P0495 - Computer-Aided Detection vs Conventional Colonoscopy for Polyp Detection: A Systematic Review and Meta-Analysis of RCTs

Award: ACG Presidential Poster Award

Anis Halimi, MD¹, Sana Rabeeah, MD², Mohammed Abu-Rumaileh, MD², Bisher Sawaf, MD³, Muhammed Elhadi, MD⁴
1Badji Mokhtar University, Annaba, Annaba, Algeria; 2The University of Toledo, Toledo, OH; 3University of Toledo Medical Center, Toledo, OH; 4College of Medicine, Korea University, Seongbuk, Seoul-t'ukpyolsi, Republic of Korea

Introduction: Colonoscopy remains the gold standard for detecting colorectal cancer and precancerous lesions. Recent advances in artificial intelligence (AI) have enabled the development of computer-aided detection (CADe) systems to assist endoscopists by enhancing polyp identification. These technologies provide real-time visual alerts to improve detection rates. We aimed to evaluate the effect of CADe-assisted colonoscopy compared to conventional colonoscopy on polyp detection.

Methods: In this systematic review and meta-analysis, we searched PubMed, Embase, Web of Science, Scopus, and Cochrane Library up to April 2025 for randomized controlled trials comparing CADe-assisted with conventional colonoscopy. Study selection, data extraction, and risk-of-bias assessment were performed independently in duplicate. Primary outcomes were polyps per colonoscopy (PPC), adenomas per colonoscopy (APC), and advanced adenomas per colonoscopy. Secondary outcomes included polyp detection rate (PDR), adenoma detection rate (ADR), advanced adenoma detection rate (aADR), polyp miss rate (PMR), and adenoma miss rate (AMR). Random-effects meta-analyses were conducted with GRADE assessment.

Results: Fifty-three RCTs (n = 44,505) were included. AI-aided colonoscopy increased polyp detection by 0.22 per colonoscopy (incidence rate difference [IRD], 95% CI, 0.07–0.37) and adenoma detection by 0.21 per colonoscopy (IRD, 95% CI, 0.18–0.25). It significantly increased the polyp detection rate (PDR) (RR, 1.24; 95% CI, 1.18–1.30) and adenoma detection rate (ADR) (RR, 1.20; 95% CI, 1.15–1.25), while reducing the polyp miss rate (PMR) (RR, 0.48; 95% CI, 0.36–0.65) and adenoma miss rate (AMR) (RR, 0.51; 95% CI, 0.43–0.60). For advanced adenomas, AI increased detection by 0.01 per colonoscopy (IRD, 95% CI, 0.00–0.02) and improved the advanced adenoma detection rate (aADR) (RR, 1.08; 95% CI, 1.01–1.16).

Discussion: AI-aided colonoscopy significantly improved polyp and adenoma detection while reducing miss rates. The improvement in detection metrics across various polyp types, including advanced adenomas, suggests potential clinical benefits. Further studies are needed to assess cost-effectiveness and long-term benefits in colorectal cancer prevention.

Disclosures:
Anis Halimi indicated no relevant financial relationships.
Sana Rabeeah indicated no relevant financial relationships.
Mohammed Abu-Rumaileh indicated no relevant financial relationships.
Bisher Sawaf indicated no relevant financial relationships.
Muhammed Elhadi indicated no relevant financial relationships.

Anis Halimi, MD¹, Sana Rabeeah, MD², Mohammed Abu-Rumaileh, MD², Bisher Sawaf, MD³, Muhammed Elhadi, MD⁴. P0495 - Computer-Aided Detection vs Conventional Colonoscopy for Polyp Detection: A Systematic Review and Meta-Analysis of RCTs, ACG 2025 Annual Scientific Meeting Abstracts. Phoenix, AZ: American College of Gastroenterology.