AI-assisted second forward-view examination of the right colon significantly improves the adenoma detection rate: a...

Adenomas in the right colon are likely to be missed during colonoscopy because of anatomical complexity and lesion characteristics. Although reinspection strategies and artificial intelligence (AI)-assisted colonoscopy improve adenoma detection, prospective randomized studies of their combined effectiveness in the right colon are limited. This multicenter randomized controlled trial aimed to compare standard colonoscopy with an AI-assisted second forward-view examination of the right colon to determine whether this strategy improves the adenoma detection rate (ADR).

Patients undergoing colonoscopy were randomly assigned in a 1:1 ratio to either the AI-assisted colonoscopy group (AI group) or the standard colonoscopy group (non-AI group). In the AI group, a second forward-view examination of the right colon was performed using real-time AI assistance. The primary outcome was the right colon ADR.

A total of 606 patients were included in the analysis (AI group, n = 305; non-AI group, n = 301). The right colon ADR in the AI group was significantly higher than that in the non-AI group (36.4% vs. 27.9%; p = 0.03; absolute increase, 8.5%). The improvement was driven by increased detection of small, flat lesions and those in anatomically challenging regions, such as the ascending colon. Subgroup analyses showed consistent improvement in the right colon ADR with AI assistance regardless of endoscopist experience.

The AI-assisted second forward-view examination significantly improved the right colon ADR. This combined strategy may improve adenoma detection in the right colon regardless of endoscopist experience. However, the independent contribution of AI assistance warrants further investigation.