A nomogram to predict submucosal fibrosis in early gastric cancer undergoing endoscopic submucosal dissection

Endoscopic submucosal dissection (ESD) is an effective treatment for early gastric cancer (EGC), but submucosal fibrosis remains a major obstacle to successful resection. This study aimed to identify risk factors and construct a clinically applicable nomogram for predicting submucosal fibrosis in EGC, and to evaluate its impact on ESD outcomes. We retrospectively analyzed 264 lesions from 251 patients who underwent ESD between January 2012 and December 2024. Patients were randomly assigned to a training cohort (n = 184) and a validation cohort (n = 80) in a 7:3 ratio. A nomogram was constructed using multivariate logistic regression. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC), calibration curves, the Hosmer-Lemeshow test, and decision curve analysis (DCA). Histologic assessment of the submucosal fibrosis was performed using Masson's trichrome staining. Independent predictors of endoscopic submucosal fibrosis included tumor size greater than 30 mm (OR  4.041, 95% CI 1.412 ~ 11.560, P = 0.009), depressed-type tumor (OR 3.713, 95% CI 1.613 ~ 8.546, P = 0.002), submucosal invasion (OR 4.804, 95% CI 1.369 ~ 16.858, P = 0.014) and tumor location in the middle (OR   11.630, 95% CI 4.243 ~ 31.879, P < 0.001) and upper third (OR  9.967, 95% CI 3.589 ~ 27.685, P < 0.001) of the stomach. The nomogram yielded an AUC of 0.880 (95% CI 0.828 ~ 0.932) in the training cohort and 0.864 (95% CI 0.786 ~ 0.941) in the validation cohort. The calibration curve showed excellent consistency between the nomogram predictions and actual observations. DCA showed a positive net benefit for the nomogram. Increased fibrosis severity was associated with lower en bloc and curative resection rates, higher rates of bleeding and perforation, and prolonged procedure time. Concordance between endoscopic and histologic fibrosis grading was high (Cohen's κ = 0.857, P < 0.001). We developed a robust and interpretable nomogram for the preoperative prediction of submucosal fibrosis in EGC. This tool can aid endoscopists in risk stratification, operator selection, and complication avoidance during ESD.