Cardiovascular disease (CVD) is the leading cause of morbidity and mortality worldwide. Medical imaging plays a crucial role in the detection, diagnosis and prognosis in CVD. Numerous studies have shown that automatic image analysis, especially analysis employing deep learning, can automate manual processes thereby shortening the analysis time and reducing the inter-observer variability.
In this talk, deep learning (DL) methods for analysis of cardiac CT and MR exams will be shown. We will demonstrate how DL methods can be used to detect subclinical signs of CVD and identify patients at risk of CVD events, such as hart infarction. Moreover, we will show how DL can be exploited for the analysis of cardiac function in CT and MR exams. Finally, we will demonstrate how DL techniques can be used for image synthesis allowing reduction of irradiation dose to subjects and improvement of image quality.