A Fully Automated Framework for Segmentation and Stenosis Quantification of Coronary Arteries in 3D CTA Imaging

Abstract

In this paper, we present a novel framework to segment and quantify stenosed coronary arteries in 3D contrast enhanced computed tomography angiography (CTA). According to our knowledge, no commercially available software package permits fully automated detection and assessment of atherosclerotic stenosis. Therefore, in clinical practice, the radiologist has to make a detailed evaluation, segment by segment, which is obviously a tedious and time consuming task. In the proposed framework, the main branches of the coronary arteries are firstly extracted from the volume datasets by using a localized region-based level sets framework. Next, calcified voxels are removed by a post processing step. The centreline of the coronaries is extracted by mesh contraction algorithm. Finally, we compute the circularity and the convex hull deficiency of the cross sections at specific locations along the coronary arteries, which are then used to automatically detect and identify the presence of calcified stenosis.