Colonoscopy with narrow-band imaging can allow real-time determination of polyp histology.

To determine whether physicians with varying levels of experience can learn and apply endoscopic criteria to distinguish between adenomas and hyperplastic polyps.

Prospective observational study.

University teaching hospital.

This study involved 37 physicians (medical residents, N = 12; gastroenterology fellows, N = 12; and gastroenterology faculty, N = 13).

Small-group, 20-minute, didactic teaching sessions in which the endoscopic criteria for determining polyp histology by using narrow-band imaging were described and demonstrated.

Learning outcomes were evaluated by using written pretests and posttests in which participants scored pathologically verified, high-definition polyp photographs as adenomas or hyperplastic polyps.

The mean overall scores increased significantly from 47.6% correct on the pretest to 90.8% correct on the posttest (P = .0001). The overall mean percentage of responses answered don't know was significantly lower on the posttest (0.6%) compared with the pretest (20.5%, P < .0001). After training, the level of agreement was substantial (κ = 0.69 for all participants, κ = 0.79 for fellows).

Our study did not assess for sustained improvement with time or in vivo accuracy of histological prediction during live colonoscopy. Further validation in a sample of community physicians is required.

A short, didactic teaching session can achieve high accuracy and good interobserver agreement in the use of narrow-band imaging for determining the histology of colorectal polyps.