Construction and validation of a machine learning-based prediction model for social isolation in patients with colorectal cancer after stoma surgery.

objectiveTo investigate the current status and influencing factors of social alienation in patients with enterostomy after colorectal cancer surgery, and to establish a risk prediction model for social alienation in this population.

methodsA total of 507 enterostomy patients treated in the General Surgery (Colorectal Division), Oncology Department, and Wound Ostomy Clinic of a tertiary hospital in Jinzhou from March 2025 to January 2026 were enrolled by convenience sampling. The dataset was randomly divided into a training set and a validation set at a ratio of 7:3. Risk factors were screened by LASSO regression combined with Logistic regression. Five machine learning algorithms, including Logistic regression (LG), support vector machine (SVM), naive Bayes, K-nearest neighbor (KNN), and extreme gradient boosting (XGBoost), were employed to establish prediction models for social alienation. The predictive performance was evaluated and compared using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, precision, and F1-score. Calibration curves and decision curve analysis (DCA) were used to assess model fitness and clinical utility.

resultsAmong 507 included patients, 211 developed social alienation, with a prevalence rate of 41.6%. The SVM model achieved the optimal performance among the five models, with an AUC of 0.835 (95%CI: 0.772-0.899), sensitivity of 0.698, accuracy of 0.782, precision of 0.759, specificity of 0.841, and F1-score of 0.727. The calibration curve indicated favorable consistency between predicted and observed probabilities. The Hosmer-Lemeshow test yielded a χ

conclusionThe SVM model shows superior performance in predicting the risk of social alienation in enterostomy patients after colorectal cancer surgery and can serve as a screening tool for early identification in clinical practice.