PhD defense of Jurica Šprem

On Thursday 11 July at 14:30, Jurica Šprem will defend his thesis entitled: “Enhanced Cardiovascular Risk Prediction by Machine Learning.”

Abstract

The detection of calcium deposits in coronary arteries is essential to reduce the high number of cardiovascular events such as heart attacks. The aim of this thesis was to investigate and improve the quantification of arterial calcifications in the coronary arteries (Coronary Artery Calcification; CAC) found in CT scans without ECG coupling (scans obtained during screenings not focusing on imaging the heart), by compensating for various image artifacts. With the developed machine learning algorithm for improved CAC scoring and the identification tool for detecting scans heavily influenced by image artifacts, we analyzed and compared differences between clinically used and proposed quantification methods to identify subjects at high risk for cardiovascular events. This dissertation demonstrates the benefits of using the newly described improved CAC validation method together with the artifact detection tool to facilitate timely identification of persons at high risk for cardiovascular events, especially in those without symptoms of cardiovascular disease.