Machine learning for coronary artery disease analysis in cardiac CT

Abstract

Cardiac CT angiography (CCTA) images of patients with suspected obstructive coronary artery disease are typically used to visually characterize coronary artery plaque and stenosis, as well as to serve as the gatekeeper for referral to invasive coronary angiography, where the fractional flow reserve is measured to identify functionally significant stenoses. The chapters of this thesis describe machine learning-based methods for automatic noninvasive identification of patients with functionally significant stenosis, and for automatic detection and characterization of coronary artery plaque and stenosis in CCTA images.