Machine learning model for predicting a high comprehensive complication index following rectal cancer surgery
Zhen Wang1, Lei Huang2, Liang He1
1Department of Gastrocolorectal Surgery, General Surgery Center, The First Hospital of Jilin University, Changchun, 130021, China.
Updates in Surgery
|September 22, 2025
Summary
This study developed a machine learning model to predict high postoperative complications after rectal cancer surgery using the Comprehensive Complication Index (CCI). The model identifies patients at high risk, enabling personalized treatment strategies and improved patient outcomes.
Area of Science:
- Surgical Oncology
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Postoperative complications significantly impact rectal cancer patient prognosis.
- The Comprehensive Complication Index (CCI) offers a more sensitive measure of severe complications than the Clavien-Dindo classification (CDC).
Purpose of the Study:
- To develop and validate a machine learning predictive model for high postoperative CCI in rectal cancer patients.
- To guide clinical practice by identifying patients at high risk for severe complications.
Main Methods:
- Utilized machine learning algorithms (Random Forest, LightGBM, Logistic Regression, Naive Bayes, XGBoost) on data from 1029 rectal cancer patients.
- Included preoperative, intraoperative, clinicopathological, and pelvic measurement data.
- Employed Shapley Additive exPlanations (SHAP) for model interpretability.
Main Results:
- The LightGBM model demonstrated optimal performance with AUCs of 0.746 (training), 0.760 (testing), and 0.709 (validation).
- Key predictors for high CCI included surgical time, interspinous distance, pelvic depth, age, diabetes, and tumor distance.
- The model showed excellent performance in predicting high CCI, validated by its superior DCA curve.
Conclusions:
- A robust predictive model for high CCI following rectal cancer anterior resection was successfully developed.
- The model facilitates personalized treatment strategies for high-risk patients, potentially improving prognosis.
- An online web tool is available for clinical application.
