Machine Learning in Predicting Cardiac Events for ESRD Patients: A Framework for Clinical Decision Support
Chien-Wei Chuang1,2, Chung-Kuan Wu3,4,5, Chao-Hsin Wu1,2
1Graduate Institute of Business Administration, Fu Jen Catholic University, New Taipei City 242062, Taiwan.
Diagnostics (Basel, Switzerland)
|May 14, 2025
Summary
Machine learning models can predict major adverse cardiac events (MACEs) in patients with end-stage renal disease (ESRD). Key predictors include antiplatelet use, left ventricular hypertrophy, and serum albumin, enabling personalized treatment plans.
Area of Science:
- Nephrology
- Cardiology
- Artificial Intelligence
Background:
- Patients with end-stage renal disease (ESRD) face a heightened risk of major adverse cardiac events (MACEs).
- Accurate risk prediction and tailored interventions are crucial for managing MACEs in ESRD patients.
Purpose of the Study:
- To develop and evaluate machine learning (ML) models for predicting MACEs in ESRD patients.
- To identify key predictive features for MACE risk.
- To enhance clinical decision-making through improved risk assessment.
Main Methods:
- Utilized CatBoost, XGBoost, and LightGBM on a dataset with 84 variables (demographics, labs, comorbidities).
- Employed feature selection, cross-validation, and SHAP analyses for model interpretability.
- Assessed model performance using AUC, sensitivity, and specificity.
Main Results:
- CatBoost achieved the highest predictive performance with an AUC of 0.745.
- Significant predictors of MACEs included antiplatelet use, left ventricular hypertrophy grade, and serum albumin.
- SHAP analysis improved model interpretability, aiding clinician-led risk stratification.
Conclusions:
- ML models show promise for enhancing MACE risk assessment in ESRD patients.
- Integrating explainable AI into clinical workflows can support personalized treatment planning.
- Future integration with EHR systems could enable real-time decision-making and improve patient outcomes.
