Growth analysis of individual abdominal atherosclerotic calcified deposits

Abstract

1 Computer Science, University of Copenhagen, Copenhagen, Denmark Abdominal calcifications can be used as a clinical indicator of atherosclerosis and can be used to predict atherosclerotic plaques in the coronary arteries. Usually the degree of severity of atherosclerosis is given by the amount of calcifications along the aorta walls (AC24 score). In this study we take a look at the development of the individual calcifications over time. Our hypothesis is that the prediction of the growth of calcifications can give insight about the progression of atherosclerosis. Therefore we investigate different growth descriptors of the calcifications. Our study is based on x-ray images due to their advantages in large scales studies e.g. a low cost, a fast examination time, and a low radiation exposure compared to other image modalities. Our data set consists of 103 women with annotated x-ray images at two time periods (baseline at 1992-93 and follow-up in 2000-01), which have lumbar aorta visible in both images and have calcifications at baseline. We made a TPS registration of the follow-up to the baseline images to align the calcifications and find suitable matches between baseline and follow-up calcifications. From the one-to-one correspondent matches we have derived simple geometric measures that describe the growth of the calcifications. The growth descriptors are derived based on idealized coordinate systems that capture the direction of the blood flow. This way we can describe the growth of the calcifications based on how they have evolved along and perpendicular to the aorta wall. We describe the growth of the calcifications by their height, width, area and the center of mass movement. Our results show that the calcifications grow longitudinally, so they will occupy more of the aorta wall. The overall growth in area is 40.35 % and the calcification grows 34.83 % longitudinally (in the direction of the blood flow) and 34.63 % in the direction perpendicular to the blood flow. This is based on 35.7 % one-to-one correspondent matches out of 633 baseline calcifications Atherosclerotic growth patterns based on X-ray are able to give an insight in the progression of atherosclerosis. The longitudinal analysis of simple geometric descriptors of individual calcified deposits can be used to increase the clinical knowledge about atherosclerosis in the calcified stages. The growth pattern analysis could also be used to give a patient specific risk assessment.