Long-Term Coronary Artery Disease Risk Prediction with Machine Learning Models

Area of Science:

• Cardiology
• Medical Informatics
• Machine Learning

Background:

• Coronary artery disease (CAD) arises from atherosclerotic plaques narrowing heart arteries, obstructing blood flow.
• Early detection and prevention are crucial for managing this life-threatening condition.
• Accurate long-term risk prediction is essential for timely intervention and patient management.

Purpose of the Study:

• To evaluate and compare various machine learning (ML) models for long-term coronary artery disease (CAD) risk prediction.
• To assess the impact of the synthetic minority oversampling technique (SMOTE) on model performance.
• To identify the optimal ML model for predicting CAD risk.

Main Methods:

• Experimentation with multiple machine learning (ML) models.
• Application and evaluation of the synthetic minority oversampling technique (SMOTE).
• Utilizing 10-fold cross-validation to assess model accuracy, precision, recall, and Area Under the Curve (AUC).

Main Results:

• The stacking ensemble model, combined with SMOTE and 10-fold cross-validation, demonstrated superior performance.
• This model achieved an accuracy of 90.9%, precision of 96.7%, recall of 87.6%, and an AUC of 96.1%.
• The results indicate the effectiveness of SMOTE in improving ML model performance for CAD risk prediction.

Conclusions:

• Machine learning models, particularly the stacking ensemble with SMOTE, show significant promise for accurate long-term coronary artery disease risk prediction.
• The findings suggest that SMOTE can enhance the predictive power of ML models in this domain.
• This approach offers a valuable tool for improving cardiovascular health outcomes through early risk identification.