Gastroenterology trainees can easily learn histologic characterization of diminutive colorectal polyps with narrow band imaging.

Background & Aims: Little is known about how teaching gastroenterology trainees polyp patterns by using narrow band imaging (NBI) affects their ability to characterize the histology of diminutive colorectal polyps. We developed and tested a tool to teach trainees to characterize the histology of diminutive polyps by using NBI.

Methods: Twelve gastroenterology trainees with varying levels of colonoscopy experience watched a teaching tool that described the NBI criteria to distinguish polyp histology. The trainees then watched 80 videos of NBI examination of diminutive polyps, recording their predictions of polyp histology and their degree of confidence. After each video, an expert provided feedback about actual polyp histology and the NBI criteria that supported each diagnosis. Twelve weeks later, without training or feedback during the interval, the trainees watched the same videos and predicted histologies of the polyps. Performance was evaluated by comparing predicted classification with actual histologic findings. Cumulative sum analysis was used to determine the learning curve for each trainee.

Results: Trainees made significant improvements in accuracy and the proportion of high-confidence predictions as they progressed through video blocks during the first session (P < .001). With active feedback, all trainees predicted polyp histologies with >90% accuracy, with a negative predictive value >90% for adenomatous histology. A median of 49 videos was required to achieve competency. For diagnoses made with high confidence, trainee performance exceeded 90% during the first and second sessions. Interobserver agreement was substantial (session 1, κ = 0.71; session 2, κ = 0.70).

Conclusions: We developed a computer-based tool, combined with short videos and active feedback, to train gastroenterologists to identify polyp histology by using NBI. After training, gastroenterology trainees characterized the histology of diminutive polyps with ≥ 90% accuracy.