A multicenter international randomized controlled trial involving 1,407 patients found that the DEEP2 AI-powered computer-aided detection (CADe) system significantly increased the number of adenomas detected per colonoscopy (APC) compared with standard high-definition colonoscopy, with the greatest benefit observed in community-based endoscopy practices, where nearly 86% of procedures were performed.
While the overall improvement in adenoma detection rate (ADR) was not statistically significant, post hoc analysis showed meaningful gains in community clinics, suggesting AI may help reduce variability in colonoscopy quality outside academic centers. As colorectal cancer screening expands to younger populations, the findings reinforce the potential of AI-assisted colonoscopy to improve detection of precancerous lesions, standardize performance across practice settings, and enhance the effectiveness of routine colorectal cancer screening.