Targeted Training Curbs Radiologist Errors in Contrast-Enhanced CT Analysis

A targeted training approach can help reduce radiologist mistakes when assessing contrast-enhanced CT for liver metastases. However, certain types of errors still persist, researchers detailed in Academic Radiology

Radiologists occasionally miss low-contrast lesions, including hepatic metastases and pancreatic adenocarcinoma, on computed tomography scans. Prior research concentrated on refining image reconstruction techniques and optimizing acquisition protocols to address this challenge. Researchers at the Mayo Clinic devised a novel education program to potentially rectify this issue, yielding a mixed outcome.

This approach led to a reduction in "search errors," instances where readers fail to focus on the lesion. However, it did not effectively address "classification" errors, where radiologists identify the problem but neglect to report it as suspicious.

The study suggests that targeted training can enhance performance even for routine tasks, like identifying low-contrast hepatic metastases. This improvement might stem from altered interaction patterns between readers and computer workstations, as noted by Scott S. Hsieh, PhD, from Mayo's Department of Radiology, and colleagues. Further research is needed to identify training methods for enhancing classification errors, assess the sustainability of reader improvements following targeted training, and compare the effectiveness of this approach with conventional training.

The Mayo Clinic enlisted 31 radiologists from a single site to participate in the program, which included abdominal imaging specialists, non-specialists, and senior residents/fellows. The sessions occurred in early 2022 and lasted around four hours. The intervention consisted of a pretest exam, search and classification training, and a subsequent test. During the search phase, radiologists interpreted up to 30 abdominal CTs, receiving eye-tracker feedback indicating where they focused and lesion locations. For classification education, radiologists reviewed up to 100 "patches" (cropped-down key CT slices) and rated lesions on a 100-point scale, comparing their ratings with expert readers' findings.

The study revealed a decline in search errors from 11% before the intervention to 8% afterward. However, there was no noticeable change in classification errors or diagnostic accuracy measures adjusted for reader confidence. A subgroup analysis indicated no evidence of change among abdominal subspecialists.

The study's authors noted that the classification training dosage might have been insufficient to yield a measurable improvement. They suggested that search training might hold greater promise in reducing the rate of missed hepatic metastases than classification training.

----