Analysis of clinicopathological characteristics of early gastric cancers with different endoscopic curability categories after endoscopic submucosal dissection and related risk factors of non-curative resection.

[BACKGROUND AND AIMS] This single-center retrospective study aims to analyze the clinicopathological characteristics of early gastric cancer (EGC) treated with endoscopic submucosal dissection (ESD), stratified by endoscopic curability (eCura) categories. Additionally, we aim to identify the risk factors associated with non-curative resection (NCR) and the prediction model before ESD.

[METHOD] We included patients with EGC undergoing ESD at Peking University Third Hospital from 2010 to 2024. Statistical analyses included chi-squared test and logistic analysis. Multiple logistic regression analysis was used to investigate independent risk factors to predict NCR. Model performance was evaluated through receiver operating characteristic (ROC) curve analysis, with area under the curve (AUC) values quantifying predictive accuracy.

[RESULTS] A total of 540 individuals with 659 lesions were included. Univariate analysis identified smoking, H.pylori infection, upper/middle gastric location, 0-I/0-IIb-type, mucosal damage, tumor size > 30 mm, undifferentiated dominated histology, and submucosal (SM) invasion as significant predictors of NCR (all p < 0.05). Multivariate analysis revealed prior-ESD independent predictors (smoking, H.pylori infection, proximal location, EUS-confirmed SM invasion). The prior-ESD model demonstrated an AUC of 0.783 (95%CI: 0.708-0.858) with 80.0% sensitivity.

[CONCLUSION] This study developed a prior-ESD prediction model for assessing NCR risk, which includes comprehensive clinicopathological data. The model might potentially improve curative resection rates and guide clinical decision-making for EGC.