How algorithms are transforming the diagnosis of ischemic heart disease—state of the art

Ischemic heart disease (IHD) is a leading cause of morbidity and mortality worldwide, highlighting the necessity for better diagnostic modalities. Artificial intelligence (AI) and machine learning (ML) are increasingly being used with multimodal cardiovascular diagnostic testing to provide standardized and reproducible assessment methodologies that have been shown to detect subtle signals beyond human recognition. This state-of-the-art review will summarize the various applications of AI across key modalities: describing its use in electrocardiography to risk-stratify patients; in coronary computed tomography angiography (CCTA) for quantitative plaque and stenosis measurements as well as measuring fractional flow reserve (FFR) derived from imaging; in cardiac magnetic resonance imaging (MRI) to automatically segment cardiac chambers and characterize tissue; and in intracoronary imaging [specifically intravascular ultrasound (IVUS) and optical coherence tomography (OCT)], where automation is evolving. We will also discuss combining these sources of data through clinical decision support systems (CDSS) that can enhance the comprehensive evaluation of IHD. We anticipate several issues for implementation, including validation, regulation, transparency, and clinical integration. Overall, AI can help reduce the number of time-consuming manual measurements used to augment quantitative features of an assessment and improve physiology-based decision-making. However, there were marked differences in performance based on the task and dataset, and AI was not always better than the human experts. Ultimately, AI must be validated prospectively, must be generalizable, and reported transparently for safe adoption in IHD care globally.