Development and validation of a nomogram for predicting moderate-to-severe complications following primary tumor resection in metastatic colorectal cancer.

Background: Colorectal cancer has high incidence and mortality. Surgery is the primary curative treatment, but postoperative complications remain common. This study developed and validated a nomogram to predict moderate-to-severe complications after primary tumor resection(PTR) in metastatic colorectal cancer (mCRC). Method: A retrospective analysis of clinical data was conducted for mCRC patients undergoing PTR at our institution between January 2022 and December 2024. All patients were randomly divided into two groups: 70% for development and 30% for validation. Univariate and multivariate logistic regression analyses were conducted to identify the independent risk factors associated with moderate-to-severe complications occurring within 30 days postoperatively. Correlation heatmaps and Lasso regression analysis were employed to systematically screen and identify the most relevant variables. Subsequently, a nomogram was developed based on the significant predictors. The area under the curve (AUC) was determined based on the receiver operating characteristic (ROC) curve for assessing the predictive probability. A calibration curve was generated to contrast the predicted probability against the observed probability. The clinical utility of the nomogram was evaluated using decision curve analysis (DCA). Internal 10-fold cross-validation was performed using bootstrapping, and boxplots as well as the average calibration curve were generated to visualize the results. Results: A total of 404 mCRC patients receiving PTR treatment were enrolled, including 282 in the development group and 122 in the validation group. Of these, 32% (90) in the development group and 39% (47) in the validation group experienced moderate-to-severe postoperative complications. Multivariate Logistic regression analysis identified age (p = 0.017, OR = 1.041, 95% CI: 1.007-1.076), preoperative albumin level (p < 0.001, OR = 0.774, 95% CI: 0.704-0.851), tumor location (p = 0.012, OR = 2.243, 95% CI: 1.216-4.906), and operation duration (p < 0.001, OR = 1.007, 95% CI: 1.003-1.011) as independent risk factors for moderate-to-severe complications after PTR surgery. Based on these findings, a nomogram was developed and validated. Conclusion: This study identified four independent risk factors for moderate-to-severe complications in mCRC patients after PTR surgery and developed a reliable predictive model to assist surgeons in optimizing perioperative management for high-risk cases.