Related Experiment Video
Updated: May 15, 2026
07:13
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
Development and internal validation of a machine learning-based model for predicting postoperative complications
Zhang Shuo1, Du Chen Hui1, Zhang Qing Long1
1Department of Liver Transplantation & Laparoscopic Surgery, The First Affiliated Hospital of Xinjiang Medical University, Urumqi, Xinjiang, 830054, China.
BMC Surgery
|May 14, 2026
Summary
A random forest (RF) model accurately predicts major complications after liver cancer surgery. Key predictors include liver stiffness, surgical approach, albumin, and blood loss, aiding clinical decision-making.
Area of Science:
- Hepatobiliary surgery
- Surgical oncology
- Machine learning in medicine
- Predictive analytics
Background:
- Primary liver cancer resection carries significant risks of major postoperative complications.
- Accurate prediction of these complications is crucial for patient management and surgical planning.
- Machine learning offers potential for developing robust risk prediction models.
Purpose of the Study:
- To identify risk factors for major postoperative complications following primary liver cancer resection.
- To develop and compare machine learning models for predicting these complications.
- To evaluate the optimal model's clinical utility for perioperative risk stratification.
Main Methods:
- Retrospective analysis of 2,389 patients undergoing primary liver cancer resection.
- Identification of robust predictors using LASSO, XGBoost RFE, and Boruta algorithms.
- Development and comparison of seven machine learning models (including Random Forest) with Bayesian optimization; performance evaluated using AUC, Brier score, and DCA.
Main Results:
- Eight key predictors identified: surgical approach, ALT, IBL, LSM, PT, total bilirubin, ALB, and intraoperative blood transfusion.
- The Random Forest (RF) model demonstrated superior performance (AUC 0.843) in predicting major complications.
- SHAP analysis highlighted LSM, surgical approach, ALB, and IBL as most influential predictors; DCA confirmed RF's clinical benefit.
Conclusions:
- The developed RF model accurately predicts major postoperative complications after liver cancer resection.
- This model serves as a valuable clinical decision-support tool for perioperative risk stratification.
- It facilitates individualized patient management strategies to improve outcomes.
