Utilization of artificial intelligence in the management of cardiovascular diseases: Promise versus reality

Artificial intelligence (AI) algorithms can detect subtle patterns in clinical data that are routinely missed by conventional diagnostic methods. The potential impact of AI-based clinical decision support software on the practice of cardiovascular medicine has inspired the development of numerous AI algorithms. Rapid evolution of AI technology has made it difficult for cardiologists to critically evaluate AI-based clinical decision support tools. This review provides an overview of modern AI models relevant to cardiovascular medicine. Also discussed are metrics for evaluating AI algorithm performance as well as known limitations of AI-based software, notably bias, limited generalizability, and lack of explainability.