A Multifactor Prediction Model for Non-curative Outcomes in Mixed-type Early Gastric Cancer

The goal of this dual-center study is to identify the most valuable predictive factors (MVPs) for non-curable mixed-type early gastric cancer (NC-MTEGC) and develop a nomogram scoring model to assist surgeons in formulating precise postoperative combined radiochemotherapy strategies in patients with mixed-type early gastric cancer (MTEGC) who have undergone radical surgical resection. The main question it aims to answer is: What are the most valuable predictive factors for NC-MTEGC, and can a nomogram scoring model developed based on these factors effectively assist in formulating precise postoperative combined radiochemotherapy strategies? Patients with MTEGC who have undergone radical surgical resection (including 160 in the training group, 151 in the internal validation set from the First Affiliated Hospital of Nanchang University, and 110 in the external test cohort from the Second Affiliated Hospital of Nanchang University) will be included in the study. The Least Absolute Shrinkage and Selection Operator (LASSO) algorithm will be used to assess key predictive indicators, a nomogram prediction model will be developed based on logistic regression, and an NC-MTEGC risk score model will be constructed. Meanwhile, the model's discriminatory ability, calibration, and clinical utility will be comprehensively validated across the three cohorts, with follow-up for relevant conditions.