Related Experiment Video
Updated: Sep 18, 2025
08:49
Author Spotlight: Enhancing Graft Viability Assessment Through Quantitative Metrics and Innovative Reservoir Systems
Published on: August 2, 2024
1.0K
Machine learning-based hybrid risk estimation system (ERES) in cardiac surgery: Supplementary insights from the ASA
Ayşe Banu Birlik1,2, Hakan Tozan3, Kevser Banu Köse4
1Department of Healthcare System Engineering, Graduate School of Engineering and Natural Sciences, Istanbul Medipol University, Istanbul, Turkey.
PLOS Digital Health
|June 23, 2025
Summary
A new machine learning model, the Ensemble-Based Risk Estimation System (ERES), accurately predicts cardiac surgery mortality risk. ERES outperforms traditional models, improving patient outcomes and clinical decision-making in cardiac surgery patients.
Area of Science:
- Cardiovascular Surgery
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Accurate prediction of postoperative mortality risk is crucial for improving patient outcomes after cardiac surgery.
- Traditional risk models like EuroSCORE I have limitations in capturing complex variable interactions, leading to suboptimal performance in specific patient groups.
Purpose of the Study:
- To develop and validate the Ensemble-Based Risk Estimation System (ERES), a novel machine learning model for enhanced mortality prediction in patients undergoing coronary artery bypass grafting and/or valve surgery.
- To compare the predictive performance and clinical utility of ERES against the established EuroSCORE I model.
Main Methods:
- Retrospective analysis of 543 cardiac surgery patients using six machine learning algorithms on preoperative clinical data.
- Application of feature selection techniques (Gini importance, Recursive Feature Elimination, Adaptive Synthetic Sampling) to enhance accuracy and address class imbalance.
- Utilized SHAP analysis and decision curve analysis to identify key predictors and assess clinical utility.
Main Results:
- The ERES model, incorporating 15 key features, demonstrated superior predictive performance compared to EuroSCORE I.
- Calibration plots showed more accurate probability estimates with ERES.
- SHAP analysis identified creatinine, age, and left ventricular ejection fraction as the most significant predictors. Decision curve analysis confirmed ERES's superior clinical utility.
Conclusions:
- The Ensemble-Based Risk Estimation System (ERES) offers enhanced accuracy and clinical utility for predicting postoperative mortality risk in cardiac surgery patients.
- Integrating advanced machine learning models like ERES into clinical practice has the potential to improve decision-making and patient outcomes.
- External validation is recommended for broader implementation of ERES in clinical settings.
