Quantitative Vascular Analysis for Non-invasive Inspection of Cardiovascular Diseases

Abstract

This thesis proposes a fully automatic technique for the quantitative analysis of thoracic aortic diseases to assess patient’s cardiovascular risk. This technique is based on the automatic segmentation of the ascending aorta beginning from the aortic arch down to the Ostia points, which is also automatically detected. Moreover, a model for the ascending aorta is proposed and built from its anatomical features.The proposed detection and segmentation algorithms are tested and validated on two imaging modalities; Computed Tomography Angiography (CTA) datasets of the Coronary Artery Framework provided by the Rotterdam Medical Center and Phase-Contrast Magnetic Resonance Images (PC-MRI) datasets.The results show the success of the proposed techniques to automatically detect and segment the ascending aorta and the ostia points with high degree of accuracy from images provided from the two imaging modalities, which enabled us to successfullydetect thoracic aortic diseases through quantitative measures performed on the segmented ascending aorta. These measures together with the ostia location would help physicians non-invasively assess the cardiovascular risk and plan for surgery.