Diagnosis of each main coronary artery stenosis based on whale optimization algorithm and stacking model

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence in Medicine

Background:

• Cardiovascular disease (CVD) poses a significant global health burden, with coronary artery disease (CAD) being a primary contributor.
• CAD involves the narrowing of major coronary arteries: left anterior descending (LAD), left circumflex (LCX), and right coronary artery (RCA).
• Accurate and early diagnosis of CAD is crucial for effective patient management and improved outcomes.

Purpose of the Study:

• To develop and evaluate a novel computer-aided diagnosis (CAD) model for the accurate detection and prediction of coronary artery disease.
• To enhance feature selection for CAD diagnosis using a k-nearest neighbor (KNN)-based whale optimization algorithm (WOA).
• To improve diagnostic performance by integrating selected features into a two-layer stacking model.

Main Methods:

• Utilized the whale optimization algorithm (WOA) with k-nearest neighbor (KNN) for optimal feature selection in coronary artery diagnosis.
• Implemented a binary conversion threshold within WOA to identify optimal feature subsets for LAD, LCX, and RCA.
• Developed a two-layer stacking model incorporating the selected features for diagnosing LAD, LCX, and RCA.

Main Results:

• The proposed KNN-based WOA method successfully selected 17 optimal features for each main coronary artery.
• Achieved high classification accuracies: 89.68% for LAD, 88.71% for LCX, and 85.81% for RCA on test sets.
• Demonstrated superior performance compared to other metaheuristic feature selection methods and machine learning algorithms on the Z-Alizadeh Sani dataset.

Conclusions:

• The KNN-based WOA effectively identifies optimal feature subsets for coronary artery disease diagnosis.
• The developed stacking model, utilizing these features, significantly improves the accuracy of CAD diagnosis for major coronary arteries.
• This computer-aided approach shows promise for enhancing the early detection and prediction of cardiovascular disease.