Effectiveness of AI-enhanced colonoscopy: A case-control study using real world evidence in a young screening age population

Le résumé fourni par la source

Background: Computer aided detection (CADe) colonoscopy has been shown in randomized controlled trials to improve detection of adenomatous colorectal polyps. However, in real-world settings the benefits are less clear. Our objective was to assess CADe colonoscopy compared to routine colonoscopy in a clinical practice setting in the Arab region with a low screening age (40 years old). Methods: After the implementation of the CADe system (GI Genius, Medtronic) in our hospital in September 2021 until July 2023, all CADe cases were registered and matched to routine colonoscopy cases with (1:4) ratio. Assignment of patients to the CADe system versus the routine colonoscopy was pragmatic design based on scheduling and availability factors. Primary endpoints were adenoma per colonoscopy (APC), and adenoma detection rate (ADR). Results: A total of 840 patients were included, out of that 72 patients were excluded due to Boston Bowel Prep <6. The final analysis included 768 patients (161 with CADe, 607 without). The CADe system was found to significantly increase the mean APC (0.602 vs. 0.293; rate ratio 1.80 [1.40–2.32]; P < 0.001) and ADR, 30.4% with CADe versus 16.8% without CADe (adjusted RR 1.68 [1.27–2.22], P < 0.001). There was no significant difference in withdrawal time with CADe. Conclusion: In our study, CADe significantly increased the detection of adenomatous polyps in a young screening age population in a clinical practice setting. These findings support recent randomized controlled trials and contrast the findings of recent studies showing lack of benefit in real world practice.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.