2023 ARRS ANNUAL MEETING - ABSTRACTS

RETURN TO ABSTRACT LISTING


E1519. Pitfalls in CT Neuroangiography: A Dozen Cases Highlighting Error Types and Biases in Diagnostic Radiology
Authors
  1. Vivek Batheja; George Washington University
  2. Jhanavi Rao; George Washington University
  3. Kathleen Johnson; George Washington University
  4. Sheida Hedjazi; Texas A&M University
  5. Ramin Javan; George Washington University
Background
The purpose of this exhibit is to highlight a dozen CTA head/neck cases that created significant clinically consequential diagnostic challenges to resident trainees during independent call. Furthermore, we believe that awareness of the various types of errors and biases in diagnostic radiology can enhance the skill of a radiologist to a level beyond knowledge-based competency. The provided examples showcase concepts such as satisfaction of search error, edge of film error, framing bias and automation bias.

Educational Goals / Teaching Points
We provide a brief introduction to the types and classifications of errors (Renfrew and Brook) and biases in radiology. Through multiple examples of CTA head/neck cases we demonstrate that one’s awareness of these could serve as an additional tool for increasing diagnostic accuracy and improved decision making. This phenomenon has been termed by some as cognitive debiasing, which includes succession of stages from precontemplation, contemplation, preparation, action, and maintenance.

Key Anatomic/Physiologic Issues and Imaging Findings/Techniques
Vertebral artery dissection only present in an unexpected course of a tortuous V3 (search error). Inferiorly projecting cavernous ICA aneurysm embedded entirely in the cavernous sinus causing cranial III nerve palsy (perceptual error). Cutoff of the mid aspect of the superior cerebellar artery, the detection of which was guided by CT perfusion images (representativeness bias). Venous contamination due to delayed contrast bolus timing leading to vein mimicking the continuation of the course of an artery despite a cutoff (technique error). Arterial cutoff at the origin of the M2 branch of MCA (perceptual error). Miscalculation of core and penumbra on CT perfusion due to excluding MCA branches near the vertex during perfusion scanning (automation bias). Congenital absence of the cervical internal carotid artery mimicking occlusion/thrombosis (zebra retreat bias); pseudolesion of pterygoid venous plexus mimicking active extravasation with clinical history of strangulation (framing bias). Acute subarachnoid hemorrhage due to ruptured anterior communicating artery aneurysm mimicking A2 segment of ACA in the setting of presence of an MCA bifurcation aneurysm (perceptual error, satisfaction of search error). Cerebellar mass despite normal neurovasculature in a patient with stroke-like symptoms (framing bias, search error). Presence of thrombus in M1 segment of MCA in addition to cervical ICA thrombus (satisfaction of search). Lung mass on trigger monitoring scout images (edge of film)

Conclusion
Timely diagnosis of critical findings on CT neuroangiography is of utmost importance with respect to clinical outcomes. Given the complexity of neurovascular anatomy and the high level of detail, these studies present some of the most challenging examinations during call for radiology trainees. In addition to the commonly taught pathologic conditions such as vascular cutoffs, dissections and aneurysms, familiarity of trainees with various types of error and bias may assist in minimizing diagnostic pitfalls in theses complex yet consequential types of studies.