Scientific Session 06 — SS06: Emergency/General RadiologyMonday, May 6, 2019
2502. Blind Spots on CT Imaging of the Head: Insights From 5 Years of Report Addenda at a Single Institution
Tu L*, Bokhari S. Yale School of Medicine, New Haven, Connecticut
Address correspondence to L. Tu (email@example.com)
Objective: The purpose of this study is to gain insight to those areas on CT imaging of the head where radiologists are most likely to produce an error of detection. We investigate reports in which a new finding was added via subsequent addendum to the radiology report. Analysis of the results allows us to produce a “heat map” of blind spots and provide recommendations for a search pattern that may increase sensitivity to the most common and significant misses on CT head imaging.
Materials and Methods: All reports for CT imaging of the head over a 5-year period were identified via CPT code within the reporting database at a single tertiary referral center. Keyword search for report addenda was used to identify reports in which contents were modified subsequent to time of initial dictation. These reports were reviewed and categorized based on the anatomic location of the detection error. The potential to alter patient management for each miss is also ascertained, and summary of the results is provided.
Results: A search of the institutional database yielded 165,943 reports for CT imaging of the head over the period of 5/1/2013–5/1/2018. Search for reports via keyword yielded 1658 (approximately 1%) of reports with addenda. Extraction and review of these yielded 394 reports (23.8% of all addenda) in which at least one error of detection was made. Of these, 35 reports described more than a single error in detection. In total there were 432 misses within the 394 reports; of these, 234 (54%) were deemed to have potential effect on management. A wide range of detection errors was seen. Common blind spots included the extracranial soft tissues (28 instances), in which misses at the incidentally imaged parotid glands (7 instances) or at the frontal scalp (9 instances) were of disproportionate incidence relative to anatomic size. Misses within the CSF spaces (51 instances) and brain/brainstem (65 instances) were well dispersed within anatomic sublocations. Among errors of detection in the boney structures (73 instances), findings at the nasal bones (15 instances) and occipital bones (15 instances) were overrepresented. Among CTA studies of the head, the distal portions of the internal carotid artery (8 instances) and portions of the middle cerebral artery distal to the M1 segment (9 instances) were the most common blind spots. Errors associated with vascular structures were also overrepresented at the sigmoid sinus and jugular bulb (together, 7 instances). Additional, fine-grained patterns of detection errors were also revealed and are discussed in full within the complete presentation.
Conclusion: These results provide insight into common blind spots in the reporting of CT of the head. A “heat map” of detection error locations is revealed, which is correlated with likelihood of clinically impactful findings. Modification of search patterns with special attention to these anatomic sites provides a model for accelerating the teaching of trainees and improving the practice of radiology.