Differentiating hypertrophic cardiomyopathy (HCM) from hypertensive heart disease (HHD) is important yet challenging. In this study, we compared 9 convolutional neural network (CNN) models based on cardiac MR cine imaging only for differentiation of the two diseases. We show that the dynamic information contained in cine about myocardial contraction and relaxation is crucial for accurate differentiation. By leveraging this information, we achieved a testing accuracy of 86.8% ± 3.5% in a cohort including 190 HCM and 113 HHD subjects. The results show that cine-based CNN is reasonably accurate for differentiation of HCM and HHD.
This abstract and the presentation materials are available to members only; a login is required.