S538 Comparative Efficacy of Diagnostic and Interventional Endoscopic Techniques in Colorectal Adenoma Management: A Meta-Regression and Parallel Meta-Analysis of Randomized Controlled Trials

Introduction: Colorectal cancer (CRC) ranks fourth globally in terms of cancer-related mortality. The most common approach for screening colon cancer is colonoscopy. Real-time, highly accurate computer-aided detection (CADe) of polyps is made possible through the deep learning algorithms. Conversely, interventional techniques such as underwater endoscopic mucosal resection (UEMR) and conventional endoscopic mucosal resection (CEMR) are used for large, non-pedunculated colorectal lesions ensuring complete excision and prevent recurrence. Little is known about the overall efficacy of therapeutic and diagnostic modalities. A parallel meta-analysis of diagnostic and interventional techniques could optimize colorectal disease management and enhance patient outcomes. This study aims to assess the efficacy of diagnostic (CAD vs standard colonoscopy) and interventional techniques (CEMR and UEMR) in colorectal adenomas. Methods: A parallel meta-analysis was conducted following PRISMA guidelines. Studies assessing diagnostic (CADe) and interventional techniques (UEMR, CEMR) for colorectal adenomas were included. Searches were performed across PubMed, Embase, and Google Scholar. Primary outcomes included adenoma detection rate (ADR) for diagnostic and en bloc resection rate for interventional studies. Random-effects models computed 95% confidence intervals (CIs) and pooled odds ratios (ORs). Heterogeneity was assessed using the I² statistic. A binary moderator variable (diagnostic vs interventional) was used in meta-regression to assess relative efficacy. Results: A significant improvement in the ADR over routine colonoscopy was shown in 5 RCTs assessing CADe (pooled odds ratio [OR]: 1.44, 95% CI: 1.28–1.61; P < 0.00001; I² = 0%). Separately, 5 RCTs comparing UEMR with CEMR showed that UEMR significantly increased en bloc resection rates (pooled OR: 2.18, 95% CI: 1.60–2.99; P < 0.00001; I² = 0%). Meta-regression revealed a significant difference in effect sizes (β = –0.44, SE = 0.17, P = 0.008). Diagnostic arm showed greater efficacy (OR = 2.22) compared to the interventional group (OR = 0.65), reflecting a 35% lower effect. Conclusion: Colonoscopy outcomes were significantly improved by both diagnostic and interventional approaches. However, diagnostic techniques (CADe) demonstrated a larger effect size compared to interventional methods (UEMR), indicating that diagnostic advancements may have a greater impact on procedural efficacy and highlights the need for tailored endoscopic strategies based on procedural goals.