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

With a slowing of cardiovascular health, the continuous effect of coronary artery disease (CAD) calls for the implementation of imaginative methods for more reliable risk segmentation and treatment. This research evaluates the extent to which explainable artificial intelligence (AI) can radically augment multimodal risk assessment of CAD. A systematic search was performed on PubMed and Google Scholar that returned 299 articles on AI, machine learning, deep learning, and CAD by using timely search terms. Adhering to PRISMA standards, 94 studies were included for a comprehensive review within a tight-rigorous selection process. The study reveals that high-level AI tools, such as convolutional neural networks and ensemble techniques, and natural language processing (NLP) help clinicians to integrate heart, electrocardiograms, and clinical record data, thus improving diagnostic accuracy and early risk prediction. AI technologies, targeted at analyzing echocardiography, computed tomography, and cardiac magnetic resonance imaging, demonstrated superior power to detect structural cardiac abnormalities, estimate myocardial blood supply, and detect markers of inflammation in fat tissues around coronary arteries. AI models for AI show better results than established traditional scoring systems such as the Framingham Risk Score, having a higher degree of accuracy, sensitivity, and specificity. In addition, reliance on explainable AI means that clinical teams are able to make sense of the underpinnings of AI predictions, meaning that the integration of AI into healthcare practice will not be complicated. The review highlights the way the AI is gradually improving the accuracy in cardiovascular risk stratification and personalization, which is an important milestone in data-driven cardiology.