Deep Learning for Quantitative Assessment of Hypertrophic Cardiomyopathy
Hypertrophic cardiomyopathy (HCM) is the most common monogenic heart disease. HCM is characterized by unexplained left ventricular hypertrophy (LVH), myofibrillar disarray, and myocardial fibrosis. There is a pressing need for rapid and robust quantification of cardiac MR markers for identification of risk. Currently the quantification relies heavily on manual image segmentation of LV, RV and scar regions, which is not only time-consuming but also suffers from significant inter-observer variability in multi-site studies. Deep learning methods have recently shown promising results in image-related tasks including medical image segmentation, because they can handle complicated imaging situations such as variations in intensity, contrast and shape. The goal of this study is to develop deep-learning automatic image analysis methods to greatly improve the efficiency and robustness of quantitative CMR markers of disease.