Predicting bowel function after diverting stoma closure in patients with rectal cancer.

Background: Patients with rectal cancer undergoing laparoscopic anterior resection and diverting stomas often suffer from bowel dysfunction after stoma closure, impairing their quality of life. This study aims to develop a machine learning tool to predict bowel function after diverting stoma closure. Methods: Clinicopathological data and post-operative follow-up information from patients with mid-low rectal cancer after diverting stoma closure were collected and analyzed. Patients were randomly divided into training and test sets in a 7:3 ratio. A machine learning model was developed in the training set to predict major low anterior resection syndrome (LARS) and evaluated in the test set. Decision curve analysis (DCA) was used to assess clinical utility. Results: The study included 396 eligible patients who underwent laparoscopic anterior resection and diverting stoma in Tongji Hospital affiliated with Huazhong University of Science and Technology from 1 January 2012 to 31 December 2020. The interval between stoma creation and closure, neoadjuvant therapy, and body mass index were identified as the three most crucial characteristics associated with patients experiencing major LARS in our cohort. The machine learning model achieved an area under the receiver operating characteristic curve (AUC) of 0.78 [95% confidence interval (CI): 0.74-0.83] in the training set (n=277) and 0.74 (95% CI: 0.70-0.79) in the test set (n=119), and area under the precision-recall curve (AUPRC) of 0.73 and 0.69, respectively, with sensitivity of 0.67 and specificity of 0.66 for the test set. DCA confirmed clinical applicability. Conclusions: This study developed a machine learning model to predict major LARS in rectal cancer patients after diverting stoma closure, aiding their decision-making and counseling.