An Integrated Circumferential Proportion and Defect Length Index for Predicting Esophageal Stricture After ESD in...

Objective: To compare the circumferential proportion of mucosal defects, defect length, and the length-to-residual mucosal proportion (LRP) index for predicting esophageal stricture after endoscopic submucosal dissection (ESD) in patients with mucosal defects ≥6 cm, and to explore LRP in defects >10 cm. Methods: This retrospective study included 378 patients undergoing esophageal ESD from February 2014 to June 2020, grouped by stricture status. LRP was defined as defect length divided by residual mucosal proportion, with residual mucosal proportion set at a lower bound of 0.05 for statistical calculation. Receiver operating characteristic analysis compared the three indicators, including subgroup analysis for defects >10 cm. Results: Postoperative stricture occurred in 82 patients (21.7%). Circumferential proportion, defect length, and LRP were higher in the stricture group than in the non-stricture group (all p < 0.001). In the overall cohort, circumferential proportion and LRP showed comparable discrimination (AUC 0.877 vs. 0.876; p = 0.825), both outperforming defect length alone (AUC 0.694; both p < 0.001). In defects >10 cm, LRP showed the highest numerical AUC (0.913), followed by circumferential proportion (0.886) and defect length (0.668). Multivariable analysis confirmed that LRP was independently associated with stricture (per 10-unit increase: OR = 1.225; 95% CI, 1.150-1.304; p < 0.001). Conclusions: LRP was comparable to circumferential proportion and may serve as an exploratory complementary metric in long-segment defects, warranting external validation.