculosis of the spine is usually a secondary infection, with spread to the spine occurring by the haematogenous route, usually from primary

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

Fig. 46.36  Pyogenic Infection of Sacroiliac Joint. (A and B) Axial T1 and coronal short tau inversion recovery sequences demonstrate bone marrow oedema at the sacral and iliac aspects of the right sacroiliac joint with articular surface irregularity, periosteal and soft-tissue oedema. (C and D) Axial T2 and axial post-gadolinium T1 weighted fat-suppressed sequences demonstrate a large abscess in the subiliacus location arising from the infected sacroiliac (SI) joint. (E) A computed tomography-guided biopsy of the sacroiliac joint was performed along with aspirate of the abscess through a trans-sacral route. This confirmed the presence of infection with Staphylococcus aureus and the patient was known to have underlying diabetes.

A

B

C

Fig. 46.37  Postoperative Infection at Site of Graft Harvest. (A and B) Pre- and post-gadolinium enhanced axial T1 weighted images demonstrate area of peripheral enhancement at the site of the graft harvest from the posterior aspect of the right iliac bone (arrows). Soft-tissue changes are seen extending posteriorly to the midline. This patient had continuing discharge from the inferior aspect of the wound in the midline. (C) Maximum intensity projection image from a computed tomography sinogram demonstrates a sinus tract extending from the skin to the graft harvest site (arrow), indicating that this was the source of infection rather than the operated site in the lumbar spine.

1217

1218

SECTION C  The Musculoskeletal System

A

B

D

E

C

Fig. 46.38  Modic Type I Degenerative Endplate Changes. (A–C) Sagittal T1, sagittal short tau inversion recovery and post-gadolinium sagittal T1 weighted fat-suppressed images demonstrate bone marrow oedema in the L4 and L5 vertebral bodies with irregularity of the endplates and asymmetric loss of disc height. (D) Axial T2 weighted image demonstrates normal appearance of the paravertebral and epidural soft tissue with no inflammation or abscess formation. (E) Sagittal reconstructed computed tomography image demonstrates vacuum phenomena within the disc space (arrow) in keeping with degenerative change and differentiating it from infective spondylodiscitis.

lung or genital tract infection. Spinal tuberculosis accounts for approximately 50% of all musculoskeletal tuberculous infections. Spinal tuberculosis is most common around the thoracolumbar junction. The incidence decreases on either side of this level but may occur at any level. Infection usually occurs at the anterior ends of vertebral bodies and spreads under the longitudinal ligament to involve contiguous vertebrae. Skip lesions may also occur due to haematogenous spread. The vertebral body is commonly affected; posterior element involvement is rare but is seen particularly in Asian patients. Three patterns of vertebral body involvement are seen. A paradiscal lesion is the most common form of involvement of spinal tuberculosis. There is involvement of subchondral bone adjacent to an intervertebral disc, with reduction in disc height. Anterior lesions occur due to spread of infection under the periosteum and anterior longitudinal ligament, resulting in loss of blood supply to the vertebral body with development of necrosis and infection. Abscess formation may occur with resultant stripping of the periosteum from the vertebral body, causing scalloping and multiple-level involvements. Central lesions involve the centre of the vertebral body, with loss of height resulting in vertebra plana. ‘Gibbus’ deformities occur due to vertebral body collapse, manifesting as acute angulation in the spine. Paraspinal abscess formation also occurs

TABLE 46.3  Differentiation Between

Pyogenic and Tuberculous Vertebral Osteomyelitis on Imaging Pyogenic Spinal Infection

Tuberculosis of Spine

1. Lumbar spine involvement common 2. Commonly single site with disc space infection and involvement of two adjacent vertebrae

More common in thoracic spine

3. Disc abscess with endplate destruction occurs 4. Significant surrounding inflammation with diffuse oedema of vertebral bodies 5. Enhanced images show diffuse vertebral body enhancement, irregular enhancement of thick-walled abscesses

Multilevel involvement and skip lesions are common. Spread may occur along anterior longitudinal ligament Intraosseous and paraspinal abscess occurs more frequently Inflammation is more localised and formation of cold abscess Enhancement of vertebral bodies more localised and show rim enhancement of thin-walled abscess cavities

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

C

D

E

F

1219

Fig. 46.39  Neuropathic Spine. (A–C) Sagittal T1, sagittal T2 and sagittal short tau inversion recovery weighted images demonstrate abnormality centred around the L3/L4 intervertebral disc space with endplate destruction, fluid signal within the disc space, along with changes in the facet joint, which is widened with associated facet effusion (arrow). (D) Sagittal T2 weighted image in the midline demonstrating chronic spinal cord injury at the level of T11/T12 (arrow) as the cause of diminished sensations leading to a neuropathic spine. (E and F) Coronal and sagittal reconstructed computed tomography spine images demonstrate vacuum phenomena, bone sclerosis and destruction with large osteophytes and new bone formation consistent with neuropathic spine.

and typically shows no significant signs of inflammation; hence, such abscesses are called cold abscesses. Table 46.3 summarises features that may help to distinguish pyogenic from tuberculous spinal infection.

Plain Radiographs In acute infections, plain radiographs may be normal. In subacute infection, bone lucency may be seen in vertebral bodies. Endplate changes and destruction occur with reduction in the intervertebral disc space. Cold abscess formation may cause paraspinal soft-tissue density on antero-posterior radiographs. Chronic infection can cause sclerosis of the bone and endplates, bone destruction with compression fractures and deformities. Gibbus deformity is an acute angulation seen in the spine on lateral views, due to vertebral compression fractures. Other deformities such as kyphosis and scoliosis also occur.

Computed Tomography The role of CT is usually limited to guiding biopsy but CT is also useful if there are contraindications, poor patient tolerance or lack of availability of MRI. Early infection tends to show bone rarefaction and destruction. Endplate changes are more accurately evaluated than on plain radiographs. Sclerosis is seen in advanced disease.

Sagittal reconstructions of thin-section acquisition show vertebral body and endplate changes clearly. Vertebral body collapse and posterior wall retropulsion can be clearly identified even in its early stages. Spinal canal stenosis and cord compression due to the bone destruction or soft-tissue inflammatory component can be clearly identified. Intravenous enhancement increases the diagnostic accuracy, outlining the inflammatory granulation tissue and also the thick irregular wall of abscess cavities. Paraspinal cold abscesses are also better seen on CT after intravenous contrast medium. Spinal canal encroachment secondary to vertebral body destruction can be assessed. Multilevel involvement and endplate changes are also well shown on CT, which is particularly used for planned intervention of spinal tuberculosis. Common indications include drainage of cold abscesses, vertebral body or intervertebral disc biopsy.

Magnetic Resonance Imaging MRI is the investigation of choice for assessment of tuberculous vertebral osteomyelitis. Usually the affected area of the spine is imaged, but in tuberculous osteomyelitis, it may be useful to perform imaging of the whole spine to exclude skip lesions. Sagittal T1, T2 and STIR images are obtained with axial T2 images through the involved vertebral bodies. MRI can also be reliably used to differentiate between tuberculous and pyogenic vertebral osteomyelitis. Patients with tuberculous vertebral

1220

SECTION C  The Musculoskeletal System

Fig. 46.40  Vertebral osteomyelitis: gallium single-photon emission computed tomography/computed tomography (SPECT/CT) showing vertebral osteomyelitis as ‘hot spots’ of increased uptake involving the vertebral bodies. (Courtesy of Dr Ewa Novosinska, Royal Free Hospital, London, UK.)

A

D

B

E

C

Fig. 46.41  Multilevel Involvement in Tuberculous Spondylitis. (A) Plain radiograph of the thoracic spine demonstrates widened paravertebral soft tissues (arrows). (B and C) Sagittal T1 and T2 weighted images demonstrate multilevel involvement with subligamentous extension. There is relative preservation of intervertebral discs. (D) Axial T1 weighted gadolinium-enhanced image demonstrates multiple paravertebral abscesses with a thin enhancing rim and subfascial spread in keeping with typical cold abscesses (arrows). (E) Sagittal reformatted computed tomography demonstrates multiple lytic destructive lesions at the sites of bone involvement with surrounding sclerosis and a focal kyphotic deformity.

CHAPTER 46  Bone, Joint and Spinal Infections

A

B

D

E

1221

C

Fig. 46.42  Tuberculous Involvement of the Sacroiliac Joint. (A–C) Coronal T1, coronal short tau inversion recovery (STIR) and axial proton density fat-suppressed sequences demonstrating destruction of the left sacroiliac joint with a fluid collection at the posterior aspect. No significant bone marrow oedema is seen around the articulation, in keeping with chronic tuberculous involvement. The coronal STIR image also demonstrates bone marrow oedema with a small fluid collection adjacent to the left greater trochanter (arrow). This demonstrates multifocal involvement with active disease in this region. (D) Coronal reformatted computed tomography (CT) image demonstrates bony erosion of the greater trochanter with associated soft-tissue change (arrow). (E) Axial CT image demonstrates extensive destructive change involving the left sacroiliac joint with areas of calcification and sequestra formation.

osteomyelitis have a significantly higher incidence of a well-defined paraspinal abnormal signal, thin- and smooth-walled abscess cavities, particularly at paraspinal or intraosseous locations, subligamentous spread to three or more vertebral levels and involvement of multiple vertebral bodies. In tuberculous spondylitis, there is also an increased incidence of thoracic spine involvement (Fig. 46.41). Involvement of vertebral bodies and extension to the epidural and paravertebral spaces are commoner than disc space involvement. Bone marrow oedema is seen as low signal on T1 images, but the presence of corresponding high T2 signal may be variable. In paradiscal lesions, T1 weighted images show low signal with loss of height of disc spaces, with high T2 signal, endplate destruction and paraspinal abscess formation. With anterior lesions there are low T1 and high T2 signals involving the vertebral body, with preservation of disc signal and height. Central lesions involve the centre of the vertebral body and MRI shows abnormal signal of the vertebral body associated with collapse and vertebra plana, classically with preservation of adjacent discs. Whole-spine fat-suppressed STIR or T2 images are extremely useful to identify high signal bone marrow oedema in vertebral bodies and also high signal in the affected discs, which should be conspicuous against the background of very low signal from adjacent normal vertebrae.

Gadolinium-enhanced images are useful for confirmation, revealing enhancement of the bone marrow oedema in affected vertebral bodies. T1 fat-saturated images before and after enhancement are useful to identify abscesses within the vertebral bodies or paravertebral soft tissues. These are typically cold abscesses and hence do not show significant surrounding inflammatory reaction, but there is uniform enhancement of the thin walls. Enhanced images allow accurate assessment of extension and also evaluate the extent of spinal cord compression. Though posterior element involvement is not common in tuberculosis of the spine, studies have been published involving only the posterior elements. Posterior element involvement is most common in the thoracic spine, and tends to affect the lamina most commonly, followed by pedicles and articular processes. Spinal cord involvement was seen in many of these patients. Sacroiliac joint involvement is rare and evidence of calcification, sequestra and joint destruction on x-ray or CT is suggestive of tuberculous infection (Fig. 46.42).

Unusual Spine Infections Candida albicans is rare but can cause spinal infections. Aspergillus fumigatus infections of the spine are extremely rare but occur in

1222

SECTION C  The Musculoskeletal System

immunocompromised people and chronic granulomatous disease. Diagnosis is difficult without microbiology and imaging features are similar to tuberculous disease. Treatment is with antifungal agents. Atypical Mycobacterium infections with M. avium-intracellulare and pneumococcal infections have also been reported but are rare.

FURTHER READING Corr, P.D., 2011. Musculoskeletal fungal infections. Semin. Musculoskelet. Radiol. 15, 506–510. Davies, A.M., Hughes, D.E., Grimer, R.J., 2005. Intramedullary and extramedullary fat globules on magnetic resonance imaging as a diagnostic sign for osteomyelitis. Eur. Radiol. 15, 2194–2199. Diehn, F.E., 2012. Imaging of spine infection. Radiol. Clin. North Am. 50, 777–798. Donovan, A., 2008. Current concepts in imaging diabetic pedal osteomyelitis. Radiol. Clin. North Am. 46, 1105–1124. Donovan, A., Schweitzer, M.E., 2010. Use of MR imaging in diagnosing diabetes-related pedal osteomyelitis. Radiographics 30, 723–736. Greenwood, S., Leone, A., Cassar-Pullicino, V.N., et al., 2017. SAPHO and recurrent multifocal osteomyelitis. Radiol. Clin. North Am. 55, 1035–1046. Hayeri, M.R., Ziai, P., Shehata, M.L., et al., 2016. Soft-tissue infections and their imaging mimics: from cellulitis to necrotizing fasciitis. Radiographics 36, 1888–1910. Helms, C., Major, N., Anderson, M., 2009. Musculoskeletal infections. In: Helms, C., Major, N., Anderson, M. (Eds.), Musculoskeletal MRI, second ed. Saunders Elsevier, Philadelphia, PA, pp. 92–110. Hong, S.H., Choi, J.Y., Lee, J.W., et al., 2009. MR imaging assessment of the spine: infection or an imitation? Radiographics 29, 599–612. Jaramillo, D., Dormans, J.P., Delgado, J., et al., 2017. Hematogenous osteomyelitis in infants and children: imaging of a changing disease. Radiology 283, 629–643.

Lalam, R.K., Cassar-Pullicino, V.N., Tins, B.J., 2007. Magnetic resonance imaging of appendicular musculoskeletal infection. Top. Magn. Reson. Imaging 18, 177–179. Lee, K.Y., 2014. Comparison of pyogenic spondylitis and tuberculous spondylitis. Asian Spine J. 8, 216–223. Leone, A., Cassar-Pullicino, V.N., Semprini, A., et al., 2016. Neuropathic osteoarthropathy with and without superimposed osteomyelitis in patients with a diabetic foot. Skeletal Radiol. 45, 735–754. Love, C., Palestro, C.J., 2016. Nuclear medicine imaging of bone infections. Clin. Radiol. 71, 632–646. Manaster, B.J., 2013. Musculoskeletal Imaging: The Requisites, fourth ed. Mosby Elsevier, Philadelphia, PA, pp. 473–488. Math, K.R., Berkowitz, J.L., Paget, S.A., et al., 2016. Imaging of musculoskeletal infection. Rheum. Dis. Clin. North Am. 42, 769–784. Naidich, T., Castillo, M., Cha, S., et al., 2011. In: Naidich, T., Castillo, M., Cha, S., et al. (Eds.), Imaging of the Spine, first ed. Saunders/Elsevier, Philadelphia, PA, pp. 407–434. Pattamapaspong, N., Sivasomboon, C., et al., 2014. Pitfalls in imaging of musculoskeletal infections. Semin. Musculoskelet. Radiol. 18, 86–100. Resnick, D., 2002. Osteomyelitis, septic arthritis, and soft tissue infection: mechanisms and situations. In: Resnick, D. (Ed.), Diagnosis of Bone and Joint Disorders, fourth ed. W.B. Saunders Company, pp. 2377–2480. Simpfendorfer, C.S., 2017. Radiologic approach to musculoskeletal infections. Infect. Dis. Clin. North Am. 31, 299–324. Tins, B.J., Cassar-Pullicino, V.N., Lalam, R.K., 2007. Magnetic resonance imaging of spinal infection. Top. Magn. Reson. Imaging 18, 213–222. Turecki, M.B., Taljanovic, M.S., Stubbs, A.Y., et al., 2010. Imaging of musculoskeletal soft tissue infections. Skeletal Radiol. 39 (10), 957–971. Zimmerli, W., 2010. Clinical practice. Vertebral osteomyelitis. N. Engl. J. Med. 362, 1022–1029.

47  Current Status of Imaging of the Spine and Anatomical Features Thomas Van Thielen, Luc van den Hauwe, Johan W. Van Goethem, Paul M. Parizel

CHAPTER OUTLINE Anatomy, 1225 Osseous Elements, 1225 Joints, 1226 Ligaments, 1228 Neural Structures—Spinal Cord, Spinal Nerves, Dura Mater, 1228 Vascular Structures, 1229 Craniocervical Junction, 1229

ANATOMY Anatomically the spine is organised segmentally, consisting of 7 cervical, 12 thoracic, 5 lumbar, 5 (fused) sacral and 3 to 5 coccygeal vertebrae. Each level, except C1, consists of the following elements: a vertebral body (corpus vertebrae) anteriorly and a vertebral or neural arch (arcus posterior) posteriorly. Together these two structures enclose the spinal canal. Functionally the spine can be divided into three so-called columns. The anterior column includes the anterior longitudinal ligament (ALL), the anterior annulus fibrosus, and the anterior two-thirds of the vertebral body. The middle column comprises the posterior third of the vertebral body, the posterior annulus fibrosus and the posterior longitudinal ligament (PLL). The posterior column includes the posterior elements with the pedicles, facet joints, laminae, and spinous processes as well as the posterior ligaments.

OSSEOUS ELEMENTS Vertebral Body The vertebral bodies have a thin rim of cortical bone and a central framework of mostly vertically oriented trabeculae. This osseous portion contains stores of phosphate and calcium and has a structural support function. Sclerotic bands can be seen in the vertebral body at the site of fusion between two vertebral components. This is typically seen at the neurocentral junction and in the dens axis. In the dens axis there may be remnants of the subdental synchondrosis. These bony structures are optimally evaluated with computed tomography (CT) imaging and to a lesser extent with magnetic resonance imaging (MRI). On CT imaging vascular channels are often visible and in a post-traumatic setting can be mistaken for small fractures. The centre of the vertebral body is composed of red bone marrow, which is haematopoietically active. Red and yellow bone marrow are

Imaging Techniques, 1229 Plain Radiography, 1230 Myelography, 1231 Spinal Angiography, 1232 Computed Tomography, 1234 Magnetic Resonance Imaging, 1235 Single-Photon Emission Computed Tomography, 1240

not entirely homogeneous and each contains elements of the other. The vertebral marrow is dynamic, changing with age, immune state, oxygenation, coagulation, and structural needs. The normal adult distribution of bone marrow is reached by the age of 25. With ageing, the bone marrow assumes a more variable appearance, with a reduction in the red cell mass and trabecular bone and increase of the fatty content. These changes appear relatively late in comparison to the changes in the bone marrow in the peripheral skeleton. The distribution of the red bone marrow in a vertebral body is predominantly seen at the metaphyseal equivalents near the endplates and the anterior part of the vertebra. Evaluation of the bone marrow is best done with MRI. In the normal spinal marrow, the distribution patterns of fatty and red marrow were categorised into four patterns by Ricci and colleagues (Fig. 47.1). Pattern 1 describes a uniform low signal on T1 weighted images with high linear signal around the basivertebral vein; this type is most commonly seen in younger patients aged 30 or less. Type 2 is a band-like high T1 signal limited to the periphery of the vertebral body. Type 3 is characterised by multiple small indistinct (difficult to visualise) high signal intensity foci on T1 weighted images throughout the vertebral body. These two patterns (type 2 and 3) are seen with increasing age and typically in persons of 40 years and older. Type 4 is a more severe form of type 3 with multiple larger high signal intensity foci (5 to 15 mm) on T1 images throughout the vertebral body.

Neural Arch The neural arch, also known as posterior arch, forms the bony lateral and posterior border of the spinal canal (Figs 47.2–47.4). It can be divided into different segments. Between the transverse and spinous process, the neural arch is called the lamina. The pedicle is the part situated between the transverse process and the vertebral body. The pedicle of each vertebra is notched at its inferior and superior edge. Together these notches form an opening called the intervertebral foramen. Through this foramen the spinal nerves exit the spinal canal.

1225

1226

SECTION D  The Spine

C

B

A

Fig. 47.1  Age-Related Bone Marrow Changes. Sagittal T1 weighted images in a 23-year-old (A), a 44-year-old (B) and a 73-year-old (C). Normal patterns of bone marrow distribution are as described by Ricci and colleagues. Image (A) shows a type 1 bone marrow pattern with uniform low signal intensity and a high linear signal around the basivertebral vein. Image (B) shows a type 2 pattern with band-like hyperintense signal intensity limited to the periphery and image C is a type 4 marrow pattern in a 73-year-old with large hyperintense foci.

O

T

S

S

P

P L

Sp

S T

P

L

L O

I

I

O

S

Sp

T

I

P

L I

A

C

B

Fig. 47.2  Normal Imaging Anatomy of the Spine on Plain Film Radiography. Lateral (A), posteroanterior (B) and oblique (C) views. I, Inferior articular process; L, lamina; O, intervertebral foramen; P, pedicle; S, superior articular process; Sp, spinous process; T, transverse process.

Spinous and Transverse Processes The spinous process is attached to the most posterior part of the neural arch. The transverse processes arise from the lateral edge of each neural arch (see Figs 47.2–47.4). The spinous as well as the transverse processes serve as important sites of attachment for the deep back muscles. As described above, they also divide the neural arch into different anatomical parts.

JOINTS Facet Joints The facet joints or the zygapophyseal joints are diarthrodial synovial joints between the inferior and superior articular processes of adjacent neural arches. These articular processes arise from the articular pillars, including the bone at the junction between pedicles and the laminae.

1227

CHAPTER 47  Current Status of Imaging of the Spine and Anatomical Features

S

Sp

S

S

T

T P L

Sp

S

T

L3

L

L

L

S

T P

O

*

T P

Sp I

I

I C

B

A

Fig. 47.3  Normal Imaging Anatomy of the Lumbar Spine on Computed Tomography. 3D reformatted CT images of L3 in posterior (A), lateral (B) and superior (C) views. *, Spinal canal; I, inferior articular process; L, lamina; O, intervertebral foramen; P, pedicle; S, superior articular process; Sp, spinous process; T, transverse process.

disc

O L

A

S

L SP

P

O

O

O

I

*

S

S

I

I

P S

1 2 3

I

SP B

C

D

Fig. 47.4  Normal Imaging Anatomy of the Posterior Arch of the Lumbar Spine on Computed Tomography. Axial images (A, B, C) at different levels as indicated on the sagittal (D) image. *, Spinal canal; I, inferior articular process; L, lamina; O, intervertebral foramen; P, pedicle; S, superior articular process; SP, spinous process; T, transverse process. 1, Fig. 47.4A; 2, Fig. 47.4B; 3, Fig. 47.4C.

The superior articular process is located relatively anterior to the inferior articular process and faces posteriorly (see Figs 47.2–47.4). In the lumbar spine the joint surface is located obliquely at an angle of about 45 degrees between the sagittal and coronal plane. In the thoracic spine the facet joints are almost oriented in the coronal plan. The inferior facets have a convex shape while the superior articular surface has a concave aspect. In a non-degenerative spine the joint surface is covered with hyaline cartilage, being the thickest in the centre of the joint. In the normal anatomy, as in most joints, the surface of the joints should be smooth and regular with an equal spacing between the two joint surfaces. The distance between the articular processes at the facet joint should be between 2 and 4 mm on plain radiography. On the posterolateral side, the facet joint is covered with a strong fibrous capsule which is composed of several layers of fibrous tissue and a synovial membrane. On the anterior side of the joint, there is no fibrous capsule. Here, the only border between the spinal canal and the facet joint is formed by the ligamentum flavum and the synovial membrane. The capsule is composed of a superior and inferior recess containing fat pads. These fat pads act as movement-compensating mechanisms and as a lubrication mechanism for the facet joint, as they are partially covered with synovial tissue.

Uncovertebral Joints The uncovertebral joints, also known as Luschka joints, are present from C3 to C7. Rarely they are absent at C7 and sometimes they can

be seen at T1. The lateral aspect of these vertebral bodies have superior projections known as uncinate processes, and with approximation with the vertebra above, they form the uncovertebral joint. The uncinate processes are rudimentary at birth and develop and evolve with age. Their length ranges from 2 to 6 mm. The uncovertebral articulation forms the medial wall of the intervertebral foramen in the cervical region below C2 and, rarely, may be seen on the first thoracic vertebra. The function of the uncovertebral joint is to maintain stability and mobility in the cervical spine by limiting the side-to-side movement. The overlapping effect allows axial rotation and lateral bending of the cervical vertebrae. The uncinate process also protects against lateral disc herniations into the neural foramen.

Intervertebral Disc—Symphysis The intervertebral disc consists of the inner nucleus pulposus surrounded by an outer layer, the annulus fibrosus. Embryologically the nucleus pulposus is formed from cells originating from the notochord; in humans these notochordial cells are lost and replaced by chondrocyte-like cells. The nucleus pulposus is macroscopically composed of soft, elastic tissue with a yellow colour. The nucleus pulposus is primarily composed of water, proteoglycans, and loose collagen fibres. With normal ageing the water content decreases. The outer annulus fibrosus is composed of multiple concentric layers of fibrocartilage tissue. The outer layers, also called Sharpey fibres, continue in the longitudinal ligament and the vertebral bodies. The

1228

SECTION D  The Spine

fibres of each layer are directed obliquely (at a 30-degree angle), forming a meshwork. In this way a very strong flexible structure is formed.

LIGAMENTS Longitudinal Ligaments A longitudinal ligament is present at the anterior and posterior part of the vertebral bodies running along the entire spine, providing stability (Fig. 47.5). The ALL is a thick ligament which is slightly thinner at the level of the vertebral bodies and wider at the intervertebral disc. The PLL is situated in the vertebral canal and runs from the dens axis (tectorial membrane) to the sacrum. It is thicker in the thoracic

spine and wider in the cervical region compared with the lumbar level. At the level of the vertebral bodies the PLL is separated from the concave posterior wall of the vertebral bodies by the anterior epidural space. This space contains epidural fat, basivertebral veins, and the anterior internal vertebral veins.

Ligamentum Flavum The ligamentum flavum or yellow ligament is a paired structure connecting the spinal laminae forming the posterior wall of the spinal canal (Fig. 47.6). At the lateral side these structures fuse with the capsule of the facet joints, forming a boundary of the intervertebral neuro­ foramina. The boundary between the two ligamenta flava in the centre is indistinguishable on imaging. The ligamenta flava provide a static elastic force to stimulate the return to a neutral position after flexion or extension. They also limit the flexion motion of the spine and help maintain a smooth posterior lining of the central spinal canal.

Interconnecting Ligaments The posterior elements are heavily reinforced with different ligaments connecting two adjacent vertebrae. The supraspinous ligament connects the tips from the spinous processes, while the interspinous ligaments connect the base of the adjacent spinous processes (see Fig. 47.5). The transverse processes are connected by the intertransverse ligaments. As discussed earlier the laminae of the adjacent vertebra are bound together by the ligamentum flavum.

3

3

NEURAL STRUCTURES—SPINAL CORD, SPINAL NERVES, DURA MATER 4 2 1

Fig. 47.5  Normal Anatomy of the Spinal Ligaments. Sagittal T1 weighted MRI of the lumbar spine. 1, Anterior longitudinal ligament, 2, posterior longitudinal ligament, 3, interspinous ligament, 4, supraspinous ligament.

A

B

The shape of the spinal canal varies from oval on the cervical and thoracic level to a triangular shape at the lower lumbar region. Anteriorly the dura mater is in contact with the PLL over the length of almost the whole spine. Laterally the dura mater is in close contact with the medial border of the pedicles. Above and below the pedicles the dura mater is in contact with the epidural fat, in continuity with the intervertebral foramina. Posteriorly the dura mater is in contact with the posterior epidural fat separating the dura from the ligamenta flava, the posterior joints, and the laminae. In the kyphotic thoracic regions, the dura can be separated from the posterior elements by a layer of epidural fat of up to 5 mm thick. This fat should not be mistaken for epidural lipomatosis.

C

D

Fig. 47.6  Ligamentum Flavum. Axial T2 weighted (A) and T1 weighted (B) images; sagittal (C) and axial (D) reformatted CT. The ligamentum flavum or yellow ligament (indicated with arrow) is a thin ligament connecting the laminae and forms the posterior wall of the spinal canal.

CHAPTER 47  Current Status of Imaging of the Spine and Anatomical Features The spinal cord and its normal internal structure (grey and white matter) is best evaluated with MR imaging (T2). It consists of a central butterfly-shaped part which is the grey matter. The measurements of the spinal cord show a wide variation according to the patient. The conus medullaris is the terminal end of the spinal cord and is in normal anatomy found at the level L1–L2. The spinal cord tapers out into a cone at this level and the nerve roots descend, forming the cauda equina. The spinal nerves exit the spinal canal through the intervertebral neuroforamina. At the cervical and thoracic levels, the nerve roots leave the spinal canal horizontally. At the lumbar level the nerve roots first descend in the lateral canal recess before exiting through the neuroforamen.

VASCULAR STRUCTURES The spinal cord is supplied by three main arteries parallel to the spinal cord: one anterior and two posterior. The blood supply can be divided in three anatomical regions. In the cervicothoracic region the blood is supplied segmentally from arteries originating from the vertebral arteries and the great vessels of the neck (i.e. aorta, carotid and subclavian arteries). The midthoracic region receives most of its blood supply from collateral circulation from superior and inferior arteries and acts as a watershed area. The segmental blood supply is received from small perforants originating from the aorta. The thoracolumbar region is supplied by segmental arteries from the aorta and the iliac arteries. Variably originating from levels T9 to L2 and in most cases from the left side of the vertebral column, the largest vessel originates from the aorta: the artery of Adamkiewicz. The venous plexus in the spine is called the Batson plexus. This plexus is unique compared with other plexuses in the body as this venous plexus does not have valves, allowing retrograde flow in the venous network.

CRANIOCERVICAL JUNCTION The craniocervical junction (CCJ) is the region connecting a spherical and a tubular structure (head and spine). To stabilise this connection the CCJ is composed of bones, ligaments, and muscles. The bony parts are formed by the occipital bone with a basilar part, squamous part (scale), and the lateral (condylar) part. At the cervical part of the connection, the atlas (C1) consists of an anterior and posterior arch and two bulky lateral masses. The axis (C2) is formed by a vertebral body, pedicles, foramina transversaria, laminae, spinous process, and the dens axis articulating anteriorly with the arcus anterior of the atlas. The complex organisation of ligaments at the CCJ provides stability but allows a complex range of motion. These ligaments can be divided into the external and internal ligaments. The anterior external ligaments are formed by the ALL running along the anterior side of the vertebral bodies from C2 downwards and the anterior atlanto-axial ligament connecting the ALL and the arcus anterior of the atlas. The ligament running between the arcus anterior of the atlas and the os occipitale is the atlanto-occipital ligament. Posteriorly the external ligaments consist of the extension of the ligamenta flava, the posterior atlanto-occipital and the posterior atlanto-axial ligament. The ligamentum nuchae runs from the external occipital protuberance to the spinous process of C7. The internal ligaments of the CCJ consist anteriorly of a thin apical ligament and thick alar ligaments. The alar ligaments attach the axis to the skull base and run from the lateral aspect of the odontoid process to the anterior part of the foramen magnum. The middle internal ligament (cruciform ligament) is formed by the transverse ligament of C1, the accessory ligaments and the superior and inferior fibres. The

1229

transverse ligament is the largest, strongest and thickest craniocervical ligament and maintains stability by locking the anterior part of the odontoid process against the posterior side of the anterior arch of the atlas. It runs posterior to the odontoid process of the dens and attaches bilaterally on the lateral tubercle of C1. The posterior internal ligament is the tectorial membrane. Superiorly this ligament continues as the dura mater; inferiorly the ligament forms the PLL. The function of the muscles in the CCJ is to initiate and maintain movement, and not to limit the movements of this joint. They can be grouped into flexion, extension, abduction, adduction and rotation. SUMMARY BOX: Anatomy • Anatomically the spine has 7 cervical, 15 thoracic, 5 lumbar, 5 sacral and 3+5 coccygeal vertebrae, each consisting of a vertebral body and neural arch. • The centre of the vertebra is composed of bone marrow. Its signal on magnetic resonance imaging (MRI) changes with age. • The neural arch forms the lateral and posterior wall of the spinal canal. • Facet joints form together with the intervertebral disc the connection between two vertebral bodies. The morphology changes depending on the location. • Uncovertebral joints are present only in the cervical spine, maintaining stability and mobility. • The vertebral connections are heavily reinforced by ligaments: an anterior and posterior longitudinal ligament at the level of the vertebral bodies and interspinous and supraspinous ligaments as well as ligamentum flavum at the posterior elements. • The spinal cord consists of grey and white matter and is best evaluated on MRI. The spinal cord ends at the level L1–L2, tapering into a cone and descending nerve roots. • The spinal cord is supplied with blood at 3 different levels by 3 main arteries (1 anterior and 2 posterior). • The craniocervical junction is a complex structure with bones, ligaments and muscle stabilising the connection between the head (spherical) and spine (tubular).

IMAGING TECHNIQUES Multiple imaging techniques are available to evaluate the spine. The choice of imaging technique depends on the indication (e.g. traumatic or non-traumatic) and on the age of the patient. As MRI has become widely available in most countries it has become the first examination to perform in a non-traumatic setting. It has the advantage of not using X-rays, which is especially important in young patients. In older patients CT can be indicated to evaluate bony degenerative changes. If a contraindication for MRI is present, other imaging techniques (e.g. CT) are indicated. In post-traumatic imaging CT is the preferred initial imaging technique in blunt spinal trauma patients. CT is also indicated in acute trauma patients when there is no optimal visualisation of the spine on plain film and in patients with unexplained focal pain or neurological deficit with a negative plain film, if unexplained soft-tissue swelling is present, or when plain film is abnormal. Plain radiography can be performed in minor injuries. CT offers outstanding information about the bony lesions of the spine and reconstruction in virtually every anatomical plane. A group at the Harborview Medical Center in Seattle, Washington, defined a series of high-risk criteria to decide whether to perform plain radiography or CT imaging as primary imaging technique in imaging of cervical spinal trauma (Table 47.1). In the evaluation of low back pain (LBP) with or without radiculopathy, careful selection of the patients to undergo imaging is indicated.

1230

SECTION D  The Spine

TABLE 47.1  Harborview High-Risk Criteria

PLAIN RADIOGRAPHY

If yes to any criteria, there is a high risk for cervical spine injury and computed tomography is indicated. • Presence of significant head injury • Presence of focal neurological deficit(s) • Presence of pelvic or multiple extremity fractures • Combined impact of accident >50 km/h (>35 mph) • Death at the scene of the motor vehicle accident • Accident involved a fall from a height of 3 m or more

Conventional X-ray imaging is a fast, easy, and inexpensive technique which offers a good overview of a large segment of the spine. It is still widely used, but cannot be justified any longer as the definitive examination. Plain radiography still has the advantage over CT and MRI for the evaluation of structural malformations and instability with or without dynamic imaging in erect position. The main disadvantage of plain radiographs is the superimposition of soft tissue and bony structures, making the interpretation difficult. In the lateral view the cervicothoracic junction can be difficult to visualise due to superimposition of the shoulders. To overcome this problem a so-called swimmer’s view is made by elevating one arm above the head and letting the other hang down by the side. At the thoracolumbar junction, the diaphragm makes it difficult to interpret the lateral view. On the anteroposterior (AP) view, superimposition of heart and mediastinum in the thoracic spine and bowel structures in the lumbar spine offer evaluation difficulties. In a post-traumatic setting, plain radiography can be used to determine the level of the injury to perform a CT of only one specific region. In the cervical spine, studies show that up to 55% of clinically significant fractures can be missed on plain radiography. In the cervical spine, AP and lateral views are the standard views of the cervical spine. To make a good assessment of the spine, a high-quality image with visualisation of the seven cervical vertebral bodies as well as the first thoracic vertebra is necessary (Fig. 47.7). The lateral view is considered the most important view as 90% of pathology can be seen on this image. To evaluate the dens and the atlanto-axial joints an AP ‘transbuccal’ view is essential (Fig. 47.8). In post-traumatic patients with neck collars, this view can be difficult to perform; if there is any doubt a CT examination should be performed. Oblique views are performed to evaluate the neuroforamina. Two views are taken with the patient turned 45 degrees to either side. The neuroforamen viewed en face is the contralateral side to which the patient has turned the head. The need for these views can be questioned, as MRI and especially CT offer a much better evaluation of the (bony) neuroforamina and may demonstrate nerve root compression. In the thoracic spine, AP and lateral views are performed. The AP view is used to assess the pedicles (possible metastasis) and the vertebral alignment (scoliosis). When performing the lateral view, the convexity of a possible scoliosis should be turned towards the film so the divergent X-rays are more parallel to the disc spaces. Autotomography (the patient breaths gently during exposure) is used to blur out the ribs and diaphragm. In the lumbar spine, the posteroanterior (PA) image of the lumbar spine is taken from posterior to anterior as the divergent x-ray beam will be more parallel to the disc spaces (Fig. 47.9) and to lessen radiation to the organs. When taken from anterior to posterior, flexion of hips and knees will reduce the lumbar lordosis. A spot image of the L5 to S1 disc space is made from anterior to posterior with a caudocranial inclination of the tube. The hips and knees are in flexion to reduce the lumbar lordosis. As in the thoracic spine, the convexity of a scoliosis, if present, should be nearer to the radiograph. The PA oblique view is acquired by turning the patient 45 degrees; a torsion of the lumbar spine should be avoided. The oblique view of the lumbar spine is especially useful for evaluating the facet joints and pedicles; this can be used in identification of a spondylolisthesis. A full-spine AP view is used to evaluate and measure scoliosis and anatomical anomalies of the spine (Fig. 47.10). The lateral view is used to evaluate the thoracic kyphosis and lumbar and cervical lordosis. Functional (extension/flexion) lateral views can be made to evaluate instability. On these images it is important to look for displacement

TABLE 47.2  ‘Red Flags’ in Low Back Pain • Cauda equina syndrome (sudden-onset urinary retention, faecal incontinence, saddle anaesthesia) • Severe unremitting (non-mechanical) worsening of pain (at night and when laying down)—infection or tumour • Significant trauma—fracture • Unexplained weight loss, fever, HIV, or history of cancer—infection or tumour • Use of intravenous drugs or steroids—infection or compression fracture • Widespread neurological signs—tumour or neurological disease • Greater than 50 years, but particularly >65 years with first episode of severe back pain • Duration longer than 6 weeks • Prior surgery

TABLE 47.3  ‘Yellow Flags’ in Low Back

Pain—Psychosocial Barriers to Recovery That May Increase the Risk of Long-Term Disability and Work Loss • Belief that pain and activity are harmful • ‘Sickness behaviours’ • Low or negative moods, social withdrawal • Treatment beliefs do not fit best practice • Problems with claim and compensation • History of back pain, time-off, other claims • Problems at work, poor job satisfaction • Heavy work, unsociable hours (shift work) • Overprotective family or lack of support

Older guidelines discouraged imaging in uncomplicated acute ( SE > FSE. Susceptibility artefacts are directly related to field strength. The factors influencing these artefacts are composition of the material, its size and orientation, and also the external magnetic field and type of pulse sequence. In order to reduce susceptibility artefacts, GRE sequences should be avoided and, instead, SE or preferably FSE sequences should be used.

Motion Artefacts These artefacts are caused by discrete movements of the patient. In the cervical spine these artefacts can also be caused by swallowing and in

the thoracic spine by cardiac motion. Motion artefacts by swallowing or cardiac movement can be reduced by applying spatial presaturation.

Truncation Artefacts Truncation artefacts are caused by imperfection in the Fourier transformations. These bright or dark lines are seen parallel to the edges of abrupt intensity changes. These artefacts are seen more on 3T imaging than at 1.5T and are a consequence of the improved SNR at 3T. In the spine these artefacts are especially seen in the interface between the spinal cord and CSF, and are a possible cause of overestimation of spinal canal stenosis on MRI. This type of artefact is one of the causes of a band of high or low signal seen near the centre of the spinal cord in mid-sagittal images and is also the source of difficulty for defining the border between the spinal cord and CSF. Their occurrence can be reduced by increasing the spatial resolution (increasing the matrix size).

Cerebrospinal Fluid Pulsation Artefacts These artefacts are caused by the pulsatile motion of CSF during the systolic phase of the cardiac cycle. In the phase-encoding direction this motion can generate linear artefacts parallel to the interface between spinal cord and CSF, producing signal changes in the spinal cord that can be mistaken for intramedullary lesions. Especially in the thoracic spine, areas of turbulent CSF flow related to arachnoidal septa can simulate intradural masses.

1240

SECTION D  The Spine

SINGLE-PHOTON EMISSION COMPUTED TOMOGRAPHY Radionuclide bone scintigraphy with single-photon emission CT (SPECT) provides metabolic imaging and is used to detect osteoblastic activity (Fig. 47.19). In the absence of other pathology, these foci indicate areas of mechanical stress, bone repair, and/or active degenerative changes in the spine. However, with just SPECT imaging there can be difficulty in precise localisation of the spinal segment affected due to anatomical variants and low spatial resolution. The use of multimodality imaging, with a combination of SPECT and CT, has a synergetic effect. The high sensitivity of SPECT combined with the specificity of CT offers the best of both imaging techniques. CT improves the image quality of the SPECT images by correcting and adding soft-tissue attenuation. CT has, as described earlier, a higher spatial resolution of 1 mm and fusion with the SPECT images allows better anatomical localisation of the pathology. SPECT/CT is a non-invasive nuclear medicine technique to evaluate disease based on functional and metabolic information. SPECT uses a gamma camera just like conventional 2D scintigraphy, but creates

cross-sectional slices that can be reconstructed into 3D images. This is combined with a low-dose or diagnostic full-dose CT depending on the suspected pathology. Prior to the examination a small dose of radioactive tracer is injected in a vein (usually the arm). The type of radioactive tracer depends on the indication. After a certain time the patient is placed in a SPECT/ CT machine. The gamma camera is rotated around the patient through a full 360 degrees while acquiring projections every 3 to 6 degrees. Additionally, a CT examination is performed through the region of interest. Using a software algorithm both types of images are fused and a motion correction is performed if necessary. Reconstructions in axial, sagittal and coronal planes are made, as well as the classic CT reconstructions as described earlier in this chapter. SPECT/CT can be useful in the degenerative spine to determine the site of active facet joint arthropathy and/or active degenerative disc disease. CT (and MR) without SPECT is excellent for detecting degenerative disc disease and facet joint arthropathy but the addition of SPECT/ CT adds the advantage of showing the active disease, as this highlights the region of bone stress. It can also be used to differentiate acute from chronic vertebral collapse, not only showing the bone activity on SPECT

A Fig. 47.19  Hybrid Single-Photon Emission Computed Tomography Imaging. Single-photon emission computed tomography images with fusion images of the nuclear examination with computed tomography (A and B) and classic computed tomography images for diagnostic evaluation of the spine (C and D). The single-photon emission images show the zones of osteoblastic activity (yellow color) with still active bone reaction in the region of the fusion L5 to S1 (A) and increased bone reaction in the right degenerative facet joint L4 to L5 (B and D).

CHAPTER 47  Current Status of Imaging of the Spine and Anatomical Features

B

C

D Fig. 47.19, cont’d

1241

1242

SECTION D  The Spine

but also allowing for good anatomical evaluation of the fracture and stability on the diagnostic CT examination. In the postoperative spine SPECT/CT is superior to MRI or CT for evaluating loosening of the fusion material as well as the position of the material. It may also be used for detection of postoperative infections.

SUMMARY BOX: Imaging Techniques • Plain radiography can no longer be seen as a definitive examination and is replaced by MRI or CT, but has the advantage of allowing dynamic imaging in erect position. • In trauma plain radiography can be used to determine the level of the injury. Always keep in mind that up to 55% of cervical fractures are missed on plain radiography. • Spinal angiography (DSA) is still the gold standard to evaluate spinal arteries, because of the greater spatial resolution. • Although MRI has become more common, CT is still adequate for evaluation of spinal cord compression and evaluation of the posterior elements and facet joints. • In the postoperative spine CT is useful for evaluation of the position of surgical materials and loosening, while MRI is superior in detecting postoperative fibrosis. • A routine MRI protocol of the spine should include sagittal T1, sagittal T2, sagittal T2 with fat suppression (e.g. STIR) and axial T2. In the postoperative spine axial T1 before and after gadolinium (Gd) should be included. • SPECT/CT offers the combination of SPECT to detect the areas of mechanical stress, bone repair or active degenerative changes with the high spatial resolution of CT to visualise soft tissue and anatomical detail. CT, Computed tomography; DSA, digital subtraction angiography; MRI, magnetic resonance imaging; SPECT/CT, single-photon emission computed tomography; STIR, short tau inversion recovery.

FURTHER READING Alyas, F., Saifuddin, A., Connell, D., 2007. MR imaging evaluation of the bone marrow and marrow infiltrative disorders of the lumbar spine. Magn. Reson. Imaging Clin. N. Am. 15 (2), 199–219, vi.

Davis, P.C., Wippold, F.J., 2nd, Brunberg, J.A., et al., 2009. ACR Appropriateness Criteria on low back pain. J. Am. Coll. Radiol. 6 (6), 401–407. Delfaut, E.M., Beltran, J., Johnson, G., et al., 1999. Fat suppression in MR imaging: techniques and pitfalls. Radiographics 19 (2), 373–382. Harreld, J.H., McMenamy, J.M., Toomay, S.M., et al., 2011. Myelography: a primer. Curr. Probl. Diagn. Radiol. 40 (4), 149–157. Lavdas, E., Vlychou, M., Arikidis, N., et al., 2010. Comparison of T1 weighted fast spin-echo and T1 weighted fluid-attenuated inversion recovery images of the lumbar spine at 3.0 Tesla. Acta Radiol. 51 (3), 290–295. Matar, H.E., Navalkissoor, S., Berovic, M., et al., 2013. Is hybrid imaging (SPECT/CT) a useful adjunct in the management of suspected facet joints arthropathy? Int. Orthop. 37 (5), 865–870. Mathen, R., Inaba, K., Munera, F., et al., 2007. Prospective evaluation of multislice computed tomography versus plain radiographic cervical spine clearance in trauma patients. J. Trauma 62 (6), 1427–1431. Nijenhuis, R.J., Mull, M., Wilmink, J.T., et al., 2006. MR angiography of the great anterior radiculomedullary artery (Adamkiewicz artery) validated by digital subtraction angiography. AJNR Am. J. Neuroradiol. 27 (7), 1565–1572. Philpott, C., Brotchie, P., 2011. Comparison of MRI sequences for evaluation of multiple sclerosis of the cervical spinal cord at 3 T. Eur. J. Radiol. 80 (3), 780–785. Reith, W., Simgen, A., Yilmaz, U., 2012. [Spinal angiography: anatomy, technique and indications]. Radiologe 52 (5), 430–436. Salgado, R., Van Goethem, J.W., van den Hauwe, L., et al., 2006. Imaging of the postoperative spine. Semin. Roentgenol. 41 (4), 312–326. Shah, L.M., Hanrahan, C.J., 2011. MRI of spinal bone marrow: part I, techniques and normal age-related appearances. AJR Am. J. Roentgenol. 197 (6), 1298–1308. Si-jia, G., Meng-wei, Z., Xi-ping, L., et al., 2009. The clinical application studies of CT spinal angiography with 64-detector row spiral CT in diagnosing spinal vascular malformations. Eur. J. Radiol. 71 (1), 22–28. Tubbs, R.S., Hallock, J.D., Radcliff, V., et al., 2011. Ligaments of the craniocervical junction. J. Neurosurg. Spine 14 (6), 697–709. Van de Berg, B.C., Lecouvet, F.E., Galant, C., et al., 2005. Normal variants and frequent marrow alterations that simulate bone marrow lesions at MR imaging. Radiol. Clin. North Am. 43 (4), 761–770, ix. Vargas, M.I., Delavelle, J., Kohler, R., et al., 2009. Brain and spine MRI artifacts at 3Tesla. J. Neuroradiol. 36 (2), 74–81. Wu, J., Lu, L., Gu, J., et al., 2012. The application of fat-suppression MR pulse sequence in the diagnosis of bone-joint disease. Int. J. Med. Phys. Clin. Eng. Radiat. Oncol. 1 (3), 88–94. Xiong, L., Zeng, Q.Y., Jinkins, J.R., 2001. CT and MRI characteristics of ossification of the ligamenta flava in the thoracic spine. Eur. Radiol. 11 (9), 1798–1802.

48  Degenerative Disease of the Spine Paul M. Parizel, Thomas Van Thielen, Luc van den Hauwe, Johan W. Van Goethem

CHAPTER OUTLINE Introduction, 1243 Degenerative Disc Disease, 1243 Pathology of the Posterior Elements, 1255

INTRODUCTION The spine is a complex anatomical structure composed of vertebrae, intervertebral discs and ligaments. All of these structures may undergo degenerative, morphological and functional changes with age. The intervertebral discs are part of the connection between two adjacent intervertebral bodies and have two main functions: allowing movement and at the same time serving as shock absorbers. Movement at a single level is limited; the combined movement of multiple levels allows a significant range of motion. The cervical and lumbar spine compared with the thoracic spine have relatively more disc height so the motion in these segments is greater. Posteriorly, the facet joints play an important role in the cause of neck and low back pain. Facet joint syndrome is a range of symptoms that cannot be linked to a single nerve root pattern.

DEGENERATIVE DISC DISEASE Nomenclature and Classification Since the first description of a ‘ruptured disc’ with monoradiculopathy by Mixter and Barr in 1934, the terminology to grade and report degenerative disease of the spine has been controversial and confusing. Some nomenclature systems are based on description of the observed morphology of the disc contour, while others include the pathological, clinical, functional and/or anatomical features. Crosssectional imaging is based on other definitions and concepts compared with myelography or discography. In 2001 a new nomenclature was proposed by the Combined Task Forces of the North American Spine Society, American Society of Spine Radiology and American Society of Neuroradiology that consists of a classification system for the reporting on imaging studies based on pathology. In this chapter we follow the general classification of disc lesions as proposed by the Combined Task Forces. The general classification as proposed by Milette is given in Table 48.1.

Age-Related Changes in the Intervertebral Disc The Combined Task Forces reserved the term ‘normal’ for young discs that are morphologically normal, without signs of disease, trauma or ageing. However, the ‘normal’ appearance of an intervertebral disc is age related due to biochemical and anatomical changes that result in a variable appearance on magnetic resonance imaging (MRI).

Degenerative Spinal Stenosis, 1262 Degenerative Scoliosis, 1265

In infants and young children, the intervertebral disc is prominent relative to the height of the adjacent vertebral bodies. With increasing age, the relative disc volume decreases. The transition between the intranuclear cleft and the outer portion of the annulus fibrosus is sharp but becomes less distinct with age. In young adults the disc contour coincides with the margins of the adjacent vertebral end plates. On MRI the normal adult disc has a low to intermediate signal on T1 weighted images and a high signal intensity on T2 weighted images relative to the bone marrow in the adjacent vertebral bodies. On T2 the bright nucleus pulposus is indistinguishable from the inner annulus. The normal adult end plates, the outer annulus fibrosus and the ligamentous structures are relatively hypointense on T1 and T2. The outer annulus is visualised on T2 and has a low signal intensity due to a high collagen content. In young adults, diurnal changes in T2 relaxation are present but these changes disappear after the age of 35 years and that is thought to be a normal aspect of ageing. In the third decade the intranuclear cleft appears as a horizontal band of decreased signal intensity on T2 in the central part of the discs, giving it a bilocular appearance on sagittal images. The intranuclear cleft represents a fibrous transformation of the gelatinous matrix of the inner nucleus pulposus. In middle-aged and elderly persons there is a gradual signal loss of the intervertebral discs on the T2 images until the disc becomes hypointense. The loss of signal correlates to a decrease in water and proteoglycan content and an increase in collagen content. Although the decrease in T2 signal of the intervertebral disc is age related, when it is more pronounced it predisposes to degenerative changes such as loss of disc height, disc herniation and annular tears (Fig. 48.1). The highest T2 values are seen near the vertebral end plates, and the lower T2 values are present in the intranuclear cleft and the peripheral annulus fibrosus due to its fibrous nature. In general, in the normal ageing disc the height is preserved, the disc margins remain regular and radial annular tears are not a usual consequence of ageing. On the basis of a series of postcontrast MRI studies of the lumbar spine, degeneration and normal ageing have been shown to be two separate processes. Resnick and Nywayama conclude there are two different processes of degeneration: a first type, which can be considered normal ageing, involving the annulus fibrosus and adjacent ring apophysis (spondylosis

1243

1244

SECTION D  The Spine

TABLE 48.1  General Classification of

Disc Lesions

• Normal (excluding ageing changes) • Congenital/developmental variant • Degenerative/traumatic lesion • Annular tear • Herniation • Protrusion/extrusion • Intravertebral • Degeneration • Spondylosis deformans • Intervertebral osteochondrosis • Inflammation/infection • Neoplasia • Morphological variant of unknown significance

deformans) (Fig. 48.2), and a second type, called intervertebral osteochondrosis, affecting the nucleus pulposus and the vertebral end plates, corresponding to a pathological ageing process. Anterior and lateral marginal osteophytes are considered as normal ageing, whereas end plate changes and reactive bone marrow changes are considered pathological. Large amounts of gas in the central disc space, as seen on conventional radiographs or computed tomography (CT), are indicative for pathological intervertebral osteochondrosis, whereas small amounts of gas near the apophyseal enthesis should be considered as spondylosis deformans (Fig. 48.3).

Degenerative Disc Disease The prevalence of degenerative disc disease is linearly related to the age. Intervertebral disc degeneration begins already early in life. Many other factors (e.g. biomechanical and the quality of collagen) are also implicated. Degeneration includes changes involving the end plates (sclerosis, defects, Modic changes and osteophytes), as well as disc changes

A

B

Fig. 48.1  ‘Black Disc’. Sagittal T1 weighted (A) and T2 weighted images in a 34-year-old man. Decrease in T2 signal intensity at the level L4 to L5 predisposes the disc to degenerative changes.

A

B

C

Fig. 48.2  Spondylosis Deformans (‘Normal Ageing’) Versus Intervertebral Osteochondrosis (‘Abnormal Ageing’) in a 50-Year-Old Woman. Three-dimensional (A), sagittal (B) and axial at L5 to S1 (C) reformatted computed tomography images. The L5 to S1 intervertebral disc is narrowed with irregular end plates, vacuum phenomenon and concentric protrusion of the disc. At the other levels, mild spondylotic changes are seen at the adjacent ring apophysis, indicating normal ageing.

CHAPTER 48  Degenerative Disease of the Spine

A

1245

B

Fig. 48.3  Vacuum Phenomenon in a 56-Year-Old Man. Sagittal (A) and axial (B) computed tomography– reformatted images. Intradiscal gas at the levels L2–L3 to L5–S1. This so-called vacuum phenomenon is a sign of advanced degeneration.

(fibrosis, annular tears, desiccation, loss of height and mucinous degeneration of the annulus). The relation between low back pain and abnormalities in the lumbar spine is controversial since these findings are often seen in asymptomatic patients on plain radiographs, CT and MRI. Degenerative changes in the disc are already seen in one-third of healthy asymptomatic persons between 21 and 40 years old. The high prevalence of asymptomatic disc ‘degeneration’ must be taken into account when MRI is used for assessment of spinal symptoms. Although the validity of disc height as an indication for degenerative disc changes is questionable, the loss of height of the intervertebral space is the earliest sign of disc degeneration on plain radiographs. Loss of disc height has been reported in asymptomatic subjects, indicating there is no direct relationship between clinical symptoms and imaging findings. The position of the patient (lying flat or standing) should also be taken into account. Other signs including sclerosis of the vertebral end plates, osteophytes, vacuum phenomenon and calcification are more reliable indicators of pathology, although they indicate late degenerative changes. Signal loss on T2 is an early indicator of intervertebral disc degeneration on MRI. As described earlier, in normal ageing the decrease in signal intensity on T2 should be uniformly distributed over the different levels. If the signal loss is seen only in one or two levels, this should be interpreted as abnormal. This finding is often referred to as ‘a black disc’ and has been applied to describe discogenic pain syndrome (see Fig. 48.1). The degenerative process typically starts at the levels with the highest mechanical stress (motion/weight bearing). In the cervical spine levels C5 to C6 and C6 to C7 are most commonly involved and in the lumbar spine the levels L5 to S1 and L4 to L5.

Annular Tears With ageing, the intervertebral disc becomes more fibrous and less elastic. The degenerative changes are accelerated when the structural integrity of the (posterior) annulus fibrosus is damaged by overload. This will eventually lead to formation of fissures in the annulus fibrosus. In the international literature the term ‘annular tear’ is most widely used, and it is also supported by the Combined Task Forces; however, the terminology ‘annular fissure’ is also used. One should take into account that ‘annular tear’ does not imply this is caused by trauma.

Annular tears can be categorised into concentric, transverse or radial tears: • Concentric tears are circumferential lesions found in the outer layers of the annulus fibrosus. Like onion rings, they represent the splitting between adjacent layers of the lamellae annulus. They are believed to be posttraumatic from torsion overload injuries. • Transverse tears or ‘rim lesions’ are horizontal ruptures of Sharpey’s fibres near the insertion in the bony ring apophysis. The clinical significance of transverse tears remains unclear, although some authors believe they influence and accelerate degeneration and are associated with discogenic pain. They are believed to be post-traumatic in origin in some cases and are often associated with small osteophytes. • Radial tears are annular tears permeating from the deep central part of the disc and extending outwards toward the annulus in either the craniocaudal or the transverse plane (Fig. 48.4). Most of these tears do not reach the pain-sensitive outer layers of the annulus. Radial annular tears are associated with disc degeneration and a complete radial tear is necessary for progressive deterioration of the disc. The clinical significance of annular tears remains unclear. Some annular tears can cause low back pain without the presence of modification of the disc contours, also known as discogenic pain. On the other hand, annular tears are often found in asymptomatic patients and can be seen as a part of the ageing process. On MRI, annular tears can be seen as an area of high signal intensity on T2 or as foci of annular enhancement on gadolinium-enhanced T1 (Fig. 48.5). On T2 the signal intensity is the same as the adjacent cerebrospinal fluid (CSF). Repetitive microtrauma may cause annular tears to enlarge and become inflamed; this can be seen as an area of increased signal intensity. This phenomenon is seen on T2 and is known as a high-intensity zone (HIZ). The HIZ is a combination of radial and concentric annular tears which merge in the periphery of the disc. The presence of an HIZ is believed to be related to discogenic pain as it involves the outer, highly innervated layers of the annulus. However, the value of this sign is limited due to a poor sensitivity and a limited positive predictive value. On T1, extradural inflammation is seen as a zone of intermediate signal intensity replacing the fat between the disc and the dural sac; on postcontrast images there is intense enhancement.

1246

SECTION D  The Spine

A

B

Fig. 48.4  Radial Annular Tear. Sagittal (A) and axial (B) T2 weighted images. The radial tear extends to the outer rim of the annulus fibrosus (A). The axial image (B) shows that, in addition, there is a concentric tear involving the outer circumference of the annulus fibrosus. There is a loss of T2 signal and decrease in disc height at the level L5 to S1.

A

B

C

D

Fig. 48.5  Enhancing Annular Tear. Precontrast sagittal T2 weighted (A) and axial (C) T1 weighted images; postgadolinium sagittal (B) and axial (D) T1 weighted images. On the T2 weighted sagittal image a posterior annular tear and central disc herniation are seen at the level L4 to L5. After gadolinium administration, there is a linear area of enhancement in the posterior annulus, indicating a concentric tear.

In the cervical spine, annular tears, rim lesions and prolapsed disc material are poorly recognised on MRI, even in severely degenerative disc. In the thoracic spine, herniated disc fragments are often associated with abnormal straight or curvelinear densities in the intervertebral space on CT, also known as the ‘nuclear trail sign’. On MRI this finding may also be associated with a comet tail configuration in the axial plane. This sign indicates advanced disc disruption and degeneration and must be distinguished from an ageing disc that has not failed.

Disc Herniation Herniation is defined as a displacement of disc material (cartilage, nucleus, fragmented annular tissue and apophyseal bone) beyond the limits of the intervertebral disc space. The definition of the intervertebral disc space is the three-dimensional volume defined by the adjacent vertebral end plates and the outer edges of the vertebral ring apophysis, excluding osteophytes. A break in the vertebral end plates or disruption of the annulus fibrosus is thus necessary for disc displacement to occur. Disc herniations through one or both vertebral end plates are called intervertebral herniations. These herniations are also called Schmorl’s nodes and are often surrounded by reactive bone marrow changes (Fig. 48.6). One hypothesis is that this type of herniation is caused by a weak spot

in the vertebral end plate caused by regression of the nutrient vascular canals, leaving a scar. When in young individuals a herniation of the nucleus pulposus through the ring apophysis occurs before bony fusion a small segment of the vertebral rim may become isolated. This is called a limbus vertebra and is most commonly found in the lumbar region and less frequently on the midcervical level. They are characterised by a defect in the anterior wall of the vertebra and usually at the anterior superior margin in the lumbar spine and at the anterior inferior margin at the cervical level. A ‘bulging’ of the disc is defined as a circumferential or generalised disc displacement involving more than 50% of the disc circumference and is not considered as being a disc herniation (Fig. 48.7). The term ‘bulging’ is not correlated with pathology or aetiology but only refers to the morphological characteristics. A bulging can be physiologically seen on the level L5 to S1 and, on midcervical level, can reflect advanced degenerative changes, can be a pseudoimage caused by partial volume effect, can be associated with bone remodelling or occur in ligamentous laxity. There are two types of disc bulging: an asymmetrical type, as frequently seen in scoliosis, or the symmetrical type, with equal displacement of the disc in all directions. Two types of disc herniations can be differentiated on the basis of the shape of the displaced disc material (Fig. 48.8). A disc herniation

CHAPTER 48  Degenerative Disease of the Spine

A

B

Fig. 48.6  Intravertebral Herniation at L3 to L4. Sagittal T2 weighted (A), T1 weighted (B) and short-tau inversion recovery (STIR) (C) images. The images show a Schmorl’s nodule in the lower end plate of L3 without a disc herniation with surrounding bone marrow oedema which is hyperintense on T2 and STIR images. There is an intravertebral herniation located in the upper end plate of L4, with little bone marrow oedema.

C

A

1247

B

C

Fig. 48.7  Bulging Disc. Symmetrical and asymmetrical bulging disc on transverse computed tomography or magnetic resonance imaging scans. Normally the intervertebral disc (grey) does not extend beyond the edges of the ring apophyses (black line) (A). In a symmetrically bulging disc, the disc tissue extends concentrically beyond the edges of the ring apophyses (50% to 100% of disc circumference) (B). An asymmetrical bulging disc can be associated with scoliosis. Bulging discs are not considered a form of herniation (C).

A

B

C

Fig. 48.8  Disc Herniations. Types of disc herniations are seen on transverse computed tomography or magnetic resonance imaging studies. Protrusions: the base of the herniated disc material is broader than the apex. Protrusions can be broad based (A) or focal (B). Extrusion (C): the base of the herniation is narrower than the apex (toothpaste sign).

1248

SECTION D  The Spine

is called an ‘extruded disc’ when the base against the disc is smaller than the diameter of the displaced disc material, measured in the same plane, which can be either axial or sagittal (Fig. 48.9). A ‘disc protrusion’ is used when the base of the disc is broader than any other diameter of the displaced disc material (Fig. 48.10). A herniated disc can be focal if 6 cm. DLBCL: discrete mass >10 cm. Note: A PA CXR is no longer required for staging purposes. Advanced Stage III Nodal regions on both sides of the diaphragm or nodes above the diaphragm and splenic involvement. Stages IIIE and IIIS (disease below the diaphragm limited to the spleen) are no longer recognised. IV Diffuse/disseminated involvement of ≥1 extralymphatic organ ± nodal involvement or noncontiguous extranodal involvement with stage II nodal disease or any extralymphatic organ involvement in stage III disease. NB: includes any involvement of CSF, bone marrow, liver or lungs except by direct extension in stage IIE disease.c a

Waldeyer’s ring, the thymus and spleen are considered nodal or lymphatic sites b The ‘X’ subscript has been eliminated c Any liver involvement by contiguous or noncontiguous spread should be regarded as stage IV disease. ‘B’ classification: used for HL only. CT, Computed tomography; DLBCL, diffuse large B cell lymphoma; FL, follicular lymphoma; HL, Hodgkin lymphoma; PA CXR, posteroanterior chest x-ray.

• Raised ESR (>50) • Hypoalbuminaemia, anaemia • Multiple sites of disease • Bulky mediastinal disease (mass >10 cm at CT) • B symptoms Treatment is almost invariably given with curative intent and there has been a remarkable improvement in survival from HL in the past 40 years, with 5-year survival rates of over 90% for patients with earlystage disease. The choice of treatment depends predominantly on stage and the presence/absence of unfavourable prognostic factors. HL is highly radiosensitive, and in the past many patients were treated with ‘mantle’ radiotherapy, encompassing the cervical nodes, the axillae and the mediastinum down to the level of T10. However, there has been a steady trend towards the avoidance of radiotherapy in young patients because of the massive increase in secondary cancers, notably of the thyroid and breast (areas included in the radiotherapy field), and death through coronary artery disease.

Early-Stage Disease (Stages IA and IIA).  Most patients with earlystage favourable disease (non-bulky) are treated with combination chemotherapy, usually with ABVD (adriamycin, bleomycin, vinblastine and dacarbazine). The use of interim FDG PET/CT imaging may allow escalation or de-escalation of therapy, with the goal of avoiding radiotherapy in patients who have a good response to combination chemotherapy and who are therefore in a very good prognostic group. Advanced-Stage Disease (Stages IIB, IIIA/B and IVA/B).  Patients presenting with a large mediastinal mass (i.e. a mass greater than 10 cm in diameter at CT) are generally treated with more intense chemotherapy initially, so as to shrink the mass. Consolidative involved site radiotherapy may then be given. For advanced-stage disease, treatment comprises more intense combination chemotherapy, with or without subsequent consolidatory radiotherapy to sites of ‘bulky’ disease. Intensity of treatment is increasingly dictated by response as assessed with an interim FDG PET/CT scan after two cycles of treatment. With a poor response, treatment may be escalated with intensive chemotherapy regimens such as BEACOPP (bleomycin, etoposide, adriamycin, cyclophosphamide, oncovin [vincristine], procarbazine and prednisolone). Failure to achieve an initial complete or almost complete response (CR) to first-line treatment and recurrence in the first year are both associated with a poor prognosis; high-dose chemotherapy followed by autologous stem cell transplant (ASCT) is the treatment of choice. The last decade has seen the advent of many new molecularly targeted therapies such as the antibody-drug conjugate brentuximab vedotin, which yields high objective response rates and is recommended by National Institute for Health and Care Excellence (NICE) in patients who relapse after ASCT.

Non-Hodgkin Lymphoma Accurate diagnosis requires adequate tissue biopsy and an experienced histopathologist. Some lesions may be amenable to ultrasound (US) or CT-guided core-needle biopsy, which may safely yield adequate tissue for histological diagnosis and immunophenotyping. But, as with HL, an entire lymph node is preferable for diagnosis.

Clinical Features and Staging Most patients present with painless lymph node enlargement, but B symptoms are less frequent compared with HL, occurring in approximately 20%. In contradistinction to HL, the histological subtype of NHL is the major determinant of treatment rather than stage. Nonetheless, the stage of the disease has strong prognostic significance, a more advanced stage being associated with a significantly worse prognosis. As with HL, the Lugano system is now used. Around 80% of patients will have advanced disease (stage III or IV) at presentation, so all newly diagnosed patients should undergo detailed physical examination, including examination of the fauces and testes. As with HL, CT or FDG-PET/CT of the neck, chest, abdomen and pelvis is mandatory, because, unlike HL, nodal involvement is haphazard and extranodal disease is far commoner. For routinely FDG-avid NHL such as DLBCL and FL, staging FDG PET/CT is indicated. If there is evidence of bone marrow disease at FDG PET/ CT, bone marrow aspirate and trephine biopsy may not be required. For all other subtypes it is generally indicated, as FDG PET/CT can miss low volume involvement and infiltration with non FDG-avid indolent NHL. For non FGD-avid NHL, contrast enhanced CT is required and depending on the pattern of symptoms, other radiological investigations such as magnetic resonance imaging (MRI) may be indicated, especially for central nervous system (CNS) lymphoma.

Prognosis and Treatment The prognosis of NHL varies tremendously depending upon the histological subtype. In order to evaluate therapies better and to

CHAPTER 64  Reticuloendothelial Disorders: Lymphoma choose the most appropriate treatment for a given patient, various prognostic indices have been developed. The International Prognostic Index (IPI) is applicable to aggressive lymphomas such as DLBCL. The following five factors are associated with significantly inferior overall survival: • Age above 60 years • Elevated serum lactate dehydrogenase (LDH) • Performance status greater than 1 (i.e. non-ambulatory) • Advanced stage (III or IV) • Presence of more than 1 extranodal site of disease A similar prognostic index (FL-IPI) has been developed for more indolent follicular lymphoma (FL), where the important factors are considered to be the following: • Age above 60 years • Elevated serum LDH • Haemoglobin less than 12 g/dL • Advanced stage (III or IV) • More than three nodal sites of disease A recent modification (FLIPI 2) includes elevated serum β2microglobulin and longest diameter of the largest involved lymph node over 6 cm. The histological subtype determines not only the type of treatment but also when treatment should start. For asymptomatic patients with FL and no adverse features, surveillance alone may be appropriate until symptoms develop or transformation to a more aggressive DLBCL occurs. By contrast, patients with DLBCL require treatment with multiagent anthracycline-containing chemotherapy immediately. Standard treatment for DLBCL and higher-grade FL comprises cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) combined with rituximab, a chimeric monoclonal antibody against the CD20 receptor, expressed by over 95% of B-cell NHLs (R-CHOP). Many new anti-CD20 monoclonal antibodies are now available, often combined with toxins or radioactive isotopes (e.g. Bexxar or Zevalin). New immunomodulatory drugs (e.g. lenalidomide) are increasingly used to treat relapsed or refractory disease. Radiotherapy alone is considered for the small proportion of patients with stage I disease and no adverse factors.

LYMPH NODE DISEASE IN LYMPHOMA In HL, lymph node involvement is usually the only manifestation of disease, whereas in NHL nodal disease is frequently associated with extranodal involvement. There are differences in the patterns of lymph node involvement in HL and NHL at presentation. Lymph nodes tend to be larger in NHL than HL; indeed, in NS and LD HL, nodal enlargement may be minimal. Typically, involved nodes tend to displace adjacent structures rather than invade them except in the case of PMLCL (PMBL), which is characterised by local invasion of adjacent structures.

Imaging Nodal Disease At present, size is the only criterion by which lymph nodes demonstrated on CT or MRI are considered to be involved, though clustering of multiple small lymph nodes—for example within the anterior mediastinum or the mesentery—is suggestive. A maximum short-axis diameter of 10 mm is taken to be the upper limit of normal, depending upon the exact site within the neck, thorax, abdomen, or pelvis. However, in the Lugano classification, a lymph node with a longest diameter greater than 1.5 cm is considered enlarged. For FDG-avid lymphomas, a node is considered positive if there is abnormal FDG uptake regardless of size. It should be noted that normal jugulodigastric nodes can measure up to 13 mm in short-axis diameter. Nodes in the gastrohepatic ligament and porta hepatis are considered abnormal if they measure more than 8 mm in diameter; retrocrural nodes greater than 6 mm are taken as

1683

enlarged. Lymph nodes at some sites, such as the splenic hilum, presacral and perirectal areas are not usually visualised on cross-sectional imaging and, when demonstrated, are likely to be abnormal. Enlarged lymph nodes in both HL and NHL are usually homogeneous and of soft-tissue density on CT. Mild or moderate uniform enhancement occurs after intravenous injection of contrast medium. Calcification is uncommon but may be seen on post-treatment images. Necrosis is occasionally seen in large nodal masses in both HL (particularly NS HL) and aggressive NHL, more frequently after treatment. On MRI, involved lymph nodes have low-to-intermediate signal intensity on T1 weighted images; they may have very high signal intensity on fatsuppressed T2 weighted and short tau inversion recovery (STIR) sequences. Though the signal intensity of involved nodes and the presence of necrosis do not appear to have much prognostic significance, there is some evidence that within large lymphomatous masses, heterogeneous T2 signal at MRI, or heterogeneous enhancement at CT is associated with a worse outcome.

Choice of Imaging Technique FDG PET/CT is now the preferred imaging technique for the staging and restaging of FDG-avid lymphomas and has superseded the use of CT in this group. FDG PET/CT is able to detect disease in normal-sized lymph nodes and can often differentiate between nodal enlargement secondary to lymphoma or reactive hyperplasia, unlike CT imaging. Numerous studies have shown that FDG PET/CT is more accurate than CT in the depiction of nodal and extranodal disease. It results in clinically significant upstaging in up to 30% of patients compared with CT, which may result in changes in therapy, particularly in HL. Most NHLs are FDG-avid, although false-negative studies can occur with low-grade lymphomas such as chronic lymphocytic leukaemia/small lymphocytic lymphoma, mycosis fungoides and extranodal marginal zone NHL. For these subtypes, contrast-enhanced CT remains the standard of care. The development of FDG PET/CT with accurate co-registration means that both morphological and functional abnormalities can be assessed simultaneously, and this has revolutionised the staging of lymphoma. It is important to recognise that lymphomatous involvement of certain organs can be very difficult to recognise with FDG PET/CT because of physiological uptake—for example, in the stomach and CNS. Debate continues as to whether it is necessary to carry out full diagnostic CT imaging as part of the FDG PET/CT study; often a low-dose CT for the purposes of attenuation correction and anatomical correlation is sufficient, especially if a baseline CE-CT at diagnosis is non-contributory. CT usually demonstrates the full extent of disease and enables localisation of the most appropriate lesion for percutaneous imageguided biopsy. US has no value in whole-body staging. Ultrasonographic appearances of lymphomatous nodal disease are non-specific, although the pattern of vascular perfusion as demonstrated by power Doppler interrogation may suggest the diagnosis, lymphomatous nodes having rich central and peripheral perfusion. The main value of US in lymphoma lies in confirming the nature of a palpable mass and assessing the major viscera. The accuracy of MRI in detecting lymph node involvement is equal to that of CT (and is better in some areas, such as the supraclavicular fossa and within the pelvis), but it has no particular advantage over CT in this respect and its role is essentially adjunctive, to solve problems or monitor response to treatment. Recent advances in MRI technology (high-field-strength magnets and parallel imaging) have enabled MRI to be used for whole-body staging: the role of whole-body diffusionweighted imaging in staging and response assessment is a field of active research. Major advantages in patients with HL in particular (who are often young) include the lack of ionising radiation.

1684

SECTION F  Oncological Imaging

Neck Up to 80% of patients with HL present with cervical lymphadenopathy. The spread of the disease is most frequently to contiguous nodal groups, with involvement of the internal jugular chain and spread to other deep lymphatic chains in the neck. Patients with supraclavicular or bilateral neck adenopathy are at increased risk of infradiaphragmatic disease. Cervical adenopathy is less common in NHL but frequently occurs in association with disease in Waldeyer’s ring, which is counted as a lymph node for staging purposes. Approximately 40–60% of patients who present with head and neck involvement will have disseminated NHL. Involved nodal groups tend to be non-contiguous. Central necrosis within a lymph node is rarely seen compared with squamous cell carcinoma nodal metastases. Imaging with FDG PET/CT, contrastenhanced CT or MRI has a useful role in evaluating the neck in patients with lymphoma, both in identification of impalpable enlarged nodes and in response assessment, particularly in patients treated with radiotherapy, where post-treatment fibrosis renders clinical assessment difficult.

A

Thorax Intrathoracic nodes are involved at presentation in 60–85% of patients with HL and 25–40% of patients with NHL. Any intrathoracic group of nodes may be affected, but all the mediastinal sites other than paracardiac and posterior mediastinal nodes are more frequently involved in HL than NHL. Nearly all patients with NS HL have disease in the anterior mediastinum. The frequency of nodal involvement in HL is as follows: prevascular and paratracheal—84% (Fig. 64.1); hilar—28% (see Fig. 64.1); subcarinal—22% (see Fig. 64.1); others—5% (aortopulmonary, anterior diaphragmatic, internal mammary) (Fig. 64.2). In NHL, involvement of the hilar and subcarinal groups is rarer, occurring in 9% and 13%, respectively, whereas superior mediastinal nodes are involved in 35%. The great majority of cases of HL show enlargement of two or more nodal groups, whereas only one nodal group is involved in up to half of the cases of NHL. Hilar nodal enlargement is rare without associated mediastinal involvement, particularly in HL. Although paracardiac and internal mammary nodes are rarely involved at presentation in HL, they may be involved in recurrent disease. In HL and NHL, large anterior mediastinal masses usually represent thymic infiltration as well as a nodal mass (Fig. 64.3). A large anterior mediastinal mass in HL (>10 cm) is recognised as an adverse prognostic feature and, as such, defines the need for more aggressive initial therapy. FDG PET/CT or CT will demonstrate unsuspected mediastinal nodal involvement despite a normal chest radiograph in 10% of patients with HL. Impalpable axillary nodal enlargement is also frequently detected on CT in HL and NHL.

B Fig. 64.1  Anterior and Middle Mediastinal Nodal Disease. (A) Contrastenhanced computed tomography showing marked confluent enlargement of the middle mediastinal nodes, extending laterally into the aortopulmonary window and also extending into the prevascular left para-aortic region. (B) Subcarinal, bilateral hilar and para-aortic nodal involvement in the same patient.

Abdomen and Pelvis At presentation the retroperitoneal nodes are involved in 25–35% of patients with HL but up to 55% of patients with NHL. Mesenteric lymph nodes are involved in more than half the patients with NHL and less than 5% of patients with HL. Other sites, such as the porta hepatis and splenic hilum, are also less frequently involved in HL than NHL (Fig. 64.4). In HL, nodal spread is predictably from one lymph node group to another through directly connected lymphatic pathways. Nodes are frequently of normal size or only minimally enlarged. Spread from the mediastinum occurs through the lymphatic vessels to the retrocrural nodes, coeliac axis and so on. Around the coeliac axis, multiple normal-sized nodes may be seen, which can be difficult to evaluate because involved, normal-sized nodes are frequent in HL (Fig. 64.4). The coeliac axis, splenic hilar and porta hepatis nodes are involved in

Fig. 64.2  Internal Mammary Lymphadenopathy. Axial computed tomography showing marked enlargement of bilateral internal mammary lymph nodes. Note the large left axillary lymph node, small superior mediastinal nodes and a left pleural effusion.

CHAPTER 64  Reticuloendothelial Disorders: Lymphoma

1685

A A

B Fig. 64.3  Mediastinal Masses in Lymphoma. (A) Contrast-enhanced computed tomography (CT) showing an anterior mass involving the internal mammary chain in a young patient with Hodgkin lymphoma. Note the pretracheal nodal disease and an abnormal left axillary node. (B) Contrast-enhanced CT in a patient with primary mediastinal large B-cell lymphoma in the anterior and middle mediastinum. Note the pericardial involvement, compressive atelectasis of the left upper lobe and large left pleural effusion.

about 30% of patients and splenic hilar nodal involvement is almost always associated with diffuse splenic infiltration (see Fig. 64.4). The node of the foramen of Winslow (portacaval node), lying between the portal vein and the inferior vena cava is often overlooked and may be the only site of disease relapse. It has a triangular shape; its normal long-axis diameter is up to 3 cm and in the anteroposterior plane is approximately 1 cm. In NHL, nodal involvement is frequently non-contiguous and bulky. Discrete mesenteric nodal enlargement or masses may be seen with or without retroperitoneal nodal enlargement. Large-volume nodal disease in both mesentery and retroperitoneum may give rise to the so-called ‘hamburger sign’, in which a loop of bowel is compressed between two large nodal masses (Fig. 64.5). Multiple normal-sized mesenteric nodes should be regarded with suspicion for the diagnosis of lymphoma, and lymphoma is a recognised cause of the ‘misty mesentery’. Regional nodal involvement is frequently seen in patients with primary extranodal

B Fig. 64.4  Upper Abdominal Lymph Node Enlargement. (A) Contrastenhanced computed tomography (CT) showing enlarged lymph nodes in the gastrohepatic ligament. Minimal lymph node enlargement (>6 mm) is also seen in the right retrocrural region. There is also a very subtle focal splenic lesion in this patient with Hodgkin lymphoma (arrow). (B) Axial contrast-enhanced CT showing lymph node enlargement around the coeliac axis and porta hepatis. There is splenomegaly.

lymphoma involving an abdominal viscus. Involved nodes tend to enhance uniformly and the presence of multilocular enhancement should suggest an alternative diagnosis such as tuberculosis or atypical infection. In the pelvis, any nodal group may be involved in both HL and NHL. Presentation with enlarged inguinal or femoral lymphadenopathy is seen in less than 20% of HL; its presence should prompt close scrutiny of the pelvic nodal groups. In patients with massive pelvic disease, MRI is helpful for delineating the full extent of tumour and the effect on the adjacent organs.

EXTRANODAL DISEASE IN LYMPHOMA Involvement of extranodal sites by lymphoma usually occurs in the presence of widespread advanced disease elsewhere. Such secondary involvement is a recognised adverse prognostic feature in HL and NHL

1686

SECTION F  Oncological Imaging

but is much commoner in the latter. However, in approximately 35% of cases of NHL, primary involvement of an extranodal site occurs, with lymph node involvement limited to the regional lymph nodes: stages IE to IIE. Primary extranodal HL is extremely rare, and rigorous exclusion of disease elsewhere is essential before this diagnosis can be

Fig. 64.5  Extensive Mesenteric Nodal Disease. Nodal enlargement in the mesentery and the retroperitoneum in a patient with non-Hodgkin lymphoma, compressing the third part of the duodenum (arrows) and resulting in the ‘hamburger sign.’

made. The incidence of extranodal involvement in NHL depends on factors such as the age of the patient, the presence of pre-existing immunodeficiency and the pathological subtype of lymphoma. Extranodal disease is commoner in children (particularly in the gastrointestinal [GI] tract and the major abdominal viscera) and in the immunocompromised host. The high incidence of extranodal involvement in these patient groups reflects the fact that these lymphomas are usually aggressive histological subtypes. The incidence of extranodal NHL rose faster than that of nodal NHL in the 1980’s onwards, particularly in the GI tract, orbit and CNS. For example, primary lymphomas of the CNS were increasing in frequency at a rate of 10% per annum until the introduction of highly active antiretroviral therapies (cART). Of the various pathological subtypes of NHL, mantle cell (a B-cell lymphoma), lymphoblastic lymphomas (80% of which are T-cell), BL and MALT lymphomas demonstrate a propensity to arise in extranodal sites. FDG PET/CT is more sensitive than CT in the depiction of extranodal disease, chiefly because of its ability to identify splenic and bone marrow infiltration (Fig. 64.6). It can upstage as many as 40% of cases, although the CT component remains essential: for example, in low-grade lymphoma and in the lungs, where small nodules may be below the resolution of PET technology. Although contrast-enhanced CT generally performs well in the depiction of extra-nodal disease, there are certain instances where MRI or US is preferable (see further on).

Fig. 64.6  Fluoro-2-Deoxy-D-Glucose Positron Emission Tomography/Computed Tomography (FDG PET/ CT) Resulting in Upstaging. FDG PET/CT scan in a patient with follicular lymphoma (FL) transformed to diffuse large B cell lymphoma (DLBCL). The latter was upstaged from stage II on computed tomography (CT) to stage IV as additional skeletal foci of disease activity are present in the spine that are not visible on CT. Note intense nodal activity above and below the diaphragm consistent with high-grade DLBCL but also lower grade activity in several supradiaphragmatic nodes in keeping with the coexistence of FL.

CHAPTER 64  Reticuloendothelial Disorders: Lymphoma

Fig. 64.7  Lung Involvement in Recurrent Hodgkin Lymphoma. Coronal reformatted computed tomography showing multiple poorly defined irregular nodules with a bronchocentric distribution and extensive interstitial changes with thickening of the interlobular septa.

Thorax

Pulmonary Parenchymal Involvement Several categories of lung involvement can be identified, including the following: • Lymphomatous involvement associated with existing intrathoracic nodal disease • Lymphomatous involvement associated with widespread extrathoracic disease • Primary pulmonary HL • Primary pulmonary NHL Some authorities also separate out AIDS-related lymphoma (ARL) and the post-transplant lymphoproliferative disorders (PTLDs), both of which commonly affect the lungs. Lung involvement at presentation occurs in just under 4% of patients with NHL and approximately 12% of patients with HL. It is usually secondary to direct extension of nodal disease into the adjacent parenchyma, hence its paramediastinal or perihilar location. In this circumstance there is no effect on stage; the ‘E’ lesion. Patients with HL presenting with an intrapulmonary lesion in the absence of demonstrable mediastinal disease are most unlikely to have lymphomatous disease of the lung unless there has been previous mediastinal or hilar irradiation, when recurrence may be confined to the lungs. Conversely, in NHL, nodal disease is absent in 50% of those patients with pulmonary or pleural involvement. As nodal disease progresses or relapses, lung involvement becomes commoner in HL and NHL, such that 30–40% of patients with HL have pulmonary involvement at some stage during the course of the disease. The radiographic appearances are extremely variable, but the commonest pattern is of one or more discrete nodules, with or without cavitation, which tend to be less well defined than those of metastatic carcinoma, which they otherwise resemble (Figs 64.7 and 64.8). The disease often spreads along lymphatic channels and involves lymphoid follicles around bronchovascular divisions, resulting in peribronchial nodulation spreading out from the hila, which can

1687

Fig. 64.8  Pulmonary Involvement in a Patient With Hodgkin Lymphoma. Computed tomography, performed at the time of presentation, showing a large poorly defined cavitating nodule in the right middle lobe.

result in streaky shadowing visible on chest radiographs and at CT (see Fig. 64.7). Less commonly, lymphomatous cells fill the pulmonary acini, producing rounded or segmental areas of consolidation with air bronchograms (see Fig. 64.8). Nodulation along the bronchial wall may enable differentiation from infective consolidation. A rare pattern of disease is widespread interstitial reticulonodular shadowing, producing a lymphangitic picture. Another rare manifestation is atelectasis, which usually results from endobronchial lymphoma rather than extrinsic compression by nodal disease. The differential diagnosis of pulmonary involvement in lymphoma is extensive and includes drug-induced changes, the effect of radiotherapy, and opportunistic infection during or following chemotherapy, particularly in patients with antecedent immunosuppression. Precise clinical correlation is essential in determining the most likely diagnosis.

Primary Pulmonary Lymphoma Primary pulmonary lymphoma accounts for less than 1% of all lymphomas and is usually low-grade B-cell NHL, arising from MALT or bronchus-associated lymphoid tissue (BALT). BALT lymphomas tend to occur in the fifth to sixth decades, have an indolent course with 5-year survival rates over 80% and tend to remain extranodal, although lymph node involvement can occur with advanced disease. Many patients will have a previous history of inflammatory or autoimmune disease, such as Sjögren’s syndrome and collagen vascular disease. The imaging findings are nonspecific. Solitary or multiple nodules or one or more rounded or segmental areas of consolidation are seen. These can persist unchanged for long periods. Ill-defined alveolar opacities are a feature and pleural effusions occur in up to 20% of cases. Variable but generally increased FDG uptake is described at PET/CT. In the remaining 15–20% of patients, primary lung lymphoma is due to high-grade NHL, usually DLBCL. The most common finding on a chest radiograph is of solitary or multiple pulmonary nodules, which characteristically grow rapidly. Chest wall and nodal involvement occurs more frequently than with pulmonary MALT-type lymphomas. Primary pulmonary HL is extremely rare. The most frequently described finding is single or multiple nodules with upper-zone predominance

1688

SECTION F  Oncological Imaging disease is usual and up to 40% have superior vena caval obstruction, which is rare with other lymphomas. On CT, differentiation of enlarged mediastinal lymph nodes from thymic involvement is often not possible. On MRI, the gland is often of mixed signal intensity similar to that of involved nodes. Cystic change within the gland can be recognised at CT and MRI with PMBL and HL. These cysts can persist or even increase in size following regression of the rest of the involved gland with successful treatment. Calcification may be present at the outset or may develop during treatment. Benign thymic rebound hyperplasia can develop after completion of chemotherapy, and can be difficult to differentiate from recurrent disease, although the typical triangular shape of the gland may help. Functional imaging with FDG PET/CT may not always differentiate between recurrence and rebound hyperplasia; clinical correlation combined with follow-up studies may be necessary.

Chest Wall

Fig. 64.9  Pleural Disease in Lymphoma. Computed tomography showing the typical appearance of pleural involvement in a patient with non-Hodgkin lymphoma. There is irregular pleural thickening with an accompanying pleural effusion on the right. Notice also bilateral paravertebral disease and nodal enlargement in the right internal mammary chain.

In HL, spread into the chest wall usually occurs by direct infiltration from an anterior mediastinal mass, especially from the internal mammary chain. However, chest wall masses can arise de novo in NHL. Bony destruction is rare and should suggest an alternative diagnosis. Chest wall disease is better shown by MRI than CT, particularly on T2 weighted or STIR sequences, where there is excellent contrast between the mass and normal low signal intensity muscle. FDG PET/CT can also demonstrate chest wall involvement.

Breast and a relatively high incidence of cavitation. Pulmonary DLBCL and HL are typically FDG avid.

Pleural Disease Pleural effusions are usually accompanied by mediastinal lymphadenopathy (Fig. 64.9) and may be detected on CT in 50% of patients with mediastinal nodal disease. They are usually exudates secondary to central lymphatic or venous obstruction and therefore clear promptly with treatment of the mediastinal disease. Pulmonary involvement need not be present. Focal pleural masses can occur at presentation but are more commonly seen in recurrent disease, when they are generally accompanied by an effusion. The rare primary effusion lymphoma is a form of DLBCL characterised by lymphoma of mesothelial spaces and occurs in association with HIV infection; infection with human herpesvirus-8 is a prerequisite for the diagnosis.

Pericardium and Heart Direct pericardial and cardiac involvement can occur with high-grade peripheral T-cell and large B-cell lymphomas. It is rare at presentation except in patients with ARL and PTLD who may present with acute onset of heart block, tamponade, or congestive cardiac failure. Pericardial effusions occur in 6% of patients with HL at the time of presentation and are associated with large masses adjacent to the heart. They are also common in primary mediastinal large B-cell lymphoma (PMBL) (Fig. 64.3B). Effusions are regarded as evidence of pericardial involvement, although this does not alter disease stage. Small pericardial effusions of uncertain aetiology are often seen at CT during treatment. They usually resolve with time, although some pericardial thickening may persist.

Thymus Thymic involvement by HL in association with mediastinal nodal disease occurs in around 30% of patients at presentation, and the thymus is regarded as a lymph node for the purposes of staging. PMBL characteristically involves the thymus, occurring typically in young women between the ages of 25 and 40 years (Fig. 64.3B). Rapidly growing bulky

Lymphoma of the breast is usually associated with widespread disease elsewhere. There may be multiple nodules, with associated large-volume adenopathy. Primary NHL of the breast is rare, accounting for approximately 2% of all lymphomas and under 1% of all breast malignancies. The age distribution is bimodal, with the first peak occurring during pregnancy and lactation, often high-grade or BL and not infrequently affecting both breasts diffusely with an inflammatory picture at US and mammography. There is a second peak at around 50 years when patients present with discrete masses, which are usually solitary, but multiple masses occur and disease is bilateral in over 10% of cases. The masses are usually fairly well defined, with little accompanying architectural distortion. Calcification is not a feature. Lesions enhance markedly at dynamic contrast-enhanced MRI and are FDG avid. In the past few years it has been recognised that T cell lymphoma can develop in the fibrous capsule that surrounds breast implants. The usual presentation of this breast implant-associated anaplastic large cell lymphoma is with delayed onset of a seroma around the implant.

Hepatobiliary System and Spleen Liver

Liver involvement is present in up to 15% of adult patients with NHL at presentation. This figure is higher in the paediatric population and in recurrent disease. In HL, liver involvement occurs in about 5% of patients at presentation, almost invariably in association with splenic HL. Pathologically, diffuse microscopic infiltration around the portal tracts is the most common form of involvement. CT and MRI are therefore insensitive in the detection of liver involvement. However, hepatomegaly strongly suggests the presence of diffuse infiltration (in contradistinction to the significance of splenomegaly). Larger focal areas of infiltration are present in only 5–10% of patients with hepatic lymphoma and are typically FDG avid. Cross-sectional imaging may demonstrate miliary nodules or larger solitary or multiple masses resembling metastases (Fig. 64.10) and with entirely non-specific features on all forms of cross-sectional imaging. At MRI, as with metastases, deposits have moderate T2 hyperintensity. Superparamagnetic iron oxide particles and hepatocyte-specific contrast agents can increase the

CHAPTER 64  Reticuloendothelial Disorders: Lymphoma

1689

A

Fig. 64.10  Lymphomatous Infiltration of the Liver. There are multiple poorly defined low-density lesions in the liver in this patient with transformed chronic lymphatic leukaemia and bowel involvement (arrow).

conspicuity of focal deposits but are not required in clinical practice. Occasionally, especially in children, periportal infiltration is manifest as periportal low-attenuation tissue at CT (Fig. 64.11). True primary hepatic lymphoma, indistinguishable radiologically from hepatocellular carcinoma, is rare, but the incidence is rising, up to 25% of affected patients being positive for hepatitis B or C. NHL of the bile ducts and gallbladder is rare but occurs with relatively high frequency in patients with ARL.

Spleen The spleen is involved in 30–40% of patients with HL at the time of presentation, usually in the presence of nodal disease above and below the diaphragm (stage III). In a small proportion, however, it is the sole focus of intra-abdominal disease (also designated stage III, as the spleen is considered to be nodal in the Cotswolds and Lugano staging classifications). In most patients, the involvement is microscopic and diffuse and thus particularly difficult to identify on cross-sectional imaging. Splenomegaly is an unreliable sign of involvement; 33% of patients have splenomegaly without infiltration and, conversely, 33% of normal-sized spleens are found to contain tumour following splenectomy. Measurements of splenic volume and splenic indices are not generally utilised. Focal splenic deposits occur in only 10–25% of cases and may be demonstrated by any form of cross-sectional imaging, including FDG PET/CT, when they are more than 1 cm in diameter (see Fig. 64.4B). Up to 40% of patients with NHL have splenic involvement at some stage. Imaging findings include a solitary mass, miliary nodules or multiple masses, all of which have a non-specific appearance. The differential diagnosis of multiple masses includes opportunistic infection and granulomatous disease. In early studies, the sensitivity of US and CT for the detection of splenic involvement was low (about 35%). Detection of small nodules has improved with the advent of contrast-enhanced multidetector CT (MDCT) with optimisation of splenic parenchymal opacification. MRI

B Fig. 64.11  Periportal Lymphoma. (A) Contrast-enhanced computed tomography in a child with non-Hodgkin lymphoma showing infiltration of low-density lymphomatous tissue from the porta hepatis, encasing the main portal vein and extending alongside the right portal vein. A solitary focal abnormality is seen posteriorly within the liver. (B) Follow-up after chemotherapy shows complete resolution of the disease.

with superparamagnetic iron oxide may improve diagnostic accuracy but is seldom undertaken outside the research arena. However, FDG PET/CT can detect splenic disease more accurately than either CT or MRI. It may present as homogeneous splenomegaly, focal lesions or a single mass. The sensitivity of PET/CT, together with the development of highly effective combination chemotherapy, had led to the abandonment of staging laparotomy with splenectomy in HL. Primary splenic NHL is rare, accounting for 1% of all patients with NHL. Patients present with splenomegaly, often marked, and focal masses are usual. Splenic involvement is also a particular feature of certain other pathological subtypes of NHL, such as mantle cell lymphoma and splenic marginal zone lymphoma. Splenomegaly is a reasonably reliable indicator of involvement with NHL and a craniocaudal splenic length of more than 13 cm at CT is taken to indicate splenomegaly. Infarction is a well-recognised complication.

Gastrointestinal Tract The gastrointestinal (GI) tract is the commonest site of primary extranodal NHL, accounting for 30–45% of all extranodal presentations

1690

SECTION F  Oncological Imaging

Fig. 64.12  Gastric Lymphoma. Axial contrast-enhanced dedicated gastric computed tomography study showing circumferential gastric antral mural thickening (arrow).

and constituting about 1% of all GI tumours. It is the initial site of lymphomatous involvement in up to 10% of all adult patients and up to 30% of children. As elsewhere, primary HL of the GI tract is exceptionally rare. Secondary involvement of the GI tract via direct extension from involved mesenteric or retroperitoneal lymph nodes is extremely common; consequently multiple sites of involvement occur. Primary lymphomas arise from lymphoid tissue of the lamina propria and the submucosa of the bowel wall; they occur most frequently below the age of 10 years (usually BL) and in the sixth decade (MALT type and enteropathy-associated T-cell type). Primary GI lymphoma is usually unifocal. Accepted criteria for the diagnosis of primary disease include the following: • No superficial or intrathoracic lymph node enlargement • No involvement of the liver or spleen • A normal white cell count • No more than locoregional lymph node enlargement. A modified Lugano staging system has been proposed, where stage I disease is confined to the bowel wall and stage II disease denotes involvement of draining lymph nodes or adjacent organs. In both primary and secondary cases, the stomach is most frequently involved (50%), followed by the small bowel (35%) and large bowel (15%).

Stomach Primary lymphoma accounts for about 2–5% of all gastric tumours. It originates in the submucosa, affecting the antrum more commonly than the body or cardia. Radiologically the appearances reflect the gross pathological findings; common appearances are multiple nodules, some with central ulceration, or a large fungating lesion with ulceration. About one-third of patients have diffuse infiltration, with marked thickening of the wall and narrowing of the lumen, sometimes with extension into the duodenum, indistinguishable from linitis plastica. Only about 10% are characterised by diffuse enlargement of the gastric folds, similar to the pattern seen in hypertrophic gastritis (Fig. 64.12). As the disease originates in the submucosa, the signs described here are best demonstrated on barium studies or endoscopically, but CT better reflects the true extent of gastric wall thickening and accompanying nodal involvement. Typically, infiltration of adjacent organs is unusual but may occur in DLBCL. In gastric MALT lymphomas, mural thickening may be minimal; CT is of limited value even with dedicated studies,

Fig. 64.13  Mucosa-Associated Lymphoid Tissue Lymphoma. Endoscopic ultrasound showing a narrow sheet of low echogenic tissue in the submucosa (arrowed). (Image courtesy of Dr A. McLean, Department of Diagnostic Imaging, St Bartholomew’s Hospital, London.)

and endoscopic US with biopsy is more useful in staging, prognostication and assessment of response (Fig. 64.13). Low-grade MALT lymphoma is more likely to cause shallow ulceration and nodularity, whereas high-grade lymphoma can produce massive gastric infiltration and polypoid masses. Most primary gastric lymphomas are FDG avid, although low-grade MALT shows variable avidity.

Small Bowel Lymphoma accounts for up to 50% of all primary tumours of the small bowel, occurring most frequently in the terminal ileum and becoming progressively less frequent proximally; duodenal lymphomas are rare (Fig. 64.14). In children, the disease is almost exclusively ileocaecal. Most bowel lymphomas are of B-cell lineage. The disease is multifocal in up to 50% of cases; mural thickening with constriction of bowel segments is typical. Patients commonly present with obstructive symptoms. Bowel wall thickening is well demonstrated on CT (Fig. 64.15; see also Fig. 64.10). With progressive tumour spread through the submucosa and muscularis mucosa, aneurysmal dilation of long segments of bowel can develop, presumably due to infiltration of the autonomic plexus. Alternating areas of dilation and constriction are a common manifestation of infiltration and are well demonstrated by CT. If lymphomatous infiltration is predominantly submucosal, multiple nodules or polyps of varying size result, mostly in the terminal ileum. It is this form of lymphoma that typically causes intussusception, usually in the ileocaecal region; lymphoma is the commonest cause of intussusception in children older than 6 years. Polypoidal disease is a particular feature of mantle-cell lymphoma. Enteropathy-associated T-cell lymphoma and immunoproliferative small intestinal disease (alpha-chain disease or IPSID) commonly present with clinical and imaging features of malabsorption, but acute presentations with perforation are common. Often the whole small intestine is affected, especially the duodenum and jejunum. In the small bowel (and colon), MALT lymphoma is manifest as mucosal nodularity, which

CHAPTER 64  Reticuloendothelial Disorders: Lymphoma

1691

Colon and Rectum Primary colonic lymphomas are usually of Burkitt or MALT subtypes, but account for under 0.1% of all colonic neoplasms, most arising in the caecum and rectum (see Fig. 64.15). The most common pattern of disease is a diffuse or segmental distribution of small nodules 0.2 to 2.0 cm in diameter, typically with intact mucosa. A less common form of the disease is a solitary polypoid mass, often in the caecum, indistinguishable from carcinoma on imaging unless there is concomitant involvement of the terminal ileum. In advanced disease, there may be marked thickening of the colonic or rectal folds, resulting in focal strictures, fissures or ulcerative masses with fistulation. Lymphomatous strictures are generally longer than carcinomatous strictures and irregular excavation of the mass strongly suggests lymphoma. Involvement of the anorectum is a feature of ARL. Patients usually present with obstruction and rectal bleeding.

Oesophagus Involvement of the oesophagus is extremely unusual and begins as a submucosal lesion, usually in the distal third of the oesophagus. Ulceration is a later phenomenon. Secondary involvement by contiguous spread from adjacent nodal disease is more common but rarely results in dysphagia. Fig. 64.14  T-cell Lymphoma Involving the Small Bowel. There are multiple loops of abnormally thickened small bowel in the pelvis on this coronal reformatted computed tomogram with a large necrotic mesenteric nodal mass (arrow).

Pancreas Primary pancreatic lymphoma accounts for only 1.3% of all pancreatic malignancies and 2% of patients with NHL. It usually presents with a solitary mass, often in the head of the pancreas, indistinguishable from primary adenocarcinoma on US, CT, or MRI. Biliary or pancreatic ductal obstruction can occur. Calcification, necrosis and vascular effacement are rare compared with pancreatic adenocarcinoma. Less commonly, diffuse uniform enlargement of the pancreas is seen. Involvement is far more common in NHL than in HL. Secondary pancreatic involvement usually results from direct infiltration from adjacent nodal masses, either focal or massive.

Genitourinary Tract The genitourinary tract is not commonly involved at the time of presentation (mediastinum but ≤liver. 4. Moderately increased uptake compared with the liver. 5. Markedly increased uptake compared with the liver >×3 and/or new lesions. X. New areas of uptake unlikely to be related to lymphoma. CR: • Score 1, 2, or 3 with or without a residual mass. • No evidence of FDG-avid disease in marrow. PR: • Score 4 or 5 with reduced uptake compared with baseline and residual mass(es) of any size. • No new lesions. • Residual bone marrow uptake higher than normal marrow but reduced compared with baseline. SD: • Target nodes or extranodal lesions score 4 or 5 with no significant change in FDG uptake from baseline at interim or end of treatment. • No change in bone marrow uptake from baseline. • No new lesions. PD: • Score 4 or 5 in target lesions/nodes with increase in intensity of uptake from baseline and/or new FDG-avid foci consistent with lymphoma at interim or end of treatment. • For new lesions: exclude other aetiologies such as infection/ inflammation. Consider biopsy or interval scan if uncertain. CR, Complete remission; FDG, fluoro-2-deoxy-D-glucose; PD, progressive disease; PR, partial response; SD, stable disease.

and peripheral T-cell lymphomas. They provide a graded scale so that different thresholds can be used to define scan positivity according to the clinical context or research question. These provide the basis of the Lugano classification where Deauville scores 1 and 2 indicate a CMR even if a residual mass is present. Deauville score 3 usually also indicates a CMR when standard therapy has been used. However, in response adapted de-escalation studies where under-treatment needs to be avoided, Deauville score 3 may be defined as a positive scan. Deauville scores 4 and 5 with reduced uptake compared to a pre-treatment scan indicate a partial metabolic response (PMR) on an interim scan but persistent disease at the end of treatment. Deauville scores 4 and 5 with an increase in uptake and/or new lymphoma lesions compared to baseline indicate progressive metabolic disease (PMD) on interim scans or at the end of treatment.

FDG-PET/CT in Response Assessment In the FDG-avid lymphomas, FDG PET/CT is more accurate than CT in response assessment, especially for patients with a radiological (CT) PR, with a PPV at least double that of CT. In one follow-up study, all PET-positive/CT-negative patients relapsed, whereas only 5% of PETnegative/CT-positive patients relapsed (see Fig. 64.29). FDG PET/CT is now routinely used for response assessment in HL and aggressive B-cell NHL. In addition to allowing de-escalation of therapy and more selective use of radiotherapy in HL, iPET response is prognostic, with large differences in progression-free survival (PFS) between those with a good

1702

SECTION F  Oncological Imaging

response compared to those with a poor PET response in HL and DLBCL. If iPET confirms a CMR, then guidelines suggest that an end of treatment scan is not required unless there are worrying clinical features. An end of treatment FDG PET/CT is now the standard of care and has been reported to show high NPVs of 90–100% in HL and 80–100% in aggressive NHL. The PPV is lower, especially in NHL (50–100%), as inflammatory cells also show FDG accumulation. It is recommended that all end-of-treatment FDG PET/CT scans be performed at least 6 weeks after completion of chemotherapy to minimise false-positive inflammatory activity.

Treatment Planning With FDG PET/CT There is evidence that FDG PET/CT has prognostic value after salvage chemotherapy and before ASCT with high-dose chemotherapy, with PFS being shorter in those with positive scans. This indication is being used more frequently in centres with transplantation services. As well as selecting patients who may omit radiotherapy in HL, there is increased interest in using FDG PET/CT to modify disease volumes when planning radiotherapy and the role of PET in defining radiotherapy volumes is under active investigation.

SURVEILLANCE AND DETECTION OF RELAPSE Relapse after satisfactory response to initial treatment occurs in 10–30% of patients with HL and up to 50% of patients with NHL. For curable lymphomas such as HL and DLBCL, relapse usually occurs within the first 2 years after treatment and patients are followed up closely during this period, although imaging is not required unless clinical features suggest the possibility of recurrence. For patients who attain a CR/CMR, there is very little evidence for routine surveillance with imaging. A number of studies have shown that relapse is rarely identified by conventional imaging before patients become symptomatic. Functional imaging is able to identify early relapse before CT and, indeed, before the development of clinical signs. However, there is as yet little evidence for the efficacy of FDG PET/CT in this role. In one series of a cohort of patients treated for HL, relapses were identified by FDG PET/CT before there was any other evidence of relapse, but the false-positive rate was high. In another study it was concluded that there was no benefit from surveillance studies for HL or aggressive NHL beyond 18 months. True positive images in the absence of clinical suspicion of relapse are rare and the false positive rate of FDG PET/CT is at least 20%. On the other hand, in clinically suspected relapse, the development of a positive FDG PET/CT scan is highly suggestive, and in this situation FDG PET/CT is likely to have a significant therapeutic impact, allowing image-guided biopsies which can target the most metabolically active lesion and thereby direct therapy by establishing relapse or transformation.

CONCLUSION Management of patients with lymphoma depends heavily on the imaging findings, which are vital in initial staging of the disease, prognostication and monitoring response to treatment. The radiologist needs to understand the fundamental aspects of tumour behaviour and must appreciate the factors that will influence therapy. The radiological report should document the number of sites of nodal disease; the presence and sites of bulky disease; the presence of extranodal disease; and factors which may influence delivery of therapy, such as central venous thrombosis. FDG PET/CT has revolutionised the imaging of lymphoma, providing unprecedented insight into the functional behaviour of this diverse group of tumours. Recent research has clearly defined the precise roles of FDG PET/CT and its place in the investigative algorithm with

ongoing research likely to define the role further. For all of these reasons, the radiologist has become a pivotal member of the multidisciplinary team managing patients with lymphoma.

FURTHER READING Barrington, S.F., Johnson, P.W.M., 2017. (18)F-FDG PET/CT in Lymphoma: has imaging-directed personalized medicine become a reality? J. Nucl. Med. 58 (10), 1539–1544. Barrington, S.F., Kirkwood, A.A., Franceschetto, A., et al., 2016. PET-CT for staging and early response: results from the Response-Adapted Therapy in Advanced Hodgkin Lymphoma study. Blood 127 (12), 1531–1538. Barrington, S.F., Kluge, R., 2017. FDG PET for therapy monitoring in Hodgkin and non-Hodgkin lymphomas. Eur. J. Nucl. Med. Mol. Imaging 44 (Suppl. 1), 97–110. Barrington, S.F., Mikhaeel, N.G., Kostakoglu, L., et al., 2014. Role of imaging in the staging and response assessment of lymphoma: consensus of the International Conference on Malignant Lymphomas Imaging Working Group. J. Clin. Oncol. 32 (27), 3048–3058. Cheson, B.D., Fisher, R.I., Barrington, S.F., et al., 2014. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J. Clin. Oncol. 32 (27), 3059–3068. Chua, S.C., Rozalli, F.I., O’Connor, S.R., 2009. Imaging features of primary extranodal lymphomas. Clin. Radiol. 64 (6), 574–588. Elstrom, R., Guan, L., Baker, G., et al., 2003. Utility of FDG-PET scanning in lymphoma by WHO classification. Blood 101 (10), 3875–3876. Gallamini, A., Barrington, S.F., Biggi, A., et al., 2014. The predictive role of interim positron emission tomography for Hodgkin lymphoma treatment outcome is confirmed using the interpretation criteria of the Deauville five-point scale. Haematologica 99 (6), 1107–1113. Illidge, T., Specht, L., Yahalom, J., et al., 2014. Modern radiation therapy for nodal non-Hodgkin lymphoma-target definition and dose guidelines from the International Lymphoma Radiation Oncology Group. Int. J. Radiat. Oncol. Biol. Phys. 89, 49–58. Johnson, P., Federico, M., Kirkwood, A., et al., 2016. Adapted treatment guided by interim PET-CT scan in advanced Hodgkin’s Lymphoma. N. Engl. J. Med. 374 (25), 2419–2429. Kwee, T.C., Kwee, R.M., Nievelstein, R.A., 2008. Imaging in staging of malignant lymphoma: a systematic review. Blood 111, 504–516. Lin, C., Luciani, A., Itti, E., et al., 2012. Whole-body diffusion magnetic resonance imaging in the assessment of lymphoma. Cancer Imaging 12, 403–408. Moog, F., Bangerter, M., Diederichs, C.G., et al., 1998. Extranodal malignant lymphoma: detection with FDG-PET versus CT. Radiology 206, 475–481. Pregno, P., Chiappella, A., Bello, M., et al., 2012. Interim 18-FDG-PET/CT failed to predict the outcome in diffuse large B-cell lymphoma patients treated at the diagnosis with rituximab-CHOP. Blood 119, 2066–2073. Radford, J., Illidge, T., Counsell, N., et al., 2015. Results of a trial of PET-directed therapy for early-stage Hodgkin’s lymphoma. N. Engl. J. Med. 372 (17), 1598–1607. Regacini, R., Puchnick, A., Shigueoka, D.C., et al., 2015. Whole-body diffusion-weighted magnetic resonance imaging versus FDG-PET/CT for initial lymphoma staging: systematic review on diagnostic test accuracy studies. Sao Paulo Med. J. 133, 141–150. Smeltzer, J.P., Cashen, A.F., Zhang, Q., et al., 2011. Prognostic significance of FDG-PET in relapsed or refractory classical Hodgkin lymphoma treated with standard salvage chemotherapy and autologous stem cell transplantation. Biol. Blood Marrow Transplant. 17, 1646–1652. Specht, L., Yahalom, J., Illidge, T., et al., 2014. Modern radiation therapy for Hodgkin lymphoma: field and dose guidelines from the International Lymphoma Radiation Oncology Group (ILROG). Int. J. Radiat. Oncol. Biol. Phys. 89, 854–862. Stewart, B.W., Wild, C.P. (Eds.), 2014. World Cancer Report 2014. IARC, Lyon. Swerdlow, S., Campo, E., Harris, N.L., et al. (Eds.), 2017. Who Classification of Tumours of Haematopoietic and Lymphoid Tissues, Revised fourth ed. IARC, Lyon.

65  Bone Marrow Disorders: Haematological Neoplasms Asif Saifuddin

CHAPTER OUTLINE Myeloproliferative Neoplasms, 1703 Polycythaemia Vera, 1703 Primary Myelofibrosis, 1703 Haematologic Malignancies, 1705 Leukaemia, 1705 Lymphoma, 1706 Primary Bone Lymphoma, 1709 Plasma Cell Dyscrasias, 1710 Solitary Plasmacytoma of Bone, 1711 Multiple Myeloma, 1712

Chapters 65 and 66 deal with various haematological conditions, both neoplastic and non-neoplastic, that may be manifest on imaging of the skeletal system. The recently updated 2016 World Health Organization (WHO) classification of myeloproliferative neoplasms is extremely complex and includes a variety of conditions of differing malignant potential. Only those that have any significant radiological manifestations in the skeletal system will be discussed.

MYELOPROLIFERATIVE NEOPLASMS POLYCYTHAEMIA VERA Polycythaemia vera (PV) is a myeloproliferative neoplasm characterised primarily by erythrocytosis. The WHO diagnosis of PV includes major and minor criteria, the 3 major being: (1) raised haemoglobin levels (>16.5 g/dL for men and >16.0 g/dL for women or haematocrit >49% in men and >48% in women), (2) bone marrow biopsy showing hypercellularity of erythroid, granulocytic and megakaryocytic elements (pan-myelosis) with pleomorphic mature megakaryocytes, and (3) the presence of JAK2 V617ZF or JAK2 exon 12 mutation. The minor criterion is reduced erythropoietin level.

Clinical Features PV has an incidence of 1.5 to 2.0/100,000 and presents at a mean age of 50 to 55 years, being twice as common in females. Typical clinical manifestations include thrombosis, neurological symptoms, pruritis, plethora, and splenomegaly. Management includes regular phlebotomy to maintain a haematocrit of less than 45%, daily low-dose aspirin, and in high-risk patients, cytoreductive therapy. Median survival is approximately 14 years, and both fibrotic and acute myeloid leukaemic transformation are recognised, the latter in just under 10% of patients at 20 years post-diagnosis.

Miscellaneous Conditions, 1714 Systemic Mastocytosis, 1714 The Histiocytoses, 1715 Langerhans Cell Histiocytosis, 1715 Erdheim–Chester Disease, 1720

Imaging Findings As with any condition resulting in marrow hyperplasia, magnetic resonance imaging (MRI) will demonstrate diffuse reduction of T1 weighted marrow signal and hyperintensity on short tau inversion recovery (STIR), although focal marrow abnormality may also be seen. The progression to acute myeloid leukaemia (AML) may not be evident, unless the patient develops a myeloid sarcoma (see later). Non-skeletal manifestations include splenomegaly and signs of thrombosis.

PRIMARY MYELOFIBROSIS Primary myelofibrosis (PMF) is a myeloproliferative neoplasm of unknown aetiology in which clonal proliferation of haematopoietic stem cells results in progressive bone marrow fibrosis, with consequent anaemia, splenomegaly, and extramedullary haematopoiesis (EMH). The diagnosis is based on demonstration of fibrosis at bone marrow biopsy, and is supported by a variety of genetic abnormalities, including JAK2 in 55%. The blood film may show pancytopenia with abnormal red blood cell morphology. Secondary myelofibrosis may also occur as a late complication of PV or, less commonly, essential thrombocytopenia, and this may be accelerated by drugs used to treat these conditions.

Clinical Features PMF has an incidence of 0.5 to 1.5/100,000 and affects men and women equally, presenting at a median age of 50 to 60 years. Clinical findings include weakness, dyspnoea and weight-loss due to progressive obliteration of the marrow by fibrosis or bony sclerosis, which leads to a moderate anaemia. There is also increased risk of infection and gout. EMH takes place in the liver and spleen, which become enlarged in 72% and 94% of cases respectively, but is also reported in lymph nodes, lung, choroid plexus, kidney, etc. The natural history is one of slow deterioration,

1703

1704

SECTION F  Oncological Imaging

A

B

D

C

with death typically occurring 2 to 3 years after diagnosis, although potential cure can be achieved with allogenic stem cell transplantation. Progression to leukaemia is also a feature.

Imaging Features Bone sclerosis is the major radiological finding, being evident in approximately 30% to 70% of cases. Typically, this is diffuse, but occasionally patchy (Fig. 65.1A), occurring most often in the axial skeleton and major long bone metaphyses. Sclerosis is due to trabecular and endosteal new bone (Fig. 65.1B), resulting in reduced marrow diameter. In established disease, lucent areas are due to fibrous tissue reaction. Periosteal reaction occurs in one-third of cases, most often

Fig. 65.1  Myelofibrosis. (A) Anteroposterior radiograph of the pelvis showing patchy sclerosis in the iliac blades, pubic bones, and proximal femora. Lytic destruction of the left anterior iliac crest is consistent with malignant transformation (arrow). (B) Axial computed tomography of the pelvis showing trabecular thickening resulting in marrow sclerosis. (C) Sagittal T1 weighted SE and (D) axial T2 weighted FSE magnetic resonance images showing diffuse marrow hypointensity. SE, spin echo; FSE, fast spin echo.

in the medial aspects of the distal femur and proximal tibia. The skull may show a mixed sclerotic and lytic pattern, and bone scintigraphy can demonstrate a ‘superscan’ appearance. MRI typically demonstrates marrow hypointensity on both T1 weighted and T2 weighted sequences, which may be diffuse or heterogeneous (Fig. 65.1C and D). Additional features include arthropathy due to haemarthrosis and secondary gout, occurring in 5% to 20% of cases. Infiltration of the synovium by bone marrow elements may result in polyarthralgia and polyarthritis. Leukaemic conversion may manifest radiologically by the development of an extraosseous soft-tissue mass (Fig. 65.1A). MRI may also show regression of marrow changes following successful stem cell transplantation.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

1705

while MRI shows masses of heterogeneous T1 weighted and T2 weighted signal intensity (Fig. 65.2B). The presence of fat signal intensity may suggest that the masses are inactive, while enhancement may suggest that they are active.

SUMMARY BOX:  Myeloproliferative Neoplasms • Polycythaemia vera and primary myelofibrosis • Clinical features include presentation in adults with splenomegaly and extramedullary haematopoiesis (EMH) • Radiographs may show marrow sclerosis in myelofibrosis • Magnetic resonance imaging demonstrates features of marrow hyperplasia/ fibrosis with diffuse or heterogeneous reduction of T1 weighted marrow SI • EMH manifests as hepatosplenomegaly, thoracic paravertebral masses, rib enlargement, and pre-sacral masses • May transform to leukaemia SI, signal intensity.

A

HAEMATOLOGIC MALIGNANCIES LEUKAEMIA Leukaemia accounts for approximately 30% of all childhood malignancy, the vast majority being the acute form, which arises from immature leukocytes. Acute lymphocytic leukaemia (ALL) accounts for 75%, AML for 20%, and other types for 5% of cases. Chronic leukaemias arise from mature cell lines and predominate in adults, but sometimes terminate in an acute blastic form.

Clinical Features ALL usually presents in children at 2–3 years of age, while AML is most commonly seen in the first 2 years of life. The acute disease is often insidious, with nonspecific malaise, anorexia, fever, petechiae, and weight loss. Limb pain and pathological fracture are common. Bone pain at presentation is five times more common in children than adults, being reported in over 33% of cases. Adults are most commonly affected by chronic lymphatic leukaemia (CLL) and AML, with a reported incidence of 3–4%. Most skeletal lesions in adults affect sites of residual red marrow, the axial skeleton and proximal ends of the femora and humeri. CLL is a disease of the elderly, characterised by enlargement of the spleen and lymph nodes with skeletal involvement being rare (Fig. 65.3), except as a terminal event.

Imaging Features B Fig. 65.2  Myelofibrosis. Extramedullary haematopoiesis. (A) Axial computed tomography study through the lower sacrum demonstrating lobular pre-sacral soft tissue masses containing areas of fat density (arrows). (B) Sagittal T2 weighted FSE magnetic resonance image of the lumbosacral junction demonstrating heterogeneous lobular pre-sacral soft tissue masses (arrows).

PMF is the commonest cause of EMH, usually resulting in hepatosplenomegaly. However, EMH may also present in the thorax as bilateral paravertebral masses or marked enlargement of the ribs, or as a pre-sacral mass. These lesions appear on computed tomography (CT) as non-calcifying soft tissue masses containing areas of fat (Fig. 65.2A),

Radiological evidence of bone involvement in acute childhood leukaemia is reported in approximately 40% of patients at presentation, the incidence of the various features being as follows: osteolysis (13.1%), metaphyseal bands (9.8%), osteopenia (9%), osteosclerosis (7.4%), permeative bone destruction (5.7%), pathological fracture (5.7%), periosteal reaction (4.1%) and mixed lytic–sclerotic lesions (2.5%). Such changes will also be seen in up to 75% of children during the course of their disease. Diffuse osteopenia is reported in 16–41% of cases and may be metabolic in aetiology, or due to diffuse marrow infiltration. The effects of corticosteroids and chemotherapy also contribute to osteoporosis. Compression fractures, including vertebra plana, occur in association with osteopenia of the spine. Metaphyseal lucent bands primarily affect the distal femur, proximal tibia and distal radius, with other metaphyses and the vertebral bodies affected later. They are typically 2–15 mm in width (Fig. 65.4A). These changes may also be seen in generalised infection, typically in infancy; in children over the age of 2 years, leukaemia is more likely.

1706

SECTION F  Oncological Imaging

A

deposits. Non-specific cortical destruction may involve the medial aspect of the proximal humerus, tibia and sometimes femur. In adults, skeletal lesions are less common and must be differentiated from metastases, or a primary malignant bone neoplasm. MRI typically demonstrates diffuse reduction of T1 weighted marrow signal intensity (Fig. 65.5A) and increased signal intensity on fat suppressed T2 weighted/STIR sequences (Fig. 65.5B), or less commonly patchy marrow infiltration (Fig. 65.5C), with marrow enhancement following contrast (Fig. 65.5D and E). A change from normal to nodular to diffuse low signal intensity can be seen with disease progression, together with an increase in the extent of signal intensity abnormality, but there is no consensus on the use of MRI for the staging of leukaemia. Response to therapy is also demonstrable with disease remission, being associated with a return to normal marrow signal intensity. Bone marrow relapse may manifest as well-defined nodules, and relapse may also occur in extra-medullary sites such as skeletal muscle. MRI can identify complications of treatment such as osteonecrosis (Fig. 65.6), which is reported in 6.5–15% of cases. Marrow necrosis may also occur due to microvascular occlusion, and manifests as a geographic area of peripheral marrow enhancement with a non-enhancing centre. Marrow necrosis may heal. Myeloid sarcoma manifests on MRI as an extra-osseous mass which is isointense on T1 weighted but hyperintense to muscle on PDW/ T2 weighted (Fig. 65.7), and which enhances avidly following contrast. Focal or diffuse increased uptake on fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT may be seen in leukaemia, but is nonspecific, since abnormal activity may also be seen following chemotherapy or the use of granulocyte colony–stimulating factor (GCSF). SUMMARY BOX:  Leukaemia

B Fig. 65.3  Chronic Lymphatic Leukaemia. (A) Sagittal T1 weighted SE and (B) axial T2 weighted FSE magnetic resonance images showing heterogeneous vertebral marrow SI and large extradural and pre-sacral lymphomatous masses (arrows), with common iliac lymphadenopathy (arrowhead-B).

More extensive involvement results in diffuse, permeative bone destruction (Fig. 65.4B), like that seen in Ewing sarcoma. The cortex becomes eroded on its endosteal surface and may ultimately be destroyed. The permeative pattern is reported in 18% of leukaemic children. Osteolytic lesions secondary to bone destruction typically have a motheaten appearance and most commonly affect long bone metaphyses. Such lesions are reported in 10–40% of patients and predispose to pathological fracture. A particular focal lesion most commonly seen in AML is myeloid sarcoma (granulocytic sarcoma, chloroma), occurring in 2.5–9.1% of cases. It is usually located in the skull, spine, ribs, or sternum of children, and is an expanding tumour caused by a collection of leukaemic cells (Fig. 65.4C). Osteosclerosis is rare, being reported in 6% of cases. Sclerotic changes in the metaphyses of long bones may occur spontaneously, or as a result of therapy. Mixed lytic-sclerotic lesions are identified in around 18% of children. Periosteal reaction is reported in 2–50% of cases and may occur in isolation or in association with destructive cortical lesions. It is due to sub-periosteal haemorrhage or proliferation of leukaemic

• Approximately 13 of childhood malignancy, usually ALL (75%) or AML (20%) • Radiological abnormality in 40% at presentation; osteolysis, metaphyseal bands, osteopenia, osteosclerosis, bone destruction, pathological fracture • Magnetic resonance imaging demonstrates diffuse reduction of T1 weighted marrow SI with corresponding hyperintensity on short tau inversion recovery (STIR) • Skeletal complications include osteonecrosis • May develop a focal mass of leukemic cells termed myeloid sarcoma ALL, Acute lymphocytic leukaemia; AML, acute myeloid leukaemia.

LYMPHOMA Lymphoma comprises a heterogeneous group of malignant neoplasms accounting for 4% to 6% of all cancers. The updated 2016 WHO classification includes mature B-cell neoplasms, mature T and NK (natural-killer) neoplasms, Hodgkin lymphoma, post-transplant lymphoproliferative disorders (PTLD), and histiocytic and dendritic cell neoplasms. However, for simplicity the terms Hodgkin and non-Hodgkin lymphoma (NHL) will continue to be used. These neoplasms primarily arise in extra-skeletal locations. Lymphoma of bone accounts for 7% of all bone malignancies and 5% of extra-nodal lymphomas and can be divided into primary bone lymphoma (PBL), multifocal PBL, or disseminated lymphoma with secondary osseous involvement, the latter usually being due to hematogenous spread or by direct extension from surrounding involved lymph nodes or soft tissues. Osseous lymphoma is usually due to NHL, with secondary skeletal involvement implying stage IV disease. It also occurs in the late stages of Hodgkin disease. Musculoskeletal soft tissues (subcutaneous fat, tendons, ligaments) may also be involved by direct extension from primary cutaneous lymphoma, which represent approximately 4% of all NHL cases, while

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

A

1707

C

B

Fig. 65.4  Leukaemia. (A) Acute leukaemia. Anteroposterior (AP) radiograph of the right ankle showing metaphyseal lucent bands (arrows). (B) Acute leukaemia. AP radiograph of the left knee showing permeative bone destruction in the distal femur and proximal tibia. (C) AP radiograph of the forearm showing destruction of the ulna with a surrounding soft tissue mass (arrows) due to myeloid sarcoma.

lymphoma may present as an extremity lymph node mass or in muscle, mimicking a soft tissue sarcoma.

Clinical Features Hodgkin disease (HD) is characterised by the Reed–Sternberg cell, which is usually a B-cell. Bone marrow involvement is detected in 5-14% of patients with HD and indicates stage IV disease, while presentation with a skeletal lesion is very rare. Three-quarters of patients present between 20 and 30 years of age, with a second peak occurring after the age of 60 years, while HD accounts for 3% of paediatric malignancy. A slight male predominance (1.4 : 1) is found and the spine is the commonest site of osseous involvement, from either direct lymph node extension or haematological spread. The pelvis, ribs, femora, and sternum are the other commonly involved sites. HD presents as PBL in 6% of cases. NHL is the commonest haematopoietic neoplasm and comprises a large variety of clinicopathological subtypes, the majority of which are of B-cell origin. The commonest cell-type overall is diffuse large B-cell lymphoma (DLBCL). Predisposing factors include AIDS and transplantrelated immunosuppression. The median age at diagnosis is 66 years, while in children NHL accounts for 5% of all malignancies. Skeletal involvement may be either primary or secondary, and bone marrow involvement is detected in 25–40% of patients with high-grade NHL.

Imaging Features Approximately one-third of skeletal lesions in HD are solitary, and whole-body staging with PET/CT or whole-body MRI is mandatory to identify further lesions. Osteolytic lesions or mixed lytic–sclerotic lesions account for almost 90% of cases, the remainder being purely sclerotic. In the spine, HD most commonly causes sclerosis, although it is an uncommon cause of ‘ivory’ vertebra. Vertebral collapse occurs early with lytic lesions, occasionally producing vertebra plana. Erosion of the anterior border of one or more vertebral bodies is also found, possibly by direct spread from affected pre-vertebral lymph nodes and is best demonstrated with MRI. In the thoracic region, paravertebral masses may precede radiographic evidence of bony involvement, while mediastinal disease is a recognised cause of hypertrophic osteoarthropathy. Rib involvement is common, with multiple lytic lesions associated with soft-tissue masses predominating. Occasionally, the ribs appear expanded. In the pelvis, involved nodes may invade the bone directly, usually in the posterior half of the ilium. Direct haematogenous involvement of the sternum is also common. The typical lesion is lytic and expanding, associated with a soft-tissue mass. CT clearly demonstrates all the described radiological features, and in addition will show the associated extra-osseous mass. Sequestrum formation is also a recognised feature of bone lymphoma (Fig. 65.8). Marrow involvement is most sensitively detected by MRI and may be

1708

SECTION F  Oncological Imaging

A

B

D

C

focal or diffuse. However, the signal intensity (SI) characteristics are nonspecific. In focal involvement, MR can be used to guide biopsy. The radiological features of NHL are similar to those described for HD. Children with generalised NHL tend to have widespread skeletal involvement manifesting as osteopenia. MRI may show multifocal marrow lesions (Fig. 65.9) before the diffuse infiltration of established disease (Fig. 65.10A and B). MRI may also be used to stage disease and will occasionally demonstrate marrow disease despite negative iliac crest biopsy. Although the MRI appearances are nonspecific, the presence of a large soft-tissue mass arising from a flat bone, or a paravertebral mass

E

Fig. 65.5  Leukaemia. (A) Coronal T1 weighted SE and (B) axial fat suppressed proton density weighted (PDW) FSE magnetic resonance (MR) images of the pelvis showing diffuse reduction of T1 weighted marrow SI (arrows) and corresponding hyperintensity on the fat suppressed sequence (arrows). (C) Coronal short tau inversion recovery MR image of the pelvis showing patchy marrow changes and areas of marrow infarction (arrows). (D) Sagittal T1 weighted SE and (E) post-contrast T1 weighted SE MR images of the thoracolumbar spine showing diffuse reduction of marrow SI with uniform mild enhancement.

with maintenance of the cortical outline of the vertebral body, is highly suggestive of lymphoma. PET/CT is sensitive in the detection of marrow involvement with 94% of all lymphomas being FDG avid, and often up-stages both HD and NHL compared to conventional CT. The uptake may be focal or multifocal (Fig. 65.11). Whole body MRI/diffusion weighted imaging (DWI) has similar sensitivity and specificity to PET/CT for detection of FDG-avid lymphomas such as HD, DLBCL and follicular lymphoma, but outperforms PET/CT in those lymphomas which show variable FDG-avidity.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

1709

Fig. 65.8  Lymphoma. Axial computed tomography of the pelvis showing extensive destruction of the left ilium with a large circumferential extraosseous mass and multiple bone sequestra (arrows).

Fig. 65.6  Leukaemia. Coronal short tau inversion recovery magnetic resonance image of the right shoulder showing osteonecrosis of the humeral head with a sub-chondral fracture (arrow), and a further marrow infarct in the scapula (arrowhead).

Fig. 65.7  Myeloid Sarcoma. Axial PDW FSE magnetic resonance image of the forearm showing extra-osseous extension of tumour from the ulna (arrows).

PRIMARY BONE LYMPHOMA PBL is defined as isolated lymphomatous involvement of bone with or without regional lymph node involvement, and no distant site of disease detected within 6 months of presentation. Multifocal PBL is less common and represents involvement of two or more bony sites with no extraosseous or nodal disease within 6 months of diagnosis.

Fig. 65.9  Lymphoma. Primary multifocal osseous lymphoma. Sagittal short tau inversion recovery magnetic resonance imaging of the spine showing multilevel vertebral (arrows) and sternal (arrowhead) marrow infiltration.

patients present in the fifth to sixth decades of life (mean age 42 years) with bone pain and/or a soft-tissue mass, and the disease is approximately twice as common in men.

Clinical Features

Imaging Features

PBL accounts less than 2% of all lymphomas in adults, 3–9% of all NHL in children, and is usually a DLBCL or follicular lymphoma. Most

Approximately 70% of PBL cases involve the major long bones (femur, tibia, humerus), usually arising in the metadiaphysis or diaphysis.

1710

SECTION F  Oncological Imaging

A

B Fig. 65.10  Lymphoma. Primary multifocal osseous lymphoma. (A) Coronal T1 weighted SE and (B) short tau inversion recovery (STIR) magnetic resonance imaging showing diffuse reduction of T1 weighted and increased STIR marrow SI.

However, epiphyseal lesions with joint involvement are also recognised, and the flat bones and spine can be involved. Disease limited to the marrow space may be radiologically occult, but the majority result in moth-eaten or permeative bone destruction (Fig. 65.12A), with only 2% being osteoblastic. Pathological fracture occurs in 17–22% of cases. An aggressive periosteal reaction is seen in 50% and approximately 50–75% of patients will have an associated soft-tissue mass, which is optimally demonstrated by MRI and is indicative of more aggressive disease with a worse prognosis. MRI signal characteristics are nonspecific, but relatively low T2 weighted SI is a recognised feature (Fig. 65.12B and C). Multifocal PBL accounts for 11–33% of cases and more commonly involves the spine, the imaging characteristics being as described above.

SUMMARY BOX:  Lymphoma • Osseous lymphoma accounts for 7% of bone malignancy and 5% of extra-nodal lymphoma • Hodgkin and non-Hodgkin (commonest diffuse large B-cell) • Primary bone lymphoma defined as isolated bone involvement +/− regional lymph node (LN) disease with no distant site detected with 6/12 • Radiologically results in moth-eaten/permeative bone destruction, or rarely osteosclerosis • A large soft tissue mass is common

Fig. 65.11  Lymphoma. Coronal fused positron emission tomographycomputed tomography image showing multifocal areas of fluorodeoxyglucose avidity (arrows).

Multifocal disease can be detected by whole-body scintigraphy, wholebody MRI, or positron emission tomography-computed tomography (PET-CT; Fig. 65.11). Diffuse skeletal involvement results in a generalised reduction of T1 weighted SI with increased marrow SI on STIR. Enlarged regional lymph nodes may also be seen.

PLASMA CELL DYSCRASIAS Plasma cell dyscrasias occur due to the clonal proliferation of plasma cells of B-cell origin in the bone marrow. This results in the overproduction of immunoglobulins which can result in varied systemic effects. Several conditions are recognised, including monoclonal gammopathy of unknown significance (MGUS), asymptomatic (smouldering) myeloma, solitary plasmacytoma, and symptomatic multiple myeloma (MM). MGUS is an asymptomatic pre-malignant condition seen in greater than 3% of the general population over 50 years of age, and is associated with a 1% per year risk of progression to MM. Smoldering myeloma is an intermediate stage between MGUS and symptomatic MM, and is differentiated from MGUS by an increased level of circulating serum M-protein and increased percentage of plasma cells in the bone marrow. It is consequently associated with a higher rate of progression to MM, estimated at 10% in the first 5 years. Solitary plasmacytoma accounts

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

A

B

1711

C

Fig. 65.12  Primary Bone Lymphoma. (A) Anteroposterior radiograph of the knee showing moth-eaten bone destruction in the distal femoral metaphysis and epiphysis (arrows). (B) Coronal T2 weighted FSE and (C) axial short tau inversion recovery magnetic resonance images showing relatively hypointense marrow replacement (arrow-B) and popliteal lymph node involvement (arrow-C).

Fig. 65.13  Plasmacytoma. (A) Anteroposterior radiograph of the right hip showing a proximal femoral lytic lesion (arrows) with pathological fracture. (B) Sagittal computed tomography multi-planar reconstruction (MPR) showing expansion of the 8th rib (arrow).

B

A

for approximately 5% of plasma cell dyscrasias and represents a focal clonal proliferation of plasma cells in either bone or soft tissue, without evidence of significant bone marrow plasma cell infiltration. Symptomatic MM is diagnosed in the presence of myeloma-related organ/tissue impairment, including hypercalcaemia, renal insufficiency, anaemia, bone lesions, and amyloidosis. In addition to the above, the rare association of polyneuropathy, organomegaly, endocrinopathy, elevated monoclonal immunoglobulin levels, and skin changes (POEMS syndrome) is recognised. Imaging with low-dose whole body CT, whole body/spinal MRI or PET/CT may be warranted in high-risk MGUS, but definitely should be performed in smouldering myeloma to look for the development of asymptomatic bone lesions. The identification of more than one focal bone lesion on MRI is considered diagnostic of MM and an indication for treatment. Patients with smouldering myeloma are also at increased risk of fractures and should be assessed for osteoporosis with dual-energy x-ray absorptiometry (DEXA) scanning.

SOLITARY PLASMACYTOMA OF BONE Clinical Features Solitary plasmacytoma of bone (SPB) has an incidence of approximately 0.15 cases/100,000 and is twice as common in males. Presentation is with bone pain or pathological fracture and the age of presentation tends to be earlier than in MM. It tends to involve the axial skeleton, with the commonest sites being the vertebrae, pelvis, ribs, upper limbs, face, skull, femur and sternum. The median survival of patients with SPB is 10 years, with just over 50% of cases progressing to MM.

Imaging Features Plasmacytoma is typically lytic and destructive, with sclerosis being rare. The lesion arises in the medullary cavity and the radiological features suggest a relatively slow growth rate. The margin is often welldefined, and cortical thinning with bone expansion is usually present (Fig. 65.13A and B). Apparent trabeculation or a ‘soap bubble’ appearance

1712

SECTION F  Oncological Imaging

A

A

B

B Fig. 65.14  Plasmacytoma. (A) Axial fused positron emission tomographycomputed tomography image showing high fluorodeoxyglucose avidity in a lytic lesion of the left posterior ilium (arrows). (B) Whole body coronal positron emission tomography multiple intensity projection (MIP) image showing additional lesions in the lower thoracic spine (arrows). (Images courtesy Dr Simon Wan)

is common, and an associated soft-tissue mass is frequently seen. The lesion has no specific MRI signal characteristics and diagnosis is made by image-guided percutaneous needle biopsy. Following the diagnosis of plasmacytoma, whole spine and pelvic MRI is indicated to identify additional lesions, which may be seen in approximately 1 3 of cases by screening the thoracic and lumbar spine. Additional lesions may also be demonstrated by PET-CT (Fig. 65.14A and B). Involvement of a vertebral body may lead to pathological collapse, and extension across the disc space is a rare feature which may aid in the diagnosis. SPB in the spine may show characteristic CT and MRI

Fig. 65.15  Plasmacytoma. (A) Coronal computed tomography MPR showing prominent vertical trabeculation within the vertebral body. (B) Axial T1 weighted SE magnetic resonance image showing thickened trabeculae extending into the vertebral body, resulting in a ‘mini-brain’ appearance.

features, particularly the presence of peripheral thickened trabeculae which protrude into the medullary cavity, resulting in a ‘mini-brain’ appearance (Fig. 65.15A and B).

MULTIPLE MYELOMA MM is the most common primary malignant neoplasm of bone and is the predominant plasma cell neoplasm, accounting for approximately 1.3% of all malignant disease and 15% of haematological malignancies.

Clinical Features Three-quarters of affected patients are over 50 years of age (median age 66 years) and approximately 3% of patients present before the age of 40 years. There is a male predominance of up to 2 : 1. Widespread

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms involvement of the skeleton is present in 80%, the axial skeleton and proximal ends of the long bones being most commonly involved. Fever, bone pain, backache, and weakness are common symptoms, and amyloidosis is reported in approximately 20% of patients. MM may also be discovered after blood tests or radiographic examination for other conditions, and in 1 3 of cases presentation may be due to pathological fracture, usually in the axial skeleton.

Imaging Features The classical radiographic appearance of MM consists of well-defined ‘punched-out’ lesions throughout the skeleton, most characteristic in the skull (Fig. 65.16A), but areas of moth-eaten or permeative bone destruction may also be seen (Fig. 65.16B). The only common differential diagnosis in this age group is metastatic disease. The presence of multiple

A

B Fig. 65.16  Myeloma. (A) Multiple myeloma. Lateral radiograph of the skull showing multiple small lytic lesions (arrows). (B) Anteroposterior radiograph of the left shoulder showing moth-eaten destruction of the proximal humerus (arrow).

1713

small (up to 20 mm), well-defined round or oval lesions is more suggestive of MM. Diffuse osteopenia usually involves the spine (Fig. 65.17A) and may result in multiple compression fractures (see Fig. 65.17B). Pathological fracture of the vertebrae affects approximately 50% of patients at some stage. Osteoblastic or mixed lesions are rare in untreated patients, but marginal sclerosis may be observed following radiotherapy. Purely sclerotic myeloma is recognised, and may be associated with POEMS syndrome (Fig. 65.18A and B). Low-dose whole-body CT (Fig. 65.19) and PET/CT (Fig. 65.20A and B) are far more sensitive than radiographic skeletal survey in detecting bone lesion. On CT, purely marrow lesions appear as focal areas of soft-tissue density, but the diffuse osteopenia of MM may be indistinguishable from other causes such as osteoporosis. Progressive disease results in endosteal scalloping, cortical destruction and soft-tissue masses. MRI is also more sensitive than radiography for detection of marrow involvement, although rib and skull metastases may be better appreciated on radiography. The MRI appearances in MM are variable, with five patterns described. (1) A normal marrow pattern may be seen in patients with low-grade plasma cell infiltration of the marrow, and occasionally in stage III disease. (2) A focal pattern consists of localised areas of decreased signal intensity on T1 weighted images with corresponding increased signal intensity on T2 weighted and STIR images which are greater than 5 mm in diameter (Fig. 65.21A), reported in 18–50%. Occasionally, focal lesions may be relatively hyperintense on T1 weighted and identified only on fat-suppressed T2 weighted images. (3) The diffuse pattern manifests as generalised reduction of marrow SI on T1 weighted, such that the intervertebral discs appear hyperintense compared with the vertebral bodies (Fig. 65.21B). The marrow appears hyperintense on fat-suppressed T2 weighted and STIR images and shows diffuse enhancement following gadolinium, these features being indicative of a high tumour burden and reported in 25–43%. (4) A combined focal and diffuse pattern may also be seen. (5) Finally, a ‘variegated’ pattern, which consists of multiple tiny foci of reduced signal intensity on T1 weighted (Fig. 65.21C) and hyperintensity on T2 weighted/STIR on a background of normal marrow is described, occurring in 1–5%. This pattern is almost always seen in early disease. These patterns have some prognostic value, in that patients with diffuse marrow abnormality on MRI will have a poorer outcome than those with a normal MRI pattern. Vertebral fractures in MM occur in 55–70% of cases and may be benign, due to diffuse osteopenia (∼66%), or pathological due to tumour infiltration (∼33%), the majority occurring in the thoracolumbar region (Fig. 65.17). MRI is valuable in the differentiation of benign versus malignant collapse. MRI may also demonstrate other complications, such as steroid-induced marrow infarction. Whole-body MRI is more sensitive than skeletal survey and wholebody CT for the detection of lesions and can result in a significant change in staging of the disease. The Durie and Salmon PLUS staging system for MM now includes whole-body imaging information from MRI and/or PET/CT, with the disease stage depending upon the number of bone lesions demonstrated. MRI may also be more sensitive than PET/CT, which may fail to identify areas of diffuse marrow involvement. MRI allows assessment of response to therapy, with a normalisation of marrow signal intensity being indicative of a good response (Fig. 65.22A and B). Partial response may manifest as a change from diffuse involvement to the focal or variegated pattern. However, MRI may also show persistent nonviable lesions which typically show fluid signal intensity characteristics and no enhancement or rim enhancement, or may develop a fatty halo (Fig. 65.22C and D). Functional imaging techniques such as DWI and PET/CT are also of value in assessing response and recurrence. DWI may demonstrate an increase in ADC values, while PET-CT may show a reduction of FDG uptake with successful treatment (Fig. 65.23A) and increased uptake with relapse (Fig. 65.23B).

1714

SECTION F  Oncological Imaging

A

B

Fig. 65.17  Myeloma. (A) Lateral radiograph of the lumbar spine showing diffuse osteopenia and collapse of the L1 vertebra (arrow). (B) Sagittal multidetector computed tomography MPR of the thoracic and lumbar spine showing diffuse osteopenia and multilevel compression fractures (arrows).

SUMMARY BOX:  Plasma Cell Disorders • Clonal proliferation of plasma cells of B-cell origin within the bone marrow • Includes MGUS, smouldering myeloma, solitary plasmacytoma and multiple myeloma (MM) • Solitary plasmacytoma presents as an expansile lytic lesion with associated soft tissue mass • MM is commonest primary osseous malignancy in adults, producing classical ‘punched-out’ lesions throughout skeleton and also generalised osteopenia • Low-dose MDCT, PET-CT and WB-MRI all more sensitive than skeletal survey for lesion detection MDCT, Multidetector computed tomography; MGUS, monoclonal gammopathy of unknown significance; PET-CT, positron emission tomography-computed tomography; WB-MRI, whole-body magnetic resonance imaging.

MISCELLANEOUS CONDITIONS SYSTEMIC MASTOCYTOSIS Mastocytosis has been reclassified in the updated 2016 WHO classification of myeloproliferative neoplasms into three subgroups: (1) cutaneous mastocytosis, (2) systemic mastocytosis (SM) which includes indolent SM (ISM), smouldering SM (SSM), SM with an associated haematological neoplasm (SM-AHN), aggressive SM (ASM) and mast cell leukaemia (MCL), and (3) mast cell sarcoma.

Clinical Features The condition presents in the fifth to eighth decades with equal frequency in men and women. Bone marrow involvement is present in approximately 90% of cases and is often asymptomatic, but may produce thoracic and lumbar spinal pain and arthralgia. The condition may be associated with myelodysplastic syndromes, myeloproliferative neoplasia, leukaemia, and lymphoma, and the prognosis is variable.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

1715

A

B Fig. 65.18  Myeloma. POEMS syndrome. (A) Axial computed tomography image showing sclerotic involvement of a lower thoracic vertebra (arrow). (B) Coronal short tau inversion recovery magnetic resonance image of the lower limbs demonstrating denervation atrophy and oedema of the anterior compartment musculature (arrows) due to polyneuropathy.

Imaging Features Imaging—including radiography, scintigraphy, bone densitometry, CT, and MRI—plays a role in the diagnosis, staging, and monitoring of the disease. Skeletal changes are due to both the direct effect of mast cells and the indirect effect of secreted mediators such as histamine, heparin, and prostaglandins. They include both osteolytic and osteosclerotic lesions, which may be either diffuse or focal. Small (4–5 mm) lytic lesions may be surrounded by a rim of sclerosis and are most commonly seen in the spine, ribs, skull, pelvis, and tubular bones. Diffuse osteopenia is a common pattern, usually involving the axial skeleton, and may be complicated by pathological fracture in 16% of cases. Differential diagnosis includes osteoporosis, Gaucher disease, myeloma, hyperparathyroidism or thalassaemia. Osteosclerosis produces trabecular and cortical thickening with reduction of the marrow spaces (Fig. 65.24A), and multifocal sclerotic lesions simulating osteoblastic metastases (Fig. 65.24B). Both multidetector CT and MRI are more sensitive than radiography in the identification of marrow involvement. In mild cases, MRI may be normal. Otherwise, there is a generalised reduction of T1 weighted signal intensity, with variable T2 weighted and STIR signal intensity, depending upon

Fig. 65.19  Myeloma. Sagittal low dose multidetector computed tomography MPR of the whole spine showing diffuse osteopenia and multiple focal lytic lesions (arrows).

the degree of associated marrow fibrosis (Fig. 65.25A and B). Sclerotic lesions appear hypointense on all pulse sequences. PET/CT can demonstrate FDG uptake in bone marrow, lymph nodes, and the spleen, this being a feature in patients with SM-AHN and mast cell sarcoma, as opposed to those with the commoner less aggressive forms of SM.

THE HISTIOCYTOSES The histiocytoses are a rare group of conditions characterised by the accumulation of macrophages or dendritic cells in various tissues and organs, affecting both children and adults. Some of these conditions involve the skeleton, particularly Langerhans cell histiocytosis (LCH) and Erdheim-Chester disease.

LANGERHANS CELL HISTIOCYTOSIS LCH is the commonest dendritic cell disorder and is divided clinically into three groups depending upon the number of lesions and the organ systems involved. In 70% of cases, presentation is limited to a single or a few bones, although the lung may also be involved. Patients present typically between 5 and 15 years of age. The second group affects 20% of patients who present between 1 and 5 years of age with multifocal bone lesions, as well as involvement of the reticuloendothelial system and often diabetes insipidus. The third group accounts for approximately

Fig. 65.20  Myeloma. (A) Coronal positron emission tomography MIP image showing widespread uptake in both the axial and appendicular skeleton (arrows). (B) Sagittal fused positron emission tomography-computed tomography image showing extensive spinal and sternal activity.

B

A

A

B

C

Fig. 65.21  Myeloma. (A) Sagittal T2 weighted FSE magnetic resonance (MR) image of the lumbar spine showing multifocal areas of increased marrow SI. (B) Sagittal T1 weighted SE MR image of the lumbar spine showing diffuse reduction of marrow SI. (C) Coronal T1 weighted SE MR image of the pelvis and hips showing the variegated pattern of marrow involvement.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

A

C

1717

B

D

Fig. 65.22  Myeloma. Treatment response. (A) Sagittal T1 weighted SE magnetic resonance (MR) image of the lumbar spine at presentation showing heterogeneous reduction of marrow SI. (B) Sagittal T1 weighted SE MR image of the lumbar spine at follow-up showing normalisation of marrow SI and vertebroplasty at T11 and T12 (arrows). (C) Coronal T1 weighted SE and (D) sagittal T2 weighted FSE MR images of the right hip showing a treated lesion which has fluid SI characteristics and a fatty rim (arrow).

10% of cases, presenting in the first 2 years of life with widespread involvement of the reticuloendothelial system, anaemia and thrombocytopenia, and is often fatal. Skeletal involvement occurs in 80% of cases of LCH and can involve any bone, but the commonest sites are the skull, mandible, ribs, pelvis, and spine. The long bones are involved in 25–35% of cases of monostotic disease, with the femur, tibia, and humerus being the most common locations. Multiple lesions are seen in 10% of cases at presentation.

Imaging Features Long bone lesions are usually located centrally within the diaphysis (∼60%) (Fig. 65.26A), followed by the metaphysis/metadiaphysis. Epiphyseal involvement is rare (∼2%). The lesions are lytic, showing an aggressive pattern of bone destruction with occasional reactive medullary sclerosis. A multi-laminated periosteal response is commonly seen (Fig. 65.26B), while endosteal scalloping and mild bone expansion are also features (Fig. 65.27A). As lesions heal, they may either resolve

1718

SECTION F  Oncological Imaging

A

B

Fig. 65.23  Myeloma. Relapse. Coronal positron emission tomography MIP images showing (A) complete remission and (B) widespread recurrence. (Images courtesy Dr Simon Wan.)

B

A Fig. 65.24  Mastocytosis. (A) Anteroposterior radiograph of the right hip showing endosteal sclerosis in the proximal femur (arrows). (B) Computed tomography of the left proximal humerus demonstrating multiple intramedullary sclerotic lesions.

Fig. 65.25  Mastocytosis. (A) Coronal T1 weighted SE and (B) sagittal T2 weighted FSE magnetic resonance images of the femora showing a mild heterogeneous reduction of marrow SI.

Fig. 65.26  Langerhans Cell Histiocytosis. (A) Lateral radiograph of the femur showing an aggressive lytic lesion in the diaphysis (arrow). (B) Anteroposterior radiograph of the forearm showing an expansile lytic lesion in the radial diaphysis with a multi-laminated periosteal response (arrow).

A

B

A

B

1720

SECTION F  Oncological Imaging

A

A

B Fig. 65.27  Langerhans Cell Histiocytosis. (A) Anteroposterior radiograph of the right elbow showing a lytic lesion in the distal humerus causing endosteal scalloping (arrow). (B) Follow-up radiograph showing almost complete resolution of the lesion (arrow).

(Fig. 65.27B) or develop sclerosis. Flat bone lesions are more commonly seen in adults. MRI shows a poorly defined lesion with intermediate signal intensity on T1 weighted and hyperintensity on T2 weighted and STIR sequences, showing enhancement following contrast. Active lesions are almost invariably associated with reactive marrow and soft-tissue oedema and periostitis (Fig. 65.28A and B). Cortical destruction and soft-tissue masses have rarely been described (Fig. 65.29A and B). Whole-body MRI should replace scintigraphy for the assessment of multifocal disease.

ERDHEIM–CHESTER DISEASE Erdheim–Chester disease is a rare form of histiocytosis usually presenting in adults at a mean age of 55 to 60 years. Males are affected three times more commonly than females and the skeleton is involved in over 95%

B Fig. 65.28  Langerhans Cell Histiocytosis. (A) Axial T1 weighted SE and (B) coronal short tau inversion recovery magnetic resonance images of the lower limb showing a lesion in the distal fibular diaphysis (arrows) with associated marrow and soft tissue oedema (arrowheads-B).

of cases. Cardiovascular involvement (50%), retroperitoneal fibrosis (∼30%), CNS involvement, diabetes insipidus and/or exophthalmos (20% to 30%) are other features.

Imaging Findings Typical features include bilateral, symmetrical cortical/medullary sclerosis of the major long bone metadiaphyses with epiphyseal sparing (Fig. 65.30A). MRI demonstrates heterogeneous reduced marrow SI on T1 weighted (Fig. 65.30B) and variable SI on T2 weighted and STIR (Fig. 65.30C). Bone scintigraphy (Fig. 65.31A) and PET-CT (Fig. 65.31B) are highly specific, showing increased uptake in affected areas.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms

A

B

Fig. 65.29  Langerhans Cell Histiocytosis. (A) Axial computed tomography of the left hip showing a lytic lesion destroying the medial acetabular wall with an adjacent soft tissue mass (arrow). (B) Coronal short tau inversion recovery magnetic resonance image shows the left acetabular lesion (arrow) and a synchronous lesion in the right proximal femur (arrowhead).

A

B

C

Fig. 65.30  Erdheim-Chester Disease. (A) Anteroposterior radiograph of the right distal femur showing chronic sclerosis of the distal femoral metadiaphysis and solid periosteal thickening (arrows). (B) Coronal T1 weighted SE and (C) short tau inversion recovery (STIR) magnetic resonance images of the femora demonstrating heterogeneous reduced T1 weighted SI and STIR hyperintensity.

1721

1722

SECTION F  Oncological Imaging

A

B

Fig. 65.31  Erdheim-Chester Disease. (A) 99m-Tc MDP bone scan and (B) fluorodeoxyglucose-positron emission tomography study showing classical features of symmetrical increased activity in the long bone metaphyses (arrows).

FURTHER READING Amini, B., Yellapragada, S., Shah, S., et al., 2016. State-of-the-art imaging and staging of plasma cell dyscrasias. Radiol. Clin. North Am. 54, 581–596. Arber, D.A., Orazi, A., Hasserjian, R., et al., 2016. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood 127, 2391–2405. Chan, B.Y., Gill, K.G., Rebsamen, S.L., et al., 2016. MR imaging of pediatric bone marrow. Radiographics 36, 1911–1930. Derlin, T., Bannas, P., 2014. Imaging of multiple myeloma: current concepts. World J Orthop 5, 272–282.

Dutoit, J.C., Verstraete, K.L., 2017. Whole-body MRI, dynamic contrast-enhanced MRI, and diffusion-weighted imaging for the staging of multiple myeloma. Skeletal Radiol. 46, 733–750. Emile, J.F., Abla, O., Fraitag, S., et al., 2016. Revised classification of histiocytoses and neoplasms of the macrophage-dendritic cell lineages. Blood 127, 2672–2681. Fritz, J., Fishman, E.K., Carrino, J.A., et al., 2012. Advanced imaging of skeletal manifestations of systemic mastocytosis. Skeletal Radiol. 41, 887–897. Lecouvet, F.E., 2016. Whole-body MR imaging: musculoskeletal applications. Radiology 279, 345–365.

CHAPTER 65  Bone Marrow Disorders: Haematological Neoplasms Lim, C.Y., Ong, K.O., 2013. Imaging of musculoskeletal lymphoma. Cancer Imaging 13, 448–457. Murphy, I.G., Mitchell, E.L., Raso-Barnett, L., et al., 2017. Imaging features of myeloproliferative neoplasms. Clin. Radiol. 72, 801–809. Murphey, M.D., Kransdorf, M.J., 2016. Primary musculoskeletal lymphoma. Radiol. Clin. North Am. 54, 785–795. Navarro, S.M., Matcuk, G.R., Patel, D.B., et al., 2017. Musculoskeletal imaging findings of hematologic malignancies. Radiographics 37, 881–900. Pawha, P.S., Chokshi, F.H., 2016. Imaging of spinal manifestations of hematological disorders. Hematol. Oncol. Clin. North Am. 30, 921–944.

1723

Roberto Silva, J., Jr., Hayashi, D., Yonenaga, T., et al., 2013. MRI of bone marrow abnormalities in hematological malignancies. Diagn. Interv. Radiol. 19, 393–399. Roberts, A.S., Shetty, A.S., Mellnick, V.M., et al., 2016. Extramedullary haematopoiesis: radiological imaging features. Clin. Radiol. 71, 807–814. Shah, S.N., Oldan, J.D., 2017. PET/MR imaging of multiple myeloma. Magn. Reson. Imaging Clin. N. Am. 25, 351–365. Singh, A., Kumar, P., Chandrashekhara, S.H., et al., 2017. Unravelling chloroma: review of imaging findings. Br. J. Radiol. 90 (1075), 20160710. Zaveri, J., La, Q., Yarmish, G., et al., 2014. More than just Langerhans cell histiocytosis: a radiologic review of histiocytic disorders. Radiographics 34, 2008–2024.

66  Bone Marrow Disorders: Non-Neoplastic Conditions Asif Saifuddin

CHAPTER OUTLINE Disorders of Red Cells, 1724 The Myelodysplastic Syndromes, 1724

This chapter deals with a variety of non-neoplastic blood-related disorders that have a major influence on imaging of the skeletal system.

DISORDERS OF RED CELLS In late fetal life and infancy, the entire bone marrow is utilised for red blood cell (RBC) production, supplemented by extramedullary haematopoiesis (EHM) in the liver and spleen. As the child becomes older and RBC life span increases, erythropoiesis is withdrawn from the liver and spleen, then gradually from the diaphyses of the long bones so that by the age of 25 years, active bone marrow is confined to the axial skeleton, the flat bones and the proximal ends of the femora and humeri. This process of withdrawal will not occur with a need for extra erythropoiesis and reverses in the presence of increased RBC destruction.

Physiological Marrow Reconversion Physiological marrow reconversion is commonly seen on musculoskeletal magnetic resonance imaging (MRI) studies in response to increased haematopoietic needs of the body, as may occur in heavy smokers, athletes (particularly those involved in high altitude training), obesity and chronic respiratory disorders including obstructive sleep apnoea. Marrow hyperplasia is most commonly noted in the distal femur and proximal tibia on magnetic resonance (MR) examinations of the knee, and in the proximal humerus on MR examinations of the shoulder. The hyperplastic marrow has a well-defined ‘geographical’ margin with normal yellow marrow, shows signal intensity (SI) which is intermediate between yellow marrow and skeletal muscle on unenhanced T1 weighted spin echo (Fig. 66.1A) and T2 weighted fast spin echo sequences, mildly hyperintense on short-tau inversion recovery (STIR) and fat-suppressed sequences (Fig. 66.1B), and is hypointense on gradient echo imaging. It should not be mistaken for marrow infiltration. Symmetrical marrow reconversion may also be seen in oncology patients who have been treated with granulocyte colony-stimulating factor (GCSF) (Fig. 66.2). Not uncommonly, focal nodular hyperplasia (FNH) of the marrow may also be seen mimicking medullary bone tumours or metastases, but the marrow SI is again typical of hyperplastic red marrow. Common sites include the vertebrae and long bone metaphyses (Fig. 66.3A and B). Rarely, FNH may show increased activity on

1724

The Haemoglobinopathies, 1724 Disorders of Blood Coagulation, 1733

fluorodeoxyglucose-positron-emission tomography (FDG-PET) studies, and biopsy may be required to exclude metastatic involvement.

The Anaemias Only chronic anaemias affect the imaging appearances of bone. Anaemias that do not produce reactive erythropoiesis, such as aplastic anaemia, do not affect the skeletal radiograph but may manifest on MRI as a generalised increase in fatty marrow SI.

THE MYELODYSPLASTIC SYNDROMES The myelodysplastic syndromes (MDSs) comprise a variety of clonal stem cell disorders which have an overall incidence of approximately 3.3/100,000, which can rise to between 15 and 50/100,000 in patients older than 70 years of age. They result in variable symptoms and signs of chronic anaemia, bleeding and frequent infections due to defective RBC, platelet and white blood cell production. MDS has been reclassified into various subtypes by the World Health Organisation (WHO) in 2016, replacing the previously used French–American–British (FAB) classification. The disease is progressive, with allogenic stem cell transplantation being the only curative treatment, and carries a 25%–40% risk of transformation to acute leukaemia. MRI may demonstrate nodular, patchy or diffuse reduction of T1 weighted marrow SI (Fig. 66.4A), with corresponding hyperintensity on fat-suppressed proton density weighted (PDW)/T2 weighted or STIR sequences (Fig. 66.4B). The patchy or diffuse pattern of marrow abnormality is associated with poorer prognosis and earlier development of leukaemia. Presacral EMH is also a recognised complication.

THE HAEMOGLOBINOPATHIES The haemoglobin (Hb) molecule consists of a protein (globin) and four haem groups, each with four pyrrole rings surrounding an iron atom. The protein moiety consists of 574 amino acids arranged in four spiral polypeptide chains. The different chains are designated by letters of the Greek alphabet (α, β), and the three normal Hbs A, A2 and F each contain two α chains, differing only in their second pairs.

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

A

1725

B Fig. 66.1  Marrow Reconversion. (A) Sagittal T1 weighted spin echo magnetic resonance imaging (MRI) showing a reduction of marrow signal intensity within the distal femur and proximal tibia (arrows) compared with the fatty marrow in the epiphyses. (B) Coronal fat-suppressed PDW fat-suppressed echo MRI showing corresponding increased marrow signal intensity (arrows).

SUMMARY BOX:  Disorders of Red Cells • Physiological marrow reconversion is very commonly seen on magnetic resonance imaging (MRI), occurring in any conditions resulting in chronic marrow stress, or chronic anaemia associated with myelodysplastic syndromes • MRI is characterised by focal or diffuse reduction of T1 weighted marrow signal intensity (SI) which is intermediate between that of medullary fat and skeletal muscle • Focal nodular hyperplasia of the marrow may mimic a bone tumour or metastasis but should be recognised based on the typical T1 weighted marrow SI

Haemoglobinopathies are genetic conditions which result in either defective or deficient globin chains and include: 1. Thalassaemia: an inherited defect of HbA synthesis with inadequate manufacture of α- or β-chains. 2. Hb variants: inherited defects of HbA synthesis producing abnormal α- or β-chains. All the variants differ from HbA by the substitution of only one amino acid in the chain: e.g. HbS (sickle cell anaemia) where valine is substituted for glutamine at residue six in the β-chain. 3. Combination of thalassaemia and abnormal Hb: for example, HbS–thalassaemia. Fig. 66.2  Granulocyte Colony-Stimulating Factor Effect. Coronal T1 weighted spin echo magnetic resonance imaging of the knee in a patient treated for Ewing sarcoma of the fibula showing reduction of metaphyseal marrow signal intensity (arrow).

Thalassaemia Thalassaemia is an autosomal recessive disorder that affects the production of either α- or β-chains, resulting in the development of α-thalassaemia or β-thalassaemia, respectively. The purely heterozygous

1726

SECTION F  Oncological Imaging

A

B Fig. 66.3  Focal Nodular Hyperplasia. (A) Coronal T1 weighted spin echo magnetic resonance imaging (MRI) showing focal reduction of marrow signal intensity in the proximal fibular metaphysis (arrow). (B) Axial fatsuppressed PDW fast spin echo MRI showing corresponding marrow hyperintensity (arrow).

B

A

Fig. 66.4  Myelodysplastic Syndrome. (A) Coronal T1 weighted spin echo magnetic resonance imaging (MRI) showing diffuse reduction of marrow signal intensity (SI) in the proximal humeral metaphysis and glenoid (arrows), the SI still being higher than that of skeletal muscle. (B) Axial fat-suppressed PDW fast spin echo MRI showing corresponding marrow hyperintensity (arrows).

form of each condition results in an asymptomatic carrier state. The homozygous form of α-thalassaemia (α-thalassaemia major; α-TM) is one of the commonest causes of hydrops fetalis and usually results in intrauterine or perinatal death. The homozygous form of β-thalassaemia (β-thalassaemia major; β-TM) results in severe childhood anaemia and

affects the skeleton in a variety of ways. Compound heterozygosity can also occur in α- and β-thalassaemia resulting in haemoglobin H disease and β-thalassaemia intermedia (β-TI), respectively, which are both typically non-transfusion-dependent conditions that can affect the skeleton due to marked marrow hyperplasia.

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

1727

Clinical Features β-TM is prevalent in those originating from the Mediterranean countries (Greece, southern Italy and the Mediterranean islands) and causes severe childhood anaemia resulting in massive hepatosplenomegaly, skeletal deformity due to chronic marrow hyperplasia, and EHM. Treatment is by regular blood transfusion commencing in infancy aiming to maintain an Hb level of >95 g/L, iron chelation with agents such as deferoxamine, deferiprone or deferasirox to prevent/reduce iron overload, and bis­ phosphonates to treat the associated osteoporosis. With the introduction of successful transfusion regimes, the skeletal changes due to the disorder itself are now less commonly encountered, but osseous complications related to repeated transfusion and the effects of chelation therapy may occur. Bone disease may also be contributed to by endocrinopathy due to iron overload, including hypogonadism, growth hormone deficiency, hypoparathyroidism, vitamin D deficiency, hypothyroidism, diabetes and renal dysfunction. The only curative treatment is allogenic bone marrow transplantation.

A

Radiological Features Skeletal changes in untreated children are essentially the result of chronic anaemia and marrow hyperplasia (15–30 times normal) and do not usually manifest before 6 months of age due to high levels of circulating fetal haemoglobin (HbF). Radiographical changes are usually evident by the second year of life and affect both the axial and appendicular skeleton. However, they are currently uncommonly seen due to early transfusion. Medullary hyperplasia results in bony expansion and cortical thinning (Fig. 66.5A), which in the long bones produces the characteristic Erlenmeyer flask appearance, also found in conditions such as Gaucher’s disease. Within the medulla, trabecular thinning initially occurs, followed by trabecular coarsening due to new bone formation (Fig. 66.5A), the changes being most marked in the metacarpals and phalanges, which become cylindrical or even biconvex. Premature fusion of the growth plates may contribute to short stature. These radiological changes are reported in approximately 15% of patients, generally occurring after 10 years of age and most commonly affecting the proximal humerus and distal femur. Growth arrest lines may also be seen. Trabecular coarsening may also be seen in the pelvis and vertebrae (Fig. 66.6). The ribs are similarly affected with club-like anterior ends, while a ‘rib-within-a-rib’ appearance also occurs due to subperiosteal extension of haematopoietic tissue through the rib cortex. With severe childhood disease, the paranasal sinuses develop poorly and often contain red marrow, accounting for the facial abnormalities (‘rodent facies’ seen in 17%) and dental malocclusion. The ethmoidal cells are spared since they contain no red marrow. The diploë of the skull vault are widened, except in the occiput. These changes occur earliest and most severely in the frontal bone, producing the classical ‘hair-on-end’ appearance. Occasionally, well-defined lytic lesions may be seen in the skull vault. Appendicular abnormalities may regress during adolescence and early adulthood, presumably due to conversion of red marrow to yellow marrow, but osteopenia will still be evident in adult patients (Fig. 66.7). The spine is commonly involved, back pain becoming commoner with increasing age and usually involving the lumbar region. Spinal changes consist of generalised osteopenia (Fig. 66.8), resulting in compression fractures and biconcavity of the vertebral bodies. Even in well-treated patients, osteoporosis may still be seen in 13.6%–50% of cases, with an additional 45% affected by osteopenia. Scoliosis is reported in 20% of children with β-TM, and early disc degeneration is also a feature, optimally demonstrated by MRI (Fig. 66.9). EMH occurs in severe cases surviving to adulthood, the most common sites being the thoracic paravertebral region by extension from the

B Fig. 66.5  Thalassaemia Major. (A) Oblique lateral radiograph of the right foot demonstrating mild bony expansion and cortical thinning in the metatarsals (arrow), and trabecular thickening in the distal tibia and hindfoot bones (arrowheads). (B) Axial T1 weighted spin echo magnetic resonance imaging showing expansion of the calcaneus and diffuse reduction of marrow signal intensity due to reconversion.

adjacent ribs (Fig. 66.10), the mediastinum and presacral region. Epidural extension from paraspinal EMH may result in spinal cord compression. These features are optimally demonstrated by MRI, which also shows diffuse reduction in marrow SI on T1 weighted images caused by marrow reconversion (Figs 66.5B and 66.9). Hypertransfusion.  Repeated transfusion therapy may produce iron overload and hyperuricaemia. Raised blood iron levels result in synovial and articular cartilage abnormalities, manifesting radiologically as symmetrical loss of joint space, cystic changes, subchondral collapse and osteophytosis. Chondrocalcinosis may also be seen, and the larger joints tend to be more commonly affected than in primary haemochromatosis. Occasionally, the radiographical changes of gout may be evident, and there is also a predisposition to osteonecrosis and osteomyelitis. Iron deposition within bone also contributes to osteoporosis and is a further cause of reduced marrow SI on MRI, particularly on T2 weighted and gradient-echo sequences. MRI may be used to estimate tissue siderosis. Iron chelation therapy.  Iron chelation therapy is recognised as causing dysplastic changes in the spine and long bones, as well as growth

1728

SECTION F  Oncological Imaging

Fig. 66.6  Thalassaemia Major. Lateral radiograph of the lumbar spine demonstrating trabecular thickening (arrows) and severe multilevel degenerative disc disease.

Fig. 66.8  Thalassaemia Major. Sagittal computed tomography multiplanar reformat of the lumbar spine showing diffuse osteopenia.

retardation, especially when treatment is started before the age of 3 years and with higher doses. The incidence is unclear since many cases are asymptomatic. Changes typically occur at the metaphysis/physis/ epiphysis of the proximal humerus, distal femur, proximal tibia and distal radius and ulna. Radiographs and MRI demonstrate irregularity of the metaphysealphyseal junction with dense sclerotic metaphyseal bands, which may then extend in a ‘flame-shaped’ manner towards the diaphysis. Splaying of the metaphysis and widening of the growth plate are later features, which resemble rickets. Severe dysplasia at the proximal femur and around the knee may result in slipped upper femoral epiphysis (SUFE), and genu varum or valgum. In the spine, platyspondyly resembling spondyloepiphyseal dysplasia and a biconvex contour to the vertebral bodies are seen and kyphosis may develop. Deferoxamine therapy also results in growth retardation, affecting both the axial and appendicular skeleton. Deferiprone is an alternative treatment and has been associated with agranulocytosis and arthropathy, most commonly affecting the knees and resulting in effusion and synovitis which can be demonstrated on MRI. Chronic changes include flattening of the femoral condyles, tibial plateau and patella. Bisphosphonate therapy. Bisphosphonates are used to treat osteoporosis in β-TM and have been associated with the development of atypical femoral fractures.

Sickle Cell Disease Fig. 66.7  Thalassaemia Major. AP radiograph of the right hip demonstrating generalised osteopenia.

Sickle cell anaemia is an autosomal recessive condition caused by a mutation in the β-globin gene which results in replacement of glutamic acid with valine at position six of the β-globin chain, producing an

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

1729

A

Fig. 66.9  Thalassaemia Major. Sagittal T1 weighted spin echo magnetic resonance image of the lumbar spine showing multilevel disc height loss due to degeneration, and diffuse reduction of marrow signal intensity due to red marrow hyperplasia.

abnormal HbS molecule. Different conditions result, depending upon the status of the second β-chain: • If the second β-chain is also HbS, then the patient has homozygous HbS-S, defined as sickle cell anaemia. • If the second β-chain is a different abnormal Hb such as HbC, HbE or thalassaemia, this results in HbS-C, HbS-E or HbS-thalassaemia. Patients with HbS-C have milder anaemia and less severe symptoms than in homozygous HbS-S, while patients with HbS-thalassaemia can present in the same way as HbS-S with anaemia and painful crises. • If the second β-chain is normal, then the patient has sickle cell trait. The clinical findings in sickle cell disease (SCD) are explained by the physical properties of the abnormal Hb. In situations of hypoxia and dehydration, intracellular polymerisation of the HbS molecule occurs, rendering the RBC less flexible. Recurrent cycles of oxygenation and deoxygenation result in irreversible membrane damage to the erythrocyte, causing the cells to become less pliable and sickle shaped. Sickle cells obstruct small blood vessels, leading to stasis and tissue hypoxia/anoxia and eventually infarction. Sickle cells also have a much shorter life (~1/10th normal), being removed from the circulation prematurely by the reticuloendothelial system, which results in chronic haemolytic anaemia. The full clinical and radiological picture occurs in the homozygous sickle cell subject (HbS-S).

Clinical Features SCD mainly affects people of African racial origin, and approximately 8%–10% of black Americans have sickle cell trait, but only 0.2% have sickle cell anaemia. People from the Middle East and Eastern

B Fig. 66.10  Thalassaemia Major. Extramedullary haematopoiesis. (A) Coronal T2 weighted fast spin echo and (B) axial PDW fast spin echo magnetic resonance imaging of the thoracic spine showing paravertebral soft-tissue masses (arrows).

Mediterranean are also affected. The anaemia of SCD is not as marked as in β-TM. Homozygous SCD reduces average life expectancy by 25–30 years, and most patients will die by the age of 50 years. The most striking clinical features are due to vaso-occlusive sickle crises, which result in infarctions. Medullary infarction involving the small bones of the hands and feet results in sickle cell dactylitis or ‘hand–foot’ syndrome and affects 50% of infants with SCD between 6 months and 2 years of age but is rare after 6 years. Presentation is with pain and swelling of the digits, together with fever. Marrow infarction may also involve the long bones, in which case it is difficult to differentiate from osteomyelitis. Infarction of the epiphyses results in avascular necrosis, most commonly of the femoral and humeral heads, which is frequently bilateral. SCD is the most common cause of osteonecrosis of the femoral head in children, while approximately 50% of all patients

1730

SECTION F  Oncological Imaging

A

B

Fig. 66.11  Sickle Cell Disease. (A) Sagittal T1 weighted spin echo and (B) axial T2 weighted fast spin echo magnetic resonance imaging of the lumbosacral spine showing diffuse reduction of marrow signal intensity due to marrow hyperplasia.

will develop osteonecrosis by the age of 35 years. Rib or sternal infarcts may simulate heart or lung disease, while soft-tissue involvement results in ulcers and muscle infarction, which are well demonstrated by MRI. Osteomyelitis in SCD typically involves the long bone diaphyses and is most commonly due to various Salmonella species, with approximately 10% of cases being due to Staphylococcus aureus. The prevalence of osteomyelitis is reported as 18%, while septic arthritis is also relatively common in SCD, with a reported prevalence of 7%. Sickle variants show less anaemia with fewer crises. However, bone infarction, especially of the femoral head, affects patients with HbS-C disease five times more often than those with HbS-S, although the overall prevalence of HbS-C is only one-third that of the homozygous disease. This may reflect the longer survival in HbS-C disease. In sickle-cell trait (HbS-A), significant anaemia and bone infarction are rare.

Radiological Features The changes in SCD and its variants are similar, differing only in degree. Changes can be divided into those due to marrow hyperplasia, bone infarction and secondary osteomyelitis. Marrow hyperplasia.  This is more severe in the homozygous disease, and the radiological features are as described for thalassaemia. Marrow reconversion is well-demonstrated by MRI, with replacement of normal fatty marrow by intermediate SI on T1 weighted images (Fig. 66.11), which may extend to involve the epiphyses. Marrow hyperplasia also results in widening of the medulla and cortical thinning, coarsening of the normal trabecular pattern with loss of corticomedullary definition in both long and flat bones (Fig. 66.12). The bones may appear osteopenic and are predisposed to fracture, which is optimally demonstrated in the vertebral end plates which assume a biconcave morphology. Persistence of red marrow predisposes to osteomyelitis and marrow infarction. EMH also occurs but is less common than in thalassaemia and HbS-thalassaemia. Bone infarction.  This is estimated to be at least 50 times more common than osteomyelitis in SCD. In children, sickle cell dactylitis

Fig. 66.12  Sickle Cell Disease. AP radiograph of the right knee demonstrating loss of corticomedullary differentiation and diffuse trabecular thickening.

results in lytic medullary lesions with associated periostitis and soft-tissue swelling (Fig. 66.13). Asymmetrical shortening of tubular bones is a common sequela to childhood sickling crises. In adolescents and adults, infarction occurs more in the metaphyses and epiphyses. The earliest radiological evidence of bone infarction is laminar periosteal reaction followed by patchy medullary destruction. Healing leads to reactive

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

Fig. 66.13  Sickle Cell Disease: Dactylitis. Infarction in several of the metacarpals and proximal phalanges has resulted in bone destruction (arrows) and swelling of the soft tissues.

sclerosis. MRI shows medullary oedema on T2 weighted and STIR sequences with associated periostitis and adjacent soft-tissue inflammation, making differentiation from osteomyelitis difficult. With healing, these areas assume a low SI due to fibrosis and medullary sclerosis. Infarction of the metaphyses on either side of the knee is common and may lead to premature fusion of the growth plates. Infarction of the vertebral body occurs in approximately 10% of patients and is usually due to venous thromboembolism in the centre of the vertebral end plate, with focal collapse producing the ‘H-shaped’ vertebra (Fig. 66.14). The floor of the depression forms a flat sclerotic margin during healing. This appearance is almost pathognomonic of SCD but has also been reported in Gaucher disease. Overgrowth of an adjacent vertebral body, producing a ‘tower’ vertebra, is also reported (Fig. 66.15). The earliest manifestation of osteonecrosis is epiphyseal oedema seen on MRI. Radiographic abnormalities include mixed lysis and sclerosis with a subchondral fracture being typical (Fig. 66.16). Eventually, secondary osteoarthritis (OA) will supervene (Fig. 66.17). Chronic ischaemia or multiple small infarctions in SCD may produce cortical thickening that is both endosteal and periosteal, with narrowing of the marrow cavity (Fig. 66.18). Splitting of the cortex may give rise to a ‘bonewithin-a-bone’ appearance (Fig. 66.17), while secondary myelofibrosis causes medullary sclerosis (Fig. 66.19). Osteomyelitis.  Osteomyelitis usually complicates bone infarction, and it may be difficult to distinguish clinically or radiologically between an infarct with infection and one without. Salmonella osteomyelitis is

1731

Fig. 66.14  Sickle Cell Disease. Sagittal T1 weighted spin echo magnetic resonance imaging of the lumbar spine showing multilevel H-shaped vertebral bodies (arrows) due to central end plate infarction.

Fig. 66.15  Sickle Cell Disease. The classical ‘stepped depression’ of the vertebral end plates is seen (arrows), producing the ‘H-shaped’ vertebra. Note increased height of the adjacent ‘tower’ vertebra (arrowhead).

1732

SECTION F  Oncological Imaging

Fig. 66.16  Sickle Cell Disease. AP radiograph of the right hip demonstrating established avascular necrosis with collapse of the right femoral head (arrow).

Fig. 66.18  Sickle Cell Disease. AP radiograph of the left shoulder showing epiphyseal sclerosis due to avascular necrosis (arrow) and endosteal sclerosis resulting in narrowing of the medullary cavity (arrowhead).

Fig. 66.17  Sickle Cell Disease. AP radiograph of the left hip and proximal femur demonstrating advanced osteoarthritis (arrow) and a ‘bone-withina-bone’ appearance (arrowhead).

common in African children. Osteomyelitis causes increased bone destruction with laminated or multilaminated periostitis (Fig. 66.20A) and eventual sequestration and involucrum formation. Early diagnosis of bone infection is important. MRI typically demonstrates poorly defined marrow oedema, periostitis and soft-tissue oedema, which are also seen in acute infarction. However, the communication between a fluid collection in the medulla and surrounding soft tissues through a cortical defect is indicative of osteomyelitis (Fig. 66.20B). In addition, the presence of geographical regions of marrow enhancement on fatsuppressed postcontrast T1 weighted images is strongly associated with infection, whereas acute marrow infarcts tend to show serpentine

Fig. 66.19  Sickle Cell Disease. AP radiograph of the left knee showing diffuse medullary sclerosis consistent with secondary myelofibrosis.

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

1733

SUMMARY BOX:  Haemoglobinopathies • A variety of genetic conditions due to deficient or defective production of haemoglobin, resulting in chronic anaemia presenting in childhood • Include thalassaemias and sickle cell disease (SCD) • Radiological manifestations primarily due to marrow hyperplasia from chronic anaemia, including bone expansion, cortical thinning and osteopenia • Magnetic resonance imaging additionally demonstrates diffuse reduction of T1 weighted marrow signal intensity • Other imaging features may be due to complications of treatment (hypertransfusion, ironchelation) or complications of the disease, including osteonecrosis/osteomyelitis in SCD

A

All produce the same radiological appearances, the diseases differing only in the frequency and severity of the observed changes. Very occasionally, a patient of either sex may develop antibodies to antihaemophilic globulin and develop an acquired form of haemophilia. The prevalence of haemophilia A is estimated at 1 : 5000 and Christmas disease 1 : 30,000.

Haemophilia (Haemophilia A) Haemophilia A is an X-linked recessive disorder resulting from deficiency of factor VIII. Bleeding, which usually follows minor trauma, may occur in the first year of life, and 70% of haemophiliacs have experienced haemarthrosis by the age of 2 years. Repeated haemarthroses result in chronic arthropathy and eventually premature OA. Soft-tissue haemorrhage, often close to muscle attachments, produces another lesion characteristic of the disease, the haemophilic pseudotumour.

Christmas Disease (Haemophilia B) This disorder is due to deficiency of factor IX and is also an X-linked recessive disease. It is less severe than haemophilia.

Von Willebrand Disease Inherited as a dominant character and affecting both sexes, von Willebrand disease is due to both a capillary defect and a deficiency in factor VIII. The coagulation defect is mild and only occasionally causes significant skeletal abnormality. The severity of the disease relates to the level of factor VIII in the blood, which is variable.

Radiological Features B Fig. 66.20  Sickle Cell Disease: Osteomyelitis. (A) AP radiograph of the right ankle showing a poorly defined lytic lesion in the distal tibial metaphysis (arrows) and associated periostitis (arrowheads) consistent with acute osteomyelitis. (B) Coronal short-tau inversion recovery magnetic resonance imaging showing the irregular distal tibial bone abscess, which communicates with a soft-tissue abscess (arrows) through a cortical defect. Note also the marrow oedema and periosteal elevation.

peripheral enhancement. Ultrasound (US) may aid the diagnosis of infection by the guided aspiration of subperiosteal fluid collections, which may be seen in more than 80% of cases of osteomyelitis but is not seen with acute infarction.

DISORDERS OF BLOOD COAGULATION Three major inherited disorders are considered: • Classic haemophilia (haemophilia A) • Christmas disease (haemophilia B) • von Willebrand disease

In patients with severe haemophilia, 85%–90% of all bleeding events involve the joints. Radiologically, acute haemarthrosis appears as a tense joint effusion (Fig. 66.21) and associated periarticular osteoporosis may indicate previous episodes. MRI is more accurate than clinical examination at identifying haemarthrosis, and US also has a role in early detection. The most common joints involved are the knee, elbow, ankle, hip and shoulder. Repeated episodes of haemorrhage result in progressive joint damage referred to as haemophilic arthropathy. Haemorrhage initially occurs into the synovium and eventually extends into the joint space. Recurrent haemarthrosis results in a chronic haemorrhagic synovitis and articular cartilage damage. Hyperaemia produces periarticular osteopenia, epiphyseal overgrowth and premature closure of the growth plates, while pannus similar to that occurring in rheumatoid arthritis causes marginal erosions. Loss of secondary trabeculae leads to a permanent coarsening of the trabecular pattern, and growth arrest lines indicate the episodic nature of the disorder. Fibrosis results in joint contracture and intraosseous haemorrhage produces subarticular cysts, which predispose to subchondral collapse. Synovial thickening together with osteopenia makes the soft tissues appear radiographically denser. An absolute increase in the soft-tissue density may also occur due to concentration of haemosiderin by macrophages in the periarticular regions.

1734

SECTION F  Oncological Imaging

Fig. 66.22  Haemophilia. AP radiograph of the left knee showing enlargement of the distal femoral and proximal tibial epiphyses (arrows). Fig. 66.21  Haemophilia. Lateral radiograph of the knee showing a prominent joint effusion (arrows) due to acute haemarthrosis.

The knee is the most common joint involved. Radiological features include enlargement of the distal femoral and proximal tibial epiphyses (Fig. 66.22), varus or valgus deformity, squaring of the inferior pole of the patella with patellar overgrowth and widening of the intercondylar notch. Eventually, advanced secondary OA develops. In the elbow, chronic hyperaemia causes accelerated appearance of the ossification centres and overgrowth of the radial head. Pressure erosion of the radial notch of the ulna, large lytic lesions of the proximal ulna and erosion of the trochlear notch are also seen (Fig. 66.23). In the ankle, asymmetrical growth of the distal tibial epiphysis results in medial tibiotalar slant. Marked flattening of the talar dome (Fig. 66.24), a variety of ankle and foot deformities and, rarely, ankylosis of the ankle or subtalar joints may be seen. Typical radiographical features of osteonecrosis may be evident in the hip and may simulate Perthes’ disease, eventually resulting in coxa magna. Protrusio acetabuli and secondary OA are also seen, while haemorrhage into the growth plate may result in slipped epiphysis. Coxa valga may result from delayed weight bearing. Staging of haemophilic arthropathy has typically been with radiography, utilising the Pettersson score. However, radiography is insensitive to the earliest changes, which occur in soft tissue. US can demonstrate acute/chronic haemarthrosis and the resulting synovitis but is less able to assess the cartilage and subchondral bone. All of the pathological features of haemophilic arthropathy can be well demonstrated by MRI, which may identify changes in the joint before any clinically evident episode of bleeding. MRI shows the earliest evidence of haemarthrosis as a low SI intra-articular blood clot within a hyperintense joint effusion. Fluid–fluid levels may also be seen. In chronic cases the haemosiderin-laden synovium appears irregularly thickened and markedly hypointense on T2 weighted

Fig. 66.23  Haemophilia. AP radiograph of the right elbow showing overgrowth of the radial head (arrow) and erosion of the proximal ulna (arrowhead). (Image courtesy of Dr Emma Rowbotham.)

CHAPTER 66  Bone Marrow Disorders: Non-Neoplastic Conditions

1735

A

Fig. 66.24  Haemophilia. AP radiograph of the left ankle showing flattening of the talar dome with irregular loss of joint space (arrow) and large subchondral cysts (arrowhead). (Image courtesy of Dr Emma Rowbotham.)

B Fig. 66.26  Haemophilia. (A) Sagittal T1 weighted spin echo and (B) short-tau inversion recovery magnetic resonance imaging of the ankle showing features of chronic haemophilic arthropathy with loss of joint space, large subchondral cysts (arrows) and marginal osteophytes due to secondary osteoarthritis. (Image courtesy of Dr Emma Rowbotham.)

Fig. 66.25  Haemophilia. Sagittal T2* weighted gradient echo magnetic resonance imaging of the ankle showing hypointense chronic haemorrhagic synovitis (arrows). (Image courtesy of Dr Emma Rowbotham.)

images, particularly gradient-echo sequences (Fig. 66.25). Other features include focal cartilage defects and subchondral cysts (Fig. 66.26). Small soft-tissue haematomas are common and, when repetitive, may lead to contractures. Of more importance is the infrequent progressive haemorrhage close to muscle attachments, usually with no history of

injury resulting in the haemophilic pseudotumour. This is more common in adults, with a reported incidence of 1.56%. Most are reported in the pelvis, thigh or calf. Subperiosteal or intraosseous haemorrhage causes pressure erosion of the bone, particularly the iliac blade in relation to the extensive origin of the iliacus muscle, and the femur. Pathological fracture may be the first manifestation of a haemophilic pseudotumour of bone. Radiographically, a haemophilic pseudotumour appears as a soft-tissue mass which may be calcified. Bone lesions may show geographical lytic destruction with cortical thickening. CT identifies the thick, relatively hyperdense pseudocapsule with a hypodense centre (Fig. 66.27). The MRI SI varies with the age of the contained blood, progressing from being

1736

SECTION F  Oncological Imaging

Fig. 66.27  Haemophilic Pseudotumour. Axial computed tomography study of the left ilium showing a calcified mass (arrows) in the left iliacus with chronic erosion of the adjacent iliac blade.

SUMMARY BOX:  Disorders of Coagulation • Include haemophilia A and B and von Willebrand disease • Clinical and imaging features typically due to recurrent haemarthrosis or soft-tissue haemorrhage • Recurrent haemarthrosis eventually results in chronic haemorrhagic synovitis, articular cartilage damage and premature osteoarthritis • Less common findings include soft-tissue haematomas and haemophilic psuedotumours

isointense to muscle in the first week on T1 weighted, and subsequently becoming hyperintense on both T1 weighted and T2 weighted images. The wall tends to be hypointense because of its contained haemosiderin. Mural nodules provide a highly characteristic appearance. Treatment is by management of the haemophilia and exploratory operation is to be avoided.

FURTHER READING Chan, M.W., Leckie, A., Xavier, F., 2013. A systematic review of MR imaging as a tool for evaluating haemophilic arthropathy in children. Haemophilia 19, e324–e334.

Cross, S., Vaidya, S., Fotiadis, N., 2013. Hemophilic arthropathy: review of imaging and staging. Semin. Ultrasound CT MR 34, 516–524. Di Minno, M.N., Ambrosino, P., Quintavalle, G., et al., 2016. Assessment of hemophilic arthropathy by ultrasound: where do we stand? Semin. Thromb. Hemost. 42, 541–549. Dinan, D., Epelman, M., Guimaraes, C.V., et al., 2013. The current state of imaging pediatric haemoglobinopathies. Semin. Ultrasound CT MR 34, 493–515. Fung, E.B., Vichinsky, E.P., Kwiatkowski, J.L., et al., 2011. Characterization of low bone mass in young patients with thalassemia by DXA, pQCT and markers of bone turnover. Bone 48, 1305–1312. Ganguly, A., Boswell, W., Aniq, H., 2011. Musculoskeletal manifestations of sickle cell anaemia: a pictorial review. Anemia 2011, http://dx.doi.org/ 10.1155/2011/794283. 794283. Gonzalez, F.M., Mitchell, J., Monfred, E., et al., 2015. Knee MRI patterns of bone marrow reconversion and relationship to anemia. Acta Radiol. 57, 964–970. Haidar, R., Musallam, K.M., Taher, A.T., 2011. Bone disease and skeletal complications in patients with β thalassemia major. Bone 48, 425–432. Hong, M., He, G., 2017. The 2016 revision to the World Health Organization classification of myelodysplastic syndromes. J. Transl. Int. Med. 5, 139–143. Inusa, B.P., Oyewo, A., Brokke, F., et al., 2013. Dilemma in differentiating between acute osteomyelitis and bone infarction in children with sickle cell disease: the role of ultrasound. PLoS ONE 8, e65001. doi:10.1371/ journal.pone.0065001. Jaganathan, S., Gamanagatti, S., Goyal, A., 2011. Musculoskeletal manifestations of hemophilia: imaging features. Curr. Probl. Diagn. Radiol. 40, 191–197. Keshava, S.N., Gibikote, S., Doria, A.S., 2015. Imaging evaluation of hemophilia: musculoskeletal approach. Semin. Thromb. Hemost. 41, 880–893. Kosaraju, V., Harwani, A., Partovi, S., et al., 2017. Imaging of musculoskeletal manifestations in sickle cell disease patients. Br. J. Radiol. 90 (1073), 20160130. doi:10.1259/bjr.20160130. Kwiatkowska-Pamuła, A., Ziółko, E., Muc-Wierzgoń, M., et al., 2013. Usefulness of spinal magnetic resonance imaging in patients with myelodysplastic syndromes. Pol. J. Radiol. 78, 42–49. Małkiewicz, A., Dziedzic, M., 2012. Bone marrow reconversion – imaging of physiological changes in bone marrow. Pol. J. Radiol. 77, 45–50. Martin, A., Thompson, A.A., 2013. Thalassemias. Pediatr. Clin. North Am. 60, 1383–1391. Seif El Dien, H.M., Esmail, R.I., Magdy, R.E., et al., 2013. Deferoxamineinduced dysplasia-like skeletal abnormalities at radiography and MRI. Pediatr. Radiol. 43, 1159–1165. Timmer, M.A., Pisters, M.F., de Kleijn, P., et al., 2015. Differentiating between signs of intra-articular joint bleeding and chronic arthropathy in haemophilia: a narrative review of the literature. Haemophilia 21, 289–296. Zhu, G., Wu, X., Zhang, X., et al., 2012. Clinical and imaging findings in thalassemia patients with extramedullary hematopoiesis. Clin. Imaging 36, 475–482.

67  Imaging for Radiotherapy Planning Peter Hoskin, Ananya Choudhury, Lizbeth M. Kenny

CHAPTER OUTLINE Types of Radiotherapy, 1737 The Radiotherapy Pathway, 1741 Image-Guided Radiotherapy, 1745

Radiation therapy has been used as a treatment for cancer for more than 100 years, with its earliest roots dating back to the discovery of x-rays in 1895. Its development in the early 1900s is largely due to the work of Marie Curie (1867–1934), who discovered the radioactive elements polonium and radium in 1898. Despite these distant origins, radiotherapy remains at the forefront of the treatment for cancer. Approximately 60% of cancer patients currently receive radiation therapy at some stage during their illness, with 75% of these treated with curative intent. Despite major advances in drug treatments for cancer, these contribute relatively little to overall cure. The use of radiotherapy continues to increase, and this trend is likely to continue as the cancer incidence rises and as both primary and metastatic cancers get smaller due to earlier detection by modern imaging. All advances in radiation treatment are driven by the desire to maximally treat cancer tissue, while avoiding dose to adjacent structures; modern imaging, multi-leaf collimation and computer technology have revolutionised our ability to achieve this goal. As a result of these advances, radiotherapy has been progressed from two-dimensional (2D) techniques to highly precise three-dimensional (3D) and four-dimensional (4D) conformal treatments that use axial tomographic images of the patient’s anatomy to guide intensity-modulated, image-guided therapy. Advances in imaging have allowed radiation oncologists to delineate and target tumours more accurately, thus achieving better treatment outcomes, improved organ preservation and fewer side effects. Software enables accurate and deformable fusion of diagnostic imaging to planning computed tomography (CT) studies, thus enabling greater accuracy in defining the tumour. With the development of functional imaging techniques such as positron emission tomography (PET), dynamic contrast-enhanced CT and multiparametric magnetic resonance imaging (MRI), it is now possible to integrate biological information (e.g. tumour oxygenation, cellular proliferation or blood flow) into the radiotherapy planning process. Therefore, imaging is critical at almost every stage in the practice of modern radiotherapy (Fig. 67.1).

TYPES OF RADIOTHERAPY External Beam Radiotherapy External beam radiotherapy constitutes the mainstay of radiation treatment. It uses a radiation beam that originates at a distance from

Functional Imaging in the Radiotherapy Process, 1748 Radiomics, 1750 Conclusion, 1750

the patient and is directed towards a defined treatment volume. Approximately 85% of all therapeutic radiation exposures are delivered using external beam techniques. Various types and energies of radiation can be delivered in this way, including electromagnetic radiation such as x-rays and γ-rays, or particles such as electrons and protons. Higher energies of radiation penetrate deeper into body tissues. As a result, low-energy x-rays (60 to 300 keV) are reserved for the treatment of skin cancers and superficial subcutaneous tumours. Most external beam radiotherapy treatments use megavoltage x-rays or electrons (6 to 18 MeV) generated by a linear accelerator, which remains the workhorse of a modern radiotherapy department (Fig. 67.2).

Conventional External Beam Radiotherapy Conventional radiotherapy refers to techniques in which the treatment volume is defined by simple geometric parameters. In general, no attempt is made to delineate the tumour outline or to shape the radiation dose distribution to conform to the tumour volume. This is commonly practised for palliative treatments where long-term normal tissue toxicity is less relevant. The irradiated volume can be defined clinically or fluoroscopically, but more often CT simulation is used (Fig. 67.3). The radiation fields tend to run parallel to each other, creating a box-like treatment volume.

Three-Dimensional Conformal Radiotherapy The incorporation of axial imaging data allows 3D reconstruction of the tumour and surrounding organs. This provides more accurate localisation of the target volume and more information regarding the amount of normal tissue that will be irradiated. The radiotherapyplanning computer software uses the attenuation coefficient information (Hounsfield units [HU]) derived from the CT image on a voxel-by-voxel basis to predict the attenuation of each therapeutic radiation beam as it passes through the body. As a result, the number and profile of the radiation beams can be orientated and shaped to fit the profile of the target from a beam’s eye–view using a multi-leaf collimator (Fig. 67.4). The resulting radiotherapy plan can be displayed as a colour map of radiation dose overlaid onto the anatomical CT images so that the radiation oncologist can determine whether the tumour volume will receive sufficient irradiation with acceptable normal tissue dose sparing (Fig. 67.5). By reducing the irradiated volume and the dose to the sensitive surrounding normal tissues, this technique facilitates the delivery

1737

1738

SECTION F  Oncological Imaging

Diagnosis

Decision for radiotherapy Consent

Planning imaging

X-ray simulator

CT

CT fusion with MR/PET

Volume definition

Dosimetry calculation

Plan verification X-ray simulator Virtual CT

Treatment delivery

refers to a variety of techniques in which the radiation beams are not only shaped and orientated to conform to the tumour volume but also the intensity of radiation is modulated across each treatment beam. This technique can produce dose distributions that conform highly to complex shapes, including treatment volumes that wrap around sensitive normal structures such as the spinal cord (Fig. 67.6), enabling high-dose delivery to the tumour volume whilst sparing the normal structures. IMRT can be achieved using a number of different technologies. Most commonly, several static radiation fields in the same plane of orientation are used, similar to the situation for 3D conformal radiotherapy but with varying dose flux across the profile of the beam, which is achieved by moving the leaves of the multi-leaf collimator across the beam at varying rates. Alternatively, arc therapies use a number of non-coplanar beam arcs in which the radiation is delivered using multiple ‘stop and shoot’ beams or as a continuously moving field that varies in intensity throughout rotation. Tomotherapy is another technique for achieving IMRT in which a megavoltage x-ray source is mounted in a similar fashion to a CT x-ray source. The treatment volume is irradiated using the machine’s continuously rotating beam that is modulated in intensity whilst the patient moves through the gantry bore. Image guidance is critical when such complex highly conformal techniques are used. Motion management using 4D planning and delivery is also incorporated, particularly when treating sites affected by cardiorespiratory motion, such as the lung and breast.

Stereotactic Body Radiotherapy Verification

Megavoltage image

kV image

Cone beam kV CT

Fig. 67.1  The Radiotherapy Pathway. Processes outlined in green involve imaging. CT, Computed tomography; MR, magnetic resonance; PET, positron emission tomography.

In stereotactic body radiotherapy (SBRT) the distribution of radiation beams is in three dimensions and not in two as in traditional radiotherapy. It may be delivered using a suitably modified standard linear accelerator or dedicated machines such as CyberKnife (Fig. 67.7). It is used for intracranial tumours where precise targeting of dose may be critical and selected, small extracranial lesions such as tumours of the lung and prostate, or small solitary metastases. The precision of SBRT allows high doses to be delivered in a very limited number of fractions. This is sometimes referred to as ultra-hypofractionation in which doses per fraction can be as high as 20 Gy (conventional radiotherapy is delivered at 2 Gy per fraction).

Brachytherapy

Fig. 67.2  External Beam Radiotherapy. A patient preparing to receive external beam radiotherapy on a linear accelerator.

of higher tumour radiation doses than would be achievable using conventional techniques.

Intensity-Modulated Radiotherapy Intensity-modulated radiotherapy (IMRT) represents a further step in the development of high-precision radiotherapy delivery. The term

Brachytherapy refers to a treatment in which a radioisotope is placed onto or inside the volume to be treated. One of the key features of brachytherapy is that the irradiation affects only a very localised area around the radiation sources because dose falls off rapidly, obeying the inverse square law. As long as the sources are placed precisely within the tumour, there is minimal exposure to radiation of healthy tissues further away from the sources. This allows very high doses to be administered to the target volume. Whilst in the early days of brachytherapy live sources were placed manually, this is no longer acceptable. The exception is low–dose rate prostate brachytherapy with 125I seeds. Here the gamma energy is only 28 keV, and the dose outside the seed in which the isotope is encapsulated is very low. In other settings the usual source is iridium 192, which is used in an afterloader. This works on the principle that non-active applicators, typically metal or plastic hollow tubes, will be placed in the volume to be treated, ideally with real-time imaging to direct them accurately. Ultrasound is particularly valuable for this. Radiation is then delivered in a protected room by remote activation of the afterloader connected to the applicators. Passage of the source along the applicator is controlled, and the rate of passage will define the amount of radiation delivered; typically the source will ‘dwell’ for a few seconds at 5-mm intervals within the tumour volume. In

CHAPTER 67  Imaging for Radiotherapy Planning

Fig. 67.3  Computed Tomography Simulator Images for Treatment of Pelvic Bone Metastases. In this example the intended target is bone metastases in the right pelvis. The bones to be targeted have been defined by the field edges; shielding has been introduced to the left pelvis to protect the bladder and bowel in that region. The cross-sectional view (upper left) shows the original computed tomography image with the radiation field simulated to demonstrate correct coverage of the target and identify other structures which may be included.

Fig. 67.4  A Multi-Leaf Collimator. This is made up of individual ‘leaves’ of a high atomic-number material, usually tungsten, which can move independently in and out of the path of a radiation beam in order to block it. This device is situated in the head of a linear accelerator to shape the treatment beam to match the borders of the target tumour. For intensity-modulated treatments the leaves of a multi-leaf collimator are moved across the field while the beam is on to create fluence modulation.

1739

1740

SECTION F  Oncological Imaging

Fig. 67.5  A Three-Dimensional Conformal Radiotherapy Plan for Treatment of the Pelvic Lymph Nodes. The planning target volume is defined by the shaded green region. Four treatment beams (anterior, posterior, left and right lateral) overlap over the target volume to provide the desired dose in this region with reduced dose to surrounding normal structures. The coloured lines represent the dose gradient in a similar way to the contour lines on a map and correspond to the numbers in the top left corner. The red line represents the 95% isodose (i.e. every point on this line will receive 95% of the prescribed radiation dose). The coronal and sagittal views show the shaping of the beam by the multi-leaf collimator (MLC). Despite the use of the static MLC, the transverse image clearly shows that the small bowel anterior to the horseshoe-shaped target volume will still receive the full radiation dose.

brachytherapy, patient setup and tumour motion are less relevant because the radiation sources move with the tumour and therefore retain their correct position, addressing one of the major challenges in accurate radiation delivery. Dosimetry in brachytherapy is based on addition of the contributions from each source position within the volume. Accurate imaging of the applicators and reconstruction of the tumour are essential. Brachytherapy applicators are CT and MRI compatible. CT is often preferred to identify accurately the position of the applicators, whereas MRI will usually give superior definition of the tumour, especially in the pelvis, where prostate and gynaecological cancers are commonly treated with this technique (Figs 67.8 and 67.9). Specific MRI sequences may be developed to enable certain applications (e.g. proton-rich sequences for identification of 125I seeds in the prostate).

Particle Therapy The most common particles used in radiotherapy are electrons. These are produced in the linear accelerator. The target, which produces x-rays when bombarded with electrons, is removed so that a beam of high-energy electrons is produced instead. Electron beams have a specific range, and the effective range is approximately one-third of the accelerating energy: for example, for a 6 MeV electron beam the effective range is approximately 2 cm. The advantage of electrons is that they have a defined range with minimal exit dose beyond that point. Therefore

Fig. 67.6  An Intensity-Modulated Radiotherapy Plan for Treatment of the Left Maxilla and Bilateral Lymph Nodes. The primary tumour region has been outlined and is defined by the solid red line. This has been expanded to produce a planning target volume shown as the dark blue shaded region and will receive a radical dose to control the macroscopic tumour. The level II lymph nodes have also been defined and will be treated to a lower overall dose to control any microscopic disease spread. The use of intensity modulation has allowed shaping of the dose distribution to avoid the brainstem. The isodose lines can be seen to bend anteriorly around the brainstem, leaving this structure in a region of low radiation dose (between the green and blue isodose lines, which correspond to 50% to 70% of the prescribed dose, thereby taking the brainstem below its tolerance threshold).

they are commonly used to treat skin tumours or other superficial structures such as the ribs. Protons have an increasing role in clinical radiotherapy, with many centres across the world being established. Protons penetrate deep into body tissues and deposit most of their energy in the last few millimetres of their range, the Bragg peak, with virtually no radiation passing beyond this distance. The position of radiation deposition in the body can therefore be defined by choosing an appropriate energy of the proton beam. In practice, a range of energies to produce the ‘spread out Bragg peak’ is used to treat a defined volume. In this way, tissues behind the target volume are spared radiation dose. This may be critical in certain sites: for example, inside the cranium or around the spinal cord. It also has major advantages in children, in whom sparing of growing tissues is an important component of minimising late effects. There are only a few centres in the world that have the capability of treating patients with other types of particle, such as fast neutrons, or carbon ions. These techniques have physical advantages similar to those of protons, with sparing of tissues at depth, but may have additional biological benefits. Both neutrons and carbon ions cause dense ionisation with much greater transfer of energy along the radiation track. This results in more radiation damage and cell death. Furthermore, there is theoretical evidence that intensely ionising radiation of this sort can overcome the detrimental effects of tumour hypoxia, which is a major cause of treatment failure with standard radiotherapy.

CHAPTER 67  Imaging for Radiotherapy Planning

1741

Fig. 67.7  A Stereotactic Body Radiotherapy Plan Using the CyberKnife Robotic Radiosurgery System to Treat a Recurrent Glioblastoma of the Occipital Lobe. Each of the blue cylinders represents a pencil beam of radiation that is controlled with extreme accuracy using image guidance. The machinery can correct for movements of the tumour caused by breathing or internal organ motion. The resulting dose distribution, which can be seen in the transverse, coronal and sagittal views, conforms extremely well with the tumour volume.

SUMMARY BOX: Types of Radiotherapy • External beam • Conventional external beam radiotherapy • Three-dimensional conformal radiotherapy • Intensity-modulated radiotherapy (IMRT) • Particle therapy • Electrons • Protons • Carbon ions • Neutron • Brachytherapy

THE RADIOTHERAPY PATHWAY Target Volume Definition As radiotherapy dose delivery becomes more precise, the need for accurate delineation of the extent of cancer becomes critical and all available resources should be harnessed to define it. Clinical examination, the histopathology and surgical findings and knowledge of the natural history of the cancer help to define our sequencing and acquisition and are essential for optimal determination of target volumes. It is now possible to deliver radiation with near-millimetre accuracy, making it necessary for diagnostic technology to define tumour boundaries with similar precision. Accurate deformable fusion of diagnostic images can be invaluable in helping to define the extent of the disease. Because of the uncertainties inherent in the delineation of tumour volume and the knowledge that correlation between imaging and pathological boundaries is not absolute, it is standard practice to incorporate a safety margin around the ‘visible’ gross tumour volume (GTV) to produce a

1742

SECTION F  Oncological Imaging

B

A Fig. 67.8  Low–Dose-Rate Brachytherapy for Low-Risk Prostate Cancer. (A) Plain radiograph of the pelvis taken a few minutes after 125I seed implantation for a patient with low-risk prostate cancer. (B) Fused computed tomography and magnetic resonance imaging of the same patient on day 28 after the implant. The iodine seeds emit 35 kV x-radiation (maximum energy) and have a half-life of 59 days. As a result, the prostate is irradiated over a period of months rather than weeks or days. Following the implant procedure there is often some migration of seed positions caused by bleeding and swelling from the trauma of implantation. It is therefore important to determine the final dose distribution from imaging a few weeks after seed insertion rather than from their initial position.

Intrauterine tube Intersal catheter Bladder with indwelling

Vaginal ring

Cervical tumour

Fig. 67.9  Magnetic Resonance–Guided Brachytherapy for Cervical Cancer. Sagittal view showing that an intrauterine tube has been placed within the uterine cavity and a ring at the top of the vagina and peripheral interstitial catheters in the periphery of the tumour to ensure adequate coverage of the tumour when the applicators are loaded with the high-dose-rate (HDR) iridium source.

clinical target volume (CTV) that accounts for extension of the tumour beyond what is visible with current diagnostic imaging. A further expansion is then made to account for set-up errors, which occur due to the inherent variability in equipment geometry and beam alignment and intrafraction tumour movement. The end result is the planning target volume (PTV) (Fig. 67.10). To capitalise on the accuracy of modern radiotherapy equipment and make additional reductions in normal tissue toxicity whilst increasing dose to the tumour, this safety margin needs to be reduced. This can occur only if (1) there is a higher level of confidence in defining tumour volumes and (2) tumour movements during radiation delivery can be visualised and accounted for. Both of these require advanced imaging capabilities.

Clinical Volume Definition (Non-Imaging Based) This is used for visible tumours or for palliative treatments where accurate tumour localisation is not required. For skin cancers, the visible tumour boundary is outlined and a margin for microscopic spread is marked on the skin surface. For treatment to deeper structures the field borders can be defined using anatomical landmarks.

Computed Tomography Simulation CT simulation has largely replaced the traditional treatment simulator which was based on a kilovoltage x-ray image defining the area to be treated using bone landmarks but with no soft-tissue definition. CT

CHAPTER 67  Imaging for Radiotherapy Planning

Fig. 67.10  Contrast-Enhanced Radiotherapy Planning Computed Tomography Image. The gross tumour volume (GTV) of a lung tumour has been outlined (blue). The treatment planning computer software has been used to expand this volume into a clinical target volume (CTV) to account for the possible microscopic spread of tumour cells into surrounding tissues (orange). This volume has, in turn, been expanded into a planning target volume (PTV) to account for setup error and tumour motion during treatment (green). The treatment plan will then be created to encompass the PTV with the required treatment isodose.

error provides excellent bone and soft-tissue imaging as well, as tissue density information, which is necessary for the accurate calculation of the dose distribution, by defining x-ray absorption characteristics. In addition to producing a CT series of images, digitally reconstructed radiographs (DRRs), mimicking a conventional radiograph in any field direction, are derived. Radiotherapy delivery is then accurately localised in one of two ways: • Virtual simulation, in which the tumour volume may be outlined formally or simply viewed in three dimensions and the fields necessary to cover the volume are simulated on the computer screen. This will typically be used for simple palliative treatments to internal organs or bones and use single or opposed fields for which the subsequent dose distribution can be calculated (Fig. 67.11). • CT conformal planning, in which the CT data are harnessed to a planning system using computer algorithms to calculate complex dose distributions. The target volume is defined by contouring directly onto the CT images and multi-field plans. Varying tissue compensation may be used to deliver a homogeneous dose to the tumour volume and minimise dose to surrounding normal organs at risk (Figs 67.5 and 67.6). The advantages of virtual simulation over conventional simulation include a faster simulation procedure, better soft-tissue imaging, potential for fusion of diagnostic images, collection of tissue density information, more precise tumour localisation, more precise organ definition, the collection of 3D anatomy data and the facilitation of complex planning techniques.

Image Fusion CT is the imaging platform for all modern radiation dosimetry planning systems. Tissue density information derived from the Hounsfield number of each voxel of the planning CT images is required for accurate estimation of the attenuation of the delivered radiation, which defines the dose distribution. However, whilst currently a CT image is always required for external beam planning, it may not be the best imaging technique for tumour volume and organ at risk definition. For example, pelvic and brain tumours are often more accurately defined using MRI, and tumours

1743

of the head and neck and lung benefit from fluorodeoxyglucose (FDG) PET/CT for locating the tumour. Image fusion can maximise the value of imaging for radiotherapy planning (Fig. 67.12). However, it is critical to understand the inherent difficulties and potential inaccuracies in fusing multiple imaging data sets and to appreciate the difference between rigid and deformable fusion. These include variations in resolution, spatial distortions that can occur with certain methods (e.g. MRI), consistency of PET standardised uptake values (SUV) thresholding, and physiological and anatomical changes that occur during the time between the imaging acquisitions alongside the increased workload and cost. Working closely with diagnostic radiologists and PET CT physicians helps greatly to outline tumours to the best possible extent. CT is still the basis of current planning systems. However, as MRI linacs become more widely available, MRI will evolve, overcoming the issues of spatial distortion. Specific examples of where multimodality imaging currently has a clear role include: • Bronchial tumours, which frequently cause collapse of some of the surrounding normal lung tissue. 18F-fluorodeoxyglucose (18F-FDG) PET/CT can distinguish active malignant tissue from collapsed lung. • Hodgkin and non-Hodgkin lymphoma, where current protocols using radiotherapy define involved node or involved site volumes by initial PET positivity. • Prostate cancer, where multiparametric MRI can define dominant tumour locations, which may be used for subvolume definition to deliver a higher dose to those areas. • Central nervous system tumours, where MRI gives superior definition to CT alone. • Head and neck cancers, where PET CT can help to locate small nodal involvement and where MRI may give much better soft-tissue delineation, especially where there is dental artefact.

Dosimetry Once the PTV and normal organs have been defined in three dimensions, the imaging data are imported into a planning system, which contains algorithms enabling complex calculation of radiation dose distribution from multiple beams. This will take account of a number of variables, including beam energy, size, shape and direction. In modern planning processes the algorithm will be driven by defining target dose for the tumour volume, and objectives for restricting dose to surrounding organs at risk. A balance between these is sought to provide an optimal dose distribution within the capabilities of the equipment.

Delivery and Verification The settings for the linear accelerator will be defined in terms of beam energy, beam number and size, beam modification in shape and dose flux and details of the setting of the beam in relation to the patient, usually based on fiducial tattoos on the skin or a plastic immobilisation shell. For external beam treatment, the radiotherapy plan is based on the planning CT images. One of the major challenges in delivering radiotherapy is accounting for the variations that occur day to day for a treatment programme that may involve daily delivery of highly localised radiation distributions over 30 to 40 fractions spanning 6 to 8 weeks. These are defined as either random or systematic errors and result from both geometric and patient-related factors. The complexity of a linear accelerator mounted on a mechanical gantry delivering treatment to a patient on a couch with 3D or 4D movement options inevitably results in a few millimetres of variation. However, of greater concern are the variations that can occur in patient position and shape together with the challenge of accounting for internal organ motion. The set of data used for producing the plan is a snapshot of the patient’s position and anatomy at a single point in time. All internal organs are subject to a degree of movement, which can occur daily

1744

SECTION F  Oncological Imaging

Fig. 67.11  Virtual Simulation for Treatment of an Invasive Bladder Cancer in a Patient With Bilateral Hip Replacements. The hip prostheses prevent the use of lateral or oblique beam angles that would normally be used for pelvic treatments, allowing only anterior and posterior fields (top right). The bladder has been contoured on each computed tomography slice (green). Anterior and posterior ‘beam’s eye view’ digitally reconstructed radiographs are shown (bottom panels). On these views the beam has been shaped by the multi-leaf collimator. Despite the beam shaping, the fact that only two beams could be used results in a column of high dose through the pelvis in the anteroposterior direction without any possibility of sparing the rectum (top left panel).

(interfraction motion) or during the treatment delivery (intrafraction motion). For some organs, such as the lung, the movement is predictable, with a repetitive cyclical motion in three dimensions. For other organs, such as the bowel, random movements occur with the passage of gas and faeces. Although an allowance for this is made when defining the PTV, it is important to ensure that at each radiation exposure (fraction) the target is within the treated volume. This is the basis of image-guided radiation therapy (IGRT).

SUMMARY BOX:  The Radiotherapy Pathway The radiotherapy pathway consists of: • Target volume definition • Dosimetry • Delivery and verification

CHAPTER 67  Imaging for Radiotherapy Planning

1745

Fig. 67.12  Computed Tomography–Magnetic Resonance Image Fusion Illustrated in a Patient With Brachytherapy Catheters in the Prostate. Automated registration software is now widely available, but manual correction for more accurate registration is still sometimes required. Inaccuracies can occur if there has been patient or organ movement between the acquisition of the two imaging sets. In addition, differences in resolution, spatial distortion and slice thickness can affect the reliability of fusion. Deformable registration programs, where one image is distorted to fit the other more precisely, are currently under evaluation.

IMAGE-GUIDED RADIOTHERAPY IGRT makes use of many different imaging strategies. IGRT techniques can be split into planar (2D) imaging, volumetric (3D) imaging and imaging over time (4D). In addition, for some anatomical sites, implanted fiducial markers can be used to localise the treatment.

Planar (Two-Dimensional) Imaging This is when two or more static images are acquired, usually at 90 degrees to each other (i.e. anteroposterior and lateral) (Fig. 67.13). It allows comparison of the bony anatomy or of a target visible by plain x-ray (such as a lung tumour) in all three axes (superoinferiorly, laterally and anteroposteriorly). However, images acquired at the energies used for radiotherapy (6 to 18 MeV) are not of diagnostic quality, with poor contrast between bone, soft tissue and air compared with low-energy (kV) imaging and hence the evolution of volumetric imaging techniques on the linear accelerator.

Volumetric (Three-Dimensional) Imaging Volumetric imaging acquires a 3D image in the treatment position on the linear accelerator before or during radiotherapy. Internal structures including the target organ and surrounding normal tissues can be identified in relation to the dosimetric map, which can be overlayed on the CT image. There are four methods for obtaining a volumetric image on the linear accelerator: • Cone beam CT (CBCT). For most standard linear accelerators, volumetric imaging is available via CBCT technology, which is a kV tube mounted at 90 degrees to the linear accelerator head and is rotated

around the patient using the linear accelerator gantry. Both the treatment head (MV) and the CBCT system (kV) have an imaging capability (Fig. 67.14). • Megavoltage CT. This technique uses a megavoltage energy fan beam to create a volumetric image for verification with helical imaging as used in conventional CT. • CT on rails. This consists of a CT unit in the same room as the linear accelerator. The patient couch can be rotated at 180 degrees to transfer from the linear accelerator to the CT. • Ultrasound. Ultrasound can provide volumetric images for IGRT in patients with prostate and bladder cancer.

Four-Dimensional Imaging This describes the process of imaging the tumour and relevant structures over a period of time. Before the advent of 4D image guidance, the GTV was expanded to create a CTV to account for microscopic spread. An additional margin was then added for setup variation (including any motion) to create the PTV (as described earlier). However, kilovoltage fluoroscopy or CBCT can be performed before the radiotherapy treatment to quantify tumour motion. The GTV can then be defined at the extremes of motion and at the midpoint of the movement. These are expanded for microscopic disease to create their respective CTVs. The union of these CTVs is used to create an internal target volume with a smaller margin added for geometric setup errors to create the PTV (Fig. 67.15). Alternatively, 4D kV or MV imaging can be used during radiotherapy to track the tumour. Fiducial markers are often used to facilitate this where there is no clear definition on imaging (Fig. 67.16). The CyberKnife system (Accuray, Sunnyvale, CA, USA) uses a combination of x-ray imaging and optical tracking in its motion tracking system. Two

1746

SECTION F  Oncological Imaging

A

C

B

D

E

F

Fig. 67.13  Planar (Two-Dimensional) Image Guidance. Two static kilovoltage images have been acquired at 90 degrees to each other using a linear accelerator-based on-board imaging device with the patient on the treatment couch and about to be irradiated (D and F). Bony landmarks have been outlined on the digitally reconstructed radiographs, which have been derived from the planning computed tomography imaging (C and E). These outlines have been transposed onto the kilovoltage images to determine whether there has been any deviation from the planned treatment (A and B). In this case there is clearly a discrepancy of 3 to 4 mm in the craniocaudal direction. This is best seen on the anteroposterior image (A) where the mandible and spinous processes are notably higher than intended. To prevent an inaccurate exposure, either the patient will be repositioned or the fields will be delivered with the appropriate correction. This situation is not uncommon for head and neck cancer treatments where there is often weight loss during the therapy period and as a result the immobilisation head shell becomes loose fitting, allowing movement to occur.

Fig. 67.14  Image Guidance Using Cone Beam Computed Tomography (CBCT). This patient is receiving treatment for a T4 laryngeal tumour. A CBCT image has been obtained with the patient on the treatment couch and about to be irradiated. The CBCT image has been reconstructed in the transverse, sagittal and coronal planes and fused with the planning CT image. The fusion can be best seen on the sagittal and coronal images because the CBCT has been obtained for a shorter craniocaudal length than the planning image. In contrast to the planar imaging, the position of the soft tissues and the tumour itself can be clearly seen and contoured. Any discrepancy between the planning image and the CBCT image can therefore be corrected in three dimensions to ensure tumour coverage and avoidance of normal tissues.

CHAPTER 67  Imaging for Radiotherapy Planning

1747

Fig. 67.15  Cone Beam Computed Tomography Being Used to Quantify the Motion of a Peripheral Small Lung Tumour. The degree of movement can be minimised by breath-holding techniques or by using mechanical methods to splint the diaphragm. Alternatively, respiratory gating can be used where the beam is automatically turned off when chest or tumour motion exceeds a predefined threshold.

A

B Fig. 67.16  Stereotactic Body Radiotherapy. Two areas of recurrent, chemotherapy refractory, metastatic testicular seminoma have been outlined in red (A). The intention is to treat these metastases with stereotactic body radiotherapy. To achieve accurate dose deposition, fiducial markers have been inserted into both tumours (arrows, B) so that any tumour motion can be tracked in real time during treatment delivery. (Image courtesy of Dr Peter Ostler, Mount Vernon Cancer Centre, UK.)

kilovoltage x-ray tube and detector pairs are mounted in the treatment room at right angles. X-rays are taken periodically during treatment to determine the location of the fiducial markers within the target. For thoracic tumours there is also continuous optical tracking of the patient’s skin to detect the breathing motion. These two data sources are then combined by the tracking software and a mapping function between the position of the external skin markers and the position of the internal target is computed. By knowing the motion of the internal target, the CyberKnife robot arm can adjust the position of the radiation beam to follow that motion.

Adaptive Radiotherapy During a course of radiotherapy, both the tumour and the surrounding anatomy can change. Strategies have been developed that take advantage of image guidance to respond to these changes by amending the plan on a daily basis. These strategies reduce the volume of the surrounding normal tissue that is irradiated to minimise toxicity. Adaptive radiotherapy has been implemented in clinical practice in certain sites, including bladder and cervix. Bladder volume varies significantly day-to-day despite using an empty bladder technique or drinking protocols. To account for this variation

1748

SECTION F  Oncological Imaging

Fig. 67.17  Adaptive Radiotherapy for Bladder Cancer. The bladder, shaded in red, forms the clinical target volume (CTV). Three incremental planning target volumes (PTV) are produced, outlined in blue, green and purple on the transverse, sagittal and coronal views. Before radiotherapy delivery, a cone beam CT image is taken and the plan, with the most appropriate PTV where the whole of the CTV is adequately covered, is chosen for treatment. This process if carried out on a daily basis during radiation treatment to ensure that the clinical target is adequately covered whilst minimising the volume of normal tissue irradiated.

when irradiating the bladder, large margins of up to 20 mm have been used when expanding from CTV to PTV, resulting in the irradiation of a large volume of normal tissue. Using daily CBCT to visualise soft tissue, the radiotherapy plan can be adapted according to the variation observed between the original plan and the CBCT. The most commonly used strategy is called ‘plan of the day’ where a library of plans is produced from the planning CT to cover a range of different bladder volumes. Based on the daily CBCT, the ‘best fit’ plan is chosen from the library for treatment on the day (Fig. 67.17). Another strategy is that of a plan based on a composite volume. Here, daily CBCT images from the first week of radiotherapy are used to calculate an average CTV, which is then expanded by a small margin of up to 10 mm, and a plan to treat this average PTV is created and used for the remaining radiation treatments. Bladder and rectal volumes also alter the position of the uterus and hence move the target when treating cancer of the cervix by as much as 3 or 4 cm. It also has a major impact on the volume of small bowel in the field. Adaptive radiotherapy has an important role in ensuring consistent delivery of the planned dose to the target during a fractionated course of 25 treatments. Typically, three planning scans are taken: one with bladder full, one empty and one mid-way. The CTV is then defined on each and an individual radiotherapy plan produced. Using daily CBCT images the ‘plan of the day’ is then chosen for each day’s delivery.

Magnetic Resonance Imaging-Guided Radiotherapy Further advances in IGRT have been seen with the advent of MRI-guided radiotherapy where a linear accelerator has been combined with an MRI to produce diagnostic-quality images both before and during radiotherapy (Fig. 67.18). At present there are two commercially available systems: the Viewray MRIdian system with a 0.35 T magnet and the Elekta Unity incorporating a 1.5 T magnet. The advantage of MRI for image-guidance compared with CBCT in anatomical sites such as the pelvis are substantial. Given that the highly controlled magnetic fields of the MRI could interfere with the complex electronics and the shielding in the linear accelerator, the combination of these two machines has been a feat of electronic engineering. By placing the linear accelerator outside the Faraday cage, both machines are effectively isolated. The

magnetic field has the potential to interfere with dose deposition. Although neutrally charged photons are unaffected by a magnetic field, negatively charged secondary electrons are deflected. This is the electron return effect or Lorentz force. Practically, the electron return effect can result in increased dose deposition at the surface or at the air–tissue interface and could be as much as 30%. Advanced radiotherapy techniques, such as IMRT, enable modulation to be used to offset this increased dose. For optimal use of MRI-IGRT, specific MRI sequences are required with high geometric fidelity and no artefacts obscuring or deforming patient anatomy. Sequences must be fast compared with diagnostic sequences to enable tumour motion tracking and reduced time in the treatment position. The ultimate goal of treatment on a combined MRI and linear accelerator is not only precise IGRT but real-time online adaptation with daily bespoke fast re-planning according to the anatomy of the day and incorporation of functional imaging in a sequential manner. The delivery of such treatment requires innovative multidisciplinary workflows, including radiology, radiography, medical physics and radiation oncology.

SUMMARY BOX: Image-Guided Radiotherapy Image-guided radiotherapy consists of: • Planar (two-dimensional) imaging • Volumetric (three-dimensional) imaging • Four-dimensional imaging • Adaptive radiotherapy

FUNCTIONAL IMAGING IN THE RADIOTHERAPY PROCESS Inclusion of Biological Information to the Treatment Process Several imaging techniques produce valid biomarkers for radiobiologically relevant tumour characteristics such as hypoxia, cellular proliferation,

CHAPTER 67  Imaging for Radiotherapy Planning

1749

Fig. 67.18  Images from Magnetic Resonance (MR) Linac. Images for a radical treatment to the cervix showing superior image quality from MR. The bottom set of images have had the relevant treatment target volume (red) and organs at risk (bladder—yellow, rectum—green) added to give precise information on the doses to be received from that linac exposure and to facilitate any necessary adjustments. (Courtesy V Valentini, LBoldrini, Gemelli Hospital, Rome.)

vascularity and clonogen density. The ability to incorporate such biological information into radiotherapy planning allows the possibility of further manipulation of the radiation dose distribution. Increasing the dose administered to relatively resistant tumour regions and moderation of the dose to sensitive areas should achieve better tumour control with less toxicity. Mathematical modelling studies have demonstrated theoretical advantages for biological conformality. However, the translation of the theory into clinical practice has been limited.

One of the most important and relevant aspects of the tumour microenvironment, which has a direct impact on the success of radiotherapy treatment, is tumour oxygenation. There is compelling evidence that hypoxic tumour regions are far less radiosensitive than those that are well oxygenated. Several imaging methods have been validated, including 18F-misonidazole PET and blood oxygenation level–dependent magnetic resonance imaging (BOLD-MRI). Although image fusion enables incorporation of information from these techniques

1750

SECTION F  Oncological Imaging

Fig. 67.19  Biologically Optimised Radiotherapy. A dominant intraprostatic lesion (DIL) in the right posterolateral peripheral zone has been defined using multiparametric magnetic resonance imaging (black outline on the transverse, coronal and sagittal images). High–dose-rate brachytherapy catheters have been inserted under general anaesthetic. The planning computer optimisation software has been programmed to maximise the radiation dose to the DIL and limit the dose to the rest of the prostate to a defined ceiling. The dose volume histograms (top left) demonstrate the dual dose levels. The DIL and DIL planning target volume (PTV) (dark blue and yellow lines) receive a higher dose as a proportion of their volume than the non-dominant prostate and the non-dominant prostate PTV (light blue and red lines).

in the radiotherapy planning process, translation into clinical radiation delivery programmes has yet to evolve. Image resolution remains a concern, as does the fact that the tumour physiology is a dynamic entity and sequential imaging is required to take account of this. Diffusion-weighted imaging depicts cell density and has been successful in defining dominant lesions in prostate cancer on T2 weighted MRI in conjunction with dynamic contrast-enhanced (DCE) sequences, this information is now being used to modulate dose distribution in both external beam and brachytherapy techniques (Fig. 67.19) but requires long-term clinical validation before it becomes widely accepted.

RADIOMICS The concept that images are sources of big data has been encompassed in the emerging field of radiomics. Algorithms have been used to extract quantitative features from medical images with the potential for improving prognostication and forming part of a predictive model for treatment outcomes. The quantitative image features can be broadly categorised into groups based on shape, size, volume, intensity and texture. The mining of radiomic data and its association with genomics is termed radiogenomics and offers the potential for combining biological imaging with the genetic footprint of a cancer for personalised treatment stratification.

CONCLUSION Radiotherapy requires close integration of both diagnostic and therapeutic radiation technology and the teams involved in both disciplines.

Capitalising on advanced and expertly acquired imaging and on the knowledge of diagnostic radiologists and PET physicians greatly enhances the ability of radiation oncologists to accurately map out the extent of known disease on imaging. A close and cohesive working relationship between the radiation oncology multidisciplinary team members (radiation oncologists, medical physicists and radiation therapists) into the planning, delivery and quality assurance of sophisticated radiation therapy greatly enhances the quality of care. Modern advanced imaging has revolutionised radiation therapy in every aspect of planning and delivery of treatment. In the future, routine integration of MRI in treatment delivery and even more sophisticated image fusion will optimise the use not only of anatomical information but also of physiological information integrating this with the morphological data. Modern radiotherapy represents sophisticated application of state-of-the-art imaging for diagnosis, target and organ at risk volume definition, tissue density measurements and verification of the therapeutic beam delivery. This translates directly into patient benefit.

FURTHER READING Dimopoulos, J.C., Petrow, P., Tanderup, K., et al., 2012. Recommendations from Gynaecological (GYN) GEC-ESTRO Working Group (IV): basic principles and parameters for MR imaging within the frame of image based adaptive cervix cancer brachytherapy. Radiother. Oncol. 103 (1), 113–122. Kerkmeijer, L.G., Fuller, C.D., Verkooijen, H.M., et al., 2016. MR-Linac Consortium Clinical Steering Committee. Quality assurance, and technical development. Front. Oncol. 6, 215. eCollection 2016.

CHAPTER 67  Imaging for Radiotherapy Planning Kibrom, Awet Z., Knight, Kellie A., 2015. J Adaptive radiation therapy for bladder cancer: a review of adaptive techniques used in clinical practice. Med. Radiat. Sci. 62 (4), 277–285. Lambin, P., Leijenaar, R.T.H., Deist, T.M., et al., 2017. Radiomics: the bridge between medical imaging and personalized medicine. Nat. Rev. Clin. Oncol. 14 (12), 749–762. McPartlin, A.J., Li, X.A., Kershaw, L.E., et al., 2016. MR-Linac consortium. MRI-guided prostate adaptive radiotherapy—a systematic review. Radiother. Oncol. 119 (3), 371–380. Padhani, A.R., 2011. Integrating multiparametric prostate MRI into clinical practice. Cancer Imaging 11 (SpecA), S27–S37. Staffurth, J., 2010. A review of the clinical evidence for intensity-modulated radiotherapy. Clin. Oncol. (R. Coll. Radiol.) 22, 643–657.

1751

The MRI-Linear Accelerator Consortium: Evidence-Based Clinical Introduction of an Innovation in Radiation Oncology Connecting Researchers, Methodology, Data Collection. Thörnqvist, S., Hysing, L.V., Tuomikoski, L., et al., 2016. Adaptive radiotherapy strategies for pelvic tumors—a systematic review of clinical implementations. Acta Oncol. (Madr) 55 (8), 943–958. Thorwarth, D., Schaefer, A., 2010. Functional target volume delineation for radiation therapy on the basis of positron emission tomography and the correlation with histopathology. Q. J. Nucl. Med. Mol. Imaging 54, 490–499. Webster, G.J., Kilgallon, J.E., Ho, K.F., et al., 2009. A novel imaging technique for fusion of high-quality immobilised MR images of the head and neck with CT scans for radiotherapy target delineation. Br. J. Radiol. 82, 497–503.

68  Functional and Molecular Imaging for Personalized Medicine in Oncology Eva M. Serrao, Avnesh S. Thakor, Vicky Goh, Ferdia A. Gallagher

CHAPTER OUTLINE Personalised Medicine in Oncology, 1752 Dynamic Contrast-Enhanced Computed Tomography, 1752 Magnetic Resonance Imaging, 1756 Positron Emission Tomography, 1761

PERSONALISED MEDICINE IN ONCOLOGY One of the major aims of oncological imaging is to detect and differentiate a tumour from normal tissue and thus it is necessary to understand the fundamental cellular changes that occur when a tumour forms, and how these can be used to generate tissue contrast. On the very simplest level, the differences in x-ray attenuation and water content between cancer and its surrounding tissues can be used to distinguish cancer from normal tissue using computed tomography (CT) and magnetic resonance imaging (MRI), respectively. The fundamental tissue, cellular and molecular changes that form the hallmarks of cancer are increasingly being understood and this knowledge is now being applied to the development of new imaging biomarkers, which will be more specific and sensitive for cancer detection than morphological information alone. Examples include the use of CT and MRI contrast agents to probe angiogenesis, as well as positron emission tomography (PET) tracers to detect alterations in cellular energetics and proliferation within cancerous tissue. In addition to identifying tumours, imaging biomarkers can be used to assess the efficacy of treatment such as chemotherapy and radiotherapy. Traditionally, this has been performed by identifying changes in tumour size using criteria such as the Response Evaluation Criteria In Solid Tumours (RECIST). Increasingly, these criteria are being modified to incorporate functional and metabolic information in addition to morphological measurements. New imaging biomarkers are being developed that are more specific and sensitive for the detection of early response to treatment by detecting early cellular or molecular changes that predict long-term successful outcome. The introduction of therapies that have specific molecular targets (such as bevacizumab and sunitinib) has been problematic for traditional imaging approaches as improved clinical outcome with these drugs is often not accompanied by a significant change in tumour size; for example, an antivascular drug may induce tumour necrosis with little change in the overall tumour diameter. Consequently, alternative imaging approaches are required to identify a successful early response to therapy in this context; the concept of combining a specific targeted drug with an imaging test that directly

1752

Emerging Molecular Imaging Techniques and Theranostics, 1763 Conclusion: Role of Functional and Molecular Imaging in Oncology, 1764

probes the cellular pathways affected by the drug as a companion biomarker is a very attractive approach for the future management of cancer patients. These specific targeted imaging biomarkers also open up the possibility of detecting subtle differences in drug response between patients: a cellular pathway may be upregulated in one patient but downregulated in another in response to the same drug at the same dose. Historically, a single treatment algorithm was used for all patients but, increasingly, this is being replaced by a personalised or patient-centred approach where drug therapy can be tailored to an individual patient. Medical practice is now underpinned by an understanding of the molecular biology of disease processes and complementing this with new imaging techniques to detect and monitor these processes is increasingly important. These molecular imaging methods can be defined as the visual representation, characterisation and quantification of biological processes at the cellular and subcellular levels within intact living organisms. Functional imaging is more loosely defined and includes techniques that probe physiological processes such as blood flow, metabolism and features of the tumour microenvironment that affect tissue function such as water diffusion. There is some overlap between the two terms and, often, the combination of functional and molecular imaging is used to define a range of imaging techniques, which are more specific than anatomical or morphological imaging and probe processes from a tissue to a molecular level. This chapter will explore the use of these functional and molecular techniques in oncological imaging.

DYNAMIC CONTRAST ENHANCEDCOMPUTED TOMOGRAPHY Dynamic contrast-enhanced CT techniques for assessing the vasculature have been possible since the 1970s. More recent technological advances have allowed contrast-enhanced acquisitions with high temporal sampling to be performed over a large volume (also known as perfusion CT), matched to a clinical need on an individual patient basis. Assessment of stroke and cancer therapy have propelled dynamic contrast-enhanced (DCE)-CT into the clinical arena (Fig. 68.1).

CHAPTER 68  Functional and Molecular Imaging for Personalized Medicine in Oncology

A

B

C

D Fig. 68.1  Dynamic contrast-enhanced computed tomography (CT) acquisition with parametric maps from a glioblastoma multiforme tumour: (A) contrast-enhanced CT, (B) regional blood flow, (C) blood volume and (D) permeability surface area product. The images demonstrate a vascular solid component with disruption of the blood–brain barrier best seen on the permeability surface area product map.

1753

1754

SECTION F  Oncological Imaging

SUMMARY BOX: Dynamic Contrast-Enhanced Imaging • Definition: serial acquisition of images following injection of an intravenous contrast agent allowing qualitative and quantitative parameters to be derived. • Exploits the alteration in vascular density, vascular permeability and blood flow present in a tumour. • Major clinical applications: assessment of antivascular drugs and vascular interventional procedures. Computed Tomography (CT): Advantages: • Wide availability • Low cost • Easy standardisation Disadvantages: • Radiation • Temporal resolution is limited by radiation dose Magnetic Resonance Imaging (MRI): Advantages: • Absence of ionising radiation • High contrast resolution • Temporal resolution is not limited by radiation dose Disadvantages: • Potential risks from repeated gadolinium use • Less reproducible than CT • Complex quantitative data analysis

Contrast Agent Kinetics CT contrast agents used in clinical practice are low molecular weight contrast agents (65%) Several large intercommunicating cysts (up to 10 cm) Mediastinal shift is common 10%–15% of cases Smaller than other types Small evenly sized cysts (up to 2 cm) Often associated with other congenital abnormalities ~8% of cases Large solid-appearing lesion containing microcysts (2.6 mm or 1). In short-segment disease, this ratio is reversed. The radiological features of Hirschsprung disease may be absent in the neonate because it takes time for the ganglionated bowel to dilate. It is common during the course of an enema to see features of mild enterocolitis such as ulceration, mucosal oedema and spasm (see Fig. 71.12B), but if a child has clinical features of severe colitis, an enema is contraindicated. Giant stercoral ulcers may also be seen in older children with a delayed presentation. Overall, the contrast enema has a reported sensitivity of 70% and a specificity of 83% for the diagnosis of Hirschsprung disease, and the negative predictive value of a normal contrast enema in patients older than 1 month of age is 98%. In TCA the contrast enema may be entirely normal. Positive findings include shortening of a normal-calibre colon, with rounding of the contours of the hepatic and splenic flexures.

Functional Immaturity of the Colon and Meconium Plug Syndrome Immature left colon (syn. small left colon) and meconium plug syndrome are relatively common causes of neonatal bowel obstruction. There is overlap in both the clinical features and radiology of the two conditions,

1814

SECTION G  Paediatric Imaging

A

B

Fig. 71.12  Hirschsprung Disease. (A) Supine lateral view, contrast enema. The rectosigmoid ratio is reversed and the sigmoid colon is shortened. The rectal wall is irregular and narrowed. (B) Supine anteroposterior view contrast enema. There is an abrupt calibre change at the junction of the descending and sigmoid colon. Features are in keeping with short-segment Hirschsprung disease.

and the terms are often used interchangeably in the literature. The former refers to a transient functional obstruction of the colon, which occurs as a result of immaturity of the myenteric plexus. It is common in the infants of diabetic mothers and in those whose mothers have a history of substance abuse. Meconium plug syndrome is a temporary colonic obstruction caused by pellets of meconium. It is associated with both cystic fibrosis (CF) and Hirschsprung disease, both of which should be excluded if a diagnosis of meconium plug syndrome is made. Premature infants and infants of mothers who received magnesium sulphate therapy have an increased incidence of meconium plug syndrome, though the latter is disputed in the literature. In both conditions, the affected infants present with symptoms and signs of bowel obstruction. There is delayed passage of meconium. The plain radiograph shows distension of both small and large bowel loops to the level of the inspissated meconium plugs. In small left colon syndrome, the contrast enema typically shows a microcolon distal to the splenic flexure, at which point there is an abrupt transition to a mildly dilated proximal colon (Fig. 71.13). The main differential diagnosis is long-segment Hirschsprung disease, and biopsy may be required if symptoms do not improve. In meconium plug syndrome, the lodged meconium plugs are the cause of the obstruction, but the findings are essentially the same. The plugs lodge in the region of the splenic flexure, proximal to which there is colonic dilatation. Unlike small left colon syndrome, a microcolon is unusual. The (water-soluble) contrast enema is also therapeutic and once the meconium plugs are passed per rectum, the infant mechanical obstruction recovers.

Meconium Ileus Meconium ileus is a form of distal intestinal obstruction caused by inspissated pellets of meconium in the terminal ileum. Approximately 80% to 90% of infants with meconium ileus have CF, and meconium ileus is the presenting feature of CF in 10% to 20% of affected patients.

Fig. 71.13  Functional Immaturity of the Colon. Small-calibre distal colon with an abrupt transition to dilated proximal colon at the splenic flexure (arrow). Note that the rectosigmoid ratio is normal and therefore Hirschsprung disease is unlikely, but biopsy may be required if symptoms do not improve.

Children with the ΔF508 mutation have an increased incidence of meconium ileus, and those who are homozygous for this mutation have a higher incidence still. More than half of the affected infants have uncomplicated (or simple) meconium ileus. In utero, these babies produce meconium that is thick

CHAPTER 71  Paediatric Abdominal Imaging

1815

Fig. 71.14  Diagrammatic Representation of Meconium Ileus. Pellets of desiccated meconium obstruct the terminal ileum, with the more proximal small bowel dilating. The unused colon is extremely narrow in calibre (a microcolon).

and tenacious, and which fills and distends the small bowel loops. The meconium desiccates in the distal ileum and becomes impacted, causing a high-grade obstruction (Fig. 71.14). Failure of meconium to pass into the colon results in a functional microcolon, whereas more proximally the small bowel loops are dilated and filled with greenish-black meconium of a toothpaste-like consistency. Meconium ileus is described as complicated when intrauterine volvulus, intestinal atresias, bowel necrosis, perforation or meconium peritonitis supervene. The presenting clinical symptoms and signs of non-complicated meconium ileus are those of a low bowel obstruction. The plain abdominal radiograph will show dilatation of small bowel loops, which are of varying calibre. Often, there is a visible ‘soap bubble’ appearance (classically in the right iliac fossa), which is caused by the admixture of meconium with gas. The contrast enema in meconium ileus demonstrates a virtually empty microcolon. Where possible, attempts should be made to coerce contrast medium into the terminal ileum. This will demonstrate multiple filling defects consistent with pellets of meconium (Fig. 71.15). More proximal reflux of contrast medium will show the dilated ileal loops. The contrast enema in uncomplicated meconium ileus is intended to be therapeutic as well as diagnostic. The contrast media alters the intraluminal osmotic pressure and increases the fluid content within the terminal ileum and colon, encouraging the passage of the obstructing meconium. Water-soluble contrast medium should be used and the choice of medium is institution specific. Dilute Gastrografin (between 1 : 2 and 1 : 4 Gastrografin:water is the suggested dilution) is commonly used but other iodine-based contrast agents can be used and may be safer in small neonates owing to the large fluid shifts that can result when using Gastrografin. If the infant’s clinical condition remains stable, the enema can be repeated, as necessary, until the obstruction is relieved. If a child with meconium ileus presents with volvulus or perforation, it is known as complicated meconium ileus. This occurs in approximately 50% of patients and can lead to intestinal stenoses, atresias and necrosis. Perforation of bowel in utero leads to chemical (sterile) meconium peritonitis. The extruded bowel contents cause an intense inflammatory reaction, with fibrosis and calcification to follow. A meconium pseudocyst is formed when there is vascular compromise in association with an intrauterine volvulus; the ischaemic bowel loops become adherent and

Fig. 71.15  Meconium Ileus. Neonate diagnosed with cystic fibrosis on sweat testing. Contrast enema demonstrates a small-calibre colon containing a few filling defects (meconium plugs). Contrast has reached the terminal ileum (arrow), which contains multiple filling defects in keeping with inspissated meconium. Contrast has advanced beyond the obstruction into dilated ileum (arrowhead).

necrotic, and a fibrous wall develops around them. The presence of complicated meconium ileus may be suggested by findings on plain radiographs such as intra-abdominal or scrotal calcifications, bowel wall calcification, prominent air–fluid levels and soft-tissue masses. The management of meconium peritonitis is surgical.

Colon Atresia Colon atresia is rare when compared with other intestinal atresias, and colonic stenosis is rarer still. Atresia has long-been thought to be caused by an in utero vascular accident; however, Baglaj and colleagues have suggested compression of the bowel wall against the closing umbilical ring as an underlying cause. The affected infant presents after several feeds with abdominal distension, failure to pass meconium and vomiting. The abdominal radiograph will show the features of a low intestinal obstruction, with the loop immediately proximal to the atretic segment being massively dilated. If multiple atresias are present, then the bowel will be distended only to the level of the most proximal atresia. A contrast enema usually demonstrates a distal microcolon, with obstruction to the retrograde flow of contrast at the point of the atresia. The management of colon atresia is surgical.

Distal Ileal Atresia Distal ileal atresia is part of the jejuno-ileal atresia spectrum. It is thought to occur secondary to a vascular event in late gestation. If the atresia is in the distal ileum, then the infant will present with abdominal distension and delayed passage of meconium. The plain radiograph will show a low obstruction with multiple dilated loops of bowel. On enema, contrast medium cannot be refluxed into the dilated small bowel, the colon is usually a microcolon, but depending on how late in gestation the vascular insult occurs, there may be some colonic contents owing

1816

SECTION G  Paediatric Imaging

to succus entericus passing into the colon before the insult. The condition is managed surgically.

Anorectal and Cloacal Malformations Anorectal malformations (ARMs) are a congenital abnormality of the terminal hindgut, known as the primitive cloaca. In the first trimester, the cloaca is divided into ventral (GU) and dorsal (rectal/anal) compartments by progressive descent of a structure known as the urorectal septum. ARMs are caused by the incomplete development of the urorectal septum and consist of an anorectal atresia/stenosis with or without a fistula between the distal colon and the GU tract. The degree of severity will depend on the point at which failure of the septum occurs. The cloacal malformation represents a very complex form of ARM and occurs exclusively in girls. ARMs occur in approximately 1 : 5000 live births and are more frequent in males. Associated congenital anomalies are very common and occur in up to 70% of patients born with an ARM. Amongst the commonest associated anomalies, the VACTERL sequence occurs in approximately 45% of patients. Between 2% and 8% of patients have Down syndrome, with the vast majority of these patients having imperforate anus without a fistula. Currarino triad is the association between an anorectal malformation, bony sacral anomalies and a presacral mass lesion. ARMs can be classified using either the Wingspread or Krickenbeck classification, and each classification has separate male and female sections. The Wingspread classification describes ARMs as low, intermediate, or high depending on the position of the rectal pouch relative to the levator sling. The Krickenbeck classification is concerned with the presence or absence of a fistula and it defines five types of fistula according to location. The Krickenbeck classification tends to be preferred in modern practice because it gives information regarding the localisation of the atretic anorectum and has an important bearing upon surgical planning. In male patients, the fistula may be to the prostatic or bulbar urethra or bladder neck. In females, the fistula may be to the vaginal vestibule, with true posterior vaginal wall fistulae being extremely rare. Perineal fistulae may be present in infants of either sex, as may imperforate anus and rectal atresia or stenosis, the latter groups being without fistulae. The diagnosis should be made clinically during the baby check immediately following delivery, and immediate referral to the local paediatric surgery team is merited. Imaging at this stage is likely to be focused on assessment of any associated anomalies. In cases with an obvious perineal fistula, the child may proceed to surgery without imaging of the ARM. The traditional cross-table lateral radiograph, with the infant in the prone position and the buttocks elevated, has been demonstrated to be inaccurate and should be interpreted with caution, especially if the child was crying or straining at the time of imaging. It is not recommended as a standard imaging test in ARM. Similarly, ultrasound has been described as a method of measuring the pouch–perineum distance in ARM, but has also been shown to be inaccurate for the same reasons as the lateral radiograph and is not considered as a standard investigation. Infants who have a perineal fistula usually undergo a posterior sagittal anoplasty within the first 24 to 48 hours of life. All other children with ARMs will have a defunctioning colostomy performed, with the aim of surgery being to separate the GI and GU tracts and stop faecal contamination of the latter. Definitive surgery is postponed until a later stage, when the infant has grown and all other imaging investigations are complete. Imaging evaluation of associated abnormalities is of greater importance in the newborn than is a detailed anatomical imaging assessment of the ARM. A detailed abdominal and pelvic ultrasound should be performed with particular attention to the kidneys and the pelvic

structures. A chest radiograph and echocardiogram are recommended, as is a spinal ultrasound. Up to 50% of children with ARM will also have spinal cord problems, most commonly a tethered cord. Spinal ultrasound is an important screening tool, but MRI will be required to further define any abnormality. In complex ARM, definitive surgery is planned for several months after birth. A diverting colostomy and, potentially, a vesicostomy (in the case of a cloacal malformation) are performed initially to allow time for the patient to grow. Imaging assessment is complex and multimodality. In terms of defining the precise anatomy of the fistulous tract, the most useful investigation remains the augmented pressure colostogram. A Foley catheter is inserted into the distal segment of the colon and its balloon gently inflated so that it seals the stoma (Fig. 71.16). With the patient in the lateral position, water-soluble contrast medium is hand-injected under mild pressure to distend the distal colon and define the fistulous tract. Interpretation of the images is made easier if there is a bladder catheter in situ, through which some contrast medium has been instilled; this gives anatomical markers for the bladder neck and the course of the urethra. It is also useful to place a skin marker at the expected site of the anus. This is important for the surgeon to assess the distance between the skin and the distal colon. The cloacal malformation only occurs in female patients. It is a complex anorectal malformation that results in one common channel draining the urethra, vagina and rectum. Examination of the perineum reveals a single opening. Defining the anatomy of these lesions is difficult, but ultrasound, pelvic MRI examinations and contrast studies of the cloaca all play a role. These girls are all managed with an initial colostomy, following which combined fluoroscopic studies are performed. These include an augmented pressure colostogram, micturating

Fig. 71.16  Recto-Urethral Fistula—Augmented Pressure Colostogram. A Foley catheter has been inserted into the distal limb of the colostomy and the balloon is inflated (arrow). Contrast has been injected under pressure and the fistula between the distal colon and the urethra is clearly demonstrable. There is retrograde filling of the bladder.

CHAPTER 71  Paediatric Abdominal Imaging

1817

cystourethrogram and a ‘cloacogram’, with images obtained in both the lateral and anteroposterior positions. This can also be achieved with MRI using saline or dilute gadolinium to outline the structures of the cloaca and the common channel. It is useful to make an assessment of the length of the common channel on imaging but this is challenging given the small size of the structures. It may be better assessed on cystoscopy. As for all ARMs, ancillary radiological investigations, as described above, are mandatory for assessment of associated anomalies in cloaca.

tool if ultrasound is unable to give sufficient diagnostic information. The availability of MRI is sometimes limited and, occasionally, an abdominal CT scan may be necessary, especially in acute abdominal pain; however, a relatively restrictive attitude towards the use of CT is important owing to the radiation exposure. Plain abdominal radiography after the neonatal period should be reserved for the queries of intestinal obstruction and free gas.

THE INFANT AND OLDER CHILD

Appendicitis is the most common cause for acute surgery in childhood. Between 30% and 40% of children do not present with the typical clinical presentation of appendicitis and, in particular, preschool patients often present with atypical features, more rapid progression and higher incidence of complications. Very young children often have a diagnostic delay and hence they have a higher risk of perforation at presentation. Consequently, imaging is often necessary to confirm, suggest or refute the clinical diagnosis and the use of imaging has dramatically reduced the false-positive appendectomy rates. Ultrasound should be the primary imaging investigation and performing a comprehensive ultrasound examination will make CT redundant in most cases. The ultrasound should be performed with a high-frequency linear transducer using a graded compression technique. The primary criteria of acute appendicitis are typically a tubular, blind-ending, non-compressible structure with maximal outer diameter over 6 mm. Other findings include wall hyperaemia or hypoperfusion (depending on the degree of inflammation/ necrosis), surrounding hyperechoic mesenteric fat and the presence of an appendicolith (Fig. 71.17). CT is rarely necessary but can be an important diagnostic tool in difficult cases where ultrasound is unable to clarify and the clinical situation enforces acute surgery. CT is also often performed when complicated periappendicular abscess formation is suspected. Note that the appendix may be retrocaecal and an inflamed retrocaecal appendix may cause subcapsular liver abscesses (Fig. 71.18). Sonographic mimics of acute appendicitis may be acute salpingitis in teenage girls or terminal ileitis (see below) (Fig. 71.19).

Abdominal Pain Abdominal pain can have numerous causes and functional abdominal pain is not uncommon in childhood; however, important entities that may require treatment must be ruled out. One of the challenges in childhood is to accurately define the type, intensity and frequency of pain because children and adolescents may have poor ability to recall episodes of abdominal pain and to localise and characterise the pain. Therefore, the role of radiology is important to establish the diagnoses. The imaging approach will be tailored by the clinical information and the age-specific entities that may cause abdominal pain in childhood. In the majority of cases, ultrasound is the primary imaging investigation of choice in both chronic and acute abdominal pain. Owing to little body fat, children are ideally suited for ultrasound and highly detailed images can be produced; Sonography is therefore a potentially powerful diagnostic tool in children. The method is fast, cheap and does not expose the child to unnecessary radiation. MRI may be a complementary

SUMMARY BOX: The Infant and Older Child • Ultrasound should always be the first technique of choice when imaging is required in the work-up of abdominal complaints. • Functional abdominal pain is common in childhood. • The main indications for plain abdominal radiographs in childhood are clinical suspicion of free air or ileus and they are otherwise rarely indicated. Plain abdominal radiographs have no role in the routine work-up of constipation and should only be performed in highly selected cases. • Abdominal computed tomography should be used only where other modalities have failed to answer the clinical question (except in trauma).

A

B

Acute Appendicitis

Mesenteric Lymphadenitis Mesenteric lymphadenitis is a common cause of abdominal pain in childhood and may present as subacute or acute abdomen. The symptoms are caused by swelling of mesenteric lymph nodes, as a reaction to a

C

Fig. 71.17  Sonographic Findings in Acute Appendicitis. (A) A longitudinal ultrasound image showing a thickened appendix (arrowheads) with an appendicolith, casting echo shadow (asterisk). (B and C) Sonographic appearance of a necrotic appendix with surrounding hyperechoic, inflamed mesenteric fat (asterisk) and some free fluid (arrow) suggestive of perforation. The axial view clearly shows a thickened appendix (arrow) with only two remaining layers of the appendix wall owing to necrosis.

1818

SECTION G  Paediatric Imaging

A

B

C

Fig. 71.18  Subcapsular Liver Abscesses. A 2-year-old with a 10-day medical history of infection with fever and diffuse symptoms, initially treated for pneumonia. At presentation he had severe abdominal pain and septicaemia. Ultrasound showed multiple subcapsular abscesses (black arrows) and a hyperechoic round structure with posterior shadowing white asterisk between the liver and the right kidney suggestive of a perforated appendix with an appendicolith (A and B). The abscesses were treated with percutaneous ultrasoundguided drainage. Computed tomography was performed after 1 week because of persistent temperature spikes. Small subcapsular abscesses (black arrows) were seen, none of which were suitable for percutaneous drainage and intravenous antibiotic treatment was continued. The appendicolith (arrowhead) was still present with adjacent slightly dilated small bowel loops (white arrow) (C).

Fig. 71.19  Sonographic findings in a 12-year-old girl with suspected appendicitis showing a thickened terminal ileum (black arrow) with reduced peristalsis and a normal appendix (white arrow).

trivial, often asymptomatic viral infection. On ultrasonography (US), multiple enlarged lymph nodes are seen in the root of the mesentery. The hyperechoic, fatty hilum is preserved and there is normal Doppler signal within the lymph nodes (Fig. 71.20). Mesenteric lymphadenitis is a diagnosis of exclusion and should be established based on US findings in the absence of any other plausible explanations for the abdominal pain.

Inflammatory Bowel Disease The diagnosis inflammatory bowel disease (IBD) encompasses Crohn disease, ulcerative colitis and unclassified IBD. IBD is thought to develop from dysregulation of the immune response to gut flora in a genetically susceptible host. The most common symptoms of IBD are chronic diarrhoea, fever and weight loss, but it may also present as acute or subacute abdominal pain. In very small children, the colon is often affected, even in Crohn disease. The imaging features are quite striking with an inflamed colon surrounded by very hyperechoic fat and enlarged lymph nodes and, sometimes, micro abscesses. There is no rectal sparing. These young children often present with fistulating disease.

The gold standard for diagnosing IBD is ileocoloscopy with biopsy. Imaging plays a role in establishing the extent of the disease, to assess possible complications and to select candidates for surgery. Ultrasonography. In children, ultrasound is the first imaging tool, especially if the diagnosis is unknown. The sensitivity and specificity for ultrasound depends greatly on the assessment of bowel wall thickness (BWT); much depends on the threshold used. A small BWT greater than 1.5 to 3 mm and a colonic wall thickness more than 2 to 3 mm is generally considered to be abnormal. Colour Doppler US may reveal hyperaemia of the inflamed bowel and this finding may further influence the diagnosis (Fig. 71.21). Other signs of IBD on US include lack of bowel stratification, altered echogenicity of the bowel wall, hyperechoic mesenteric fat and enlarged lymph nodes. One should always look for complications of the disease, such as abscesses and fistulas, but the sensitivity for these features on US is low. US is a quick, radiation-free and readily available investigation but is highly operator dependent and the findings are not necessarily reproducible. Furthermore, some studies show that the sensitivity of detecting terminal ileitis is limited. The technique also has limited value in obese children and with the presence of bowel gas. Conventional barium studies. Conventional barium studies have been replaced by other imaging techniques, such as MRI and US, and play a very limited role in imaging of IBD in children because the techniques are stressful, give a high radiation dose to the patient and are unable to demonstrate extraluminal disease. Small bowel follow-through studies (SBFT) may still play a role in the assessment of bowel obstruction but ultra low-dose multidetector computed tomography (MDCT) and capsule endoscopy have replaced SBFT for this indication. Magnetic resonance imaging. The preferred technique for small bowel assessment on MRI is MR enterography. A comprehensive MRI has a high specificity and sensitivity for bowel inflammation and does not involve ionising radiation. This technique is normally well tolerated by children from 6 to 7 years and older. Sufficient bowel distension is important for a proper assessment of the bowel wall. MR enteroclysis may be an alternative if the child is unable to drink the relatively large amount of fluid required to distend the small bowel. A low-residue diet

CHAPTER 71  Paediatric Abdominal Imaging

A

1819

B Fig. 71.20  Sonographic Appearances of Mesenteric Lymphadenitis. (A) Multiple, unremarkable mesenteric lymph nodes (arrowheads) with a preserved hyperechoic fatty hilum (arrows). (B) Normal hilar flow is seen on Doppler examination.

B

A

C

should be given 3 days before the examination, with no food per mouth from 24 hours before the examination. MR enteroclysis can estimate the length and localisation of the affected bowel and detect both intra- and extraluminal disease; however, if there is clinical suspicion of a perianla fistula and/or abscess, dedicated pelvic MR fistulography may be required for detection and delineation of the fistula. A dynamic MRI sequence should be included to assess possible bowel strictures because this may change the treatment approach from medical to surgical. Controversies exist regarding MRI’s ability to determine disease activity, both owing to the lack of gold standard

Fig. 71.21  Terminal Ileitis. Ultrasound shows marked thickening of the terminal ileum and increased echogenicity of the bowel wall (A). (B) A slightly thickened distal ileum with reduced peristalsis. Colour Doppler examination revealed hyperaemic bowel wall in the inflamed bowel segment (C).

and because acute and chronic disease may coexist in the same bowel loop. Diffusion-weighted imaging has been shown to increase the sensitivity and specificity of both intraluminal and extraluminal disease in IBD (Fig. 71.22) and may be even more sensitive than contrastenhanced sequences, therefore making the latter superfluous in many cases. MRI signs of IBD are listed in Table 71.5. Computed tomography. CT enteroclysis and CT enterography have become widely used techniques for small bowel investigation in adults. These techniques should be avoided in children owing to the high radiation dose. CT should be reserved for investigations of acute

1820

SECTION G  Paediatric Imaging

A B

C

D Fig. 71.22  Inflammatory Bowel Disease. Magnetic resonance imaging (MRI) of Crohn disease. (A) Axial hydrographic sequence showing thickening of the small bowel adjacent to the ileosomy (arrowhead). (B) MRI in inflammatory bowel disease should include hydrographic cine images to assess bowel peristalsis. (C) Diffusion-weighted image of the same bowel segment with high signal suggestive of restricted diffusion (white arrow). (D) Contrast-enhanced T1 weighted image showing contrast enhancement of thickened bowel loops (arrowheads).

TABLE 71.5  Magnetic Resonance Imaging

Signs of Inflammatory Bowel Disease Bowel Loop Appearance • Fixed • Dilated • Pseudo-sacculation appearance • Strictures

Bowel Wall • Thickness • Focal lesions (ulceration, pseudo-polyps, mural abscess) • Enhancement pattern after gadolinium injection (mucosal alone, layered, global, serosal hypervascularity) • Restricted diffusion • Extramural signs: fibro-fatty proliferation, distended and enhancing mesenteric vessels fistula, abscess, enlarged lymph nodes

complications where US is insufficient, or for drainage of complex abscesses, or when the abscess is inaccessible for US-guided drainage.

Intussusception Intussusception is a common surgical emergency in infants and young children and consists of a telescoping of a segment of bowel

(intussusceptum) into a more distal segment (intussuscipiens). This condition usually occurs in children under 1, with a peak incidence between 5 and 9 months of age; however, it may occur up to school age. Ileocolic intussusceptions are the most common type. Ileoileocolic, ileo-ileal, and colocolic are much less common. Most (more than 90%) have no focal lead point and are caused by lymphoid hypertrophy, usually following a viral infection. Secondary lead points (which include nasojejunal tubes, Meckel diverticulum, intestinal polyp, duplication cyst and lymphoma) are present in 5% to 10% of patients. In young infants or children more than 6 to 7 years of age, intussusception is more likely to be caused by a secondary lead point. The clinical presentation of intussusception varies. The classical clinical signs are of colicky abdominal pain and bloody stools; a palpable abdominal mass appears in less than 50% of the children. Intussusception must be treated as a surgical emergency. The clinical situation may deteriorate rapidly, particularly in infants, and may become life threatening with hypovolemia and shock. Prolonged symptom duration will reduce the likelihood of successful reduction. Intussusception is now diagnosed by US, with a sensitivity and specificity of 100% in several reported studies, even when performed by less-experienced radiologists, if properly trained. The characteristic appearance of intussusception makes diagnosis or exclusion very easy. The intussusceptum is usually found just deep to the anterior abdominal wall, most often on the right side of the midline. The sonography must be performed with a high-frequency, linear array transducer. The intussusception forms a mass of 3 to 5 cm in diameter

CHAPTER 71  Paediatric Abdominal Imaging with a ‘target appearance’ in the transverse plane, and a ‘sandwich appearance’ in the longitudinal plane. The characteristic ‘crescent in doughnut’ sign—a hyperechoic semilunar structure, caused by the mesenteric fat pulled into the intussceptum—facilitates the differentiation from mimickers of intussusception, such as bowel wall thickening, faeces and the psoas muscle (Fig. 71.23). Lymph nodes and fluid may be seen within the intussuceptum and, in some studies, have been found to be associated with decreased hydrostatic-reduction rate. Bowel necrosis is difficult to assess by US, even with power Doppler examination of the bowel wall. Free intraperitoneal fluid is commonly seen in patients with intussusception and is, therefore, an unreliable indirect sign of bowel ischaemia. US should not only be performed to establish the diagnosis but also to look for secondary lead points and other intra-abdominal problems unrelated to the intussusception (Fig. 71.24); however, no sonographic features, including the presence of a secondary lead point,

A

should preclude an attempt of reduction. In a well-hydrated, haemodynamically stable child the only contraindications for hydrostatic or pneumatic reduction are the presence of free intraperitoneal air or clinical signs of peritonitis. The role of plain radiography in the diagnosis of intussusception is controversial. The radiographic features of intussusception include a soft-tissue mass contrasting an air-filled bowel loop, the so-called ‘meniscus sign’ (Fig. 71.25). There may be dilated, gas-filled bowel loops proximal of the intussusceptum and the absence of gas within the caecum may suggest a ileocaecal intussuception; however, the caecum may be difficult to localise in a child, and the sigmoid is located on the left side of the abdomen in almost 50% of children and may be indistinguishable from the caecum on a plain radiograph. Abdominal radiographs should therefore not be routinely used in intussusception. The presence of free intraperitoneal gas is extremely rare in children with intussusception,

B Fig. 71.23  Intussusception Is Easily Appreciated on Ultrasound. (A) The axial view shows the ‘doughnut’ or ‘target sign’ (arrowheads) caused by the multiple layers of bowel and the pathognomonic hyperechoic semilunar appearance of the mesenteric fat within the intussusceptum (asterisk). (B) The longitudinal view reveals the typical ‘sandwich’ appearance caused by the multiple layers of bowel wall and mesenteric fat (asterisk).

A

1821

B Fig. 71.24  Colocolic Intussusception. Children may have difficulties to localise and characterise abdominal pain. This 5-year-old child presented with a limp and was initially referred to ultrasound of the hip, which was normal. The radiologist then performed an ultrasound of the abdomen and an intussusception was diagnosed; however, it was seen only in the left lower quadrant (A). The transverse diameter was more than 5 cm and the ileocaecal pole was normal. A barium enema was performed and showed a colocolic intussusception, which was successfully reduced. The immediate follow-up ultrasound showed the lead point, a colonic polyp (arrow), which was later removed surgically (B). (Image courtesy of Dr. Behzad Khoshnewiszadeh.)

1822

SECTION G  Paediatric Imaging are rarely seen within small bowel intussusceptums. If the bowel wall is oedematous the small bowel intussusception is less likely to reduce spontaneously (Fig. 71.27).

Constipation

Fig. 71.25  Intussusception. An abdominal radiograph in a child with intussusception may show a soft-tissue mass contrasting an air-filled bowel loop, the so called ‘meniscus sign’ (arrow); however, the diagnosis is made by ultrasound and the role of radiography in intussusception is controversial.

and may be assessed by a quick fluoroscopy performed before the fluoroscopically guided reduction. In intussusception involving the colon, image-guided reduction is the first-line therapy of choice and can be performed using a pneumatic technique or by contrast enema, under fluoroscopy or ultrasound guidance (Fig. 71.26). Most centres in the UK use pneumatic reduction under fluoroscopy guidance, but the choice of technique varies across Europe and should be based on the experience and expertise of the radiologist who performs the reduction. The use of sedation is also controversial; however, a prospective clinical study found an increased success rate and no difference in complications when using deep sedation in pneumatic reduction of intussusception in children. Regardless of the technique used, one should aim at a reduction rate of more than 80%. Small bowel intussusception. The reported frequency of small bowel intussusception varies. It may be seen as an incidental finding and will usually reduce spontaneously; however, small bowel intussusception sometimes causes ileus and must be treated surgically because this condition is not suited for enema reduction. It may be challenging to differentiate small bowel from ileocolic intussusception on ultrasound; however, this is crucial to choose the right treatment. Small bowel intussuceptum diameter tends to be smaller than that of ileocolic, but measurements of lesion diameter alone cannot enable reliable differentiation between ileocolic and small bowel intussusception. Some authors have found that the presence of a fatty core in the lesion, in combination with lesion diameter and wall thickness, and especially the ratio between the diameter of the fatty core to the thickness of the outer wall, can be used to differentiate between the types of intussusception. Lymph nodes

Constipation is probably the most common GI problem in infants and children. Childhood functional constipation has an estimated prevalence of 3% in the Western world. The symptoms are typically infrequent painful defecation, faecal incontinence and abdominal pain. This can lead to encopresis or faecal soiling and, occasionally, can cause acute, severe abdominal pain. Less than 5% of children with constipation have an underlying disease. The diagnosis of constipation is essentially clinical and radiological investigations play a very limited role in the work-up of constipation and should not routinely be performed in children with functional constipation. The plain abdominal radiograph will demonstrate the degree of faecal loading and dilatation of the large bowel; however, the presence of faecal loading on the plain radiograph does not necessarily indicate constipation and several studies show that plain radiographs have a low sensitivity and specificity for diagnosing constipation, and may even lead to misdiagnosis. Radiological investigations should only be performed in a carefully selected group of patients where an underlying cause is suspected. US should be the first imaging tool in chronic abdominal pain when radiological work-up is indicated. It gives a good overview over the bowel and intra-abdominal organs and can help differentiate a faecal mass from a true mass. Other imaging techniques described for the evaluation of constipation include the measurement of colonic transit time using radiopaque markers, fluoroscopy and MRI defaecography. These are only indicated in highly selected cases.

Intestinal Motility Disorders Intestinal motility disorder is a term used to describe a variety of abnormalities that have reduced motility of the bowel and no organic occlusion of the bowel lumen in common. They can be divided into acute and chronic disorders. Acute dysmotility includes paralytic ileus in which there is temporary cessation of peristalsis in the gut. This simulates intestinal obstruction because there is failure of propagation of intestinal contents. Acute gastroenteritis can simulate small bowel obstruction by causing a local paralytic ileus, with dilatation of the affected segment of bowel and multiple fluid levels on an erect plain radiograph of the abdomen. Chronic motility disorders include primary abnormalities of the bowel—Hirschsprung disease (aganglionosis) and neuronal intestinal dysplasia, which is a defect of autonomic neurogenesis characterised by an absent or rudimentary sympathetic ganglion innervation of the gut, or by hyperplasia of cholinergic nerve fibres and hyperplasia of neuronal bodies in intramural nerve plexuses. The definitive diagnosis of Hirschsprung is based on rectal biopsy. A positive contrast enema portends a high probability of Hirschsprung disease, but inconclusive or negative studies do not exclude the disease and neither positive nor negative barium enemas make rectal biopsy superfluous if there is clinical suspicion of Hirschsprung disease. The main role of barium enema in this disease is to find the transitions zone to assess the length of the aganglionic segment in children with a positive rectal biopsy, which is important in order to choose the correct surgical technique (Fig. 71.28). Most children with neuronal dysplasia present with severe constipation but tend to spontaneously recover colonic motility between 6 and 12 months of age. Plain radiographs of the abdomen may show signs of obstruction of bowel and the barium enema may show dilatation of the bowel. Diagnosis is by biopsy and barium enema may be performed to establish the level of obstruction in order to plan surgery.

CHAPTER 71  Paediatric Abdominal Imaging

A

C

1823

B

D

Fig. 71.26  Hydrographic Reduction of Intussusception. (A) Contrast defect in the proximal transverse colon owing to the intussusceptum. (B) Further reduction of the intussusception into the ascending colon. (C) Small contrast defect in the caecum. Normally the last part of the reduction is the most difficult owing to oedema of the ileocaecal valve. (D) Contrast filling of the terminal ileum (arrowhead) as a sign of successful reduction. Proper filling of proximal small bowel loops is advised to ensure complete reduction.

Chronic intestinal pseudo-obstruction (CIP) is rare and represents a spectrum of diseases that have in common clinical manifestations consisting of recurrent symptoms mimicking bowel obstruction over weeks or years. The age of presentation varies from newborn to adulthood. The condition is caused by a visceral neuropathy or myopathy, which can be familial or non-familial, resulting in a lack of coordinated intestinal motility. Megacystis–microcolon–intestinal hypoperistalsis syndrome is the most severe form of CIP and is usually fatal in the first year of life. Plain radiographs of the bowel will show loops of bowel with pronounced dilatation. The diagnosis is made by intestinal manometry and biopsy. A contrast medium enema can exclude mechanical obstruction in children with acute symptoms.

Henoch–Schönlein Purpura Henoch–Schönlein purpura (HSP) is an acute, small vessel vasculitis that occurs almost exclusively in childhood. The manifestations are purpuric skin lesions (without thrombocytopenia), GI manifestations, arthritis or nephritis. In most cases the changes of HSP are completely reversible and healing takes place in 3 to 4 weeks. Abdominal pain is a common symptom, and the GI involvement is caused by oedema, bleeding, ulceration and intussusception of the intestine. Ultrasound is the imaging investigation of choice in HSP and will detect most surgical cases. The sonographic features of HSP are uni- or multifocal thickening of the bowel wall accompanied by reduced peristalsis, with normal or slightly dilated bowel loops between the

1824

SECTION G  Paediatric Imaging

A

B

Fig. 71.27  Small Bowel Intussusception. A 4-month-old girl presented with clinical and radiographic signs of ileus. Ultrasound showed intussusception. The sparse mesenteric central fat and the small transverse diameter (2 cm) was strongly in keeping with a small bowel intussusception. The persistent finding after referral from the local hospital and the oedematous bowel wall (arrowhead) suggested that no spontaneous reduction would occur and that this in fact was the cause of ileus (A). At surgery a Meckel diverticulum was found as the lead point (arrow) (B).

Fig. 71.29  Henoch–Schönlein Purpura. Sonographic features of bowel wall thickening (arrow) caused by intramural haemorrhage (arrowhead) in a child with Henoch–Schönlein purpura. Fig. 71.28  Barium enema is useful to show the transition zone (arrowheads) between the aganglionic narrow distal segment of the rectum and the normal dilated bowel. To increase the sensitivity and specificity of this examination, barium enema should ideally be performed without emptying the bowel before the examination and one should not inflate the catheter balloon to prevent distension or perforation of the aganglionic segment (also see the ‘Neonate’-section).

thickened segments (Fig. 71.29). Some patients have a small amount of intraperitoneal free fluid. Intussusception is easily seen on US with a sensitivity of up to 100%. Intussusceptionin HSP is usually ileo-ileal, hence not amenable to pneumatic or hydrostatic reduction; however, small bowel intussusception often reduces spontaneously. Ileocolic intussusception may also occur and should be treated as idiopathic intussusception (see above).

Abdominal Distension

Enteric Duplication Cysts Enteric duplication cysts are congenital anomalies caused by abnormal canalisation of the GI tract. They can occur anywhere along the length

of the gut but are most frequent in the ileum, where they lie along the mesenteric border and share a common muscle wall blood supply. They have a mucosal lining and 43% contain ectopic gastric mucosa. The majority of duplication cysts do not communicate with the GI tract. Duplication cysts may be diagnosed antenatally. If small and nonobstructing, they may not cause any symptoms. Clinically, they usually present in the first year of life with vomiting or abdominal pain. Infection or haemorrhage into the cyst can cause it to enlarge and suddenly cause pain. A duplication cyst may act as a lead point of intussusception. An abdominal radiograph is useful to assess bowel obstruction or signs of bowel obstruction and may show displacement of bowel loops, or even a soft-tissue mass; however, the cyst itself is normally not seen on plain radiographs (Fig. 71.30). US will demonstrate the cyst, which is usually spherical in shape and less often tubular. The cyst is typically anechoic but may have echogenic contents if there has been bleeding into the cyst. The classical feature of an intestinal duplication cyst is the presence of bowel wall lining the cyst, with an inner echogenic mucosal layer and an outer hypoechoic muscular layer (Fig. 71.31). This ‘double-layer’

CHAPTER 71  Paediatric Abdominal Imaging sign on US is thought to be characteristic for intestinal duplication cysts (see Fig. 71.31B); however, the sonographic appearance of other intra-abdominal cysts may mimic the double-layer sign. Therefore, thorough examination with a high-frequency transducer to identify the split hypoechoic muscularis propria layer, or all five layers of the cyst wall, increases the specificity in making the sonographic diagnosis

1825

of a true duplication cyst (see Fig. 71.31C). 99mTc-pertechnetate is taken up by ectopic gastric mucosa, when present, and is helpful in diagnosing duplication cysts presenting with GI bleeding.

Mesenteric Cysts Mesenteric cysts (intra-abdominal lymphangioma) are relatively rare and may be found in the mesentery, omentum or retroperitoneum. Pathologically there is lack of communication of small bowel or retroperitoneal lymphatic tissue with the main lymphatic vessels, resulting in formation of a cystic mass. They most often present in childhood and children are more likely to present acutely with pain, abdominal distension, fever or anorexia caused by haemorrhage into the cyst, infection or torsion. Large cysts may compress the ureters or lead to bowel obstruction. Plain abdominal radiographs show a soft-tissue mass, which displaces adjacent bowel loops. Occasionally, the cyst wall is calcified. US examination demonstrates a thin-walled, uni- or multilocular cystic mass that may be adherent to the solid organs and bowel. The cyst wall consists of a single layer, which contrasts with the double-layered wall seen with enteric duplication cysts. If the intracystic fluid is chylous, infected or haemorrhagic, then echogenic debris will be present. MRI or CT may define the anatomical margins of the cyst more precisely, but again, MRI is the preferred imaging method (radiation protection). MRI can also characterise the cyst, which will vary according to the cyst contents. Occasionally, large mesenteric lymphangiomas may be misinterpreted as septated ascites; however, the thin septae contain small vessels and may enhance following the administration of intravenous contrast material (Fig. 71.32).

Non-Bilious Vomiting

Fig. 71.30  Intestinal Duplication Cyst. The initial radiograph in a child with an intestinal duplication cyst showed a round soft-tissue mass on the right side of the abdomen (arrow) and slightly dilated bowel loops. The child presented with abdominal distension and increased regurgitation.

A

B

Vomiting is not a disease but a symptom that can be caused by numerous conditions, both GI and extra-gastrointestinal (Table 71.6). Vomiting is the forceful ejection of gastric contents from the stomach up the oesophagus and through the mouth and is never physiological, but the underlying cause may be harmless. GOR is the backflow of undigested food from the stomach and up the oesophagus and is common in infants, and may be normal in children up to 18 months of age. GOR can occur in healthy children without causing any symptoms; however, GOR may mimic or trigger vomiting. Radiological investigations should be performed when there are warning signals requiring investigation in infants with either GOR or vomiting; they can help to identify and rule out underlying causes that may require treatment (Table 71.7).

C

Fig. 71.31  Intestinal Duplication Cyst. (A) Pathology specimen of a resected duplication cyst. (B) Sonographic features of the duplication cyst showing an anechoic cyst with a ‘double-layer’ sign of the cyst wall (arrow). (C) Magnified image of the cyst wall revealing multiple layers (arrowhead) in keeping with an intestinal duplication cyst.

1826

SECTION G  Paediatric Imaging

TABLE 71.6  Causes of Vomiting

in Children

Obstructive Gastrointestinal Causes • Pyloric stenosis • Malrotation with intermittent volvulus • Intestinal duplication cyst • Antral or duodenal web • Severe constipation • Foreign body • Incarcerated hernia Non-Obstructive Gastrointestinal Causes • Gastroenteritis • Achalasia • Gastroparesis • Peptic ulcer • Eosinophilic oesophagitis or gastritis • Food allergy • Inflammatory bowel disease • Appendicitis or pancreatitis Neurological Disorders • Increased intracranial pressure • Childhood migraine Infections • All infections in childhood may cause vomiting, particularly in the younger child Metabolic and Endocrine Disorders Renal • Obstructive nephropathy • Renal failure Cardiac • Heart failure • Vascular rings Others • Induced illness by carers • Child neglect or abuse • Self-induced vomiting • Cyclical vomiting syndrome

TABLE 71.7  Warning Signals Requiring

Investigation in Children With Regurgitation or Vomiting • Bilious vomiting • Gastrointestinal bleeding • Consistently forceful vomiting • Onset of vomiting after 6 months of life • Failure to thrive • Diarrhoea • Severe constipation • Fever • Lethargy • Hepatosplenomegaly • Bulging fontanelle • Macro-or microcephaly • Seizures • Abdominal tenderness or distension • Documented or suspected genetic or metabolic disorder

ligaments anchoring the stomach are dynamic to allow expansion. Absence of one or more ligaments and ligamentous laxity increase the risk of gastric volvulus. In organoaxial torsion, the stomach flips upward along its long axis and the ggastro-oesophageal junction and pylorus maintain their normal position. Even though there is little risk of ischaemia, there may be full or partial obstruction of the gastric outlet. Gastric distension or gas-filled colon may predispose to this condition. Fluoroscopic barium study will reveal a distended stomach where the greater curvature is positioned superior to and to the right of the lesser curvature (Fig. 71.34). The condition is normal in infancy and can be seen as an incidental finding. In symptomatic older children, gastropexia may be required. Mesenteroaxial volvulus is a rare entity, where the stomach twists transversely around its mesenteric axis, causing close approximation of the gastro-oesophageal junction and pylorus (Fig. 71.35). This is always a surgical emergency, causing gastric obstruction with high risk of ischaemia. The child presents acutely with vomiting and abdominal pain and distension.

Malrotation With Chronic Intestinal Obstruction or Intermittent Volvulus

Gastro-oesophageal reflux disease (GORD) is present when the reflux of gastric contents causes troublesome symptoms and/or complications. Diagnosing GORD may be difficult, particularly in infants and young children. There is no agreed perfect method for detecting GORD, but the diagnosis is usually made on the basis of questionnaires, 24-hour pH monitoring and impedance measurements. Both ultrasound and fluoroscopy with contrast may show the presence of GOR, but radiological investigations play no role in establishing the diagnosis of GORD. Barium contrast study of the upper GI tract is useful to confirm or rule out anatomical abnormalities of the upper GI tract in the presence of GORD, like malrotation, with intermittent volvulus, or hiatus hernia (Fig. 71.33).

Malrotation with midgut volvulus is described in detail within the neonatal section. It is, however, important to emphasise that even though symptomatic malrotation most frequently presents in the neonatal period with midgut volvulus, it may also occur in the older child with either acute or chronic symptoms. The chronic presentation is a diagnostic challenge. The most common symptoms are crampy abdominal pain, failure to thrive, recurrent vomiting and signs of malabsorption. The symptoms may be non-specific and so diagnostic delay is common. Pathophysiology of these chronic symptoms may relate to intestinal obstruction from Ladd’s bands or from venous and lymphatic congestion in cases with intermittent volvulus. Surgical treatment is recommended in all patients, even when asymptomatic, owing to the lifelong risk of potential serious complications. The diagnosis can normally be made from an upper GI barium examination (Fig. 71.36) but is occasionally seen on cross-sectional imaging investigations for chronic abdominal complaints.

Organoaxial Torsion and Gastric Volvulus

Hypertrophic Pyloric Stenosis

Sometimes, particularly in younger children, organoaxial torsion of the stomach may be the cause of forceful regurgitation or vomiting. The

Hypertrophic pyloric stenosis (HPS) is the commonest cause of vomiting, requiring surgery in infants. It typically presents with projectile vomiting

Gastro-Oesophageal Reflux Disease

CHAPTER 71  Paediatric Abdominal Imaging

A

B

1827

C

Fig. 71.32  Imaging Characteristics of Mesenteric Lymphangioma. Ultrasound of the abdomen shows a large, septated fluid-filled structure (arrow) with debris (A). T2 weighted magnetic resonance imaging (MRI) shows a large, septated cystic mass, displacing the bowel cranially and towards the mid-abdomen (arrow) (B). T1 weighted, fat-saturated MRI after administration of intravenous contrast agent shows subtle enhancement of the intracystic septae (arrowheads) (C).

A

B

Fig. 71.33  Hiatus hernia can be diagnosed on ultrasound or an upper gastrointestinal contrast study. (A) Fluoroscopic image of a sliding hiatus hernia (arrow). (B) Sonographic appearances of the hiatus hernia with widening of the oesophageal hiatus (arrowheads) through which the stomach slides to protrude into the thoracic cavity (asterisk).

2 to 8 weeks after birth, but sometimes it may also present in the neonate. It is caused by hypertrophy of the pyloric muscle and mucosa with an elongated, narrow pyloric canal that fails to relax, which leads to a gastric outlet obstruction. Boys are more frequently affected than girls, and there is a five-fold increased risk with a first-degree relative with this condition. The classic clinical presentation is a dehydrated, cachectic child with a palpable olive-shaped mass in the upper epigastrium; however, there has been a change in the epidemiology over the last decades because children are admitted to hospital and examined earlier, before they get severe symptoms; thus the ‘olive’ mass is now found in less than 30% of the patients. Ultrasound is the investigation of choice when there is clinical suspicion of HPS, and has replaced barium studies in diagnosing this condition; US has a reported sensitivity and specificity of up to 98% and 100%, respectively. It allows assessment of the pyloric morphology and pyloric behaviour. The morphological features include a

hypoechoic, thickened pyloric muscle and an elongated pyloric canal (Fig. 71.37A and B). The obstructed pyloric canal is lined with hypertrophic, hyperechoic mucosa. The hypertrophic muscle typically bulges into the antrum of a fluid-filled stomach, creating the so-called ‘shoulder sign’ and the double-layered, hypertrophic mucosa protrudes into the stomach, creating the so-called ‘nipple sign’ (see Fig. 71.37C). The sonographic appearance of the hypertrophied pylorus resembles that of the uterine cervix and is sometimes referred to as the ‘cervix sign’ (see Fig. 71.37C). Abnormal, exaggerated peristaltic waves caused by the stomach trying to force its contents past the narrowed pyloric outlet can also be seen on real-time US. The length of the pyloric canal may vary considerably in children with HPS, and may be difficult to measure exactly owing to its slightly curved course. It is reported that a length of the pyloric canal of greater than 17 mm is only seen in patients with HPS and a length of up to 15 mm is normal. Measurement between 15 and 16 mm

1828

SECTION G  Paediatric Imaging

is borderline abnormal and should be considered closely together with the transverse thickness of the pyloric muscle. The latter is considered the most reliable measure in the diagnosis of HPS. Although the pyloric muscle may vary with patient age and weight, studies have shown that on repeat ultrasound the transverse diameter of the pyloric muscle in true HPS will be above 3 mm, regardless of age and weight; thus follow-up ultrasound should be performed when the sonographic measurements

are equivocal or when there is a strong clinical suspicion, particularly in pre-term or early-onset symptoms. The treatment is surgical pylorotomy, most often performed laparoscopically.

Omphalomesenteric (Vitelline) Duct Remnants The omphalomesenteric (vitelline) duct is a normal fetal structure that connects the midgut to the extra-embryonic yolk sac. The omphalomesenteric duct usually involutes in the mid first trimester. Its persistence can give rise to a variety of congenital malformations. Most symptomatic omphalomesenteric ducts occur in boys and 60% present before the age of 10 years. The clinical presentation depends on the underlying malformation abnormality. Meckel diverticulum is the most common end result of the spectrum of omphalomesenteric duct anomalies. Other presentations include umbilicoileal fistula, umbilical sinus and umbilical cyst. There may also be a fibrous cord running from the ileum to the umbilicus. Small bowel obstruction from this cord is the most common cause of ileus in otherwise healthy children and adolescents. Rarely, the entire duct remains patent. The symptoms present in the neonatal period with discharge of faeces from the umbilicus, or the ileum can prolapse onto the anterior abdominal wall.

Meckel Diverticulum The most common type of omphalomesenteric duct remnant is the Meckel diverticulum, which arises on the antimesenteric border of the ileum (as opposed to enteric duplications cysts, which develop on the mesenteric border of the small bowel). The diverticulum is present in 2% to 4% of the population. The size of Meckel diverticula varies, with those greater than 5 cm in length being considered ‘giant’. Most are located within 60 cm of the ileocoecal junction. All the layers of the intestine are contained within their walls and they frequently contain islands of gastric and/or pancreatic mucosa. The most common presentations are melaena caused by haemorrhage from peptic ulceration, ileus caused by intussusception or volvulus around a Meckel diverticulum. Meckel diverticulum may also become entrapped in an inguinal hernia, which has become known as a Littre hernia. Patients may present with abdominal pain and occasionally peritonitis caused by diverticulitis or perforation. The diagnosis of Meckel diverticulum is difficult to establish preoperatively, and the investigations should be tailored by the clinical presentation. The Meckel diverticula that haemorrhage contain ectopic

Fig. 71.34  Organoaxial Torsion of the Stomach. Fluoroscopic contrast study showing organoaxial torsion of the stomach. The greater curvature is in an inverse position (arrowheads), and the antrum of the stomach is flipped caudally (arrow). Antrum/pylorus in the normal position (thin arrow).

A

B

Fig. 71.35  Mesenteroaxial Gastric Volvulus. (A and B) On plain radiographs a double bubble may be seen on top of a fluid-filled stomach. The uppermost bubble represents the antrum (arrow) (A). On fluoroscopy there is proximity of the pylorus and the gastro-oesophageal junction (long arrows) with the latter located below and to the right of the pylorus. The nasogastric tube was helpful to locate the oesophagus and the gastro-oesophageal junction in this patient.

CHAPTER 71  Paediatric Abdominal Imaging

1829

Fig. 71.36  Malrotation Without Volvulus. An upper gastrointestinal contrast study performed in an 18-monthold girl with intermittent abdominal pain and failure to thrive. The initial, lateral image shows contrast passing to the D2, which turns caudally (arrowhead). Normally the duodenojejunal flexure (DJF) should be placed to the left of the spine, at the level of the duodenal bulb (asterisk). In this patient the duodenum did not have the normal U-shape and the DJF was located low, on the right side of the abdomen. Late images show the entire small intestine on the right side of the abdomen. The small image shows a normal, U-shaped duodenum with a normal DJF (arrowhead).

A

B

C

Fig. 71.37  Hypertrophic Pyloric Stenosis. Sonographic appearances diagnostic of hypertrophic pyloric stenosis with an elongated and thickened pyloric muscle (A and B). A longitudinal scan of the pyloric canal shows the hypertrophic mucosa (arrowhead) and the hypertrophic muscle (arrow) protruding into the fluid-filled stomach (C).

gastric mucosa in 95% of cases. 99mTc-pertechnetate scintigraphy has a very high sensitivity and specificity for bleeding Meckel diverticula, but is of little value in children with no history of rectal bleeding. Plain radiographs can diagnose the presence of ileus and occasionally a soft-tissue mass may be seen (Fig. 71.38); however, the diagnosis is most often made intraoperatively. Occasionally, enteroliths may be seen

with peripheral calcification around radiolucent centres. This is the most specific feature for Meckel diverticulum on plain radiographs. An ultrasound is often performed if the clinical situation allows for further investigation. While the diverticulum itself may not be identified on US, abscess formation and signs of inflammation with hyperechoic mesenteric fat may be seen in acute diverticulitis. Intussusception caused

1830

SECTION G  Paediatric Imaging by a Meckel diverticulum will also be identified with US (see Fig. 71.27). CT is occasionally performed in the evaluation of intestinal obstruction or peritonitis, but again the diverticulum itself is usually difficult to identify. Conventional barium studies may be useful in patients with chronic, persistent symptoms and negative cross sectional- or nuclear medicine imaging. The characteristic feature of a Meckel diverticulum is a saccular, blind-ending pouch on the antimesenteric border of the ileum with a triradiate fold pattern converging with the ileum. Neoplasms arising in Meckel diverticula are very rare and do not occur in childhood.

Gastrointestinal Malignancies

Fig. 71.38  Meckel Diverticulum. Abdominal radiograph of a 2-year-old boy with acute abdominal pain and clinical signs of peritonitis. The image shows dilated bowel loops and gas-fluid levels suggestive of mechanical ileus. The Meckel diverticulum can be seen as a saccular soft-tissue shadow in the mid-abdomen (arrows).

A

Primary GI tumours are rare in childhood. Colorectal cancer or gastrointestinal stromal tumours (GISTs) may occasionally be seen in children but are most often part of a syndrome. Lymphomas account for 10% to 15% of all childhood cancers. Extranodal involvement is frequently seen in non-Hodgkin lymphoma (NHL) and more frequently seen in children compared with adults. The GI tract is the most common site of manifestation. The distal ileum, caecum, appendix and ascending colon are most commonly affected, and the involvement may be multifocal (Fig. 71.39). There is often marked bowel wall thickening with stenosis or dilatation of the affected segment. The lymphomatous infiltrate is hypoechoic on ultrasound and shows soft-tissue attenuation, with sparse contrast enhancement on CT or MRI, and diffusion-weighted sequences show restricted diffusion (Fig. 71.40). In contrast to inflammatory causes of bowel wall thickening, loss of stratification appears early in lymphomatous involvement of the bowel wall. Mesenteric and retroperitoneal lymph node involvement may be seen in both Hodgkin lymphoma

B

Fig. 71.39  Small Bowel Lymphoma. (A and B) A coronal and sagittal reformatted contrast-enhanced abdominal computed tomography examination of a 12-year old boy with non-Hodgkin lymphoma presenting with abdominal pain, showing multifocal thickening of the intestinal wall and ascites. Pleural effusion is also present.

CHAPTER 71  Paediatric Abdominal Imaging

A

B

1831

C

Fig. 71.40  Focal Lymphomatous Involvement of the Small Bowel. (A) Ultrasound revealed a very thick, hypoechoic bowel wall with loss of normal stratification (arrowheads). Magnetic resonance imaging of the same patient showed a slightly dilated loop of bowel with marked bowel wall thickening, which returned intermediate signal on a T2 weighted image (B) and high signal on diffusion-weighted imaging in keeping with restricted diffusion (C).

(HL) and NHL. Ascites is sometimes seen. The spleen is frequently involved in both HL and NHL. Liver involvement rarely occurs without splenic involvement and is most commonly seen in NHL. The imaging findings in diffuse lymphomatous involvement of the liver and spleen are non-specific. On both MRI and US the parenchyma may be normal or show a hazy, ‘salt-and pepper’ appearance. There may or may not be hepatosplenomegaly.

TABLE 71.8  Abdominal Complications of

THE IMMUNOCOMPROMISED CHILD

Liver • Steatosis • Siderosis • Veno-occlusive disease

Various abdominal complications can be encountered in young cancer patients following chemotherapy. They often present with diffuse abdominal complaints but may also be asymptomatic. The most common abdominal complications of chemotherapy in childhood are listed in Table 71.8. Some of the entities are rare, and some are specific for children undergoing chemotherapy. It is therefore important to recognise the signs of these complications to improve a patient’s outcome. Gallstones: sludge or crystals may be seen as an incidental finding and may be related to the specific drug used or to prolonged illness. The findings may disappear spontaneously. Cholecystitis may occur as a complication of cholelithiasis; however, acalculous cholecystitis is more commonly seen. The gall bladder is best evaluated with US using the high-frequency, linear transducer. Gallstones are freely moving, hyperechoic objects with posterior acoustic shadowing within the gall bladder. Sludge is seen as hyperechoic, liquid material, and crystals or sludge balls may resemble gallstones but without posterior shadowing. Acalculous cholecystitis is diagnosed by the presence of painful gall bladder showing wall oedema with or without distention of the gall bladder. Liver steatosis is a common finding, particularly in children undergoing treatment for acute lymphatic leukaemia. The steatosis may be focal or diffuse. It is important not to differentiate fatty liver infiltration from diffuse leukaemic liver infiltration. Focal liver steatosis will follow a typical distribution with hyperechoic areas on US, owing to fat deposition around the main branches of porta hepatis. Steatosis is easily diagnosed on MRI where there will be loss of signal on the opposed-phase images relative to the in-phase images on proton shift sequences. Liver fibrosis and siderosis may also occur as a consequence of treatment for childhood malignancies. In both liver fibrosis and siderosis the liver will be hyperechoic relative to the kidney. Siderosis is caused by haemosiderin deposition within the liver, which, owing to the ferromagnetic effect of iron, will return low signal on both T1 and T2 sequences.

Chemotherapy in Childhood Gall Bladder • Sludge or sludge balls • Stones • Acalculous cholecystitis

Typhilitis/Neutropenic Enterocolitis Pancreatitis Fungal Oesophagitis

Veno-occlusive disease is a rare entity, but a severe complication resulting from chemotherapy with indirect, non-specific findings on imaging. It affects the small hepatic vessels and sinusoids, leading to high resistance within the liver, and thrombi within the large vessels are not seen. On imaging there may be hepatomegaly, periportal oedema, slow or reversed portal venous flow, high resistive indices (RI) within the hepatic artery, and ascites. The diagnosis is made by liver biopsy. Neutropenic enterocolitis is an opportunistic infection of the bowel caused by the patients’ own intestinal flora, which most often affects the caecum (typhlitis) but may extend to the distal ileum and is sometimes seen in other parts of the bowel. It presents as bowel wall thickening, with or without dilatation of the proximal bowel (Fig. 71.41). Ultrasound is the investigation of choice to diagnose and follow-up patients with typhlitis but CT may be used in an acute setting to look for complications like bowel wall ischaemia and perforation.

ABDOMINAL MANIFESTATIONS OF CYSTIC FIBROSIS The GI manifestations of CF are primarily caused by the abnormally viscous luminal secretions within hollow viscera, and the excretory ducts of solid organs. Abdominal complications of CF can present at any age from neonates to adolescence. The first manifestation of CF may be meconium ileus, which presents in the neonatal period (see the

1832

SECTION G  Paediatric Imaging

A

B

Fig. 71.41  Neutropenic Enterocolitis in a 6 Year Old Boy. Ultrasound shows a longitudinal section of a grossly thickened caecum (arrow), this is often called Typhlitis. (A) Coronal CT image shows generalised ascites and hyperenhancing small bowel loops throughout the abdomen. The caecum is thickened and shows reduced enhancement (arrowhead). Caecal perforation was demonstrated at laparotomy (B).

of the colonic mucosa. Enlargement of the appendix is seen in most asymptomatic patients with CF. This is caused by intraluminal mucus stagnation. US will show an increased diameter of the appendix but without thickening of the appendiceal wall and there is absence of periappendiceal inflammation. Appendicitis occurs less frequently in patients with CF. The symptoms are often misinterpreted as DIOS, or masked owing to the use of antibiotics. The diagnostic delay leads to increased risk of complications. Exocrine pancreatic insufficiency is seen in up to 95% of children with CF at 1 year of age. The most common findings are fatty pancreatic parenchymal replacement or fibrosis; hence the imaging findings vary from a large, lobulated pancreas with fatty infiltration to a small fibrotic pancreas. Pancreatic cysts are also a frequent finding and are a result of obstructed exocrine ducts. Pancreatic cystosis, where the pancreas is completely replaced by cysts, may be seen. Hepatobiliary disease is also common in CF and ranges from asymptomatic gallstones to biliary cirrhosis. Hepatosteatosis is the most common disorder of the liver parenchyma in CF.

ABDOMINAL TRAUMA Fig. 71.42  Distal intestinal obstruction syndrome (DIOS) in a 2-year-old child with cystic fibrosis, presenting with abdominal pain. The plain abdominal radiograph shows faecal impaction (asterisk) in the distal small bowel and proximal colon with dilated small bowel proximal to the obstruction.

neonatal section). An equivalent to meconium ileus, appearing later in life, is the distal intestinal obstruction syndrome (DIOS). Plain radiographs will reveal faecal impaction in the distal ileum and right colon with various degrees of proximal small bowel dilatation (Fig. 71.42). Therapeutic gut lavage with balanced electrolyte solutions opacified with contrast agents may be required. Intussusception occurs in 1% of all patients with CF and may manifest as recurrent abdominal pain. Intussusception is diagnosed with US (see above) and is most often ileocolic. Symptomatic small bowel intussusception that necessitates surgery occurs more frequently in children with CF. Fibrosing colonopathy is related to the use of pancreatic enzymes replacement therapy and is also seen more frequently in children with DIOS. A fluoroscopic barium enema will reveal multiple colonic strictures and irregularities

Blunt abdominal traumas account for 80% of traumatic injuries in childhood. CT is the imaging method of choice in the evaluation of abdominal and pelvic injury following blunt trauma in haemodynamically stable children. CT classification systems for grading intra-abdominal injuries apply to children as well as adults; however, children are more often treated conservatively. Ultrasound may play a role in follow-up of abdominal trauma, and the application of intravenous ultrasound contrast media enhances the sensitivity and specificity of demonstrating organ injuries and ongoing haemorrhage. Plain radiographs have low sensitivity in detecting intraabdominal injuries, but may reveal free intraperitoneal gas when bowel perforation is suspected; however, even here CT is the investigation of choice. Multiphase CT should be avoided in children owing to the high radiation burden and is now replaced by split bolus contrast technique to achieve both arterial and portal venous contrast phases in one acquisition. The morphological characteristics of the paediatric abdomen, abdominal wall and rib case may lead to different injuries following blunt trauma than the injuries seen in adults. The liver is the most commonly affected viscus in blunt trauma in children, followed

CHAPTER 71  Paediatric Abdominal Imaging

A

1833

B

Fig. 71.43  Ultrasound in a Neutropenic Child With Spiking Temperature. (A) The focal fungal lesion (arrowhead) in the liver is easily missed with a low-frequency (5–2 MHz) curvilinear transducer. (B) The lesion (arrowhead) is more conspicuous with a high-frequency (9.5 MHz) linear transducer.

by splenic injuries. The kidneys are often affected and injuries to the vessels or collecting system typically result from deceleration because the kidneys in children have greater laxity than in adults and there is less protective fat. Bowel injuries and pancreatic injuries are rare in children but can be seen owing to compression of the rib cage against the spine. Another mechanism for pancreatic and bowel injury in children is the direct impact from a bicycle handlebar to the abdomen. Associated rib fractures may not always be seen owing to the high elasticity of the growing skeleton. Lap-belt ecchymoses represent an important high-risk marker for injury, particularly to the lumbar spine, bowel and bladder. Young children have a relatively high centre of gravity, which produces shearing forces by the belt. The hypoperfusion complex or ‘shock bowel’ is caused by poorly compensated hypovolaemic shock, which results in dilated, fluid-filled loops of bowel, and is a more frequent finding in children than adults. On CT there is intense contrast enhancement of the bowel wall mucosa, and thickening of the bowel wall. The major abdominal blood vessels and kidneys also show intense enhancement, and the calibre of the aorta and inferior vena cava are reduced. Enhancement of the spleen and pancreas is decreased owing to splanchnic vasoconstriction. It is always important to be aware of the possibility of non-accidental injury (NAI) in children with abdominal trauma.

THE LIVER Imaging Techniques Ultrasonography Grey-scale US allows detailed assessment of the liver parenchyma provided that (1) every section is examined systematically and (2) a high-frequency linear probe is used. The linear probe is particularly helpful for evaluation of the liver surface (e.g. undulations seen in cirrhosis), small focal lesions (e.g. small fungal foci, small or diffuse metastases, Fig. 71.43) and diffuse parenchymal processes (e.g. congenital hepatic fibrosis, Fig. 71.44). A curved or vector probe is helpful for assessing the deeper liver in older children. Lower-frequency probes may be necessary if there is dense hepatic fibrosis. Colour Doppler is a mandatory adjunct to grey-scale imaging and is used to assess the hepatic and portal veins and the hepatic artery, and to assess for the presence of collateral vessels (e.g. portal cavernous transformation following extrahepatic portal vein occlusion, or

Fig. 71.44  Ultrasound in an infant with congenital hepatic fibrosis shows lace-like hyperechoic bands throughout the liver. There is associated polycystic kidney disease. Arrows, enlarged polycystic right kidney; Gb, gall bladder.

recanalisation of fetal veins in portal venous hypertension) and varices. Pulsed-wave Doppler is routinely used to evaluate flow in the portal vein. When traces are difficult to obtain in the moving child, colour Doppler will at least establish the flow direction in most cases. Twinkling (artefact) seen on colour Doppler is useful to detect calcifications, gallstones and bile duct hamartomas. The sensitivity increases when the focal zone is set deep to the abnormality and with higher colour write priority.

Magnetic Resonance Imaging Provided there is adequate preparation, MRI allows high-resolution (submillimetre isotropic) high-contrast imaging of the liver. Respiratory gating is used whenever possible. Useful pulse sequences include short tau inversion recovery fast spin-echo (STIR), volumetric T2 weighted spinecho with variable refocusing pulse (CUBE, SPACE, VISTA), in-phase and opposed-phase spoiled gradient-echo, diffusion-weighted imaging and

1834

SECTION G  Paediatric Imaging

A

B Fig. 71.45  Imaging in a 3-year-old with Beckwith–Wiedemann syndrome and previously resected hepatoblastoma. (A) During off-treatment surveillance, a 15-mm subcapsular nodule (between markers) is seen with highfrequency ultrasound in the left lobe of the liver. (B) Axial spoiled gradient-echo magnetic resonance imaging 20 minutes following intravenous injection of gadoxetic acid (a hepatocyte-specific contrast agent) demonstrates normal enhancement of surrounding liver, but no enhancement in the nodule (arrowheads). Recurrent hepatoblastoma was therefore suspected, and confirmed histopathologically. Note contrast material in the common bile duct (arrow).

T1 weighted gradient-echo before and after intravenous administration of contrast medium (possibly with several post-contrast acquisitions, as used in adults to detect features such as centripetal enhancement in infantile haemangioma and rapid washout in neoplasms). Non-contrast-enhanced angiographic techniques include selective inversion followed by balanced steady-state free precession acquisition (NATIVE, TRANCE, FBI). This technique depends on inflow of fresh spins, and may therefore require several acquisitions with varying placement of the inversion volume. The technique is promising for portal venography. Hepatocyte-specific contrast agents are promising for detection of small lesions, lesion characterisation, and for functional assessment of biliary drainage, but are currently restricted to off-label use (Fig. 71.45).

Computed Tomography CT should be restricted in children owing to (1) its general poor softtissue contrast, (2) because paucity of body-fat in young children hampers identification of tissue planes and (3) for radiation protection. If used, protocols need to be in place to minimise the radiation dose. Ideally there should be no unenhanced CT imaging with only one intravenous contrast-enhanced CT data set—with very few exceptions.

Angiography Angiography is very rarely indicated for diagnosis and is, as such, only performed in specialist centres—usually as a precursor before possible intervention.

Fig. 71.46  Duplex Doppler sonography in a neonate demonstrates a patent ductus venosus (arrowhead) between the portal vein (P) and the left hepatic vein (L) just distal to the inferior vena cava. Anatomical variation is common in this fetal umbilico-systemic shunt.

Imaging Anatomy The neonatal liver is hypoechoic relative to the kidneys. This reverses during infancy. The umbilical vein and the ductus venosus are patent in the premature and early newborn child (90%) for extrahepatic biliary atresia. Radioisotope studies have high sensitivity but a false-positive rate above 20%. The poor specificity is mostly caused by poor biliary excretion/drainage in other conditions, such as metabolic disease, infection, persistent intrahepatic cholestasis, total parenteral nutrition and neonatal hepatitis.

Choledochal Malformation (Choledochal Cyst; Fig. 71.51) Choledochal malformations comprise a spectrum of conditions with overlapping expressions in which there is abnormal widening of the biliary tract without acute obstruction. Approximately 80% manifest clinically during childhood. Malignant transformation in children is

not documented. The overall incidence is 1 : 100,000 to 200,000, but as high as 1 : 1000 in Japan. In infants younger than 3 months with extrahepatic biliary dilatation and conjugated hyperbilirubinaemia, a bile duct diameter less than 3 mm suggests a non-surgical cause, whereas a diameter greater than 4 mm suggests choledochal malformation. The intermediate cases are often associated with inspissated bile syndrome. Most commonly there is spherical or fusiform dilatation of the extrahapatic ducts (type 1). A common pancreaticobiliary channel may be seen sonographically and with MRCP; flux of pancreatic excretions into the common bile duct is thought to be a pathogenetic factor. The clinically most important differential diagnosis for a cyst at the porta hepatic is cystic biliary atresia. Other differential diagnoses include duodenal duplication cyst and lymphangioma. Other variants demonstrate a cystic diverticulum from the common bile duct (type 2), cholodochocele into the duodenum (type 3) or a combination of intra- and extrahepatic dilatation (type 4). Choledochal malformations may be part of the fibropolycystic spectrum (see Liver).

1840

SECTION G  Paediatric Imaging

Caroli disease is an association between intrahepatic segmental duct dilatations (type 5 choledochal malformation), (congenital) hepatic fibrosis and cystic kidney disease. Intrahepatic focal biliary dilatation can be recognised by the central dot sign, which represents the encased adjoining portal vein branch. Differential diagnoses to intrahepatic choledochal malformation include sclerosing cholangitis (see later) and recurrent pyogenic cholangitis. There may be complications such as cholangitis (caused by stagnant bile), biliary obstruction (calculi), cirrhosis (secondary to obstruction) and cholangiocarcinoma in adults. Surgical treatment is usually hepatoenterostomy.

Inspissated Bile In infants and young children, obstruction (partial or total) by stagnantformed bile may be associated with prematurity, haemolysis, functional intestinal obstruction (Hirschsprung disease), CF and total parenteral nutrition. It is usually idiopathic in infants (inspissated bile syndrome). US demonstrates the (slightly) echogenic bile with no acoustic shadowing, and the secondary biliary dilatation (mainly extrahepatic). There may be increased periportal echogenicity if long-standing. Choledochal malformation is the main differential diagnosis in infants.

Persistent Intrahepatic Cholestasis This is a spectrum of inherited disorders that need to be considered in infantile jaundice. It includes Alagille syndrome (paucity of intrahepatic bile ducts associated with butterfly vertebrae, congenital heart disease, ocular and/or facial anomalies). There is no documented role for imaging in these conditions; however, this may change with the introduction of hepatocyte-specific contrast agents. Other causes of biliary obstruction include external compression by cystic (e.g. duplication cysts) or solid (e.g. enlarged lymph nodes) masses. Biliary neoplasms are rare in children (see Neoplasia).

Sludge and Gallstones The prevalence of gallstones increases with age. Gallstones are rare in children unless there is risk factor, such as a haemolytic disorder, obesity, CF, small bowel disease, choledochal malformation or total parenteral nutrition. There may not necessarily be a developmental continuum from sludge to stone formation.

Spontaneous Perforation of the Bile Ducts

usually after 2 years of age, but a neonatal form is well-known. The differential diagnoses include primary biliary cirrhosis, autoimmune hepatitis, biliary atresia and graft-versus-host disease. Sclerosing cholangitis leads to progressive cholestasis and cirrhosis. The risk a person with primary sclerosing cholangitis has for developing cholangiocarcinoma is estimated at 0.6% to 1.5% per year.

Neoplasia Biliary neoplasms are rare in children. Biliary rhabdomyosarcoma is suspected when there is a mass at the porta hepatic and associated proximal biliary dilatation. It rarely invades the portal vein. Cystic variants are known. Cholangiocarcinoma is an important differential diagnosis in children with sclerosing cholangitis. It is unlikely to arise in childhood from choledochal malformations.

PANCREAS Imaging Techniques US usually provides complete depiction of the pancreas in children. The pancreatic tail is usually seen through a splenic acoustic window. Islet cell neoplasm and other small lesions, however, are not commonly visible, and may also not be seen on MRI. Complete imaging in suspected cases therefore involves radioisotope studies. The pancreatic ducts may be difficult to visualise on MRCP unless the exocrine pancreas is stimulated using intravenous secretin; however, care should be taken in case of recent pancreatitis, which may be exacerbated.

Imaging Anatomy Pancreatic size is variable, volumetric references unavailable, and sonographic measurements have unknown reliability. The pancreas is large relative to the size of the child, and hypoechoic. Suggested antero­ posterior dimensions on US are (infants to teenagers, cm) 1 to 2 cm (head and tail) and 0.6 to 1.1 cm (body) with standard deviations of around 0.4 cm. Suggested normal pancreatic duct diameters are 1.1 mm in toddlers to 2.1 mm in late teens with standard deviations of about 0.2 mm.

Congenital Abnormalities and Associations Pancreas Divisum (Fig. 71.52)

This is a very rare condition. The clinical presentation is of increasing ascites, irritability and variable mild jaundice in a young infant. Hepatobiliary radioisotope scintigraphy demonstrates extrabiliary pooling; US or MRI may verify coinciding loculated fluid and possibly an underlying cause, such as a choledochal malformation, gallstone or biliary stenosis.

This is a common, and often uncomplicated, variant. The embryonal ventral and dorsal anlage have separate ducts that normally connect. If such fusion fails, two separate ducts persist. As the smaller ventral anlage develops into the larger part (body, tail, part of the head), its duct (of Santorini) may provide inadequate draining capacity through the minor papilla, which may predispose to recurrent acute pancreatitis; however, an association is now disputed (in adults).

Cholangitis

Annular Pancreas (Fig. 71.53)

Cholangitis in children prompts a search for predisposing conditions. These include anatomical abnormalities (choledochal malformation), congenital hepatic fibrosis, biliary obstruction (gallstone, inspissated bile) and complications of surgery (e.g. transplantation).

Annular pancreas, i.e. pancreatic tissue encasing the second part of the duodenum, results from erroneous migration of the anlagen. It may result in duodenal obstruction and is a differential diagnosis of the double-bubble sign on neonatal radiographs. Associated anomalies are common in children (71%) and differ from those in adults, the most common being trisomy 21, intestinal and cardiac anomalies that often require surgery.

Sclerosing Cholangitis The pathology and imaging findings are similar to those in adults: a beaded appearance of alternating strictures and dilatations of intra- and extrahepatic bile ducts. Primary sclerosing cholangitis occurs in up to 80% of patients with ulcerative colitis (the clinical presentations may be metachronous). Conversely, IBD is found in approximately half of children with sclerosing cholangitis. Secondary sclerosing cholangitis may be seen in Langerhans cell histiocytosis. Clinical presentation is

Other Agenesis of the dorsal anlage: short rounded head, no body/tail. It is associated with insulin-dependent diabetes mellitus, possibly by a gene mutation. Ectopic pancreatic tissue is most frequently located in the wall of the stomach, duodenum or jejunum.

CHAPTER 71  Paediatric Abdominal Imaging

1841

Systemic Disorders

Cystic Fibrosis (Fig. 71.54) High-viscosity secretions cause distal obstruction and inevitable destruction of the pancreas in children with CF. The main imaging findings are atrophy and/or fat replacement; however, fibrous tissue and foci of calcification may be seen. Pancreatitis is a less common complication, occurring in just over 1% of patients.

Other Diffuse enlargement of the pancreas may be seen in children with Beckwith–Wiedemann syndrome, which also predisposes to pancreatoblastoma, albeit seen very rarely. Both autosomal dominant polycystic kidney disease and von Hippel–Lindau disease may manifest with pancreatic cyst(s). Fig. 71.52  Thick-slab heavily T2 weighted fast spin-echo in a child following acute pancreatitis. A pseudocyst (p) is seen. The pancreatic duct from the tail, body and proximal head appears to drain exclusively through the duct of Santorini, which narrows abruptly near the minor papilla (arrow).

Pancreatitis Underlying conditions differ in children. In acute pancreatitis they include congenital biliary abnormalities, viral infections, systemic disease (e.g. Henoch–Schönlein purpura and other vasculitides, metabolic and other hereditary disease) and pancreatic duct abnormalities; however, approximately one-third are idiopathic. Apart from playing a role in the acute stage (as in adults), imaging in children is directed towards uncovering any underlying anatomical abnormality. Conditions associated with chronic pancreatitis include CF, Shwachman–Diamond syndrome and other hereditary disorders.

Trauma Children are more prone to pancreatic injury, which may manifest as laceration, transection and/or acute pancreatitis. As with any injury in childhood, non-accidental causes need to be considered, particularly in the youngest.

Congenital Hyperinsulinism

Fig. 71.53  Contrast-enhanced computed tomography in a toddler with symptoms of partial upper gastrointestinal obstruction demonstrates annular pancreas. The second part of the duodenum (arrows) is completely encircled by pancreatic tissue.

A

Congenital hyperinsulinism is caused by diffuse or focal inappropriate secretion of insulin. 18F-Fluoro-DOPA PET scintigraphy has an important role in differentiating the two, which is important for planning surgical options (partial or complete resection of the pancreas). Co-registration with MRCP images allows assessment of the relation between a focal lesion and the common bile duct.

B Fig. 71.54  Fatty Replacement of the Pancreas in a Child With Cystic Fibrosis. (A) Sonogram demonstrates a small, hyperechoic pancreas (arrowheads). (B) Contrast-enhanced computed tomography shows a fatattenuated pancreas. Gradual destruction happens during childhood owing to stagnant secretions, and may also manifest as fibrosis and calcification.

1842

SECTION G  Paediatric Imaging

TABLE 71.10  Suggested Upper Limit of

Normal Sonographic Measurement of the Spleen in Children Age (months)

Fig. 71.55  Solid cystic and papillary neoplasm (Frantz tumour; arrowheads) in an adolescent girl. Axial T2 weighted fast spin-echo magnetic resonance imaging shows a heterogeneous hyperintense mass with a thick hypointense (fibrotic) rim.

Suggested Upper Limit for the Sonographic Longitudinal Diameter (Coronal Orientation) of the Spleen (mm)

1–3 4–6 7–9 12–30 36–59 60–83 84–107 108–131 132–155 156–179 180–200

70 75 80 85 95 105 105 110 115 120 120

The limits are based on an ethnically homogeneous material. Considerable intra- and inter-observer variation need to be accounted for.

MRI is preferred for further investigation of larger (>1 cm) uncertain lesions.

Imaging Anatomy Accessory spleens are common and of no real clinical importance. In the neonate the spleen is hypointense both on T1 and T2 weighted images. With increasing white pulp to red pulp ratio, its MRI appearance is similar to that in adults by 8 months of age. Table 71.10 suggests upper size limits. High-frequency US may demonstrate the spotted appearance of a reactive spleen, and this should not be mistaken for multifocal infection. Fig. 71.56  Magnetic resonance imaging in a 9-year-old boy with pancreatic insulinoma. Axial T2 weighted MRI (grey scale) with overlay of diffusion-weighted (b, 1000) magnetic resonance imaging (red tones) shows the insulinoma (arrowheads) because of its mildly restricted water diffusion. Note also how tumour deflects the main pancreatic duct (arrow) posteriorly.

Imaging Findings

Neoplasms

Wandering Spleen

Pancreatoblastoma is rare, even in predisposed children with Beckwith– Wiedemann syndrome. Presentation is usually in infants or young children with a heterogeneous solid/cystic mass and variable enhancement. Solid and papillary epithelial neoplasm (Frantz tumour; Fig. 71.55) has relatively higher incidence in adolescents and in females. Imaging often demonstrates a large well-defined lesion with cystic/haemorrhagic degeneration, calcification and a low-intensity (fibrous) rim on MRI. Islet cell tumours are associated with multiple endocrine neoplasia type 1 and with von Hippel–Lindau disease. Functioning entities in children include insulinoma (most common; Fig. 71.56), gastrinoma, VIPoma and glucagonoma. These are often small and may be undetectable on US, CT and MRI; thus, imaging in suspected cases needs to include radioisotope studies.

Wandering spleen is associated with prune-belly syndrome, surgery and gastric volvulus, and predisposes to splenic torsion and infarction.

SPLEEN Imaging Techniques US (high-frequency linear transducer) is the best investigation for detecting small parenchymal lesions, such as focal fungal infection.

Splenomegaly

Splenomegaly has a wide differential, as in adults. In children one also needs to consider mononucleosis, depositional disorders (Gaucher disease, mucopolysacharidosis, Niemann–Pick disease), Langerhans cell histiocytosis and other conditions.

Focal Lesions Most focal lesions (solitary or multifocal) in the spleen are infectious (Fig. 71.57): abscesses and fungal infection (Fig. 71.58); granulomata and hydatid cysts often contain calcification. (See also under Neoplasia.)

Calcifications Granulomatous disease, infarcts, haemangioma, hamartoma, hydatid

Lateralisation Disorders (Fig. 71.59) Entities are overlapping. Both asplenia and polysplenia are associated with congenital heart disease. Asplenia is more commonly associated with immunodeficiency and polysplenia with azygos continuation, preduodenal portal vein, bilateral left-sidedness of the lungs and biliary atresia. On US the spleen should be located near the greater curvature of the stomach, regardless of its site. Radioisotope studies may be required to confirm asplenia.

CHAPTER 71  Paediatric Abdominal Imaging

Infarction Splenic infarction may occur in disorders with massive sequestration (sickle-cell anaemia), deposition (storage disorders) or infiltration (leukaemia).

Trauma Trauma epidemiology and mechanisms are different in children than in adults, but imaging is similar. Focused abdominal sonography for trauma (FAST) has low (50%) negative predictive value for abdominal injury in haemodynamically stable children post trauma, as compared with CT. More comprehensive, but still fast (median imaging time 5 minutes), US performed by sonographers in non-selected children post trauma soon after arrival in an emergency department was reported as highly accurate, but less sensitive, in a prospective study, for any abdominal traumatic injury compared with a combination of CT, peritoneal lavage and laparotomy, with a negative predictive value of 91% for haemoperitoneum.

Fig. 71.57  Ultrasound in a 10-year-old with falciparum malaria shows an enlarged spleen with a spotted appearance, which is commonly seen in reactive states and should not be mistaken for focal lesions.

A

1843

NAI is always an important differential, particularly in the younger age groups. In children with abdominal injury following abuse, the frequency of injury to the spleen has been reported to come third, after small bowel and liver.

Neoplasia Most splenic neoplasms are benign. Cystic lesions are usually epidermoid/ dermoid cysts, lymphangioma or splenic (epithelial) cysts. Solid lesions include haemangioma and hamartoma (Fig. 71.60). Malignant tumours are most commonly lymphoproliferative disease. Metastases and primary splenic angiosarcoma are rare.

Fig. 71.58  Ultrasound with a high-frequency linear transducer in a child with acute lymphoblastic leukaemia and febrile neutropenia shows several hypoechoic (arrowhead) and target lesions (arrow) suggestive of fungal infection.

B Fig. 71.59  Polysplenia on Computed Tomography. (A) Multiple spleens (between arrowheads) in this child are associated with minor congenital heart disease and azygos (arrow) continuation of interrupted inferior vena cava. (B) Visceral situs inversus, multiple spleens (S) which in this child with primary ciliary dyskinesia was associated with major congenital heart disease and bronchial isomerism (Kartagener syndrome).

1844

SECTION G  Paediatric Imaging

A

B

C

D

Fig. 71.60  Splenic Hamartoma as an Incidental Finding in a Child. Sonogram (A) and duplex Doppler ultrasound (B) demonstrate a vascularised well-demarcated, hypoechoic lesion (arrowheads). (C) Coronal T2 weighted magnetic resonance imaging shows the lesion (h) as an almost geometrical hexagon with a fine hypointense perimeter and centre (resembling focal nodular hyperplasia of the liver). (D) There is homogeneous enhancement of the lesion (h) 3 minutes after intravenous administration of gadoteric acid.

FURTHER READING Ahle, M., Ringertz, H.G., Rubesova, E., 2018. The role of imaging in the management of necrotising enterocolitis: a multispecialist survey and a review of the literature. Eur. Radiol. 28, 3621–3631. Alamo, L., Meyrat, B.J., Meuwly, J.Y., et al., 2013. Anorectal malformation— finding the pathway out of the labyrinth. Radiographics 33, 491–512. Berger, M.Y., Tabbers, M.M., Kurver, M.J., et al., 2012. Value of abdominal radiography, colonic transit time, and rectal ultrasound scanning in the diagnosis of idiopathic constipation in children: a systematic review. J. Pediatr. 161, 44–50.e1-2. Bolmers, M.D., van Rossem, C.C., Gorter, R.R., et al., 2018. Imaging in pediatric appendicitis is key to a low normal appendix percentage: a national audit on the outcome of appendectomy for appendicitis in children. Pediatr. Surg. Int. 34, 543–551. Brancatelli, G., Federle, M.P., Vilgrain, V., et al., 2005. Fibropolycystic liver disease: CT and MR imaging findings. Radiographics 25, 659–667. Carroll, A.G., Kavanagh, R.G., Ni Leidhin, C., et al., 2016. Comparative effectiveness of imaging modalities for the diagnosis of intestinal obstruction in neonates and infants. Acad. Radiol. 23, 559–568. Cheng, G., Soboleski, D., Daneman, A., et al., 2005. Sonographic pitfalls in the diagnosis of enteric duplication cysts. AJR Am. J. Roentgenol. 184 (2), 521–525. Cuna, A.C., Reddy, N., Robinson, A.L., et al., 2018. Bowel ultrasound for predicting surgical management of necrotizing enterocolitis: a systematic review and meta-analysis. Pediatr. Radiol. 48, 658–666.

Daneman, A., Navarro, O., 2003. Intussusception. Part 1: a review of diagnostic approaches. Pediatr. Radiol. 33 (2), 79–85. Daneman, A., Navarro, O., 2004. Intussusception. Part 2: an update on the evolution of management. Pediatr. Radiol. 34 (2), 97–108, quiz 87. Fike, F.B., Mortellaro, V.E., Holcomb, G.W., et al., 2012. Predictors of failed enema reduction in childhood intussusception. J. Pediatr. Surg. 47, 925–927. Frykman, P.K., Short, S.S., 2012. Hirschsprung-associated enterocolitis: prevention and therapy. Semin. Pediatr. Surg. 21, 328–335. Goldin, A.B., Khanna, P., Thapa, M., et al., 2011. Revised ultrasound criteria for appendicitis in children improve diagnostic accuracy. Pediatr. Radiol. 41 (8), 993–999. Goodwin, S.J., Flanagan, S.G., McDonald, K., 2015. Imaging of chest and abdominal trauma in children. Curr. Pediatr. Rev. 11, 251–261. Gordon, P.V., Swanson, J.R., 2014. Necrotizing enterocolitis is one disease with many origins and potential means of prevention. Pathophysiology 21 (1), 13–19. Greer, M.C., 2018. Paediatric magnetic resonance enterography in inflammatory bowel disease. Eur. J. Radiol. 102, 129–137. Hochart, V., Verpillat, P., Langlois, C., et al., 2015. The contribution of fetal MR imaging to the assessment of oesophageal atresia. Eur. Radiol. 25, 306–314. Ilivitzki, A., Shtark, L.G., Arish, K., et al., 2012. Deep sedation during pneumatic reduction of intussusception. Pediatr. Radiol. 42, 562–565.

CHAPTER 71  Paediatric Abdominal Imaging Iqbal, C.W., Rivard, D.C., Mortellaro, V.E., et al., 2012. Evaluation of ultrasonographic parameters in the diagnosis of pyloric stenosis relative to patient age and size. J. Pediatr. Surg. 47 (8), 1542–1547. Irvine, I., Doherty, A., Hayes, R., 2017. Bleeding Meckel’s diverticulum: a study of the accuracy of pertechnetate scintigraphy as a diagnostic tool. Eur. J. Radiol. 96, 27–30. Khachab, F., Loundou, A., Roman, C., et al., 2018. Can diffusion weighting replace gadolinium enhancement in magnetic resonance enterography for inflammatory bowel disease in children? Pediatr. Radiol. 48, 1432–1440. Lee, E. (Ed.), 2017. Pediatric Radiology: Practical Imaging Evaluation of Infants and Children, first ed. Lippincott Williams and Wilkins. Makin, E., Davenport, M., 2012. Understanding choledochal malformation. Arch. Dis. Child. 97, 69–72. Niedzielski, J., Kobielski, A., Sokal, J.,Krakos M., 2011. Accuracy of sonographic criteria in the decision for surgical treatment in infantile hypertrophic pyloric stenosis. Arch. Med. Sci. 7, 508–511. Nievelstein, R.A.J., Robben, S., Blickman, J.G., 2011. Hepatobiliary and pancreatic imaging in children—techniques and an overview of non-neoplastic disease entities. Pediatr. Radiol. 41, 55–75. Putnam, L.R., John, S.D., Greenfield, S.A., et al., 2015. The utility of the contrast enema in neonates with suspected Hirschsprung disease. J. Pediatr. Surg. 50, 963–966. Riccabona, M., Lobo, M.L., Ording-Muller, L.S., et al., 2017. European Society of Paediatric Radiology abdominal imaging task force recommendations in paediatric uroradiology, part IX: imaging in anorectal and cloacal malformation, imaging in childhood ovarian torsion, and efforts in standardising paediatric uroradiology terminology. Pediatr. Radiol. 47, 1369–1380.

1845

Roebuck, D., 2008. Focal liver lesions in children. Pediatr. Radiol. 38, S518–S522. Roebuck, D., Olsen, Ø., Pariene D., 2005. Radiological staging in children with hepatoblastoma. Pediatr. Radiol. 36, 176–182. Rosen, R., Vandenplas, Y., Singendonk, M., et al., 2018. Pediatric gastroesophageal reflux clinical practice guidelines: joint recommendations of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN) and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN). J. Pediatr. Gastroenterol. Nutr. 66, 516–554. Rozel, C., Garel, L., Rypens, F., et al., 2011. Imaging of biliary disorders in children. Pediatr. Radiol. 41, 208–220. Silva, C.T., Daneman, A., Navarro, O.M., et al., 2013. A prospective comparison of intestinal sonography and abdominal radiographs in a neonatal intensive care unit. Pediatr. Radiol. 43, 1453–1463. Strafrace, S., Blickman, J.G. (Eds.), 2016. Radiological Imaging of the Digestive Tract in Infants and Children, second ed. Springer Switzerland. Tabbers, M.M., DiLorenzo, C., Berger, M.Y., et al., 2014. Evaluation and treatment of functional constipation in infants and children: evidencebased recommendations from ESPGHAN and NASPGHAN practice: diagnosis and treatment of functional constipation. Eur. J. Pediatr. 58 (2), 258–274. Tong, S.C., Pitman, M., Anupindi, S.A., 2002. Best cases from the AFIP. Ileocecal enteric duplication cyst: radiologic-pathologic correlation. Radiographics 22 (5), 1217–1222. Tsai, T.L., Marine, M.B., Wanner, M.R., et al., 2017. Can ultrasound be used as the primary imaging in children with suspected Crohn disease? Pediatr. Radiol. 47, 917–923.

72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children Owen Arthurs, Marina Easty, Michael Riccabona

CHAPTER OUTLINE Overview, 1846 Imaging Techniques, 1846 Congenital Anomalies, 1855 Renal Anomalies, 1855 Urethral Anomalies, 1860 Uterus and Vagina, 1862 Undescended Testis, 1862 Antenatal Diagnosis of Pelvicalyceal and/or Ureteral Distention, 1863 Urinary Tract Infection and Vesicoureteric Reflux, 1865 Renal Cystic Disease, 1869 Nephrocalcinosis, 1873

Renal Calculi, 1873 Tumours, 1875 Benign Tumours, 1875 Malignant Tumours, 1876 Inflammatory Diseases of the Scrotum, 1878 Scrotal Masses, 1878 Ovarian Masses, 1879 Presacral Masses, 1879 Hypertension, 1880 Trauma, 1883 Renal Transplantation, 1883

OVERVIEW

staging and prognostication as well as for imaging complications. CT is used much less frequently in children than in adults for urolithiasis, although it does have a place when imaging children with US proven calculi who present in pain; CT is used in children when they sustain trauma. The role of MRI is increasing both for anatomical and functional diagnostic information, particularly in cooperative older children, and where nuclear medicine is not available. Conventional angiography is reserved for specific clinical indications, and is invasive with a high radiation burden. CT and magnetic resonance angiography (MRA), however, have emerged as replacements for angiography for many diagnostic purposes. Here we review the relative strengths and weaknesses of these techniques, illustrated using specific pathologies and recommend imaging algorithms.

In this chapter, we cover the important areas of renal, urinary tract and pelvic imaging in children, emphasising the importance of congenital abnormalities and the need for minimising radiation burden and optimising image quality. Ultrasound (US) is the preferred method for imaging the paediatric population, due to the ease of availability, lack of ionising radiation, reproducibility and because it is usually well tolerated. Modern US is often the only investigation required to make a diagnosis, with the high frequency probes and new technology producing exquisite anatomical details in children who are ideal subjects. Alternatively, US can be used to direct other imaging techniques, such as functional assessment of the urinary tract by 99mtechnetium mercaptoacetyltriglycine (99mTcMAG3) dynamic imaging. Intravenous urography (IVU) is now rarely used. Fluoroscopic micturating cystourethrography (MCUG) is essential to exclude bladder outflow obstruction such as in posterior urethral valves (PUV) and urethral abnormalities, and to assess for vesicoureteric reflux (VUR). Contrast-enhanced voiding urosonography (ce-VUS) with micro-bubble contrast is used in some European countries rather than a fluoroscopic examination for VUR assessment, avoiding the radiation burden from conventional MCUG. Direct isotope cystography using 99mTcpertechnetate is a low-dose functional study used to assess VUR, particularly in girls, where urethral anatomy is usually normal. Older, continent and cooperative children may benefit from an indirect radionuclide cystogram (IRC) as part of their dynamic 99mTc-MAG3 renogram, as a non-invasive means of assessing for VUR. Cross-sectional imaging—computed tomography (CT) and magnetic resonance imaging (MRI)—is crucial in tumour imaging, for disease

1846

SUMMARY BOX Ultrasound is the preferred imaging method for children. There are specific anatomical and physiological indications for fluoroscopy, nuclear medicine studies, computed tomography and magnetic resonance imaging. Consider the relative strengths and weaknesses of each technique, and refer to published imaging algorithms.

IMAGING TECHNIQUES Plain Radiography Radiographs still have a role to play in babies with congenital renal anomalies, particularly in those cases associated with skeletal anomalies, such as vertebral segmentation anomalies and pubic diastasis (in bladder

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

1847

Well-hydrated patient, full bladder, adequate equipment & transducer, training, etc.

Urinary bladder: size (volume), shape, ostium, wall, bladder neck, include distal ureter & retrovesical space/ internal genitalia Optional: CDS for urine inflow, perineal US, scrotal US ...

Kidneys: lateral and/or dorsal, longitudinal and axial sections, parenchyma? Pelvicalyceal system? Standardised measurements in 3 planes & volume calculation. If dilated: + max. axial pelvis & calyxdiameter, narrowest parenchymal width + ureteropelvic junction Optional: (a)CDS & duplex Doppler

Post-void evaluation Bladder: residual volume, bladder neck, shape & configuration Kidneys: dilatation of pelvicalyceal system/ureter changed? Optional: ce-VUS, 3DUS ... Fig. 72.1  Renal Calculus. Plain abdominal radiograph of a large staghorn calculus in the right kidney.

exstrophy). They may show renal tract calculi (Fig. 72.1) where the exposure should be coned to the kidneys, ureters and bladder (so-called ‘KUB film’). Age-appropriate exposure settings and electronic filtering (in digital radiography equipment) are essential.

Note: Cursory US of entire abdomen is recommended for first study, and in mismatch of findings and query Fig. 72.2  Ultrasound of the Urinary Tract. European Society of Paediatric Radiology imaging recommendations for standard paediatric sonography of the urinary tract. CDS, Colour Doppler sonography; ce-VUS, contrastenhanced voiding urosonography; 3DUS, three-dimensional ultrasound; US, ultrasound. (Adapted from Riccabona et al. Pediatr Radiol. 2008; 38:138–145.)

Ultrasound US is the most useful way of providing anatomical information about the intra-abdominal, pelvic and retroperitoneal structures. The ability to delineate and recognise normal and abnormal findings is directly related to the skill of the ultrasonographer and the equipment used, including high-frequency transducers, and familiarity examining children in a conducive environment, with knowledge of the spectrum of diseases in childhood being paramount. Several European guidelines regarding standard paediatric urinary tract US are available.

Standard Technique In the young child, ideally a full bladder is necessary, which usually requires at least 30 to 60 minutes of encouraged fluid intake to allow adequate hydration. Any US examination of the abdomen should begin by imaging the bladder, in an attempt to capture the bladder when full (as the infant in nappies may void at any time!). The distended bladder provides an acoustic window for the lower urinary tract, bladder neck and ureteric orifices (vesicoureteric junction), distal ureters, internal genitalia, retrovesical space, pelvic musculature and vessels (Fig. 72.2). It is customary to measure pre- and post-micturition bladder volumes using shape adapted correction factors, as incomplete voiding may be related to bladder dysfunction and urinary tract infections (UTIs), and the presence of pre- and/or post-micturition upper tract dilatation. Normal age-related changes in the kidney must be appreciated. The neonatal kidneys lack renal sinus fat in the first 6 months of life, and the medullary pyramids are typically large and hypoechoic relative to the cortex (the opposite to that found in older children and adults), which may be mistaken for pelvicalyceal dilatation (PCD) or ‘cysts’ (Fig. 72.3). The normal neonatal renal cortex is also hyper- to iso-echoic relative to the adjacent normal liver, which again can often be reversed

Fig. 72.3  Normal Neonatal Kidney. Oblique ultrasound image of a normal neonatal kidney. The medullary pyramids are hypoechoic relative to the cortex (the opposite to that found in older children and adults), which may be mistaken for pelvicalyceal dilatation.

in adults. The neonatal renal pyramids may be echogenic, a transient physiological appearance in up to 5% of newborns, and should not be mistaken for nephrocalcinosis, although it can be seen in older infants with dehydration. The average newborn kidney is approximately 4.5 cm in length, and measurements of bipolar renal lengths can be compared against age/height/weight indexed charts (Fig. 72.4). As the paediatric kidney is more spherical than the ellipsoid adult kidney, renal volumes may be a better assessment using the equation:

1848

SECTION G  Paediatric Imaging 13 12 11

Renal length (cm)

10 9 8 7 6 Fig. 72.5  Ovarian Cysts. Ultrasound of a neonatal physiologically large ovarian cyst with daughter cysts.

5 4 Predicted mean

3

95% prediction limits 2

0

2

4

6 Months

8 10 12

5

Age

10

15

Years

Fig. 72.4  Renal Growth Chart. Normal reference sizes and 95th centiles given by age of child.

Kidney volume = 0.523 (π 6) × length × width × depth Colour Doppler sonography (CDS) of the renal vessels is particularly important in assessing perfusion in a variety of conditions, including infection (segmental perfusion impairment), trauma (hilar vascular injury), biopsy (post-biopsy complications), renal failure, urolithiasis tumours, renal transplants and hypertension, and for the vascular anatomy in pelvicalyceal dilatation (PCD); or urinary tract dilatation (UTD); previously called pelvicalyceal dilatation (PCD) although this term is now outdated). High-frequency linear transducers yield better images in smaller children, especially in the prone position and should be performed for detailed analysis in all examinations. Diuretic sonography applies diuretic stress by administering furosemide and thereafter evaluating the dynamics of the collecting system and the renal arterial flow patterns to differentiate between somewhat obstructed from only dilated systems.

Normal Gonadal Imaging in Girls Normal pelvic structures can be difficult to visualise in children: visualisation of the ovaries by US depends on their location, size and the age of the girl—they are more easily seen in the first few months of life. Ovarian volume is usually under 1 mL in the neonate, and 2 to 4 mL in the prepubertal child. Ovaries typically look heterogeneous due to the presence of follicles, and larger follicles can appear as small ovarian ‘cysts’ which are normal at all ages (Fig. 72.5). After puberty, ovarian volumes of 5 to 15 mL are normal, with normal primordial follicles less than 10 mm in diameter, and stimulated follicles 10 to 30 mm in diameter. The normal uterus also changes dramatically with hormonal changes, and is the most useful guide to pubertal staging. The neonatal uterus

is prominent due to circulating maternal oestrogens, typically measuring 2 to 4.5 cm in length, with thickened and clearly visible endometrial lining (Fig. 72.6A). By 1 year of age it becomes smaller, has changed to the prepubertal tubular appearances: the fundus and cervix are the same size and the endometrium is no longer visible (see Fig. 72.6B). At puberty, the fundus starts to enlarge, becomes up to three times the size of the cervix, with a total uterine length of 5 to 7 cm and the typical adult pear-like shape. The endometrial appearances will clearly vary with the phase of the menstrual cycle. The vagina may be visualised by US if air-filled (shown as a linear bright echo), or if fluid filled. On MRI the vagina is best seen on sagittal T2 weighted spin-echo sequences. As with the uterus, the appearance and the thickness of the vaginal epithelium and the signal from the vaginal wall change with the age and the phases of the menstrual cycle. US genitography with saline filling of the vagina, or 3D US can be used for uterine anomalies, and perineal US for vaginal and urethral problems.

Normal Gonadal Imaging in Boys The prostate has an ellipsoid homogeneous appearance, but is difficult to see in newborns, as are the normal seminal vesicles. As the processus vaginalis remains open for some time after birth (and may never close completely), hydroceles are considered a normal physiological finding in the newborn. Cryptorchidism is discussed later in this chapter. The normal testis changes in appearance during childhood. It has a homogeneous hypoechoic echotexture and is spherical/oval in shape during the neonatal period, measuring less than 10 mm in diameter (Fig. 72.7). The epididymis and mediastinum testis are usually not seen at this point, but are clearly identified by puberty. Testicular size in adolescence ranges from 3 to 5 cm in length and from 2 to 3 cm in depth and width (2 to 4 mL in total). Testicular flow, as measured by Doppler US, also changes with age. The testis in infants shows very low-velocity colour flow, which can be difficult to see despite optimised slow-flow settings, and even normal prepubertal testes may not demonstrate low-velocity flow even on power Doppler US. Technically it may be difficult to identify abnormalities in a single testis given the wide range of normal values, and thus a side-by-side comparison can be useful.

Cystography There are several ways to visualise the bladder in the paediatric population. The method of choice depends both on the type of problem

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

A

1849

B Fig. 72.6  Normal Uterus. (A) Sagittal ultrasound (US) image of normal neonatal uterus, which is prominent due to circulating maternal oestrogens with a clearly visible endometrial lining. (B) Sagittal US image of the infantile uterus which has a prepubertal tubular appearance: the fundus and cervix are the same size and the endometrium is no longer visible.

Indications: febrile and recurrent UTI, particularly in infants, suspected PUV, UT malformation, PCD or UTD > IIº or ‘extended criteria’

Preparations: no diet restriction or enema, urine analysis, after AB are completed... Catheterisation: feeding tube, 4-8Fr or suprapubic puncture, anaesthetic lubricant or coated plaster Latex precaution: neuro tube defect, bladder exstrophy

Fig. 72.7  Normal Testes. Axial ultrasound image of normal prepubertal testes.

suspected and the age of child. All methods, apart from the 99mTc-MAG3 IRC, require a bladder catheter, which becomes more unpleasant for both child (and the carer who observes) as the child gets older. The modern MCUG, using pulsed fluoroscopy, digital image amplifiers and last image hold, means that high-quality imaging is now available at an acceptably low radiation dose (Table 72.1). The direct isotope cystogram is useful in the assessment of VUR in young babies (before toilet training), particularly in girls where there is no need to demonstrate the urethral anatomy, or for screening other family members when the index of suspicion for reflux is high.

Micturating (Voiding) Cystogram (MCUG/VCUG) Indications.  If the male baby is found (on US) to have significant bilateral pelvicaliceal dilatation (PCD; previously termed hydronephrosis), a dilated ureter, a pathological urethra or a thick-walled bladder, an MCUG is then performed. MCUG is also indicated for VUR assessment, such as after (recurrent or complicated) UTI, or for assessing complex malformations that involve the urinary tract. The MCUG is the only accepted method of lower urinary tract imaging to demonstrate the urethral anatomy clearly, although perineal US performed during voiding also allows assessment of the urethra. MCUG is essential in boys when a urethral lesion is suspected: for example, in boys with suspected PUV (Figs 72.8 and 72.9), cloacal anomaly, or anorectal anomaly and suspected colovesical fistula (Fig. 72.10). However, a fistula may not be seen on

Fluoroscopic view of renal fossae and bladder, initial + early filling Bladder filling with radio-opaque contrast medium gravity drip; bottle 30-40 cm above table, watch dripping, AB?

Fluoroscopy: signs of increased bladder pressure, imminent voiding, urge: bilateral oblique views of distal ureters, include catheter, document VUR, include kidney (spot film: intra-renal reflux)

When voiding: remove catheter, unless cyclic VCUG = 3 fillings, 1st y(s) female: 2 spot films of distended urethra (slightly oblique) male: 2–3 spot films during voiding (AP & steep oblique/lateral) include renal fossae during voiding, if VUR spot film

After voiding: AP view of bladder and renal fossae assess contrast drainage from kidney if refluxed Note: VUR staging, minimise fluoroscopy time and spot films; no control film Fig. 72.8  Voiding Cystourethrography. European Society of Paediatric Radiology imaging recommendations for voiding cystourethrography (VCUG). AB, Antibiotics; PCD, hydronephrosis (now pelvicalyceal dilatation); PUV, posterior urethral valve; UT, urinary tract; UTI, urinary tract infection; VUR, vesicoureteral reflux. (Adapted from Riccabona M, et al. Pediatr Radiol. 38, 138–145.)

1850

SECTION G  Paediatric Imaging

TABLE 72.1  Comparison of Relative Radiation Doses from Urological Examinations

AXR MCUG in girls MCUG in boys DIC DMSA MAG3 renogram MAG3 transplant DTPA transplant CT abdomen/pelvis

DRL (MBq)

Effective Dose (mSv)

Equivalent CXR (0.02 MSv)

N/A N/A N/A 20 MBq 80 MBq 100 MBq 200 MBq 330 MBq N/A

0.7 0.9 1.5 0.3 1.0 0.7 1 2 10

35 45 75 15 50 35 50 100 500

Equivalence to NBR (2.6 mSv/year)

Equivalence to a Return Transatlantic Flight (0.1 mSv)

3.3 months 4.2 months 6.9 months 1.4 months 4.6 months 3.3 months 4.6 months 9.2 months 3.85 years

7 9 15 3 10 7 10 20 100

AXR, Abdominal radiograph; CT, computed tomography; CXR, chest radiograph; DIC, direct radio-isotope cystogram; DMSA, dimercaptosuccinic acid; DRL, dose reference level; DTPA, diethylene triamine pentaacetic acid (Tc-99m DTPA); MBq, megabecquerel; MCUG, micturating cystourethrography; N/A, not applicable or available; NBR, national background radiation dose (approximate for UK).

A

B

Fig. 72.9  Posterior Urethral Valves on Micturating Cystourethrography. (A) There is acute calibre change in the posterior urethra caused by posterior urethral valves, although in this case the bladder is not trabeculated. Note bilateral high-grade vesicoureteral reflux. (B) Bilateral high-grade reflux is demonstrated.

an MCUG. A distal loopogram (i.e. intubating the distal limb of the colostomy and injecting radio-opaque water-soluble iodine-containing contrast under pressure) is the best technique to delineate these connections. Micturating cystourethrography technique.  A sterile narrow feeding tube is used to catheterise the neonatal urethra, secured with tape. A suprapubic catheter may be used in individual cases, or in-dwelling Foley catheter, provided that the balloon is carefully deflated in order to prevent bladder pathology from being obscured, or obstruction to bladder emptying increasing the chance of bladder rupture. Warmed water-soluble iodinated contrast is then dripped from a height of no greater than 60 cm (physiological filling pressure of 30 to 40 cm water). Rapid bladder filling using a syringe may generate high pressures leading to bladder over-distension and artificial VUR, as well as altered bladder capacity values. Bladder capacity increases during the first 8 years of life and normal bladder capacity for children 0 to 8 years can be estimated using the following equation:

Fig. 72.10  Colo-Urethral Fistula. On micturition in this patient with an anorectal malformation, contrast medium passed retrogradely into the colon via the fistula.

Bladder capacity = (age + 1) × 30 mL. Early bladder filling views are obtained with the child supine, with tight coning to reduce radiation dose. Oblique views are then obtained to assess the vesicoureteric junction and urethra on voiding. In the first few years of life, cyclical filling is advocated, as there is an increased chance of detecting VUR with consecutive voiding cycles. The bladder is refilled and on the second or third void, the catheter is removed so that a well-distended urethral view is obtained. Prophylactic oral or intravesical antibiotics are used, and MCUG is contraindicated in the presence of a UTI. In a modified MCUG, the contrast infusion is monitored for stopping or backflow in drip rate, which may represent dysfunctional sphincter or detrusor contractions, indicating functional disturbances.

Contrast-Enhanced Ultrasonography ce-VUS is a non-ionising alternative used throughout Europe and in the USA, but less commonly used in the UK. Recommended indications for ce-VUS presently include screening populations, in girls, bedside investigations and for follow-up. Study recommendations and VUR grading are available (Fig. 72.11). In experienced hands ce-VUS also

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

No diet restriction or enema, urine analysis… Accepted indications: VUR follow-up, girls, family screening, bedside Catheterisation: feeding tube, 4–8Fr, or suprapubic puncture anaesthetic lubricant or coated plaster Latex precaution: neural tube defect, bladder exstrophy

Standard US of bladder and kidneys (supine, ± prone) Bladder filling with normal saline (only from plastic containers)

US contrast medium, e.g., SonoVue, 0.2 to 1% of bladder volume, slow injection, no filters, US monitoring, potentially fractional administration

Peri-/post-contrast US of bladder and kidneys US techniques: fundamental, HI, CDS, dedicated contrast imaging alternate scans of right and left side during and after filling

1851

TABLE 72.2  Indications for 99m

Tc-Dimercaptosuccinic Acid Examination

• Assessment of differential renal function: • e.g. Assessment of functioning renal tissue when renal anatomical variants are encountered, such as duplex kidneys, horseshoe kidneys and cross-fused kidneys, as well as ectopic kidneys • Assessment for focal parenchymal defects, typically 4–6 months following urinary tract infection • Acute dimercaptosuccinic acid scans may be performed to confirm pyelonephritis • Assessment of which kidney to biopsy in bilateral disease • Assessment of function in children with cystic renal disease • Assessment of focal defects in a hypertensive child, particularly before catheter angiography • Assessment of a transplant kidney in those children with an unfavourable bladder who thus may be prone to reflux and silent infection • Assessment of differential renal function in children undergoing abdominal radiotherapy where the kidneys are in the therapy field

During and after voiding: US of bladder and kidneys supine ± prone, sitting or standing potentially one cycle for perineal US (Urethra!)

VUR diagnosis: echogenic microbubbles in ureters or renal pelves Fig. 72.11  Contrast-Enhanced Voiding Urosonography. European Society of Paediatric Radiology imaging recommendations for contrastenhanced voiding urosonography. CDS, Colour Doppler sonography; HI, harmonic index; US, ultrasound; VUR, vesicoureteral reflux. (Adapted from Riccabona M, et al., 2014. Pediatr Radiol. 44, 496–502.)

allows for assessment of the urethra during voiding using a perineal access. Technique.  Typically, a small quantity of an US microbubble contrast agent in a saline drip solution of (actual or estimated) bladder filling volume is given via urinary catheter as for MCUG, following standard renal tract US views. Dedicated low-mechanical index (MI) (low acoustic power) US of both kidneys and the bladder, including the retrovesical space during and after filling, and the urethra during voiding is then performed, with the demonstration of echogenic microbubbles in the ureters or renal pelvis indicating VUR. Intrarenal VUR can also be demonstrated with the appropriate technique, as well as fistulas between vagina and the urethra using a perineal access during voiding.

Nuclear Medicine

Direct Radio-Isotope Cystogram A direct radio-isotope cystogram (DIC) is predominantly used to detect VUR in baby girls, or as VUR follow-up in baby boys. Technique.  Catheterisation using a 6 F feeding tube can usually be performed with the baby lying on the gamma camera, wearing a double nappy. Twenty megabecquerels of 99mTc-pertechnetate is introduced into the bladder followed by warmed saline. The baby is restrained with sandbags and Velcro straps, and bladder filling is performed twice. The baby will spontaneously void when the bladder is full and VUR evaluated. The kidneys are kept in the field of view at all times to detect VUR.

Indirect Radio-Isotope Cystogram This is a useful, well-tolerated and physiological procedure to assess bladder function and for the presence of VUR. It is performed at the end of dynamic renography in cooperative and toilet-trained children who can void on demand. After a dynamic 99mTc-MAG3 renogram,

without frusemide, the gamma camera is turned vertically. The child is seated on a commode with their back to the camera. The acquisition is started just before voiding starts and continues for 30 s after voiding, up to approximately 2 minutes total acquisition time, if required. The study can be repeated if there is still tracer present in the bladder when bladder emptying is incomplete, or when refluxed tracer re-enters the bladder from the upper renal tract. Bladder dysfunction can be assessed and VUR can be seen on the dynamic study. Guidelines are available from the European Association of Nuclear Medicine.

Static Renal Scintigraphy Using 99m Tc-Dimercaptosuccinic Acid Dimercaptosuccinic acid (DMSA) is used as a 99mTc tracer; it is filtered by the glomeruli and reabsorbed, binding to the proximal convoluted tubules to give a static image over several hours. Approximately 10% of the tracer is excreted in the urine. Routinely, three posteriorly acquired views are obtained (posterior, right and left posterior oblique views), with anterior views used in abnormal renal anatomy (transplants, pelvic and ectopic kidneys) or scoliosis. Anterior and posterior views may then be used to estimate the differential renal function (DRF) by the geometric mean. As renal DMSA uptake relies on sufficient glomerular clearance, adequate renal function is essential for meaningful results, as well as sufficient urinary drainage from the renal pelvis to avoid tracer pooling artefacts. The main use of the 99mTc-DMSA scan is in the assessment of the DRF and cortical abnormalities, such as renal scarring, fusion defects, ectopic or duplex kidneys and in hypertension (e.g. Fig. 72.12). DRF may also be helpful for pre- and post-transplant assessment and in abdominal tumours, where the renal blood supply may be at risk, or where the kidney lies in the radiotherapy field. All indications are given in Table 72.2. There are no contraindications.

Dynamic Renography Dynamic renography is used to assess split renal function and drainage from the renal collecting systems in suspected obstruction. In Europe, the most common isotope used in paediatric dynamic imaging is 99m Tc-MAG3, which reflects tubular function. 99mTc-DTPA is also used, and uptake reflects glomerular function; thus it is more commonly used to calculate the glomerular filtration rate (GFR) and is the least expensive renal imaging agent capable of dynamic assessment. After

1852

SECTION G  Paediatric Imaging

Fig. 72.12  Horseshoe Kidney. 99mTc-dimercaptosuccinic acid scan of a horseshoe kidney showing fusion at the lower poles in the midline.

intravenous administration, about 50% of the 99mTc-MAG3 in the blood is extracted by the proximal tubules with each pass through the kidneys. The 99mTc-MAG3 is then secreted into the lumen of the tubule. As 99m Tc-DTPA is filtered by the glomerulus, with only 20% extraction fraction, it is not a good agent to use in neonates, children with impaired renal function or in the presence of significant obstruction. In 99mTc-MAG3 studies, if there is dilatation of a collecting system, a diuretic is used such as furosemide at a dose of 1 mg/kg (maximum dose 20 mg). Timing of diuretic administration varies widely, but may be given just after the tracer to try to prevent loss of venous access later in the study, in a distressed child. Time activity curves are generated which demonstrate uptake and excretion. Analogue images in the uptake phase may demonstrate renal scarring, albeit less clearly than on the 99m Tc-DMSA static renal scan. Split renal function can also be calculated— being even more accurate than static renography in a significantly obstructed kidney. The indications are given in Table 72.3. Where VUR is suspected, and the child is toilet-trained, cooperative and continent, an IRC may be performed. Technique.  The children are encouraged to drink plenty so that they are well hydrated upon arrival at the department (which is essential for meaningful results). The administered dose is scaled on a body surface area basis, with a maximum dose of 100 MBq of 99mTc-MAG3. The child empties the bladder, lies supine on the gamma camera face, distracted by television or a film, and is immobilised with sandbags. Ten- to 20-second frames are acquired for 20 minutes following isotope injection, imaging the heart, kidneys and bladder. The child then voids again before returning to the gamma camera for images following postural change and micturition, at about 40 minutes post injection.

TABLE 72.3  Indications for

Dynamic Renography

99m

Tc-MAG3

• To assess divided renal function and urinary drainage where the collecting system is significantly dilated • To assess differential renal function following surgery or procedure, for example post pyeloplasty, or post removal of a double J stent • Following renal transplant, to assess for urinary leak or possible obstruction

The DRF estimation is calculated from the renogram between 60 and 120 seconds from the peak of the vascular curve, expressed as a percentage of the sum total. The recommended methods to evaluate the DRF are the Patlak-Rutland plot and the integral method. Interpretation of the renogram in a dilated kidney, such as pelviureteric junction obstruction (PUJO) needs to be performed carefully. The shape of the time activity curve, response to furosemide and drainage of the collecting system following a change of posture and micturition are evaluated for significant stasis of tracer. There are four classic drainage patterns, described as normal (I), obstructed (II), dilated unobstructed (IIIa) and equivocal (IIIb). Ideally the renogram should be assessed with the US images available so that the degree of PCD and the renal parenchyma can be evaluated. Normal kidneys with DRF below 45% should be monitored using US. If dilatation increases, and a 99mTc-MAG3 demonstrates stepwise fall in function, pyeloplasty may prevent further deterioration in renal function.

1853

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

Urography (Plain Radiograph and Intravenous Urogram) With advances of sophisticated US techniques, and availability of crosssectional imaging, and widespread use of renal scintigraphy, the use of IVU has decreased with very few indications remaining (Table 72.4). Except for suspected urolithiasis, an initial ‘control’ full radiograph is rarely indicated. The radiation exposure should be minimised, using 1 to 3 properly timed and coned views (KUB film) based on the individual query (see also European Society of Paediatric Radiology—ESPR task force recommendations).

Computed Tomography CT urography (CTU) may be performed where specialised US or MR urography (MRU) is unavailable. Childhood conditions that may require imaging by CT include three main areas: calculi, tumours and trauma (Table 72.5).

Method Unenhanced CT has almost no role in paediatrics, except for calculi or calcifications, and these are typically best seen using US. For all other indications, standard non-ionic iodinated contrast medium is required at an age-adapted dose. Modern imaging techniques using automatic exposure control and low age-adapted kVp and size-based milliampere (mA)s settings will keep the dose to a diagnostic minimum, as will optimal age-dependent timing delays and avoiding multiphase acquisitions. In cases of abdominal trauma, topographic delayed imaging after 10 to 15 minutes (or a split bolus technique) can be useful in selected cases to detect contrast medium extravasation from the genitourinary tract. Conventional cystography or retrograde urethrography is preferred

in suspected urethral injury (see also ESPR recommendations). Fig. 72.13 shows a normal CT urogram.

Magnetic Resonance Imaging Anatomical MRI is now the imaging investigation of choice for abdominal and pelvic masses in children, as it gives excellent soft-tissue contrast resolution in any imaging plane, decreasing imaging times with modern equipment, improving tissue characterisation without the use of radiation. In most cases, MRI can replace CT, for example in the assessment of Wilms tumour, although thoracic CT is still be required to assess for pulmonary metastases. MRI can also give superior anatomical information to delineate pelvic anatomy, for example ambiguous genitalia in older children, or where spinal MRI is needed to evaluate tumour spread or suspected spinal cord abnormalities (neuropathic bladder). The most common use of MRI/ MRU is to evaluate complex urinary tract malformations and urinary tract obstruction, having practically replaced IVU, for example Fig. 72.14. MRA is particularly useful to delineate the major renal vessels in preoperative (tumour) imaging, complicated hypertension, or as pre- or post-transplant assessment. The indications are given in Table 72.6.

TABLE 72.4  (Rare) Indications for

Intravenous Urography in Children

• Suspected ureteral and severe renal trauma, only if computed tomography (CT) or MRI is not available • As a delayed kidney, ureter and bladder view after contrast-enhanced CT (avoiding a second CT) • In rare settings where CT is impossible, for instance in intensive care • Urolithiasis, where ultrasound is inconclusive and CT unavailble, or before lithotrypsy if necessary (with reduced number of films) • Distinct pelvicalyceal or ureteral pathology (e.g. calyceal diverticula, early stages of medullary sponge kidney) using an individually tailored single acquisition protocol

A

TABLE 72.5  Indications for computed

tomography of the Urinary Tract in Children • Diagnosis and follow-up of suspected malignant tumour, although MRI is now regarded as equally reliable in assessing abdominal masses • Major abdominal trauma with suspicion of serious pelvic fracture(s), bladder rupture or if an ureteral injury is suspected • Complicated infection, such as suspected abscess if magnetic resonance imaging (MRI) is unavailable and ultrasound (US) is inconclusive • Less common indications may include a suspicion of chronic renal infection or tuberculosis, or of nephrocalcinosis when a US study is inconclusive • Large calculi or xanthogranulomatous pyelonephritis (XPN) in an older child where surgery may be planned, and where MRI is not available or detailed stone definition is requested • CT angiography (CTA) may potentially replace conventional angiography in certain circumstances, discussed further in the section ‘Hypertension’

B Fig. 72.13  Computed Tomography Urography. (A) Computed tomography urography (coronal MIP) in a child with multiple infundibular stenoses secondary to abnormal anatomy and multiple instrumentations due to recurrent renal calculi. (B) VRT (volume rendered reconstruction) of abnormal right collecting system on computed tomography urogram.

1854

SECTION G  Paediatric Imaging

Fig. 72.14  Magnetic Resonance Urography. Heavily T2 weighted thick-slice coronal magnetic resonance imaging in a 3-year-old girl with right-sided duplex kidney which drains normally, but pelviureteric junction obstruction on the left.

TABLE 72.6  Indications for Magnetic

Resonance Imaging of the Renal Tract in Children • Diagnosis and follow-up of abdominal or pelvic mass/suspected tumour • In acute pyelonephritis and its complications, if ultrasound inconclusive • Where spinal imaging is required in children with suspected neuropathic bladder • Anatomical magnetic resonance urography for assessment of the anatomy in complex urogenital tract malformations • Functional magnetic resonance urography—upper urinary tract obstruction and renal dysplasia • Contrast-enhanced magnetic resonance angiography may have a role in the assessment of hypertension • Complex urogenital (or claocal) malformation to complete work-up if US is inconclusive or pre-operatively • Contrast enhanced MR urography after trauma (e.g. query urinoma)

Method Specific MR sequences and protocols are now available in most institutions for different clinical scenarios. Axial T1 and T2 weighted imaging is the mainstay of any imaging protocol, and either single-shot coronal heavily T2 weighted imaging or a 3D T2 sequence can give a ‘urogram’ overview of the entire urinary system (Fig. 72.15). Serial imaging following contrast application is useful to delineate the enhancement of tumours, or arterial enhancement for anatomy. MRU is the only imaging technique to give detailed functional as well as anatomical detail. Full MRU requires a dedicated protocol, including hydration, sedation, catheterisation and diuresis, with prolonged T1 gradient-echo dynamic sequences demonstrating contrast uptake, elimination and drainage similar to radionuclide renograms. Diffusion-weighted MRI (DWI) is rapidly evolving and can give an index of tumour cellularity (but does not differentiate benign from malignant tumours), identify metastases, assess acute pyelonephritis and evaluate tumour treatment response.

Fig. 72.15  Magnetic Resonance Urography. Three-dimensional T2 magnetic resonance urography image demonstrates a grossly dilated collecting system of the right kidney; cerebrospinal fluid (CSF) is also demonstrated in the background.

TABLE 72.7  Indications for Angiography

of the Renal Tract in Children

• Hypertension with a high suspicion of renovascular disease, including suspected vasculitis, especially polyarteritis nodosa • Renal vein sampling for renin values to evaluate which kidney is causing the hypertension • Before interventional procedures, for example embolisation for arteriovenous malformations or balloon dilatation for renal artery stenosis • Rarely in bilateral Wilms tumours before surgery • Testicular vein embolisation for varicocele obliteration

Techniques such as blood oxygen level-dependent (BOLD) MRI are being assessed to demonstrate changes in oxygenation of the kidney during acute obstructive episodes and in transplant imaging.

Interventional Procedures Angiography

Angiography is reserved for specific clinical situations (Table 72.7) and needs to be undertaken by an experienced operator. Selective arteriography with magnification and oblique views is necessary to detect lesions in small renal vessels.

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

Antegrade Pyelogram This investigation should be carried out by an experienced operator in the radiology department or in theatre, usually before surgery, to provide even better anatomical detail of the renal pelvis and/or ureter than that available from US or IVU. Occasionally antegrade studies are combined with pressure flow measurements with or without urodynamic studies to determine the physiological significance of a dilated upper urinary tract (the ‘Whitaker test’).

Nephrostomy The placement of a pigtail or J-J catheter in a dilated renal pelvis or ureter should be undertaken by an experienced operator under US guidance, with similar techniques to adults. US-guided needle placement into an appropriate lower pole calyx with injection of contrast agent gives delineation of the collecting system before insertion of the tube. The complications of placing a nephrostomy tube are generally those of catheter placement, extravasation of contrast medium, and leakage of urine; thus a combined sonographic-fluoroscopic approach is often recommended.

Retrograde Pyelogram The instillation of dilute contrast medium into a ureter via a catheter inserted into the distal ureteric orifice is usually undertaken by a urologist in the operating theatre. With modern flexible ureteroscopes, the contrast medium may be instilled into the upper ureter or even the renal pelvis, to outline the ureter and its drainage.

Renal Biopsy Many disorders affecting the kidney need a biopsy for histological confirmation or diagnosis, particularly glomerular disease, nephrotic syndrome and Immunoglobulin A (IgA) nephropathy. Surgical exploration (in situations where a percutaneous US-guided access is risky or contraindicated) or percutaneous US-guided needle biopsy should be considered, with complications such as subcapsular haematoma and arteriovenous (AV) fistula being fairly rare. Guidelines have recently been formulated for renal biopsy in children.

1855

initially forms from an interaction between the mesonephric duct/ ureteric bud and the metanephros at around 4 weeks of gestation. The primitive kidney ‘ascends’ by differential growth, rotating 90 degrees from horizontal to medial, taking its blood from the aorta and draining via the inferior vena cava. An ectopic kidney may result from excess, incomplete or abnormal ascent. Abnormalities of fusion may occur if the kidneys touch each other in the process of ascending.

Renal Ectopia Approximately 1 in 1000 kidneys is ectopic, and in 10% this is bilateral. The commonest is a pelvic kidney, which normally lies anterior to the sacrum just below the bifurcation of the aorta (Fig. 72.16). The true intrathoracic kidney (entering via the foramen of Bochdalek) is rare. Occasionally the kidney may be a superior ectopic kidney lying below a very thin membranous portion of the diaphragm. The adrenal glands are usually normally sited (but sometimes abnormally shaped) in the presence of renal ectopia, and there are many adrenal anomalies unrelated to renal variation.

Abnormalities With Renal Fusion The commonest renal fusion abnormality is the horseshoe kidney. The lower poles of the kidneys are fused in the midline, possibly due to malposition of the umbilical arteries causing the developing nephrogenic masses to come together. The isthmus of the horseshoe commonly lies anterior to the aorta and vena cava, at the level of the inferior mesenteric artery, with malrotated collecting systems lying anteriorly, which may lead to PUJO and associated infections or calculi. The abnormal axis of the lower poles of the kidneys in a horseshoe should not be missed on US: if bowel gas obscures the anterior view, the loss of the normal medial renal contour and graded compression is useful, and the kidney itself can be used as a window to visualise the parenchymal bridge. The anterior view of a 99mTc-DMSA may be helpful to assess functioning renal tissue in a horseshoe. MRI also demonstrates abnormalities of renal and (often associated) vascular anatomy and their complications. The horseshoe kidney may become more easily damaged in a road traffic accident because of the position of the lap belt in relation to the

CONGENITAL ANOMALIES SUMMARY BOX Congenital abnormalities—range from asymptomatic simple disorders to complex multisystem disorders with systemic manifestations. Renal anomalies may be an indicator of multisystem disease.

RENAL ANOMALIES Renal Agenesis Unilateral agenesis occurs in approximately 1 in 1250 live births. Antenatal diagnosis is uncommon, suggesting that agenesis may be the result of an involuted multicystic dysplastic kidney (MCDK). Most functioning kidneys should be identifiable by US wherever they are located, and either 99mTc-DMSA or MR can be used to detect poorly functioning kidneys. VUR is more common in a non-functioning kidney, and associated ipsilateral abnormalities, including uterine/seminal vesicle or gonadal abnormalities are common, and are easier to detect in the neonatal period before these structures involute.

Abnormal Migration and Fusion of the Kidneys Appreciating renal embryology will assist in the understanding of the various forms of renal malformation and vascular anomalies. The kidney

Fig. 72.16  Ectopic Kidney. 99mTc-dimercaptosuccinic acid scan of a pelvic ectopic kidney which demonstrates scarring from recurrent urinary tract infections.

1856

SECTION G  Paediatric Imaging

isthmus crossing the spine. Power Doppler US provides excellent assessment of minor trauma to a horseshoe kidney rather than resorting to CT. There is a slightly increased risk of Wilms tumour in horseshoe kidneys and the anomaly may be associated with Edwards syndrome (trisomy 18) and Turner syndrome.

Cross Fused Renal Ectopia Crossed fused renal ectopia is seen in 1 in 7000 post-mortem examinations. The crossed ectopic kidney lies on the opposite side to its ureteral insertion into the bladder. The left kidney is more commonly the ectopic kidney and the commonest pattern is fusion of the upper pole of the crossed kidney with the lower pole of the normally positioned kidney, in an L shape (Fig. 72.17). Occasionally the crossed kidney remains unfused. Imaging by US demonstrates a unilateral large mass of renal tissue, with contralateral absence of renal tissue. Because of abnormalities in rotation, PUJO is common and a 99mTc-MAG3 (or dynamic MRU)

study is often helpful in assessment of drainage. VUR is common. Associated anomalies within the VACTERL group may be seen (V vertebrae, A imperforate anus/atresia, C cardiac anomalies, TE stands for tracheoesophageal fistula, R renal anomalies, L limb anomalies).

Duplex Kidneys The commonest renal anomaly (2% of the population) is an uncomplicated duplex kidney which is considered a normal variant. Complete duplication is caused by two separate ureteral buds presenting onto the mesonephric duct. The ureter draining the lower moiety will come to lie more superior to and lateral to the ureter draining the upper moiety, increasing the VUR risk to the lower moiety—the latter constituting a ‘complicated duplication’ with potential disease implications. The upper moiety ureteric insertion into the bladder is usually more distal and medial, and may even be ectopic, or may be associated with an ureterocoele, causing varying degrees of dysplasia and obstruction in the upper moiety (Figs 72.18 and 72.19). Sometimes the upper (hypodysplastic) moiety is difficult to see or may mimic an upper pole renal cyst, and if the upper moiety ureter drains ectopically into the vagina in a girl, continuous wetting or recurrent UTIs may result. Careful interrogation by US, complemented by MRI, is essential in order to pick up both echogenic and atrophic upper moieties. Incomplete ureteric duplication (‘incomplete duplication’) describes ureteric duplication above the bladder. Yo-yo VUR may be demonstrated by dynamic renography with tracer passing down one ureter and back up the partially duplicated second ureter into the respective renal moiety, but the kidney in these children looks the same as in a complete duplication. A ‘septated renal pelvis’ is the mildest variant of this condition. An ureterocoele prolapsing into the vagina may present as a perineal mass in the newborn, or bladder outlet obstruction due to an ureterocoele prolapsing into the posterior urethra (mimicking PUV).

Imaging

Fig. 72.17  Crossed Fused Ectopia. 99mTc-dimercaptosuccinic acid scan of crossed-fused ectopia.

A

B

A duplex kidney may be diagnosed by US, with a larger than normal kidney with two distinct collecting systems being separated by a bridge of renal tissue. The appearances of a duplex kidney vary with the pathology of each moiety. The upper moiety is usually dilated, particularly when associated with a ureterocoele or ectopic ureteric insertion, or may be atrophic (Figs 72.20 and 72.21). (Power) Doppler can estimate the degree of dysplasia and may demonstrate the ectopic ureteric jet. The lower moiety may be dilated in PUJO, or demonstrate scarring and uroepithelial thickening in the collecting system to indicate VUR.

C

Fig. 72.18  Duplex Kidney. (A) Ultrasound shows grossly dilated upper moiety of the right kidney, with milder dilatation of the lower moiety. (B) Selected images from the MAG3 shows a duplex left kidney with upper moiety obstruction. (C) Micturating cystourethrography showing bilateral vesicoureteric reflux with a right-sided duplex (not the same patient).

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

1857

Fig. 72.19  Pelvicalyceal Dilatation. Ultrasound grading of pelvi-calyceal distension or dilatation with in neonates and infants. Modified from the European Society of Paediatric Radiology ‘Hydronephrosis’ grading (Riccabona et al. 2008)—based on Hofmann’s US grading, and the SFU classification (Fernbach et al. 1993).

Fig. 72.20  Ureterocoele. A large ureterocoele is identified on ultrasound, clearly visible on the early filling view of the micturating cystourethrography where contrast is seen circumferentially outside the ureterocoele within the bladder.

99m

Tc-MAG3 scintigraphy helps in assessing renal function, scarring and dysplasia, allowing the relative contribution to function from each moiety to be calculated. The uncomplicated duplex on a functional study may demonstrate up to 58% DRF compared with a simplex contralateral kidney, giving the false impression of reduced function

in the simplex kidney. The axis of the duplex kidney on a DMSA may point towards the ipsilateral shoulder rather than the contralateral shoulder on the image, prompting the radiologist to look carefully on US for the hallmark finding of two separate collecting systems. The 99m Tc-MAG3 renogram is also used to assess function and drainage in

1858

SECTION G  Paediatric Imaging Anatomical

is sufficient ureteral distension, dependent upon good hydration and bladder distension.

Anomalies of the Renal Pelvis and Ureter US may demonstrate calyceal diverticula or calyceal dilatation due to an infundibular stenosis, either congenital or acquired (secondary to infection or obstruction by calculi). MRI/MRU may further evaluate these appearances. Megacalycosis is a poorly understood condition where the dysplastic calyces are dilated in the absence of obstruction. It is probably due to underdevelopment of the medullary pyramids or maturation defects (resembling the shape of a pelvic kidney), and there may be an association with a megaureter and renal ectopia. Thus ureteric or pelvicaliceal dilatation does not indicate obstruction, and while the condition may predispose to stone formation, corrective surgery for ureteric dilatation is not beneficial.

Pelviureteric Junction Obstruction

Urographic 1

2 3 4

5

6

7

Fig. 72.21  Duplex Kidney. Diagrammatic representation of duplex kidneys with an ectopic ureterocoele of the left upper moiety without function. Diagnosis depends on recognition of indirect signs: 1 = increased distance from the top of the visualised collecting system to the upper border of the nephrogram; 2 = abnormal axis of the collecting system; 3 = impression upon the upper border of the renal pelvis; 4 = decreased number of calyces compared with the contralateral kidney; 5 = lateral displacement of the kidney and ureter; 6 = lateral course of the visualised ureter; and 7 = filling defect in the bladder.

a complicated duplex kidney, and an IRC can be used in the toilet-trained child to assess for lower moiety VUR. Correct interpretation and calculations by scintigraphy rely heavily on accurate anatomical knowledge from US or MRI. An MCUG can be useful both to assess for presence of possible ureterocoele and for VUR into the lower moiety. Anatomical MR sequences can delineate the upper and lower poles, and MRU can help distinguish between an obstructed and non-obstructed dilated system. High-resolution isotropic 3D data sets may be reconstructed to clearly demonstrate ectopic insertion of the ureters, provided there

PUJO is the commonest cause of renal tract dilatation, comprising up to 40% of cases. Antenatal and then postnatal US of the renal pelvis is useful in suspected PUJO or pseudo-obstruction. Poorly understood, there may be an anatomical abnormality, which can present early with antenatal unilateral PCD, or a crossing vessel causing extrinsic compression, which usually presents later with intermittent pain or infection. Abnormal ureteric peristalsis or VUR (with infections) may lead to ureteric kinking, fibrosis or there may be delayed recanalisation of the foetal ureter. Secondary PUJO may occur due to scarring after UTI and particularly with high-grade VUR, in elongated tortuous megaureters, as well as with obstructing tumours or retroperitoneal fibrosis.

Imaging Diagnosis of PUJO is important in order to highlight those severe cases which may progress, causing loss of renal function, which may require urgent surgical intervention. The initial US is performed approximately seven days after birth so that the baby is well hydrated and the degree of renal pelvic dilatation (RPD) is not underestimated. The prone transverse view of the renal pelvis is used to assess the antero-posterior (AP) pelvic diameter: between 7 and 10 mm on antenatal US requires follow-up in the newborn period, and greater than 1 cm thereafter; PCD/UTD can be graded accordingly. Function and drainage is assessed on the diuretic 99mTc-MAG3 renogram or functional magnetic resonance urogram (fMRU) (Fig. 72.22). Such studies are ideally performed when the baby is over 6 to 8 weeks of age, preferably over 3 months, when there is some renal maturity. If RPD is severe and bilateral, and the significantly thinned renal parenchyma is undifferentiated and echogenic, with reduced vascularity on power Doppler, earlier assessment of function is required. Diuretic sonography with serial measurements after furosemide application may help to differentiate dilated from (partially) obstructed kidneys that then require further imaging. In severe bilateral disease, 99mTc-DMSA may be used as an assessment of the relative contribution of each kidney to total renal function. The study must be performed with careful attention to detail. The baby must be well hydrated and furosemide should be given immediately or soon after the isotope, although timing protocols are varied. Following the dynamic renogram, a delayed image, following postural change and micturition (if the bladder emptied) should be performed to assess the contribution of gravity to drainage from the dilated upper urinary tract. The degree of urinary stasis cannot be assessed on the supine imaging alone, and a post-micturition/post-catheterisation view is essential. If there is reduced function in the affected kidney and the degree of RPD on US is changing, urologists may be more inclined to operate. Less than 25% of cases of antenatally diagnosed PUJO undergo surgery.

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

1859

B

A

Fig. 72.22  (A) Pelviureteric junction obstruction, on MAG3 renogram showing normal uptake and drainage of the left kidney with stasis of tracer in the right kidney at the level of the pelviureteric junction both in the supine position (20) and following micturition (M). (B) Time activity curve shows the abnormal right sided drainage

Megaureter and Hydroureter In utero, if the foetal ureter is visualised, then it is dilated: it may indicate a primary megaureter, refluxing megaureter, non-refluxing nonobstructive hydroureter or secondary hydroureter (e.g. associated with PUV and ureterocoele).

Imaging US is used to assess for a ureterocoele, or secondary signs of VUR may be evident, for example a laterally positioned or gaping ostium, bladder wall trabeculation and thickening, significant post-voiding residual, or M-mode assessment of peristalsis. Bilateral disease is assessed and extrinsic compression leading to secondary hydroureter is readily seen. An MCUG (or ce-VUS) is performed to assess whether ureteric dilatation is caused by VUR, either causing dilatation or coexisting with megaureter. MRU is used to assess anatomy in more complex situations (Fig. 72.23), as the dynamic sequences may demonstrate ureteric peristalsis, and to assess drainage and function.

Bladder Anomalies Bladder exstrophy–epispadias–cloacal exstrophy complex represents a spectrum of anomalies, with an incidence of around 1 in 20,000 live births (typically male). Absence of the normal bladder or the failure to see bladder filling on antenatal US may suggest bladder exstrophy. This results from a failure of closure of the abdominal wall during foetal development, leading to protrusion of the anterior wall of the bladder through the lower abdominal wall defect, and there may be an associated omphalocele. There is an open defect of the anterior abdominal wall or perineal wall and widening of the pubic symphysis, with epispadias in male babies (dorsal cleft in penis exposing the urethral mucosa). Antenatal diagnosis is difficult and the condition most often presents at birth with the exposed bladder. An estimation of the degree of severity of the condition can be found by measuring the extent of pubic symphysis diastasis on foetal US or plain radiographs (Fig. 72.24). This is normally

up to 10 mm, and 10 to 25 mm suggests epispadias with greater than 25-mm bladder exstrophy. The upper renal tracts are initially normal. Treatment involves staged repair of the bladder and genitalia. Following treatment, careful US and 99mTc-MAG3 is advised as VUR and obstruction are common following bladder closure.

Prune-Belly Syndrome The combination of absence/hypoplasia of the abdominal wall musculature, UTD and bilateral undescended testes is known as ‘prune-belly syndrome’, or abdominal musculature deficiency syndrome. The incidence of prune-belly syndrome is approximately 1 in 40,000 live births, again predominantly in male babies; female babies cannot have the complete triad, and the urological manifestations may be less severe. The entire spectrum of prune-belly syndrome is difficult to explain but thought to be either a defect of abdominal wall mesoderm formation early in embryogenesis, or severe bladder outlet obstruction leading to overdistension of the abdominal wall and urinary tract. Clinical presentation is with a lax abdominal wall with thin wrinkled skin and a protuberant abdomen. US may be challenging but will demonstrate dilated ureters with a large-capacity bladder often with little intrarenal dilatation, as well as absent or hypoplastic abdominal wall muscles. The bladder neck is wide with a dilated proximal posterior urethra and more distal conical narrowing, leading to a poor urinary stream through the anterior urethra (‘pseudo-valve’). Prognosis depends on the degree of associated renal dysplasia. These babies are prone to VUR and, consequently, infection. A MCUG is indicated to assess for VUR and the appearances of the urethra. Dynamic diuretic 99mTc-MAG3 renography may be helpful in long-term follow-up to assess function and drainage in those children without renal compromise.

Functional Bladder Disturbance and Neurogenic Bladder Children with neurogenic bladder have uncontrolled voiding and incomplete bladder emptying, due to inappropriate detrusor muscle

1860

SECTION G  Paediatric Imaging

Fig. 72.24  Bladder Exstrophy. Abdominal radiograph showing a radioopaque calculus in a patient with bladder exstrophy with characteristic diastasis of the pubic symphysis.

with IRC is helpful in the assessment of ongoing VUR. MCUG may demonstrate a small-volume trabeculated bladder with diverticula, and VUR into the kidneys; using a modified MCUG protocol or videourodynamic study will give additional functional bladder information. Non-neurogenic bladder and voiding disorders are usually treated by bladder training, often using biofeedback techniques, supported by medications.

URETHRAL ANOMALIES

Fig. 72.23  Megaureter. Three-dimensional T2 weighted magnetic resonance urography image/rotated reconstruction demonstrates a megaureter with an additional kink at the ureteropelvic junction and a markedly dilated collecting system of the kidney.

contraction and external sphincter relaxation, for example caused by spinal dysraphism, such as a myelomeningocele. Initial spinal US/MRI is essential to rule out spinal abnormality. When no spinal abnormality is found, the term ‘non-neurogenic neurogenic bladder’ is used. Nonneurogenic voiding disorders (or dysfunctional voiding) is becoming an increasingly important entity that is identified by modified MCUG or videourodynamics, but remains poorly recognised. Treating neurogenic bladders is aimed towards continence, and preventing deterioration in renal and bladder function. Kidney function deteriorates secondary to poor bladder emptying, leading to PCD and VUR, with subsequent complications of infection. Clean intermittent catheterisation, surgical procedures including bladder augmentation, continence procedures and artificial urinary sphincters and medication all have a role in preventing renal damage. Specialised urodynamic clinics are integral in the management of these patients. Regular US of the renal tract with catheterised bladder emptying is necessary to assess renal tract abnormalities. Follow-up dynamic 99mTc-MAG3 renography

The urethra is best demonstrated in the oblique lateral projection of the MCUG or by retrograde urethrography. This gives the best view of the bladder neck and posterior urethra, without which an anatomical cause of bladder outlet obstruction cannot be excluded (see Fig. 72.9). Bladder outlet obstruction may be anatomical or functional such as in neurogenic bladder. Anatomical causes of outlet obstruction include PUV, ureterocoele prolapsing into the posterior urethra causing obstruction, urethral dysplasia, anterior valve/diaphragm/syringocele, meatal stenosis, paraurethral cysts, urethral diverticula or duplication and post-traumatic or infective urethral strictures as well as (severe) hypospadias. Pelvic tumours or urethral polyps can cause obstruction, as can bladder masses such as haemangioma or neurofibroma. These lesions may be demonstrated by (perineal perimicturitional) US, but MRI will be the imaging of choice for assessment and staging of a pelvic mass.

Posterior Urethral Valves Congenital urethral obstruction creates a spectrum of disease, with the timing and severity determining the presenting symptoms. Boys with high-grade obstruction present as neonates with urosepsis, renal insufficiency and pulmonary hypoplasia with severe respiratory distress. They may be detected on antenatal US exhibiting urinary ascites, oligohydramnios, enlarged bladders and renal dilatation. Less severe obstruction may lead to presentation in childhood with UTI. PUV consist of abnormal folds of mucosa between the wall of the urethra and the verumontanum. There are three types, ranging from type I, slit-like orifice between 2 folds at the verumontanum, to type III, valve with eccentric pinpoint aperture causing the valve to balloon forward on micturition (‘wind in the sail’ appearance on the MCUG). Type III valve is associated with renal dysplasia, even without significant upper

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

A

1861

B Fig. 72.25  Normal Urethra. Micturating cystourethrography demonstrating full-length views of the urethra with a catheter in situ (A) and with the catheter removed (B).

tract dilatation. All newborn boys with significant antenatal PCD/UTD (grade III or higher) and thick bladder wall, ascites or oligohydramnion should be assessed for the presence of PUV. If valves are suspected, the bladder should be drained suprapubically until valves are ablated; with secondary obstruction at the VUJ, a nephrostomy may become necessary to ensure sufficient urinary drainage. Note that even with optimal treatment, the associated severe congenital renal dysplasia may still result in renal failure requiring subsequent renal transplantation.

Imaging In PUV, US may show a dilated upper urinary tract, renal dysplasia, or urinomas due to calyceal rupture secondary to increased pressure (the ‘pop off’ mechanism), and bladder wall thickening with a dilated posterior urethra and hypertrophied bladder neck. If the US is performed before the baby is well hydrated, the degree of UTD may be underestimated. MCUG is essential (but contraindicated in acute sepsis) and may demonstrate bladder wall trabeculation, a very large (or small) bladder with diverticula, and either unilateral or bilateral VUR, associated with ipsilateral poor renal function (although there are cases with no VUR and ‘only’ UVJO with secondary PUJO). The valve is visualised as an acute-calibre transition on the oblique/lateral urethral projection at micturition. If there is a urethral catheter in situ, a ‘catheter out’ view when micturating is essential in order to exclude a small valve leaflet compressed by the catheter (Fig. 72.25). Upper tract drainage by nephrostomy or ureterostomy may be required to preserve renal function. Treatment comprises early cystoscopic valve ablation followed by bladder drainage. A follow-up MCUG after valve ablation may be useful but cystoscopy is required, as remnant valve leaflets may be missed on urethrography. 99mTc-DMSA is initially used to provide DRF with 99mTc-MAG3 renography and US follow-up. Antireflux procedures are performed in PUV patients in order to preserve renal function. In boys with severe dysplasia, end-stage renal disease usually occurs in the first 20 years of life.

Anterior Urethral Abnormalities Anterior urethral valves are around 10 times less common than PUV. The anterior valves may be located anywhere along the anterior urethra,

Fig. 72.26  Syringocele. Urethral syringocele demonstrated on micturating cystourethrography.

and are often associated with a urethral diverticulum. Severe valves are often diagnosed in the newborn period, but mild valves may present later in childhood. Typical ballooning of the urethra with deviation of the penis may occur during micturition on MCUG. A syringocele is a dilated Cowper’s gland or duct that may cause urethral obstruction in the neonate (Figs 72.26 and 72.27). Sometimes non-specific symptoms of dribbling or haematuria may occur later in childhood. Spontaneous rupture or transurethral incision is curative, but may lead to a urethral stricture.

1862

SECTION G  Paediatric Imaging

Fig. 72.28  Urethral Duplication. There is duplication of the urethra almost from the bladder neck along the whole length of the penis.

Fig. 72.27  Scaphoid Congenital Megaurethra. Massive dilatation of the urethra is caused by non-development of the penile erectile tissue, particularly the corpus spongiosum.

Urethral Stricture Strictures may be congenital or acquired and may occur in boys or girls. Seventy-five per cent of male congenital urethral strictures occur in the bulbous urethra (where the embryological proximal urethra merges with the urogenital membrane), and at the urethral meatus in girls. Children with congenital strictures present either as neonates or in the postpubertal period. Early presentation as a neonate may demonstrate significant upper tract dilatation secondary to urethral obstruction. Postpubertal boys present with irritation, urinary dribbling, haematuria and UTI as well as prostatitis or epididymitis. Diagnosis is by MCUG and treatment is usually urethrotomy. Traumatic straddle injuries lead to bulbar strictures. Pelvic trauma and urethral rupture, or catheter induced injuries, tend to cause strictures either at the bladder neck or at the membranous portion of the urethra.

Rectourethral Fistula Rectourethral fistula is usually associated with an imperforate anus and is best demonstrated by a high-pressure loopogram via a distal colostomy which should demonstrate passage of contrast from the bowel to the posterior urethra, or by high-pressure urethrogram during voiding on a MCUG.

Duplication of the Urethra Duplication of the urethra is a rare congenital abnormality usually discovered in childhood with one of the urethras ending as a hypospadias. Duplication of the urethra may be complete and the child presents with a double urinary stream, UTI, sometimes outflow obstruction and incontinence. The classification of urethral duplication is complex and beyond the scope of this chapter (Fig. 72.28).

UTERUS AND VAGINA Differentiation of the gonads into ovaries or testes depends on the presence or absence of the Y chromosome. In the absence of hormonal secretion of anti-Müllerian hormone from the fetal testis, the Müllerian ducts meet and differentiate into the uterus, cervix, fallopian tubes and

proximal two-thirds of the vagina. The urogenital sinus forms the distal vagina. Uterine development depends on the formation of the Wolffian or mesonephric duct. Uterovaginal malformations are classified embryologically into either Müllerian agenesis, a developmental defect of the caudal portion of the Müllerian ducts (Mayer–Rokitansky–Küster–Hauser (MRKH) syndrome); disorders of lateral fusion caused by the failure of the two Müllerian ducts to fuse in the midline; and disorders of vertical fusion that are caused by abnormal union between the Müllerian tubercle and urogenital sinus derivatives (leading to disorders of the hymen, cervical agenesis and transverse vaginal septa). Disorders of lateral fusion are heterogeneous and have been classified into six groups. Lateral fusion and vertical fusion abnormalities often coexist and thus congenital vaginal abnormalities can be considered as those with or without obstruction. Anomalies of the uterus and vagina may present as an abdominal or pelvic mass in the neonatal period. Female genital anomalies are usually suspected on US and diagnosed using MRI, or US genitography using intravaginal saline in centres with experience. Fifty per cent of cases will have an associated renal anomaly and approximately 12% may have associated vertebral segmentation anomalies. Congenital anomalies of the vagina (septation, stenosis, imperforate hymen) often present in pubertal girls with menstrual symptoms but no bleeding due to obstruction (haematometrocolpos; Figs 72.29 and 72.30). Other causes of primary amenorrhoea include Turner’s syndrome. MRKH syndrome, which affects 1 in 4000 to 5000 girls, and includes vaginal atresia, uterine anomalies and malformations of the upper urinary tract. Associated renal abnormalities include PCD, dysplasia, unilateral ectopia and renal agenesis. Three-dimensional US and MRI are helpful in imaging the spectrum: all girls with genital anomalies or adreno-genital syndrome should have careful US evaluation of the renal tract and adrenal glands.

UNDESCENDED TESTIS Cryptorchidism refers to the absence of a testis in the scrotum, affecting 4% of full-term newborns and 30% of preterm newborns, falling to around 0.8% after the first year. It is usually right-sided, but may be bilateral. During embryogenesis, the testes form beside the mesonephric kidneys and descend via the inguinal canal to the scrotum. This normal process may halt anywhere along its descent, causing an undescended testis, or the testis may become ectopic or absent. Early diagnosis and

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children treatment are important to prevent infertility and a risk of malignancy in the undescended testis. Unilateral testicular agenesis is associated with ipsilateral renal agenesis. There is current debate regarding whether US can confidently localise/ exclude undescended testes. US has estimated sensitivity and specificity of around 45 and 80%, respectively, in accurately localising non-palpable testes. US is particularly helpful for testicular locations in the inguinal canal or next to the bladder/close to the abdominal wall—deeper positions are more difficult. US can also detect other scrotal pathology, such as hydrocele or cystic dysplasia of the rete testis, which mimics a testicular tumour and is associated with a MCDK. MRI is far superior in localising near-normal, non-palpable testes, and is preferable in ambiguous genitalia or hypospadias, but small and dysplastic testes may be indistinguishable from non-specific nodules. The testis is typically hypoplastic with low T2 signal. Diagnostic laparoscopy is the definitive investigation and allows for concurrent biopsy or surgical correction—thus the use of testicular US for localisation is questionable, with little demonstrable impact on patient management or treatment.

1863

ANTENATAL DIAGNOSIS OF PELVICALYCEAL AND/OR URETERAL DISTENTION Some anatomical renal abnormalities can now be detected antenatally, which begs the question as to what should be done for the child in the immediate postnatal period? The most obvious abnormalities are anatomical disorders, such as renal agenesis and horseshoe kidneys. The next issue is that of PCD and/or ureteric dilatation. Imaging should try to distinguish between an obstructed and non-obstructed system; the goal of imaging is to direct treatment to preserve renal function and growth potential. There are many imaging algorithms now developed by the European paediatric radiology community, which have been widely publicised.

Prenatal Diagnosis of Congenital Abnormalities of the Renal/Urological Tract (CAKUT) and Differential Diagnosis Newborns with a prenatal US diagnosis of a renal tract abnormality do not form a homogeneous group. Transient RPD is thought to be physiological during fetal development, but persistent or enlarging RPD during the antenatal period will typically be referred for postnatal investigation, although most will be determined to be normal. The differential diagnosis is wide (Table 72.8). Bilateral disease must be distinguished from unilateral disease, as unilateral normal kidney and ureter implies that normal renal function can be achieved. The most important diagnosis to make immediately (US and MCUG within the first 24 hours of life) is that of PUV, potentially causing bilateral obstructive PCD and renal damage, so that perinatal surgery can be performed for this indication. For those children with less marked abnormalities, repeating US at 1 and 4 to 6 weeks of life can often eliminate those in whom the functionally immature system has improved, and no further imaging is required (Fig. 72.31). The urgency of postnatal imaging is heavily dependent on the prenatal US findings and the quality/availability of prenatal US.

Bilateral Renal Pelvic Dilatation

Fig. 72.29  Haematocolpos. Sagittal ultrasound image of thickened endometrium with spill of blood into the obstructed, distended vagina lying behind the normal bladder.

Renal PCD, former called ‘pelvicalyceal dilatation (PCD)’, may be defined as calyceal dilatation plus a renal pelvis of greater than 10 to 15 mm in its AP diameter with no US evidence of a dilated ureter. This was referred to previously as PUJO or stenosis. The measurement is gestation dependent and equates to approximately 5-mm RPD at 20 weeks’

Fig. 72.30  Haematometrocolpos. Only one, that is the right of two uterine cavities in this didelphys uterus, demonstrates haematometrocolpos on coronal and axial magnetic resonance imaging.

1864

SECTION G  Paediatric Imaging

TABLE 72.8  Differential Diagnosis of

Prenatal US: mild or moderate dilatation = UTD II + III

Prenatal Urinary Tract Dilatation or Pelvicaliceal Distention

US: 1st US around day 5

Unilateral Pathology Renal pelvicalyceal dilatation (PCD) Vesicoureteric reflux (VUR) Dilated ureter (with or without reflux) Multicystic dysplastic kidney and other cystic kidney disease Complicated duplex kidney Upper moiety dilatation—either ureterocoele or ectopic drainage Lower moiety dilatation—usually VUR, but rarely PCD (even secondary to pelviureteric junction obstruction (PUJO)) only

Abnormal: pelvis ≥7 mm + dilated calyces, or other anomalies (PCD > IIº)*1

US at 1 month

MCUG*3

Bilateral Pathology Bilateral PCD Bilateral VUR Bilateral dilated ureter (with or without reflux) Bladder pathology (e.g. neurogenic bladder) Bladder outlet pathology (e.g. posterior urethral valves) Bilateral complicated duplex kidneys Multicystic kidney on one side and cystic dysplastic kidney on opposite side Cystic kidney disease

gestation and a 10-mm pelvis in the third trimester. PCD is commonly unilateral but may be bilateral, in which case investigation is essential. PCD/UTD can be graded according to severity. High-grade or bilateral PCD is frequently associated with severe abnormality, that is severe obstructive uropathy and PUV, which may deteriorate rapidly, and need urgent US and MCUG (Fig. 72.32). The main question in severe PCD is whether the child needs early bladder drainage and intervention, after which further imaging work-up may be delayed until physiological maturity of the kidneys occurs. In mild-to-moderate foetal PCD, initial postnatal US is best postponed at least until after 1 week of age; these US findings should then dictate additional imaging. Treatment and further imaging should then be performed according to the severity of the initial findings (see Fig. 72.1). In children with suspected PUJO or ureterovesical junction (VUJ) obstruction (e.g. primary obstructive megaureter, obstructive ureterocoele) a more sophisticated imaging algorithm is proposed (Fig. 72.33), with US findings determining subsequent imaging. MCUG is recommended to differentiate obstructive and refluxing dilatation, and MAG3 renography (increasingly dynamic functional MRU) to assess function and urinary drainage. The exact timing of follow-up will depend on the size of the PCD and DRF: the functionality is much more important than a simple anatomical abnormality which may have no functional consequence. The results should be used as guidance, as satisfactory grading may be difficult, immaturity of renal function may improve, and there is no convincing benefit of early surgical release of obstruction to improve renal function. Those children with functional impairment clearly require closer follow-up than those with an isolated mild PCD alone. Long-term follow-up is recommended until at least 15 to 20 years of age. The largest group of children seen with a prenatal diagnosis of PCD will have a normal postnatal US, or only mild residual dilatation. Those with persistent significant dilatation require functional and VUR assessment, using both radionuclide investigation of antegrade function and an MCUG, particularly in boys. However, the use of MCUG is falling out of favour, due to the lack of evidence that low-grade VUR has any consequence on the child’s outcome or prevention of renal scarring later in life, and the high incidence of VUR resolving spontaneously. It

Normal (≤ PCD IIº)

Normal

US at 3 months

Abnormal

Normal Abnormal pelvis ≥10 mm, PCD >II° other malformation*2 ‘extended criteria’*1

Pelvis 10 mm, PCD > IIº

Stop follow-up

Further morphological and functional evaluation: Scintigraphy, IVU, MRU

Stop follow-up

*1

use extended US criteria considering urothelial sign, kidney size & structure, etc. US genitography: in all patients with single kidney. MCDK, ectopic kidneys. etc. *3 ce-VUS can be used in girls and for screening populations *2

Fig. 72.31  Pelvicalyceal Dilatation. European Society of Paediatric Radiology imaging algorithm for mild or moderate urinary tract dilatation (UTD). HN, Hydronephrosis; IVU, intravenous urography; MCDK, multicystic dysplastic kidney; MCUG, micturating cystourethrography; MRU, magnetic resonance urography; US, ultrasound. (Adapted from Riccabona M, et al., 2008. Pediatr Radiol. 38, 138–145, update 2017.)

is important to differentiate these children from the group who present later with UTI, on whom much of the data on VUR and renal scarring are acquired. Both VUR and PCD may simply be a marker of unilateral renal dysplasia in these children, which results in abnormal renal function and scarring later in life. The use of prophylactic antibiotics in children with unilateral abnormalities detected antenatally is not evidence based, and it is becoming increasingly difficult to justify invasive imaging without being able to offer therapeutic treatment. There is ongoing uncertainty about the management of bilateral PCD. The investigative protocol in this situation should be as for the unilateral PCD, but should include an MCUG as well as formal sequential glomerular GFR estimation.

Unilateral Renal Pelvic Dilatation The natural history of unilateral PCD is a relatively benign condition which frequently resolves spontaneously. True obstruction will typically lead to parenchymal compression and eventually atrophy. If the child has any of these features, assessment of renal function is imperative, although these changes do not predict progressive renal deterioration.

CHAPTER 72  Imaging of the Kidneys, Urinary Tract and Pelvis in Children

1865

Prenatal US: gross dilatation = UTD ≥IV MCUG in all boys, particularly if uterer dilated or thick bladder wall ce-VUS in girls, potentially delayed

+ Narrowed or dysplastic parenchyma, dilated ureter particularly if bilateral or single kidney*1

Early US + MCUG (ce-VUS)

PUV

Drainage? Renal function? + isotopes*2, MRU*3

High-grade VUR

Obstructive uropathy

Others*4

US follow-up 6 months: isotopes*2, MRU*3?

PUJO, MRU*5

As indicated*5

*1 (US) genitography: in patients with single kidney MC1DK, ectopic kidney, suspected genital anomaly *2 MAG3-better than DMSA in dilated systems and neonates, DMSA usually after 3–6 months, not before 6 weeks;

+ open bladder catheter to avoid VUR-induced errors

*3 MRU-complex anatomy, function, obstructive component, etc. *4 e.g. MCDK, cystic dysplasia, duplex or horseshoe kidney, other malformations, non-obstructive HN, cysts/cystic tumour, etc. *5 see relevant algorithm

Fig. 72.32  Pelvicalyceal Distention (PCD). European Society of Paediatric Radiology imaging algorithm for postnatally diagnosed severe/high-grade foetal hydronephrosis. ce-VUS, Contrast-enhanced voiding urosonography; HN, hydronephrosis; IV, intravenous; MCUG, micturating cystourethrography; MRU, magnetic resonance urography; PUJO, pelviureteric junction obstruction; PUV, posterior urethral valves; US, ultrasound; UTD, urinary tract dilatation; VUR, vesicoureteric reflux. (Adapted from Riccabona M, et al., 2009. Pediatr Radiol. 39, 891–898.)

There is no test that will predict which kidney with a prenatal PCD will deteriorate. US assessment within the first 24 hours may underestimate PCD due to neonatal dehydration and renal immaturity, and thus US at around 7 days of life should be used to more reliably assess calyceal and pelvic dilatation, measure the transverse diameter of the pelvis and the calyces, assess renal parenchyma and renal volume, and confirm the structural normality of the bladder and opposite kidney.

Megaureter PCD with a dilated ureter is dealt with in the same way as pelvic dilatation above, but suggests that VUR is more likely. The imaging protocol therefore is similar to that of PCD plus a MCUG (ce-VUS) to document or exclude VUR. Careful US of the bladder and VUJ must be performed to exclude a small ureterocoele or other bladder abnormality, and ureteric peristalsis. Reimplantation of the ureter into the bladder in a child younger than 1 year of age may result in abnormal bladder function in later life; urodynamics may not be able to clarify this further. In the neonate, a spinal US is useful to exclude spinal cord problems when a neurogenic bladder may be suspected. Stenting of the ureter or temporary diversion by uretero-cutaneostomy may be necessary to preserve renal function, assessed by diuretic 99mTc-MAG3 drainage studies.

Renal Failure Acute renal failure in the newborn is a common problem. It may have an antenatal onset, in congenital disease such as renal dysplasia and genetic disorders such as ARPCKD, or hypotensive or hypoxic event at or around delivery, resulting in renal vein thrombosis (RVT), medullary or acute tubular necrosis (ATN), or any combination of these conditions. US with Doppler is the first investigation, as all of these disorders result in an echogenic kidney with loss of the normal corticomedullary differentiation. The main task of US is to differentiate between prerenal, intrinsic or postrenal cause of the renal insufficiency.

Renal Vein Thrombosis ATN is typically symmetrical and RVT typically asymmetrical, but can be bilateral. In RVT, swollen echogenic kidneys with prominent interlobular arteries are demonstrated by US. Often no venous flow can be identified at the renal hilum, and echogenic linear streaks of venous thrombi may be seen in the periphery. Seventy-five per cent of RVT is unilateral, starting in the periphery: power Doppler is helpful to depict early stages, around 50% will have inferior vena cava (IVC) involvement, and 10% associated adrenal haemorrhage. Doppler US spectral flow display will show the typical high resistance flow profile with reverse diastolic flow in the affected renal artery. Contrast-enhanced imaging examinations are rarely used in this context; follow-up US (split renal volume assessment) with 99mTc-DMSA scintigraphy (4 to 12 months of age) should be performed to assess the long-term effects.

URINARY TRACT INFECTION AND VESICOURETERIC REFLUX UTI is a common (bacterial) infection causing illness in infants and children, often with non-specific symptoms. The long-term goal of imaging in UTI is to preserve renal function and growth potential: this is often misinterpreted as (a) identifying VUR on imaging, and (b) minimising future UTIs, although there is limited evidence that these are synonymous. Many children investigated will be normal; thus, an effort must be made to only image those at high potential clinical risk, and to keep radiation doses to a minimum. Even in those imaged, there is limited evidence to suggest that treatment affects the natural history of renal damage. There is continuing controversy over the significance of VUR in the setting of UTI. In particular, neonatal VUR is now thought to be a transitory condition that, in the majority, diminishes or disappears

1866

SECTION G  Paediatric Imaging

US (+ DDS/CDS)

PCD ≥grade III

Mild (PCD 3-6 months)

Equivocal also, if 180–200 cm/s (note age variations) •RAR >3.5, RI >0.05 •Acceleration time >80 ms •Tardus-parvus pattern

Diagnosis evident

Renal artery stenosis

Other

DSA with simultaneous PTA Consider renal vein sampling

Further Imaging As appropriate2

Inconclusive, or no sign of renovascular disease

Stage 1 hypertension or BP well controlled on 1–2 drugs

Stage 2 hypertension BP not controlled by ≥2 drugs, stage 1 age short trunk Narrow thorax with respiratory distress in infancy Bowed legs Prominent forehead with depressed nasal bridge Hydrocephalus and brainstem and spinal cord compression Inheritance: AD Gene: FGFR3

Hypochondroplasia (see Fig. 73.3)

Variable short stature Prominent forehead Milder phenotype than achondroplasia. Inheritance: AD Gene: FGFR3

Group 2 (Type 2 Collagen Group) Spondyloepiphyseal dysplasia Short stature with short trunk at birth congenita (see Fig. 73.4) Cleft palate Myopia Maxillary hypoplasia Thoracic kyphosis and lumbar lordosis Barrel-shaped chest Inheritance: AD Gene: COL2A1

Group 8 (TRPV4 Group) Metatropic dysplasia (see Fig. 73.5)

Spondylometaphyseal dysplasia, Kozlowski type (see Fig. 73.6)

Short limbs Relatively narrow chest Small appendage in coccygeal region (‘tail’) Progressive kyphoscoliosis Progressive change from relatively short limbs to relatively short trunk (hence ‘metatropic’) Inheritance: AD Gene: TRPV4 Short trunk, progressive scoliosis. Inheritance: AD Gene: TRPV4

Radiological Features Short ribs Severe platyspondyly Trident acetabula Irregular metaphyses Short broad tubular bones of the hands and feet Small scapulae Type 1—Normal skull, marked shortness and bowing of the long bones (‘telephone receiver femora’) Type 2—‘Clover leaf’ skull. Limb shortening milder. ‘Bullet-shaped’ vertebral bodies with short pedicles Decrease of the interpedicular distance of lumbar spine caudally, and posterior vertebral body scalloping (in older child and adult) Squared iliac wings with small sciatic notch Flat acetabular roofs Short ribs Short wide tubular bones Large skull vault Small foramen magnum V-shaped notches in growth plates (‘chevron deformity’) ‘Trident’ hands Absence of normal widening of the interpedicular distance of the lumbar spine caudally Short, relatively broad long bones, variable brachydactyly Elongation of the distal fibula and of the ulnar styloid process

Oval, ‘pear-shaped’ vertebral bodies. Anisospondyly; i.e., L5 vertebral body smaller than L1 in infancy Odontoid hypoplasia and cervical spine instability Short long bones Delayed ossification of epiphyses of knees, shoulders, hindfoot, pubic and ischial bones Severe coxa vara developing in early childhood Horizontal acetabulum Delayed ossification of pubic rami Relative sparing of hands Short long bones with marked metaphyseal flaring (‘dumb-bell’) Platyspondyly with wide intervertebral spaces Flat acetabular roofs Short iliac bones Short ribs with anterior widening Progressive kyphoscoliosis Hypoplastic odontoid peg Irregular platyspondyly with anterior wedging ‘Overfaced’ pedicles Metaphyseal irregularity (especially of proximal femora) Coxa vara Continued

1888

SECTION G  Paediatric Imaging

TABLE 73.1  Clinical and Radiographic Features of Selected Osteochondrodysplasias and

Dysostoses—cont’d

Clinical Features Group 9 (Ciliopathies With Major Skeletal Involvement) Ellis–van Creveld (see Short stature Fig. 73.7) Short limbs, more marked distally Polydactyly Hypoplasia of the nails and teeth Sparse hair Congenital cardiac defects (ASD, single atrium) Fusion of upper lip and gum Inheritance: AR Gene: EVC1, EVC2 Asphyxiating thoracic Often lethal dysplasia—Jeune (see Respiratory problems with long narrow thorax Fig. 73.8) Short hands and feet Nephronophthisis in later life in survivors Inheritance: AR Gene: IFT80, DYNC2H1, WDR34, TTC21B, WDR19, IFT172, IFT140

Radiological Features Short ribs (in infancy) Short iliac wings; horizontal ‘trident’ acetabula (pelvis becomes more normal in childhood) Premature ossification of proximal femoral epiphyses Laterally sloping proximal tibial and humeral epiphyses Polysyndactyly; carpal fusions (90% cases) Cone-shaped epiphyses of middle phalanges

Small thorax with short ribs, horizontally orientated High clavicles Short iliac bones Trident acetabula Premature ossification of proximal femoral epiphyses Cone-shaped epiphyses of phalanges Polydactyly (10% cases) Cystic renal disease

Group 10 (Multiple Epiphyseal Dysplasia and Pseudoachondroplasia Group) Platyspondyly with ‘tongue-like’ anterior protrusion of the Pseudoachondroplasia (see Short limbs with normal head and face vertebral bodies Fig. 73.9) Accentuated lumbar lordosis Biconvex upper and lower vertebral end plates Genu valgum or varum Atlantoaxial dislocation Joint hypermobility Small proximal femoral epiphyses with cartilage overgrowth Inheritance: AD Wide triradiate cartilage Gene: COMP Small pubis and ischium Pointed bases of the metacarpals Short tubular bones with expanded, markedly irregular metaphyses, small irregular epiphyses with delayed bone age Relatively long distal fibula Multiple epiphyseal dysplasia Joint stiffness ± limp Small, irregular epiphyses (see Figs 73.10 and 73.11) Early osteoarthritis Delayed ossification Mild limb shortening Delayed bone age Inheritance: AD Short tubular bones of the hands and feet Gene: COMP, MATN3, COL11, COL9A1, COL9A2, COL9A3 Mild acetabular hypoplasia Early osteoarthritis a Note, multilayered patella only seen in autosomal recessive MED due to mutations in the DTDST gene (Group 4, sulphation disorders) Group 11 (Metaphyseal Dysplasias) Metaphyseal dysplasia, Short limbs, short stature presenting in early childhood Schmid type (see Fig. 73.12) Waddling gait Genu varum Inheritance: AD Gene: COL10A1 Cartilage-hair hypoplasia (metaphyseal dysplasia, McKusick type) (see Fig. 73.13)

Short stature, sparse hair, immunodeficiency, increased malignancy risk Inheritance: AR Gene: RMRP

Group 12 (Spondylometaphyseal Dysplasias) Spondylometaphyseal Inheritance: AR, AD dysplasia, Sutcliffe type Gene: Some cases are linked to Col2A1 (see Fig. 73.14)

Metaphyseal flaring irregularity and increased density— differential for rickets Most marked at hips Large proximal femoral epiphyses Coxa vara; femoral bowing Normal spine Metaphyseal irregularity. Brachydactyly with delta epiphyses

Corner fracture-type appearance of metaphyses: a differential for suspected physical abuse Platyspondyly with oval vertebral bodies and anterior beaking

1889

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

TABLE 73.1  Clinical and Radiographic Features of Selected Osteochondrodysplasias and

Dysostoses—cont’d

Clinical Features Group 18 (Campomelic Dysplasia and Related Disorders) Campomelic dysplasia (see Neonatal Fig. 73.15) Respiratory distress Cleft palate Prenatal onset of bowed lower limbs Pretibial dimpling Survivors Short stature, scoliosis Learning difficulties Inheritance: AD (sex reversal) Gene: SOX9 Group 21 (Chondrodysplasia (CDP) Group) Chondrodysplasia punctata Phenotype dependent on genetics (see Fig. 73.16) Flat nasal bridge, high arched palate Cutaneous lesions, e.g. ichthyosis Asymmetrical or symmetrical shortening of long bones Joint contractures Cataracts Inheritance: XLD, XLR, AR, AD Gene: XLD—EPP, NHDSL XLR—ARSE AR—LBR, AGPS, DHPAT, PEX2 AD—Unknown (also some AR types) Group 22 (Neonatal Osteosclerotic Dysplasias) Caffey disease (infantile Usually present in the first 5 months of life cortical hyperostosis) Hyperirritability Soft-tissue swelling Inheritance: AD, AR Gene: AD—COL1A1 AR—Unknown Group 23 (Osteopetrosis and Related Disorders) Osteopetrosis (see Fig. 73.17) Several types Enlargement of liver and spleen Bone fragility with fractures Cranial nerve palsies Blindness Osteomyelitis Anaemia Inheritance: Severe types—AR Milder/delayed types—AD Gene: AR—TCIRG1, CLCN7, RANK, RANKL, CAII AD—LRP5, CLCN7 Pyknodysostosis (see Short limbs with a propensity to fracture Fig. 73.18) Respiratory problems Irregular dentition Inheritance: AR Gene: CTSK Osteopoikilosis (see Fig. 73.19)

Often asymptomatic May be associated with skin nodules (Buschke–Ollendorff syndrome) Inheritance: AD Gene: LEMD3

Radiological Features 11 pairs of ribs Hypoplastic scapulae Bowing of femora and tibiae Short fibulae Progressive kyphoscoliosis Dislocated hips Deficient ossification of the ischium and pubis Hypoplastic patellae

Stippled calcification in cartilage, particularly around joints and in laryngeal and tracheal cartilages. Disappears later on in life Shortening, which may be asymmetrical, of the long bones and/ or digits Coronal cleft vertebral bodies Punctate calcification is also seen in some chromosomal disorders, fetal alcohol syndrome, neonates of mothers with autoimmune disorders, warfarin embryopathy

Commonly affects mandible, clavicle, ulna May be asymmetrical Periosteal new bone and cortical thickening Abnormality limited to diaphyses of tubular bones

Generalised increase in bone density Abnormal modelling of the metaphyses, which are wide with alternating bands of radiolucency and sclerosis ‘Bone-within-bone’ appearance Rickets Basal ganglia calcification (in the recessive form)

Multiple Wormian bones, delayed closure of fontanelles Generalised increase in bone density Straight mandible Deficient ossification of terminal phalanges Resorption of lateral clavicles Pathological fractures Sclerotic foci/bone islands, particularly around pelvis and metaphyses

Continued

1890

SECTION G  Paediatric Imaging

TABLE 73.1  Clinical and Radiographic Features of Selected Osteochondrodysplasias and

Dysostoses—cont’d Melorheostosis (see Fig. 73.20)

Clinical Features

Radiological Features

Asymmetry of affected limbs Sclerodermatous skin lesions over affected bones Vascular anomalies Muscle wasting and contractures Inheritance: Sporadic

Dense cortical hyperostosis of affected bones with ‘dripping candle wax’ appearance Long bones most commonly affected

Group 24 (Other Sclerosing Bone Disorders) Diaphyseal dysplasia Muscle weakness (Camurati–Englemann Pain in the extremities disease) (see Fig. 73.21) Gait abnormalities Exophthalmos Inheritance: AD Gene: TGFβ

Sclerotic skull base Progressive endosteal and periosteal diaphyseal sclerosis Narrowing of medullary cavity of tubular bones Isotope bone scan: increased uptake

Group 25 (Osteogenesis Imperfecta and Decreased Bone Density Group) See Table 73.4 Osteogenesis imperfecta (see See Table 73.4 Figs 73.22 and 73.23) Inheritance: Types I and V—AD Types II, III and IV—AD, AR Genes: Type I—COL1A1, COL1A2 Type II—COL1A1, COL1A2, CRTAP, LEPRE1, PPIB Type III—As for type II plus FKBP10, SERPINH1 Type IV—COL1A1, COL1A2, CRTAP, PKBP10, SP7 Type V—Unknown Group 26 (Abnormal Mineralisation Group) Hypophosphatasia (see Perinatal severe, infantile and Fig. 73.24) juvenile forms: Poor dentition with premature loss of teeth, including roots Inheritance: AR Gene: ALPL Juvenile and adult forms: Inheritance: AD Gene: ALPL

Low alkaline phosphatase Phenotypic variability More severe forms—deficient ossification, with ‘missing’ bones Metaphyseal spurs

Group 27 (Lysosomal Storage Diseases With Skeletal Involvement (Dysostosis Multiplex Group)) Macrocephaly, thick skull vault Typically present in early childhood Mucopolysaccharidoses (see ‘J-shaped’ sella turcica Variable clinical manifestations Fig. 73.25) Wide (oar) ribs, short wide clavicles, poorly modelled scapulae Short stature Abnormality of Ovoid, hooked vertebral bodies with gibbus at thoracolumbar Distinctive coarse facies mucopolysaccharide and junction Intellectual impairment (in some) glycoprotein metabolism. Odontoid hypoplasia Corneal opacities (in some) Differentiation between the Flared iliac wings with narrow base of iliac wings Joint contractures types is dependent upon Small irregular proximal femoral epiphyses, coxa valga, poorly laboratory analysis (of urine, Hepatosplenomegaly modelled long bones with thin cortices, coarse trabecular Cardiovascular complications leucocytes and fibroblastic pattern Inheritance: AR, except for MPS type 2 which is XLR cultures) Short wide phalanges with proximal pointing of second to fifth Gene: Hurler/Scheie (type 1H/1S)—IDA metacarpals Hunter (type 2)—IDS Neurological changes include hydrocephalus, leptomeningeal Sanfilippo (type 3)—HSS, NAGLU, HSGNAT, GNS cysts and a variety of abnormalities best demonstrated by MRI Maroteaux–Lamy (type 6)—ARSβ Sly (type 7)—GUSβ

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

1891

TABLE 73.1  Clinical and Radiographic Features of Selected Osteochondrodysplasias and

Dysostoses—cont’d

Clinical Features

Radiological Features

Morquio syndrome (MPS type 4) (see Fig. 73.25)

Normal intelligence Joint laxity Knock knees Short stature Corneal opacities Inheritance: AR Gene: GALNS, GLβ1

Mucolipidoses type II (I-cell disease) (see Fig. 73.26)

Symptoms may be apparent in neonatal period Craniofacial dysmorphism Gingival hyperplasia Joint stiffness Inheritance: AR Gene: GNPTα/GNPTβ

Hypoplastic/absent odontoid peg (cervical instability may lead to cord compression) Platyspondyly with posterior scalloping of vertebral bodies Anterior ‘beak’ or ‘tongue’ of vertebral bodies Flared iliac wings and constricted iliac bases Progressive disappearance of the femoral heads Coxa valga, genu valgum Irregular ossification of metaphyses of long bones Small irregular epiphyses Proximal pointing of second to fifth metacarpals Osteopenia with coarse trabeculae Periosteal cloaking Pathological fractures Stippled/punctate calcification Metaphyseal irregularity Flared iliac wings Ovoid vertebral bodies Broad ribs

Group 29 (Disorganised Development of Skeletal Components Group) Multiple cartilaginous Multiple bony prominences, particularly at the ends of long exostoses (see Fig. 73.27) bones, ribs, scapulae and iliac bones Secondary deformity and limitation of joint movement Inheritance: AD Gene: Type 1—EXT1 Type 2—EXT2 Type 3—Unknown Enchondromatosis (Ollier Asymmetrical limb shortening disease) (see Fig. 73.28) Expansion of affected bones Occasional pathological fracture Absence of vascular malformation (Ollier disease) Presence of vascular malformation (Maffucci syndrome) Malignancy rare in Ollier disease Malignancy relatively common in Maffucci syndrome (at least 15%) Inheritance: Non-genetic Gene: Non-genetic (PTHR1 and PTPN11 mutations found in a few patients—significance unknown) Fibrous dysplasia (see Pain and deformity of involved bones Fig. 73.29) Monostotic—only one bone involved Polyostotic—multiple bones involved McCune–Albright syndrome consists of polyostotic fibrous dysplasia, patchy café au lait skin pigmentation and precocious puberty (usually in girls) Inheritance: Sporadic Gene: GNAS1 (polyostotic)

Multiple flat/protuberant, polypoid/sessile exostoses Secondary joint deformities Reverse Madelung deformity (short distal ulna) Vertebral bodies rarely involved Skull vault spared

Typically asymmetrical Shortening of affected long bones Radiolucencies, particularly in metaphyses, with expansion of bone, cortical thinning and internal calcification Pathological fractures Joint deformity Calcified phleboliths within vascular malformations (in Maffucci syndrome, but not usually seen until adolescence)

Asymmetrical thickening of skull vault, with sclerosis of the base, obliteration of the paranasal sinuses, facial deformity ‘Ground-glass’ or radiolucent areas of trabecular alteration in the long bones associated with patchy sclerosis and expansion, with cortical thinning and endosteal scalloping. No periosteal reaction without pathological fracture Progressive deformities due to fracture/ bone softening: e.g. ‘shepherd’s crook’ femoral necks

Group 32 (Cleidocranial Dysplasia and Isolated Cranial Ossification Defects Group) Frontal bossing, wide skull sutures, delayed fontanelle closure, Cleidocranial dysplasia (see Macrocephaly Wormian bones Fig. 73.30) Large fontanelle with delay in closure Multiple supernumerary teeth Supernumerary teeth Excessive shoulder mobility Variable clavicular hypoplasia/pseudoarthrosis Narrow chest Absent or delayed pubic ossification Inheritance: AD Gene: RUNX2 Continued

1892

SECTION G  Paediatric Imaging

TABLE 73.1  Clinical and Radiographic Features of Selected Osteochondrodysplasias and

Dysostoses—cont’d

Clinical Features Group 33 (Craniosynostosis Syndromes) Pfeiffer syndrome (see Craniofacial dysmorphism Fig. 73.31A) Broad, medially deviated thumbs and first toes Soft-tissue syndactyly of fingers and toes Inheritance: AD Gene: FGFR1, FGFR2 Apert syndrome (see Fig. 73.31B)

Craniofacial dysmorphism present from birth Proptosis High arched/cleft palate Bifid uvula ‘Mitten/sock deformity’ of hands/feet Inheritance: AD Gene: FGFR2

Radiological Features Sagittal/coronal/squamous temporal craniosynostosis (‘cloverleaf skull’) Dysplastic proximal phalanges of thumbs/first toes Hypoplastic or absent middle phalanges two-thirds and/or three-quarters soft-tissue syndactyly of fingers and toes Carpal fusions Coronal craniosynostosis Bony and soft-tissue syndactyly of hands and feet Progressive carpal/tarsal and large joint fusion Progressive fusion of cervical spine (commonly C5/C6) Dislocated radial heads

AD, Autosomal dominant; AR, autosomal recessive; ASD, atrial septal defects; XLD, X-linked dominant; XLR, X-linked recessive.

TABLE 73.2  Clinical and Radiographic

Features Used in the Diagnosis of Constitutional Disorders of Bone and Malformation Syndromes

Family history Stature—proportions and symmetry Abnormal body proportion—short trunk (suggests abnormal spine (e.g. platyspondyly) or short limbs Short limb segments—proximal (rhizomelic), middle segment, i.e. forearms/ lower legs (mesomelic), distal (acromelic) Long bone modelling—under/overtubulation, cortical/medullary texture and density Evidence of epiphyseal dysplasia (small/delayed epiphyses) or metaphyseal dysplasia (irregular metaphyses, spared/large epiphyses) Pattern of vertebral changes—anisospondyly (collagen-2), non-widening interpedicular distance/short pedicles (FGFR3), overfaced pedicles/ irregular platyspondyly (TRPV4) Local deformities—polydactyly, focal bone lesions, Madelung deformity Facies—macrocephaly, microcephaly, dysmorphology Other—hearing, sight, learning difficulties, hepatosplenomegaly, immunological abnormalities Bloods/biochemistry—bone biochemistry (hypophosphatasia/rickets), thyroid function tests, storage disorders Change of phenotype and radiographic changes over time Fig. 73.1  Thanatophoric Dysplasia Type 1. Micromelia with bowed femora and metaphyseal spurs. Short ribs with a small thorax. Platyspondyly. Trident acetabula.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

1893

B

A

D

E

C

Fig. 73.2  Achondroplasia in a Neonate. Narrow interpedicular distances of the lumbar spine and posterior scalloping of the vertebral bodies develop with age and are not seen in neonates. (A) Micromelia, short ribs with a small thorax and sloping metaphyses of proximal humeri. (B) Small square iliac wings, short sacrosciatic notches, horizontal trident acetabula and sloping proximal femoral metaphyses. (C) Platyspondyly and bullet-shaped vertebral bodies. (D) Bullet-shaped phalanges. (E) Magnetic resonance imaging may be used to demonstrate cervical spinal cord impingement due to stenosis at the foramen magnum.

A

B

C

D

Fig. 73.3  Hypochondroplasia. The milder end of the FGRF3 spectrum, with (A) narrow interpedicular distances which do not widen inferiorly in the lumbar spine, (B) pedicular shortening and mild posterior vertebral scalloping, (C) squaring of iliac wings, horizontal acetabula and coxa vara, and (D) mild shortening and broadening of long bones.

1894

SECTION G  Paediatric Imaging

C

D

A

B

E

Fig. 73.4  Spondyloepiphyseal Dysplasia Congenita. Anisospondyly (varying shape and size of the vertebral bodies, with L5 smaller than L1) in a neonate (A) and at 5 years of age (B); this appearance can be seen in normal individuals in utero and occasionally in premature infants, but is abnormal at older ages. (C and D) Delayed ossification of the upper cervical spine, pubic rami and epiphyses in a neonate; with continuing abnormal epiphyseal ossification at 4 years of age (E).

1895

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

B

A

D

C

Fig. 73.5  Metatropic Dysplasia. (A) Narrow thorax, short ribs with prominent anterior ends and marked platyspondyly. Expanded metaphyses (B to D) seen here of the right proximal humerus, proximal femora and tibiae, with narrow diaphyses gives rise to the so-called ‘dumb-bell’ appearance of the long bones.

B

Fig. 73.6  Spondylometaphyseal dysplasia Kozlowski type—demonstrates radiological features associated with TRPV4 gene mutations; (A) overfaced pedicles (where the lateral aspect of the vertebral body projects beyond the pedicle) and (B) irregular platyspondyly. The epiphyses are small and irregular (C).

A

C

A

B

C

Fig. 73.7  Ellis–van Creveld Syndrome. (A) Short ribs with a narrow thorax may contribute to respiratory distress. (B) Trident acetabula. (C) Postaxial polydactyly.

Fig. 73.8  Jeune Asphyxiating Thoracic Dystrophy. Short ribs, trident acetabula. No platyspondyly.

A

B

C

D

Fig. 73.9  Pseudoachondroplasia. (A and B) Mild platyspondyly with anterior protrusions of the vertebral bodies. (C) Plain film and (D) magnetic resonance imaging images of the knee demonstrate how the overgrowth of the cartilaginous anlage contributes to the reshaping of the epiphysis. (Images courtesy Dr A Calder, Great Ormond Street Hospital, London.)

A

Fig. 73.10  Multiple Epiphyseal Dysplasia. A child with a proven DTST mutation. (A) Small proximal femoral epiphysis with irregularity of the metaphysis. (B) Delayed bone age with ‘melting snow’ appearance of the metacarpal heads. (C) Flattened epiphyses of the knee and (D) layered patella.

A

B

C

D

B Fig. 73.11  Multiple Epiphyseal Dysplasia. A child with a proven COMP mutation. Mild metaphyseal and epiphyseal irregularity at the age of 9 years old (A and B), becomes more established with age, with established secondary joint damage by the age of 15 years (C and D). Relative sparing of the spine helps differentiate this from pseudoachondroplasia. Continued

1898

SECTION G  Paediatric Imaging

C

D Fig. 73.11, cont’d

A

B

C

Fig. 73.12  Metaphyseal Chondrodysplasia, Schmidt Type. Irregular cupped metaphyses of upper and lower limbs make this a differential for rickets (A, B and C). Bowed femora with coxa vara (B) is also a feature. Bone density is normal.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

A

B

D

C

Fig. 73.13  Metaphyseal Chondrodysplasia, McKusick Type (Cartilage Hair Hypoplasia). There are mild metaphyseal irregularities with relative epiphyseal enlargement (A, B and C). Cone-shaped epiphyses may be seen, particularly in the hands, although they are relatively mild in this case (D). Associated sparse hair and immunodeficiency support the diagnosis.

1899

1900

SECTION G  Paediatric Imaging

B

A

C

Fig. 73.14  Spondylometaphyseal Dysplasia, Sutcliffe Type. There is platyspondyly with oval vertebral bodies and anterior beaking. The metaphyses appear irregular and fragmented (A, B and C), at times reminiscent of corner fractures, making physical abuse a radiological differential.

Fig. 73.15  Campomelic Dysplasia. Narrow iliac wings, flared iliac bones, mesomelic shortening, angulation of the femora and hypoplastic scapula.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

A

C

B

Fig. 73.16  Chondrodysplasia Punctata. A neonate with the rhizomelic form. The punctate stippling at the epiphyses (A and B) and coronal cleft vertebral bodies (C) are still detectable at this age.

A

B

C

Fig. 73.17  Osteopetrosis. Generalised increase in bone density, with a bone-in-bone appearance (A and B). Increased bone fragility is illustrated by the concomitant proximal phalangeal fractures (A). On magnetic resonance imaging the bony sclerosis is demonstrated as diffuse low T1 signal throughout the bones, including the epiphyses which should normally contain fatty marrow (C).

1901

1902

A

SECTION G  Paediatric Imaging

C

B

Fig. 73.18  Pyknodysostosis. (A and B) Generalised increase in bone density. (C) Hypoplastic terminal phalanges of the left hand help differentiate this from osteopetrosis.

A

B

Fig. 73.19  Osteopoikilosis. Multiple sclerotic bone islands.

Fig. 73.20  Melorheostosis. Dense cortical bone (‘dripping candle wax’) in a ray distribution affecting (A) the left humerus, radius, ulna, (B) carpals and phalanges.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

A

C

B

D

Fig. 73.21  Diaphyseal Dysplasia (Englemann Syndrome). There is marked undertubulation of the long bones, with sclerotic diaphyseal expansion (A to C). There is accompanying diffuse sclerosis of the skull vault and base (D)

1903

1904

SECTION G  Paediatric Imaging

A

A

B

B Fig. 73.22  Osteogenesis Imperfecta Type II. (A) This is generally lethal in the perinatal period, with extensive fracturing causing a beaded appearance of the ribs and accordion appearance of the femora. Computed tomography (CT) may help to better define the bony anatomy, although the resolution of conventional CT may be limited in small fetuses (B), leading to growing interest in micro-CT as a research tool. (CT image courtesy Fetal and Perinatal Skeletal Dysplasias: An Atlas of Multimodality Imaging (Fig. 6q page 349), Amaka C. Offiah, Christine M. Hall, Deborah Krakow, and Michelle Fink. CRC Press 2012.)

Fig. 73.23  Osteogenesis Imperfecta Type III. (A to C) Reduced bone density, gracile long bones, with bowing deformity due to multiple fractures. ‘Popcorn calcification of the metaphyses’. (D) Healing rib fractures, scoliosis with spinal fixation and a portacath for bisphosphonate infusions.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

1905

A

C

B

D Fig. 73.23, cont’d

C Fig. 73.24  Hypophosphatasia. There is a wide spectrum of abnormalities, from almost complete absence of bony ossification in the severe perinatal form (A) to mild femoral (B) and humeral (C) bowing.

1906

SECTION G  Paediatric Imaging

A

A

B

B

C

Fig. 73.25  Mucopolysaccharidosis Type 4 (Morquio), Attenuated Form. (A) Small and sclerotic femoral heads (progressive fragmentation over the preceding few years), irregular acetabula. (B) Platyspondyly, anterior beaking and posterior scalloping of the vertebral bodies. (C) Flattened metacarpal heads and irregular carpal bones. (Images courtesy Dr Jörg Schaper Pediatric Radiology Institute of Diagnostic & Interventional Radiology, Heinrich Heine University, Düsseldorf, Germany.)

C Fig. 73.26  Mucolipidosis Type II (I-Cell Disease). (A) (B) Coarse trabeculation and periosteal cloaking of the long bones. (C) Stippling is seen at the ankle.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

A

B

C

Fig. 73.27  Multiple Cartilaginous Exostoses. Multiple exostoses at the proximal humerus (A), around the knee (B) and ankle (C). Note that they point away from the growth plates.

1907

1908

SECTION G  Paediatric Imaging

A

B

Fig. 73.29  Fibrous Dysplasia. Expansile radiolucent lesion in the proximal femur (A) increasing in size over time, with mild bowing—‘shepherd’s crook ‘deformity’ (B). There is ground-glass density and cortical scalloping. A

A

B Fig. 73.28  Multiple Enchondromatosis (Ollier Disease). (A) Multiple expansile lytic lesions with associated soft-tissue swelling of some fingers. (B) Enchondromata of the distal radius appear as metaphyseal striations and stippled calcification.

B Fig. 73.30  Cleidocranial Dysplasia. (A) Hypoplastic lateral ends of the clavicles, bell-shaped thorax. (B) Hypoplastic pubic rami. Bilateral coxa valga.

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

1909

Fig. 73.32  Acrodysostosis. Marked shortening of the metacarpals and phalanges, particularly distally.

A

B Fig. 73.31  Acrocephalosyndactyly Syndromes. (A) Pfeiffer syndrome. Craniosynostosis with a clover-leaf skull. Prominent sutural markings and wide anterior fontanelle. Hypoplasia of middle phalanges, bilateral two-thirds soft-tissue syndactyly. (B) Apert syndrome. Failure of normal distal bony segmentation in association with craniosynostosis. Single dysplastic phalanx of broadened great toes and only two phalanges for second to fifth toes. Fusion of the metatarsophalangeal joints of second to fourth toes. Fusion of metatarsal bases. Bony bar between first and second metatarsals bilaterally (left more prominent). Soft-tissue syndactyly of second to fifth toes (‘sock foot’).

changes (e.g. distal (acro-) (Fig. 73.32) or proximal (rhizo-) shortening); and looking for patterns of bone changes that suggest involvement of a particular gene (e.g. narrow interpedicular distances in FGFR3 conditions). Rapid advances are being made in the field of gene mapping, with many conditions being localised to abnormalities at specific loci on individual chromosomes. Identification of genetic mutations allows ‘families’ of conditions to be recognised, with some common clinical and radiological features. This is reflected in the groupings in the international nosology (see Table 73.1) with groups including those

with mutations in FGFR3 (the ‘achondroplasia’ group), or the TRPV4 group of disorders, consisting of autosomal dominant (AD) brachyolmia, metatropic dysplasia and spondylometaphyseal dysplasia, Kozlowski type (see Figs 73.5 and 73.6). Although classification based on genetic mutations is of value in determining an underlying causative defect, this diagnostic approach does not necessarily arrive at a precise clinical diagnosis. With the greater accessibility of genetic testing using ‘panels’ of gene tests established in recent years for particular skeletal phenotypes, it is increasingly common for the radiologist to be presented with a child with a known genetic mutation and asked if this genetic mutation corelates with the child’s imaging. Anticipatory genetic testing such as this can be not only a challenge for radiologists but also a great asset in cases where, after extensive clinical (including radiological) work-up, the diagnosis is still uncertain. Currently, National Health Service (NHS) patients in England with undiagnosed dysplasia may benefit from whole genome sequencing as part of the ‘100,000 genomes’ research project. Because of the large number of relatively rare conditions, it is difficult or impossible for an individual radiologist to be familiar with every feature of every disorder. In addition, many of these conditions may have age-dependent features such that the radiological findings evolve or even resolve with time. For example, in Morquio disease (mucopolysaccharidosis, MPS type 4) the capital femoral epiphyses are well ossified at the age of 2 years, but at 8 years are small and flattened, and by 10 years have typically disappeared (see Fig. 73.25). Each radiologist’s personal experience of the individual conditions will be limited because of the vast numbers of conditions involved. Textbooks are also of limited value because of the necessarily restricted number of illustrations and obsolescence. For these reasons, skeletal dysplasias, dysostoses and malformation syndromes as a group lend themselves to computer and web-based applications. Web-based resources include those that allow individuals to refer cases to a group of experts— for example, the European Skeletal Dysplasia Network (ESDN)—or those that allow individuals to attempt to make a diagnosis themselves (e.g. the dynamic Radiological Electronic Atlas of Malformation Syndromes [dREAMS]). A specific diagnostic label should only be attached to a patient when it is secure. An inaccurate diagnosis may have a profound effect upon the family in terms of genetic counselling, and upon the patient in terms of management and outcome. There may be a need to monitor and re-evaluate the evolution of radiological findings over time before establishing a diagnosis. A significant proportion of cases (approximately 30%) are unclassifiable because the combination of findings does not

1910

SECTION G  Paediatric Imaging

conform to any recognised condition. It is important that data, both clinical and radiological, and specimens (e.g. bone, tissue and blood) are stored to allow the diagnosis to be revisited as new research emerges. Only in this way will it be possible to ‘match’ conditions and to establish the natural history of a disorder.

Imaging

Prenatal Screening and Investigation In the UK, all pregnant women are now offered prenatal ultrasound (US) screening at between 18 and 21 weeks’ gestation. Neonatally, lethal skeletal dysplasias may be diagnosed at this stage by demonstrating short limbs, bowed limbs (see Figs 73.1 and 73.22) or a narrow thorax. Where there has been a previously affected sibling, specific malformations such as polydactyly, polycystic kidneys or micrognathia may be assessed. Skeletal US findings are highly significant but are not very specific, and in general it is unwise to offer a precise diagnosis on the basis of US alone. Pregnancy terminations offered on the grounds of such findings should subsequently have radiological and histopathological evaluation to determine the precise diagnosis and before genetic counselling is offered. Many non-lethal conditions, which may present at birth, also can be ascertained on prenatal US. Prenatal US has high precision for predicting lethality. The early diagnosis of a lethal dysplasia can prevent unnecessary and distressing prolongation of life and help to reduce parental expectations and anguish. This practice is leading to a change in the incidence of certain conditions formerly presenting at birth. Occasionally, other imaging techniques may help to confirm a suspected prenatal diagnosis. Maternal abdominal radiographs are now obsolete; the poor diagnostic quality does not justify the radiation risk to either the fetus or the mother. Low-dose prenatal computed tomography (CT) is being successfully performed in some centres for the evaluation of skeletal anomalies (see Fig. 73.22). Magnetic resonance imaging (MRI) is increasingly used for in utero evaluation of specific anomalies, particularly of the central nervous system but more recently of other systems, including the musculoskeletal system, and for the assessment of lung volumes. Amniocentesis or chorionic villus sampling can be used for biochemical evaluation of fetuses at risk from storage disorders when a previous pregnancy has been affected, but at the cost of a 0.5%–1% risk of miscarriage. Cell-free fetal DNA testing, or non-invasive prenatal testing, is now increasingly available although not yet part of the routine NHS diagnostic armamentarium. This involves the testing of free fetal DNA within maternal blood and although not as accurate as the more invasive tests, does not carry the same risk of miscarriage.

Imaging for Diagnosis In addition to prenatal US, a skeletal survey is recommended for pregnancy terminations performed for suspected malformation and for spontaneous abortions and stillbirths, with parental consent. Although this may involve anteroposterior (AP) and lateral ‘babygrams’ of the entire fetus/infant, ideally additional views of the extremities should be performed. Cross-sectional post-mortem imaging is increasingly available, including CT, MRI, and (in a research setting) micro-CT. After live birth, a standard full skeletal survey is indicated when attempting to establish a diagnosis for short stature or for a dysmorphic syndrome. This should include: • AP and lateral skull (to include the atlas and axis) • AP chest (to include the clavicles) • AP pelvis (to include the lumbar spine and symphysis pubis) • lateral thoracolumbar spine • AP one lower limb • AP one upper limb

• posteroanterior (PA) one hand (usually the left; allows bone age assessment). Occasionally, additional views will be required, particularly with specific clinical abnormalities, and these may include views of the feet, e.g. if polydactyly is present, or views of the cervical spine if cervical instability is suspected with specific diagnoses (Table 73.3), or both upper and lower limbs if asymmetry or deformity is a clinical feature. If a diagnosis cannot be established, then (limited) follow-up imaging is indicated (e.g. at 1 and 3 years), to evaluate progression and evolution of radiographic appearances. Conditions with decreased bone density may be assessed and monitored by means of dual-energy x-ray absorptiometry (DXA). The roles of quantitative US and high-resolution peripheral quantitative CT for assessment of bone density are yet to be fully determined in children.

Imaging for Complications When a confident diagnosis is established, further imaging is essential to monitor the progress of potential complications. Complications may result as part of the natural evolution of the condition, but may also be iatrogenic. Radiography and MRI of the cervical spine in flexion and extension will monitor instability; AP and lateral views of the spine will monitor kyphosis and scoliosis. In children with reduced bone density, a lateral spine DXA should replace radiographs to diagnose vertebral fractures. Long limb radiographic views, CT scanograms or increasingly images from low-dose, upright biplanar systems such as the ‘EOS’ system (EOS Imaging, France), will help to assess asymmetry, genu varum and genu valgum, and to monitor progression of limb length discrepancy. CT or

TABLE 73.3  Disorders With Instability in

the Cervical Spine

Cervical Spine Instability With Odontoid Peg Absence or Hypoplasia Achondroplasia Chondrodysplasia punctata Diastrophic dysplasia Dyggve Melchior–Clausen disease Hypochondrogenesis Infantile hypophosphatasia Kniest dysplasia Metaphyseal chondrodysplasia, McKusick type Metatropic dysplasia Morquio disease (MPS type 4) and other mucopolysaccharidoses (MPS) Mucolipidoses (MLS) Multiple epiphyseal dysplasia Neurofibromatosis type 1 Opsismodysplasia Pseudoachondroplasia Pseudodiastrophic dysplasia Spondyloepiphyseal dysplasia congenita Trisomy 21 Cervical Spine Instability With Cervical Kyphosis (C2/C3) Diastrophic dysplasia Spondyloepiphyseal dysplasia congenital Lethal Atelosteogenesis Campomelic dysplasia

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant MRI may monitor the development of hydrocephalus or the presence of neuronal migration defects and structural defects, such as the absence of the corpus callosum. CT may also demonstrate encroachment on the cranial nerve foramina and both CT and MRI are of value in assessing spinal cord compression. US can demonstrate associated organ anomalies (e.g. cystic disease of the kidneys or hepatosplenomegaly) and echocardiography may reveal associated intracardiac abnormalities. Arthrography, US, CT and MRI are of value in the assessment of joint problems, particularly when surgical intervention is proposed or following surgery. Technetium skeletal scintigrams are occasionally used to determine the extent of bony involvement in specific disorders (e.g. asymptomatic lesions identified in chronic recurrent multifocal osteomyelitis [CRMO]). However, in patchy disorders such as fibrous dysplasia, radiographically affected areas may not demonstrate abnormal uptake of radionuclide. In recent years, whole-body MRI has become an increasingly feasible, radiation-free option for global assessment of multifocal disease, as technological advances in rolling table platforms, phased array coils and the use of multiple radiofrequency input channels have allowed for faster, higher-resolution imaging. Fat-suppressed images (short tau inversion recovery [STIR]) and diffusion-weighted imaging (DWI) can increase the sensitivity for detection of focal lesions, although ongoing research is still required in this area to fully understand the range of normal findings in healthy children.

1911

in each subtype is supplemented intravenously. This can improve symptoms; however, it cannot reverse already established changes. In some conditions, cure may be achieved: for example, in the severe form of osteopetrosis (which is lethal in childhood unless treated), by means of a compatible bone marrow transplant in the first 6 months of life. Bone marrow transplantation has also been used with some success in selected patients with mucopolysaccharidoses. Growth hormone therapy is used in selected disorders to influence final height. Growth hormone stimulates type I collagen production and is being used, in particular, to augment growth rate in children with osteogenesis imperfecta (OI) (Table 73.4). Bisphosphonates are pyrophosphate analogues that inhibit osteoclast function. They have been used in OI to improve bone density and have been used to treat bone pain and osteopenia in a variety of rheumatological and dermatological conditions. The radiological hallmark of bisphosphonate therapy, so-called ‘bisphosphonate lines’, are now well recognised (Fig. 73.33). In many conditions, orthopaedic procedures are invaluable in maintaining or improving mobility. For example, osteotomies prevent or correct dislocations or long bone bowing deformities. Patients with OI may require multiple osteotomies to correct severe deformities, as well as intramedullary rodding to reduce fractures, to maintain alignment and to provide support and stability (see Fig. 73.23). Joint replacements may be necessary, especially in those dysplasias (e.g. multiple epiphyseal dysplasia) in which major involvement of the epiphyses may result in premature osteoarthritis. In some conditions, limb-lengthening procedures may be appropriate to improve mobility. This is usually offered in disorders with asymmetric shortening (Table 73.5), but is sometimes offered to selected patients with achondroplasia or other short-limbed dysplasias for cosmetic reasons. With the identification and localisation of specific chromosomal abnormalities associated with particular disorders, the development of gene therapy for clinical use poses many challenges and offers great potential for the future.

Management Only when an accurate diagnosis has been established can the prognosis and natural history be given. For example, myopia can be corrected and retinal detachment prevented in Stickler syndrome (hereditary arthro-ophthalmopathy); cord compression can be prevented in conditions with instability of the cervical spine (see Table 73.3) or with progressive thoracolumbar kyphosis and spinal stenosis, as in achondroplasia. Conditions previously considered lethal in the perinatal period can now be treated with pharmacological therapies: for example, the success of enzyme replacement therapy with asfotase alfa in selected cases of severe perinatal hypophosphatasia. Enzyme replacement therapy is also an option in mucopolysaccharidoses, where the specific enzyme deficient

Genetic Counselling When an accurate diagnosis has been made, meaningful genetic counselling can be given, both to the parents and to the affected individual. Most conditions are inherited in an AD or autosomal recessive (AR) manner. In conditions with an AD inheritance, the affected individual

TABLE 73.4  Osteogenesis Imperfecta Clinical (Based on the Sillence Classification) and

Radiological Findings I

II

III

IV

Va

1 : 30,000 Lethal Stillborn Blue — — — —

Rare Severe By 30 years Blue, then grey Rare DI Short

Unknown (rare) Mild/moderately severe Old age White Rare IVA normal IVB DI Normal/mildly short

Unknown (rare) Moderate Old age White Rare Normal

Stature

1 : 30,000 Mild Old age Blue Frequent IA normal IB DI Normal

Normal/mildly short

Radiological Findings Fractures at birth Osseous fragility Deformity

60 >60 50–59 50–59 43–49 43–49 55 >55 77 >77

Good Good Deficient Deficient Deficient Severely deficient Poor

Sharp Usually blunt Rounded Rounded Rounded/flat Rounded/flat Flat

Mature Mature 3 months of age

IIIb

77

Poor

Flat

IV

77

Poor

Flat

Covers femoral head Covers head Covers head Covers head Covers head Compressed Displaced up Echo poor Displaced up Reflective Interposed

The earliest radiographic feature is that of a radiolucent subchondral fissure—the crescent sign (Fig. 73.39). Disease progresses with loss of height, fragmentation and sclerosis of the femoral head (see Fig. 73.39). A coxa magna deformity may ensue, with lateral uncovering of the capital femoral epiphysis. There may be associated irregularity of the acetabular margin. The extent of subchondral fracture is said to be a good predictor of the final outcome. Several radiological classification systems, such as the Herring lateral pillar and (modified) Catterall systems, have been developed and shown to be reliable when used by an experienced observer. While skeletal scintigraphy is highly sensitive and specific for detecting avascular necrosis (AVN), MRI has now largely replaced it. In those with normal signal intensity in the epiphysis, or with marked loss of signal on both T1 and T2 weighted sequences (dead bone), intravenous enhancement is not necessary. In those with more equivocal findings, intravenous contrast medium may assist in identifying areas of viable bone. Treatment options in the acute phase are largely supportive, aiming to reduce impact on the fragile femoral head. If there is significant deformity later, corrective osteotomy/leg lengthening may be considered.

On point of dislocation Dislocated Dislocated Dislocated

A

Slipped Capital Femoral Epiphysis This is the commonest hip disorder of adolescence. Anterolateral and rotational forces of the hip muscles on the femoral shaft result in anterosuperior translation of the proximal femoral metaphysis relative to the epiphysis: that is, technically it is the metaphysis rather than the epiphysis that slips (Fig. 73.40). It is more common in boys, in AfroAmericans and in the obese. It most commonly occurs at the time of the pubertal growth spurt at Risser grade 0 (see Fig. 73.49 below). Slipped capital femoral epiphysis (SCFE) is extremely rare before the onset of puberty (adrenarche) and is uncommon in girls after menarche or in boys after Tanner stage 4. Bilateral slips occur in about 25% of Caucasian and up to 50% of Afro-American children. When unilateral, the left side is more often involved (65%). Endocrine disorders associated with SCFE include hypothyroidism, growth hormone deficiency, hypogonadism and panhypopituitarism. Radiography remains the investigation of choice. The frog leg lateral is more sensitive for detection of a slipped epiphysis, and in many centres is the radiograph of choice in patients of this age presenting with hip pain. Not performing a frog leg lateral makes SCFE easier to miss, and delayed presentation/diagnosis of slips have worse outcomes. Unstable slips—where the child is unable to weight-bear on the hip—also have a higher rate of complications. A Klein line is drawn on the AP projection while the slip angle is measured on frog lateral radiographs (Fig. 73.41). Once detected, most SCFEs are fixed in situ. Complications of SCFE include chondrolysis (narrowing of the joint space), AVN and osteoarthritis.

B Fig. 73.39  Idiopathic Avascular Necrosis of the Femoral Head (Perthes Disease). (A) Frog lateral on day of presentation showing the characteristic lucent crescent sign of early avascular necrosis of the right capital femoral epiphysis. (B) Rapid progression with irregularity, loss of height and sclerosis of the femoral head 8 months later.

1916

SECTION G  Paediatric Imaging

A

A

B

B

Fig. 73.41  Slipped Capital Femoral Epiphysis Measurements. Left slipped capital femoral epiphysis. (A) Klein line is shown on the right. This line should normally intersect approximately the lateral sixth of the capital femoral epiphysis in the anteroposterior projection. (B) The slip angle—between a line (BD) perpendicular to the plane of the growth plate (AA), and a line (CD) parallel to the longitudinal axis of the femoral shaft in the frog lateral projection—is 17.4 degrees. The severity of slip may be graded using the slip angle as mild (51) although this is not the primary determinant of treatment options.

Idiopathic Coxa Vara In this condition, there is coxa vara (reduction of femoral neck/shaft angle). A separate fragment of bone (Fairbank triangle)—from the inferior portion of the femoral neck—is characteristic. If the neck/shaft angle is less than 100 degrees, then without surgical intervention the varus deformity will progress.

Proximal Focal Femoral Deficiency

C Fig. 73.40  Slipped Capital Femoral Epiphysis. Plain radiographs, demonstrating potentially subtle abnormality on the anteroposterior view with mild opening of the growth plate (A), more obvious on the frog leg lateral (B). A left-sided slip is seen on magnetic resonance imaging in a different patient (C).

Femoral Dysplasia (Idiopathic Coxa Vara/Proximal Focal Femoral Deficiency Spectrum) The spectrum of femoral dysplasia encompasses all conditions from the mild idiopathic coxa vara, through moderate forms with deficiency of the proximal femur (Fig. 73.42), to severe forms in which only the distal femoral condyles develop.

Proximal focal femoral deficiency (PFFD) is bilateral in only 10% of cases. Varying degrees of agenesis of the proximal femur occur (see Fig. 73.42); there is an association between severity of femoral dysplasia and severity of acetabular dysplasia. Based on the presence or absence of the femoral head, connection between the proximal femur and femoral head, and the morphology of the acetabulum and shortened femur, Aitken classified PFFD into four groups of increasing severity from A to D. In addition to the femoral shortening, the lower leg also may be short, and the fibula absent or hypoplastic. Radiography demonstrates the degree of aplasia and, particularly in younger children, MRI and/or arthrography is useful for the visualisation of unossified cartilage. Because PFFD is associated with absence or deficiency of the cruciate ligaments, MRI also has a role in imaging the knee(s) of affected patients.

Tibia Vara and Tibial Bowing Tibia vara refers to unilateral or bilateral outward bowing of the legs at the level of the knee joint or proximal tibia. Causes include

CHAPTER 73  Skeletal Radiology in Children: Non-Traumatic and Non-Malignant

1917

Fig. 73.42  Proximal Focal Femoral Deficiency. Hypoplastic proximal femur. Absent ossification of the femoral head with a normal acetabulum.

Fig. 73.44  Blount Disease. Fragmentation of the medial half of the left proximal tibial epiphysis. Radiograph showing varus angulation.

Blount disease (see Fig. 73.44) affects the medial aspects of the proximal tibial epiphyses. It is unilateral in 40%. There are infantile and adolescent presentations. The infantile form of the disease occurs between the ages of 1 and 3 years. Adolescent Blount disease has a higher post-surgical recurrence rate than the infantile form. Initial beaking of the medial proximal tibial metaphysis progresses to irregularity, fragmentation and premature fusion of the medial aspect of the proximal tibial growth plate. Posteromedial bowing of the tibia is most often physiological, but despite its often worrisome radiographic appearance, it resolves spontaneously. Anterolateral bowing of the tibia is most frequently associated with neurofibromatosis type 1 (found in 50%–55% cases of bowing) and fibrous dysplasia (in 15% of cases). Bowing may be secondary to or in combination with a pseudoarthrosis, which tends to progress without treatment.

Talipes

Fig. 73.43  Focal Fibrocartilaginous Dysplasia. Pathognomonic appearance of the proximal tibia with bowing deformity and a cortical radiolucent band. This condition is usually self-resolving and should be managed conservatively. The proximal tibia is the commonest site to be affected.

physiological bowing (bilateral and self-resolving), rickets, trauma, infection, neurofibromatosis, Ollier disease, Maffucci syndrome, fibrous dysplasia, focal fibrocartilaginous dysplasia (Fig. 73.43) and Blount disease (Fig. 73.44). Focal fibrocartilaginous dysplasia characteristically affects the proximal tibia but may also affect other bones, appearing as a linear radiolucency extending inferolaterally from the proximal tibial metadiaphysis. It causes bowing of the affected bone, but is benign and usually self-resolving. Surgery is required in those children with severe bowing or in whom the bowing does not resolve with time.

Talipes equinovarus (congenital clubfoot) consists of varus (inversion) and equinus (fixed plantar flexion) of the hindfoot, and varus of the forefoot. It results from abnormal development around the ninth week of gestation. Aetiological considerations include genetic factors and early amniocentesis (before 11 weeks), with recent work providing tentative evidence for aetiologically distinct subtypes. It occurs two to three times more commonly in boys. Useful measurements are summarised in Table 73.7.

Madelung Deformity This condition results from premature fusion of the medial half of the distal radial physis. The radii are short and bowed. There is reduction of the carpal angle, with wedging of the carpal bones between the distal radius and ulna. Madelung deformity may be inherited when it occurs as an AD mesomelic dysplasia (Léri–Weill syndrome/dyschondrosteosis) or it may present as an isolated disorder: for example, following trauma or infection (see Fig. 73.35). In reverse Madelung deformity, there is bowing of the forearm bones, in association with a short (abnormal) ulna. Causes of reverse Madelung

1918

SECTION G  Paediatric Imaging

TABLE 73.7  Diagnosis of Talipes Deformity

DP Radiograph

Lateral Radiograph

Hindfoot varus

Talocalcaneal angle: 50° in newborns; >40° in older children Midtalar angle medial to first metatarsal base — — Narrow with increased overlap of metatarsal bases Broad with reduced overlap of metatarsal bases

Talocalcaneal angle: 50° in newborns; >45° in older children Calcaneotibial angle: >90° plantar flexion of calcaneus Calcaneotibial angle: