Comparative Efficacy of Different AI Systems for Polyp Detection by Size During Colonoscopy: Systematic Review and Network Meta-Analysis.
J Med Internet ResSep 9, 2026 (epub)
Clinical ResearchGastroenterology
Ling Ba, Yaxin Qi, Xinrui Lv et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and network meta-analysis evaluated the efficacy of AI-assisted colonoscopy compared to standard colonoscopy for polyp detection by size in adults. Analyzing 13 RCTs with 4156 participants, the study found that AI showed a modest advantage in detecting diminutive polyps (≤5 mm) with a standardized mean difference of 0.21, but effects on larger polyps were minimal and uncertain.
Key findings
- AI-assisted colonoscopy showed a modest advantage for diminutive polyps (≤5 mm) with an SMD of 0.21.
- Effects on small (6-9 mm) and large (≥10 mm) polyps were minimal, with SMDs of 0.02 and 0.01, respectively.
- EndoScreener demonstrated the most consistent evidence for diminutive polyp detection with an SMD of 0.36.
- GRADE certainty was very low for diminutive polyps and low for small and large polyps.
Why it matters
Understanding the efficacy of AI systems in polyp detection can guide clinicians in optimizing colonoscopy practices and prioritizing resources for screening, particularly for diminutive polyps, which may enhance early colorectal cancer prevention efforts.