Application and progress of machine learning in coronary computed tomography angiography

Cardiac computed tomography angiography (CCTA) has become an important non-invasive method to evaluate coronary artery disease. With the extensive application and increased image analysis features, more demands on operational technique and efficiency are asked. Machine learning (ML) is the subarea of artificial intelligence (AI), which is completely data driven, by computer algorithm to identify and analyze the potential relationship of centralized variables in large data sets for realizing the prediction of external data. In the field of cardiac CT, the application of various ML algorithms would improve the efficiency and quality of CCTA, helping accurate lesion assessment and risk stratification. It also brings new applications in cardiac functional imaging. The applications of ML in cardiac CT have been reviewed in present paper including CT-image analysis, risk stratification, CT-myocardial perfusion and CT-fractional flow reserve. DOI: 10.11855/j.issn.0577-7402.2021.03.12