S1274 Diagnostic Accuracy and Clinical Utility of Detachable String Magnetically Controlled Capsule Endoscopy (DS-MCE) Versus Conventional Esophagogastroduodenoscopy (EGD): A Systematic Review and Meta-Analysis

Introduction: Esophagogastroduodenoscopy (EGD) is the standard for upper GI evaluation, but it is invasive, often requiring sedation, and is poorly tolerated. Detachable String Magnetically Controlled Capsule Endoscopy (DS-MCE) offers a non-invasive alternative, enabling controlled esophageal assessment before transitioning to free gastric navigation. Its comparative diagnostic performance and feasibility remain underexplored. Methods: We conducted a systematic review and meta-analysis per PRISMA 2020. Databases (PubMed, Embase, Scopus, Cochrane) were searched through May 2024. Included were prospective studies comparing DS-MCE to EGD in adults. Primary outcomes were sensitivity, specificity, diagnostic odds ratio (DOR), AUC, and likelihood ratios for detecting esophagogastroduodenal disease, esophageal lesions, and esophagogastric varices. Secondary outcomes included technical success, Z-line visibility, recording time, discomfort, satisfaction, and adverse events. Bivariate and random-effects models were used. Results: Five prospective diagnostic accuracy studies involving 792 patients were included. For the primary outcome of overall esophagogastroduodenal disease detection, DS-MCE demonstrated a pooled sensitivity of 95.9% (95% CI: 90.9-98.2%) and specificity of 94.6% (87.4-97.8%), with an AUC of 0.971 and DOR of 233.3 indicating excellent diagnostic performance. Similar high accuracy was observed for esophageal lesion detection (sensitivity 96.0%, specificity 95.2%, AUC 0.973) and esophagogastric varices (sensitivity 97.3%, specificity 96.1%, AUC 0.979, DOR 475.4). The positive and negative likelihood ratios were favorable across all diagnostic endpoints. Technical success was 96.6%. Mean esophageal recording time was 4.91 minutes. Z-line visibility was 60.55% in all 4 quadrants, 51.16% in 3, and 57.47% in 2 quadrants. Visibility improved in newer studies despite heterogeneity (I² > 90%). DS-MCE was well tolerated, with a discomfort score of 1.71/10 and rare adverse events (0.16%). Descriptive satisfaction data showed consistent patient preference for DS-MCE over EGD. Conclusion: DS-MCE provides diagnostic accuracy comparable to EGD with excellent feasibility and tolerability. It is a promising non-invasive tool for screening and outpatient diagnostics. Further large-scale validation is needed. ChatGPT (OpenAI) was used for language clarity and grammar refinement. All scientific content, data, and analysis are original and solely by the authors.