Document resource
Introduction Endoscopy remains standard-of-care for Barrett’s oesophagus (BO) surveillance. Capsule-sponge devices are a less operator-dependent, cost-effective alternative. We previously developed a capsule sponge-based risk stratification algorithm for BO surveillance (Lancet Onc 2022). Here, we compared neoplasia detection between sponge-based and standard endoscopic surveillance.Methods We included patients with non-dysplastic BO who underwent sponge and had endoscopic follow-up (2020–24) across 13 UK hospitals. Samples were processed centrally (Cyted Medical). Patients were stratified to low, moderate, or high-risk based on clinical (age, sex, segment length) and biomarker (p53, atypia, AUS) features. Follow-up endoscopy was guided by sponge risk and clinician discretion. Using 1:1 propensity-score-matching (nearest-neighbour, calliper 0.2) on age, sex and BE length, we compared HGD/cancer (primary) and any dysplasia (secondary outcome) detection between moderate/high-risk sponge patients and controls undergoing routine surveillance in 6 community hospitals. Low-risk patients were not considered to have sponge-detected dysplasia, even if dysplasia was found at follow-up endoscopy.Results 910 patients underwent sponge surveillance, and 966 controls underwent standard endoscopic surveillance. After 1:1 matching, 883 patients were included in each group with <10% standardized differences across matching variables, indicating good balance. Sponge groups detected 3.9% (34/883) HGD/cancer versus 1.8% (16/883) in standard surveillance (OR 2.17, 95% CI 1.19-3.96, p=0.012). For any dysplasia, detection rates were similar: 8.5% (75/883) vs 8.5% (75/883), respectively (OR 1.00, 95% CI 0.72–1.49, p=1.0).Conclusion Sponge-based surveillance detected more HGD/cancer than standard surveillance. While not designed to show superiority, this demonstrates that sponge-based surveillance can effectively detect HGD/cancer compared to current strategies.Abstract O69 Table 1Odds of primary and secondary outcomes in the propensity-score matched cohort*