Machine Learning Estimation of Myocardial Ischemia Severity Using Body Surface ECG
Rui Jin1,2, Jake A Bergquist1,3,2, Deekshith Dade1
1Scientific Computing and Imaging Institute, University of Utah, SLC, UT, USA.
Computing in Cardiology
|September 26, 2025
Summary
Machine learning accurately predicts ischemic tissue volume from electrocardiograms (ECGs) in an acute myocardial ischemia (AMI) animal model. This breakthrough offers potential for improved early detection and risk stratification of AMI patients.
Area of Science:
- Cardiovascular Research
- Biomedical Engineering
- Artificial Intelligence in Medicine
Background:
- Acute myocardial ischemia (AMI) is a major global cause of cardiovascular mortality.
- Current early detection and risk stratification of AMI using electrocardiograms (ECGs) have limitations.
- Machine learning (ML) shows potential for advanced ECG analysis and disease detection.
Purpose of the Study:
- To apply ML techniques for predicting ischemic tissue volume directly from body surface ECGs.
- To address the lack of high-quality training data for ML in AMI.
- To enhance clinical risk stratification for AMI patients.
Main Methods:
- Development and application of ML networks using ECG recordings from an AMI animal model.
- Training ML models to correlate ECG signals with ischemic tissue volume.
- Validation of ML model performance using R-squared values.
Main Results:
- The developed ML networks achieved a robust prediction performance.
- An average R-squared value of 0.932 was obtained, indicating strong predictive accuracy.
- The study demonstrated the feasibility of predicting ischemic tissue volume from ECGs.
Conclusions:
- ML tools can effectively predict ischemic tissue volume from body surface ECGs in AMI.
- This approach holds promise for improving early detection and risk stratification of AMI.
- Further development and utilization of ML in cardiology can enhance patient outcomes.
