Narrative Review on Explainable Artificial Intelligence for Multimodal Risk Stratification in Coronary Artery...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Imaging Analysis

Background:

• Coronary artery disease (CAD) poses a significant global health challenge, necessitating advanced risk stratification methods.
• Current risk assessment tools require augmentation for improved accuracy and personalized treatment strategies.

Purpose of the Study:

• To evaluate the impact of explainable artificial intelligence (AI) on multimodal risk assessment for coronary artery disease (CAD).
• To determine if AI can enhance diagnostic accuracy and early risk prediction in CAD management.

Main Methods:

• Systematic literature review of 299 articles on AI, machine learning, deep learning, and CAD, adhering to PRISMA standards.
• Inclusion of 94 studies focusing on AI applications in analyzing cardiac imaging (echocardiography, CT, MRI) and clinical data.
• Analysis of AI techniques including convolutional neural networks, ensemble methods, and natural language processing (NLP).

Main Results:

• Advanced AI tools, including deep learning and NLP, effectively integrate diverse data sources (ECG, clinical records, imaging) for improved CAD diagnosis.
• AI excels in detecting structural abnormalities, assessing myocardial blood supply, and identifying inflammatory markers from cardiac imaging.
• AI models demonstrate superior accuracy, sensitivity, and specificity compared to traditional risk scores like the Framingham Risk Score.

Conclusions:

• Explainable AI offers a transparent approach, facilitating clinical adoption and trust in AI-driven cardiovascular risk stratification.
• AI represents a significant advancement in data-driven cardiology, enabling more accurate risk prediction and personalized patient care for CAD.
• The integration of AI promises to revolutionize cardiovascular medicine through enhanced diagnostic capabilities and tailored treatment strategies.