JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Development and validation of a machine learning model to predict postoperative complications following radical gastrectomy for gastric cancer

  • 0Department of Gastrointestinal Surgery, Clinical Oncology School of Fujian Medical University and Fujian Cancer Hospital, Fuzhou, China.
Frontiers in Oncology +

|

September 24, 2025

|

September 24, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

Machine learning models and a dynamic nomogram were developed to predict postoperative complications after gastric cancer surgery. The Random Forest model and nomogram accurately identify high-risk patients for personalized interventions.

Area Of Science

  • Oncology
  • Medical Informatics
  • Surgical Oncology

Background

  • Postoperative complications significantly impact recovery and prognosis in gastric cancer patients undergoing radical gastrectomy.
  • Accurate prediction of these complications is crucial for optimizing patient outcomes.

Purpose Of The Study

  • To develop and validate machine learning (ML) models for predicting postoperative complications after radical gastrectomy.
  • To construct a clinically applicable dynamic nomogram for risk stratification and personalized intervention.

Main Methods

  • Retrospective analysis of 1,486 patients (training) and 498 (validation) undergoing radical gastrectomy.
  • Feature selection using Lasso regression, Boruta algorithm, and Recursive Feature Elimination (RFE).
  • Development and evaluation of six ML models: TreeBagger, Random Forest, SVM, XGBoost, GNB, and ANN, with subsequent nomogram construction.

Main Results

  • Random Forest (RF) model showed superior performance in both cohorts.
  • Identified independent risk factors: age, BMI, diabetes mellitus, ASA grade, operative time, and surgical approach.
  • The dynamic nomogram achieved AUCs of 0.805 (training) and 0.856 (validation), demonstrating reliable predictive capability.

Conclusions

  • The RF model offers optimal predictive accuracy, while the interpretable nomogram provides comparable discrimination and clinical utility.
  • Both tools enable early identification of high-risk patients, facilitating personalized interventions to improve postoperative recovery.

Keywords: