The Missed Diagnosis in CT and The Role of Optimizing Scan Protocols in Preventing Them Elliot K. Fishman MD Professor of Radiology, Surgery and Oncology In the daily radiology practice, the rate of interpretation error is between 3% and 4%; however, of the radiology studies that contain abnormalities, the error rate is even higher, averaging in the 30% range. Fool Me Twice: Delayed Diagnoses in Radiology With Emphasis on Perpetuated Errors Kim YW, Mansfield LT AJR 2014;202:465-470 In our study, the majority of errors made were errors of underreading (42%), where the finding was simply missed. Fool Me Twice: Delayed Diagnoses in Radiology With Emphasis on Perpetuated Errors Kim YW, Mansfield LT AJR 2014;202:465-470 The majority of errors are false-negative interpretations and occur during interpretation of CT examinations. Recurring false-negative CT errors include failure to appreciate unexpected bowel or pancreatic malignancy, incidental pulmonary emboli, abnormality of vascular structures, bone lesions, omental disease, incidental abnormality present on targeted examinations on the periphery of the field of view. Eight CT lessons that we learned the hard way: an analysis of current patterns of radiological error and discrepancy with particular emphasis on CT McCreadie G, Oliver TB Clinical Radiology (2009) 64, 491-499 Common False Negative Errors During CT Reporting n Gastrointestinal tract tumor missed n Pancreatic tumor missed n Pulmonary embolus missed n Vascular lesion missed n Significant bone lesion missed n Omental metastatic disease missed n Incidental abnormality missed on targeted exam n Lesion missed on periphery of the field of view n Eight CT lessons that we learned the hard way: an analysis of current patterns of radiological error and discrepancy with particular emphasis on CT McCreadie G, Oliver TB Clinical Radiology (2009) 64, 491-499 Our data then indicate that interpretive errors, rather than communication errors, are by far the most generic cause of malpractice suits against radiologists. The Causes of Medical Malpractice Suits Against Radiologists in the United States Whang JS et al. Radiology 2013; 266:548-554 1
At our busy academic institution, we have noticed repeated examples of certain misdiagnosis, even by experienced abdominal imagers, both in our own department and at outside institutions. This is likely related to many factors. Discussed here are a variety of common diagnostic errors on body CT examinations. MDCT of the Abdomen: Common Misdiagnosis at a Busy Academic Center Horton KM, Johnson PT, Fishman EK AJR 2010; 194:660-667 For each diagnostic error, we explore the reasons for the misdiagnosis and provide experience based advise to avoid these mistakes. MDCT of the Abdomen: Common Misdiagnosis at a Busy Academic Center Horton KM, Johnson PT, Fishman EK AJR 2010; 194:660-667 Why is pathology missed on a CT scan? n Poor reader search strategy (i.e. miss a PE on an abdominal CT scan in the upper most scan sections) n Poor reader understanding of pathology (i.e. overcall or undercall of bowel pathology) n Assumptions made on review of the dataset (i.e. assume a well defined 2 cm renal mass is a cyst when it is a hypovascular renal mass) Why is pathology missed on a CT scan? n Unsuspected pathology not related to the primary cause for the examination n Incidental findings of clinical importance can occur in every organ and every anatomic zone n In academic institutions checking out the residents and/or fellows (added 10-2014) What is a key caused of missed diagnosis on CT scan? n CT scan protocols n IV and oral contrast n Single vs dual phase vs three phase studies n Patient positioning n Scanning volume selected Do you need to look at the full field of view on a CT scan and if so when? n Cardiac CTA or Cardiac Calcium Scoring Study n Spine CT especially T-spine, L-spine and Sacrum 2
Chest Pain and Triple Rule Out Targeted vs Full FOV Reviewing the full-fov from lumbar spine CT examinations will result in the detection of a small number of substantial extraspinal pathologic findings in addition to many benign incidental findings. Extraspinal Findings at Lumbar Spine CT Examinations: Prevalence and Clinical Importance Lee SY et al. Radiology 2012; 263:502-509 3
The full field of view (FOV) abdominal images were required o best visualize extraspinal abnormalities in 79.4% of cases. Spine CT Examinations: Prevalence and Clinical Importance Lee SY et al. Radiology 2012; 263:502-509 Extraspinal findings were present in 40.5% of adult outpatients undergoing lumbar spine CT examinations for low back pain and/or radiculopathy, most (62.3%) of whom had findings classified as benign and not requiring further work-up. Spine CT Examinations: Prevalence and Clinical Importance Lee SY et al. Radiology 2012; 263:502-509 Substantial extraspinal pathologic findings, consisting of an early stage renal cell carcinoma and transitional cell carcinoma, chronic lymphocytic leukemia, sarcoidosis, and 13 abdominal aortic aneuysms, were present in 4.3% of our cohort of 400 patients. Spine CT Examinations: Prevalence and Clinical Importance Lee SY et al. Radiology 2012; 263:502-509 Do you need to look at the topogram (scoutview) on all cases of CT scanning? Do you need to look at the topogram (scoutview) on all cases of CT scanning? Editorial: Reviewing the CT scout view: medicolegal and ethical considerations. Berlin L AJR 2014 Jun;202(6):1256-63 Methods: 2032 scout views were reviewed retrospectively by 2 radiologists blinded to history and CT findings. All cases with major findings (defined as any abnormality that would prompt additional diagnostic tests or require management) were correlated with the CT, other imaging or medical record when necessary by a 3 rd radiologist to determine (1) validity of the scout view finding, and (2) whether it was identifiable on the current CT scan. 4
The CT scout view showed a significant finding in up to 23% of cases, usually in an anatomic region imaged by CT. In as many as 2% of cases, the abnormality disclosed on the scout view may not be included in the CT FOV. The CT Scout View: Does It Need to Be Routinely Reviewed as Part of the CT Interpretation Johnson PT, Scott WW, Gayler BW, Lewin JS, Fishman EK AJR 2014; 202:1256-1263 In a small percentage of cases, review of the CT scout view will disclose significant pathologic findings not included in the CT FOV. The results of this study support the routine inspection of the scout view, especially for the detection of pathologic findings in anatomic regions not imaged by CT. The CT Scout View: Does It Need to Be Routinely Reviewed as Part of the CT Interpretation Johnson PT, Scott WW, Gayler BW, Lewin JS, Fishman EK AJR 2014; 202:1256-1263 Our results support routine review of the anatomic regions on the scout view that are not imaged on CT and suggest that interpreting the scout view will not generate unnecessary testing if the limitations of the scout view are recognized. The CT Scout View: Does It Need to Be Routinely Reviewed as Part of the CT Interpretation Johnson PT, Scott WW, Gayler BW, Lewin JS, Fishman EK AJR 2014; 202:1256-1263 Although 2% appears to be a low percentage, extrapolating Johnson et al. s data to the 85 million patients who undergo CT each year means that as many as 1.7 million patients may have a major abnormal finding that is seen on a scout view but not on the related axial CT images. Reviewing the CT Scout View: Medicolegal and Ethical Considerations Berlin L AJR 2014;202:1264-1266 Johnson et al. concluded that their findings support the routine review of the scout view when interpreting a CT study. Reasonable medical practice, logic, and medicolegal as well as ethical considerations confirm their conclusion. Reviewing the CT Scout View: Medicolegal and Ethical Considerations Berlin L AJR 2014;202:1264-1266 Bladder Cancer: Facts n Estimates of 72,570 new cases and 15,210 deaths in the US in 2013 n Most cancers are transitional cell carcinoma while others are squamous cell carcinoma and adenocarcinoma n Data from NCI (National Cancer Institute) 5
Bladder Cancer as an Incidental Finding n How often is it an incidental finding? n How often is it missed on a routine CT scan? n What is the legal liability of missing the diagnosis of bladder cancer in an asymptomatic patient? n What do you look for on CT for the routine evaluation of the bladder on a contrast enhanced abdominal CT? Bladder Cancer as an Incidental Finding: Thoughts n Incidental bladder cancers are often best seen on arterial phase imaging n Any enhancement of the bladder wall or off the bladder wall should be investigated further n Do not assume a zone subtle bladder enhancement is of no clinical significance n Coronal and sagittal may be helpful in many cases Bladder Cancer TCC of the Bladder 6
Subtle Bladder Cancer (Early and Late) Incidental Bladder Cancer Incidental Bladder Cancer Early and Late Phase Images Incidental Bladder Cancer 7
In other words, although TCC has typically been regarded as a hypovascular tumor, these lesions have considerable urothelial hypervascularity and are typically most conspicuous on early phase images. As a result, any focal hyperenhancement of the bladder urothelium must be considered suspicious for malignancy. Malignancies on CT: The Underrated Role of CT in Diagnosis Raman SP, Fishman EK AJR 2014; 203:347 354 Evaluation of the bladder has been largely considered the domain of cystoscopy, and the bladder regularly goes ignored by the radiologist. However, several imaging findings should strongly suggest the presence of malignancy whether CT is performed as CT urography for hematuria or routinely in the emergency setting. Malignancies on CT: The Underrated Role of CT in Diagnosis Raman SP, Fishman EK AJR 2014; 203:347 354 Common Sources of Error n Failure to review a select portion of the exam such as lung bases (i.e. lung nodule, PE) n Failure to review all sets of images (lung windows, bone windows) Unsuspected Pulmonary Embolism n Scans thru lower lung fields allow detection of unsuspected PE s but they may be missed especially if thicker sections are reviewed ( 5 x 5 mm or 3 x 3 mm vs.75 x.5 mm) n Situation most common in oncology patient for tumor staging or for follow-up n We have seen this most commonly with pancreatic cancer patients 8
Unsuspected Pulmonary Embolism- Solution Incidental PE n High index of suspicion especially in oncology patients ( 1-5% in the published literature) n Routine review of thin section CT scans and not just thicker slices Incidental PE This study shows that missed PE can occur on abdominal CT. It is recommended that interpretation include a careful search of the lower pulmonary arterial vasculature on contrast-enhanced abdominal CT scans. Missed Pulmonary Embolism on Abdominal CT Lim KY, Kligerman SJ, Lin CT, White CS AJR 2014; 202:738-743 9
The challenge in identifying PE is clearly greater on abdominal CT than on chest CT. In addition to the multiple pitfalls described already, the lungs are typically not a primary target of abdominal CT interpretation, and only a limited part of the pulmonary anatomy is included. Missed Pulmonary Embolism on Abdominal CT Lim KY, Kligerman SJ, Lin CT, White CS AJR 2014; 202:738-743 Gastric Tumor Detection n Overcall of tumor is common in fundus and antrum especially if the stomach is not well distended n Detection is optimized by data review in MPR or 3D display Gastric Polyp Gastric Adenocarcinoma 10
Stomach Not Distended vs Gastric Malignancy? Gastric Polyps (benign) 11
Suspected Mass in Gastric Fundus Hepatoma in Cirrhotic Liver (only seen arterial phase) 12
Hx-Hepatitis C: Rule Out HCC Hepatoma Best Seen on MIP Hepatoma Best Seen on MIP Pancreatic Mass Detection n Pancreatic mass vs peripancreatic tumor n Undiagnosed islet cell tumor n Misdiagnosis of splenic artery aneurysm as an islet cell tumor or splenule as a islet cell tumor 13
Dx: Suspected Pancreatic Mass 5 mm Neuroendocrine Tumor HOP 14
Do you see a mass? Neuroendocrine Tumor HOP Seen Only On Arterial Phase Hx- Renal Cell Carcinoma Do you see a pancreatic mass? Metastatic RCC to Pancreas only Seen on Arterial Phase 15
Renal Pathology-Solution n Understand the limitation of a renal stone protocol and what a normal stone study means An extremely common cause of perceptual error in genitourinary radiology (and other areas of diagnostic radiology as well) is improper technique. Medicolegal Issues in Genitourinary Radiology Berlin JW RSNA Categorical Course in Diag Radiol: Genitourinary Radiology 2006;pp 95-101 What are the sources of pitfalls, pratfalls and misdiagnosis of CT of the kidneys? Pitfalls in Renal Imaging n Phase of data acquisition n Image display format n Rendering technique used for display Pitfalls in Renal Imaging n Phase of data acquisition n Image display format n Rendering technique use for display 16
Phase of Acquisition: Non-Contrast CT n Ideal for detection of renal calculi n Ideal for defining CT attenuation of a suspected renal lesion pre-contrast to help determine whether a lesion is a high density renal cyst n Can specifically allow diagnosis of high density renal cyst Hign Density Renal Cyst There is no one perfect phase for imaging all renal masses arterial phase venous phase excretory phase Stone Protocol Studies n Remember why we do stone protocol studies n Remember what stone studies can say and what they can not say 17
Phase of Acquisition: Non-Contrast CT n But non-contrast CT will miss n Small renal tumors especially when not changing renal contour n Acute pyelonephritis n Vascular pathology including AVM There is no one perfect phase for imaging all renal masses arterial phase venous phase excretory phase Subtle Right Renal Cell Carcinoma 18
Pitfalls in Renal Imaging n Phase of data acquisition n Image display format n Rendering technique use for display 19
Do you see the lesion 20
Transitional Cell Carcinoma of the Ureter However, evaluation of the ureters using CT can be particularly problematic. Not only are the ureters often poorly opacified and distended due to poor CT technique, but the majority of ureteral TCC s are found in the distal third of the ureters, a segment that is particularly difficult to completely distend. Moreover, even when well distended, ureteral tumors can be extremely subtle and difficult to appreciate, particularly when relying primarily on the source axial images. MDCT Evaluation of Ureteral Tumors: Advantages of 3-D Reconstruction and Volume Visualization Raman SP, Horton KM, Fishman EK AJR 2013 Dec;201(6):1239-47 Proper diagnosis hinges not only on appropriate interpretation of the source axial images (with recognition of several suggestive CT features of malignancy), but also upon the utilization of 3-D technique (CR and MIP technique) as an ancillary diagnostic tool. In our experience, proper utilization of 3-D technique can be incredibly useful in the diagnosis of subtle tumors that are barely perceptible on the source axial images, and which may be missed otherwise. MDCT Evaluation of Ureteral Tumors: Advantages of 3-D Reconstruction and Volume Visualization Raman SP, Horton KM, Fishman EK AJR 2013 Dec;201(6):1239-47 CT of the Ureter: Role of 3D Imaging in Lesion Detection MR for Back Pain n Accentuation of subtle strictures and sites of narrowing n Accentuate subtle abnormal urothelial enhancement and thickening n Better visualization of the distal ureter n Better visualization of flat polypoid lesions 21
TCC Renal Pelvis/Proximal Ureter 22
TCC Proximal Ureter Subtle TCC Left Ureter 23
Proper diagnosis hinges not only on appropriate interpretation of the source axial images (with recognition of several suggestive CT features of malignancy), but also upon the utilization of 3-D technique (CR and MIP technique) as an ancillary diagnostic tool. In our experience, proper utilization of 3-D technique can be incredibly useful in the diagnosis of subtle tumors that are barely perceptible on the source axial images, and which may be missed otherwise. MDCT Evaluation of Ureteral Tumors: Advantages of 3-D Reconstruction and Volume Visualization Raman SP, Horton KM, Fishman EK AJR 2013 Dec;201(6):1239-47 Occlusion of the Mesenteric Artery Unsuspected mesenteric arterial abnormality may elude diagnosis when axial MDCT sections are interpreted without 3D renderings. Unsuspected Mesenteric Arterial Abnormality: Comparison of MDCT Axial Sections to Interactive 3D Rendering Chen JK, Johnson PT, Horton KM, Fishman EK AJR 2007;189:807-813 24
On a per-patient basis, the axial and 3D interpretations were equivalent in 24% (10/41) of the cases. Axial CT partially agreed with 3D CT in 10% (4/41), and no mesenteric arterial lesion was reported on axial CT in 66% (27/41). The 3D CT findings were supported by other imaging, surgery, clinical findings, or management in 49% (20/41) of the cases. The mesenteric lesions identified resulted in a change in patient management in 15% (6/41) of the subjects. Unsuspected mesenteric arterial abnormality:comparison of MDCT axial sections to interactive 3D rendering Chen JK, Johnson PT, Horton KM, Fishman EK AJR 2007 Oct;189(4):807-813 Arterial Mesenteric Occlusion: facts n 60-75% of all bowel ischemia cases n Can be arterial embolism or arterial thrombosis n Arterial embolism usually in mid vessel and proximal involvement usually due to thrombosis SMA Thrombus Pitfall: In cases of suspected ischemia make sure you examine the entire vessel not just its proximal portion. If you can t define it make that clear in your report. 25
To emphasize the point-you can t assume that because you don t see a thrombus that a thrombus is not present unless you have optimal vessel visualization. In most case you probably will be correct but when you are wrong SMA Occlusion in Patient with RLQ Pain 26
5 Days Later The midline sagital view on CT:a guide to pathology n SMA syndrome n Vascular stenosis n Median arcuate ligament syndrome (MALS) n Staging pancreatic cancer n Mesenteric aneurysms n Mesenteric collaterals Axial and coronal reformations of 64-section multidetector row CT have equal sensitivity and specificity for the diagnosis of acute abdominal pathology. However, coronal reformations improved the diagnostic confidence for all readers but most significantly for the least experienced. Therefore, radiology departments with residents should consider routinely generating coronal images in patients with acute abdominal pain. Acute abdomen: Added diagnostic value of coronal reformations with 64-slice multidetector row computed tomography. Zangos S et al. Acad Radiol. 2007 Jan;14(1):19-27. For the most inexperienced reader, the coronal reformations were helpful in 95% of cases, while for the most experienced reader, the coronal reformations were helpful in 35% of the cases. The coronal images were deemed helpful in an average of 62.3% of the cases for the four readers. However, diagnosing subtle pathology in the abdominal wall was difficult on coronal reformations alone. Overall, coronal reformations improved diagnostic confidence and interobserver agreement over axial images alone for visualization of normal abdominal structures and in the diagnosis of abdominal pathology. Acute abdomen: Added diagnostic value of coronal reformations with 64-slice multidetector row computed tomography. Zangos S et al. Acad Radiol. 2007 Jan;14(1):19-27. Sixteen-section multi-detector row CT transverse and coronal reformations are equally sensitive and specific for diagnosis of appendicitis. Coronal reformations improve confidence in visualization of appendix (whether diseased or normal) and in diagnosis or exclusion of appendicitis. Acute appendicitis: added diagnostic value of coronal reformations from isotropic voxels at multi-detector row CT. Paulson EK et al. Radiology. 2005 Jun;235(3):879-85. 27
Helical CT has proven to be an excellent tool in the work-up of acute abdominal pain with a diagnostic accuracy for acute appendicitis of 93-99%. However, occasionally there are equivocal or false positive or negative cases, often due to nonvisualization of the appendix. The development of multi-detector row CT and recent advancements in reconstruction software has allowed rapid, high-resolution imaging of the entire abdomen and pelvis resulting in multiplanar reformations (MPR) with a spatial resolution similar to that of the axial plane. This article reviews the utility of CT in suspected acute appendicitis and the potential added diagnostic value of coronal reformations in confirming or excluding the diagnosis. MDCT of acute appendicitis: value of coronal reformations. Neville AM, Paulson EK Abdom Imaging. 2009 Jan-Feb;34(1):42-8. Coronal and sagittal CT head reformations improve the sensitivity and diagnostic confidence in the clinical setting of acute trauma. Overall, coronal and sagittal reformations improved diagnostic confidence and interobserver agreement over axial images alone for visualization of normal structures and in the diagnosis of acute abnormality. Subtle pathology detection with multidetector row coronal and sagittal CT reformations in acute head trauma. Emerg Radiol. 2010 Mar;17(2):97-102 Zacharia TT, Nguyen DT. Bone or Soft Tissue Pathology: Problem n Pathology may be overlooked as it is often edge of film diagnosis n Pathology may be overlooked as it was not part of the exam history n Axial imaging is limited for detecting spine pathology Bone or Soft Tissue Pathology: Solution n Attention to the bone and soft tissues as part of the image review n Routine sagital reconstruction created at the scanner and sent to PACs for rapid radiologist review 28
Most clinically important vertebral body compression fractures in nontrauma patients at risk for low bone mineral density may go unreported at abdominal multidetector CT if sagittal reconstructions are not routinely evaluated. Unreported Vertebral Body Compression Fractures at Abdominal Multidetector CT Carberry GA et al. Radiology 2013; 268:120-126 After review of 2015 abdominal multidetector CT scans in patients who underwent dual-energy x-ray absorptiometry (DXA) within 6 months of CT, prospective diagnosis of a moderate or severe vertebral body compression fracture was not determined in 84% (81 of 97). Unreported Vertebral Body Compression Fractures at Abdominal Multidetector CT Carberry GA et al. Radiology 2013; 268:120-126 Future Challenges in CT: Radiation Dose Issues n The need to minimize dose to patients is our highest priority but it does have certain consequences that we must all be aware of. Some that come to mind are; n While low dose protocols may not impact lesion detection in select protocols (i.e. lung nodule followup, prone virtual colonoscopy) they may pose significant challenges in other protocols (i.e. detection of liver metastases) Future Challenges in CT: Radiation Dose Issues n Although multiphase acquisition (i.e. various combinations of non-contrast, arterial, venous, delayed phase imaging) is not always necessary, in many cases it is critical to lesion detection and analysis. n Single phase acquisitions may limit our ability to define/detect disease In the daily radiology practice, the rate of interpretation error is between 3% and 4%; however, of the radiology studies that contain abnormalities, the error rate is even higher, averaging in the 30% range. Fool Me Twice: Delayed Diagnoses in Radiology With Emphasis on Perpetuated Errors Kim YW, Mansfield LT AJR 2014;202:465-470 In our study, the majority of errors made were errors of underreading (42%), where the finding was simply missed. We advocate the use of checklists for different types of radiologic examinations, depending on the body part imaged, to facilitate active search patterns to decrease the incidence of this type of error. Fool Me Twice: Delayed Diagnoses in Radiology With Emphasis on Perpetuated Errors Kim YW, Mansfield LT AJR 2014;202:465-470 29
We advocate the use of checklists for different types of radiologic examinations, depending on the body part imaged, to facilitate active search patterns to decrease the incidence of this type of error. Fool Me Twice: Delayed Diagnoses in Radiology With Emphasis on Perpetuated Errors Kim YW, Mansfield LT AJR 2014;202:465-470 Pearls to remember n Proper patient protocols is critical n Reconstruction of data into non-axial planes is critical n The Radiologist and the Radiologic Technologist need to work as a team n Errors will happen but we need to try to be 100% perfect 30