Deep learning architectures for automatic detection of viable myocardiac segments

by OPEN SCIENCE

Cardiovascular Diseases (CVDs) are the leading causes of mortality worldwide, with an estimated 17.9 million deaths in 2019, mainly due to myocardial infarction (MI), commonly known as heart attack and stroke1 . Radiologically diagnosing MI in its early phases plays a crucial role in treating and improving clinical outcomes. In recent decades, significant research works have been developed to improve the prognosis of cardiac diseases and therefore reduce CVDs deaths. MI is a medical condition preventing blood supply to the cardiac muscle, caused by the blocking of coronary arteries (see Figure 1.1). The size of infarcted myocardial tissue is impacted by the time taken until treatment is performed. Acute complications may progress towards Heart Failure (HF) if the affected heart cannot pump blood sufficiently to the circulatory system’s lungs and body [27]. MI occurs as a result of atherosclerosis, in which plaque builds up inside the artery walls. This build-up makes the arteries progressively narrower and slows blood flow, causing angina. Finally, an area of cholesterol plaque can tear inside of a coronary artery. This rupture results in a blood clot forming on the plaque’s surface, which can then completely block blood flow through arteries. If the blockage isn’t remedied fast, the heart muscle begins to die. The healthy heart area is substituted with the infarct area. Medical studies have shown that in AMI, the infarct tissue mainly contains heterogeneous

128 pages, Paperback

Published March 7, 2022