S3086 AI-Assisted Versus Conventional Video Capsule Endoscopy for Detection of Small Intestinal Lesions: A Systematic Review and Meta-analysis

Introduction: Video capsule endoscopy (VCE) is the gold standard for evaluating the small intestine but is limited by its time-consuming nature and reader fatigue, leading to variability in interpretation. Artificial intelligence (AI) has shown promise in enhancing lesion detection and diagnostic performance. This meta-analysis aimed to compare the diagnostic performance of AI-assisted versus conventional VCE in detecting small intestinal lesions. Methods: A systematic search of MEDLINE, Embase, and Cochrane databases was conducted from inception through May 2025 for studies comparing diagnostic performance of AI-assisted VCE to conventional VCE in small bowel pathologies. The primary outcome was diagnostic sensitivity. Secondary outcomes included accuracy, specificity, positive predictive value (PPV), and negative predictive value (NPV). Pooled estimates and 95% confidence intervals (CIs) were calculated using a random-effects model with the metafor package in R. Between-study heterogeneity was assessed using the I² statistic. Results: Sixteen studies met the inclusion criteria. AI-assisted VCE demonstrated a higher pooled diagnostic sensitivity of 0.94 [95% CI: 0.90–0.97] compared to 0.84 [95% CI: 0.75–0.93] with conventional VCE. Pooled diagnostic accuracy was 0.92 [95% CI: 0.88–0.96] for AI-assisted VCE and 0.96 [95% CI: 0.94–0.98] for conventional VCE. Specificity was higher with conventional VCE (0.99 [95% CI: 0.99–1.00]) than with AI-assisted VCE (0.93 [95% CI: 0.86–0.99]). AI-assisted VCE showed higher negative predictive values, while conventional VCE had higher positive predictive values. Heterogeneity was moderate to high (I²: 45%–75%), likely due to variations in AI algorithms and lesion types. Conclusion: This meta-analysis shows that AI-assisted video capsule endoscopy significantly improves sensitivity and NPV, reducing missed lesions and supporting its integration into clinical workflows to enhance diagnostic accuracy and efficiency in small bowel evaluations.