Abstract 134: Can Coronary Calcification Scores Be Approximated by Routine Clinical CT?

Introduction: Coronary calcification scores (CAC) scores are established predictors of coronary atherosclerosis and are useful, independent measures of cardiovascular risk. More recently, medical image based phenotyping has gained traction in the medical literature for cardiovascular and other disease processes. CAC scores are limiting as they are only obtained when a physician suspects cardiovascular disease, are expensive, and only capture data relative to the heart. Routine clinical tomography (CT) scans can be processed using Analytic Morphomics to obtain quantified calcification in the descending aorta. Calcification of the descending aorta can be quantified through processing of routine clinical CT. We hypothesize that CT-based morphometric assessment of calcification in the descending aorta will be a useful approximation of the CAC score. Methods: Analysis was performed on 210 participants who had a CAC score within 30 days of a chest and/or abdominal CT. The scans were retrospectively obtained from the University of Michigan radiology database. Processing of CT images was performed with Analytic Morphomics, a semi-automated system for obtaining granular measurements of body composition recorded using a vertebral level coordinate system. The data capture included aortic segmentation and quantification of calcification. Aortic wall calcification percent was established at each vertebral level. Kendall’s correlations, averaged across 100 bootstrapped samples, establishes relationships of aortic wall calcification and CAC score. Results: A boxplot reporting Kendall’s correlation coefficients at each vertebral level are reported in Figure 1. Correlations among CAC and aortic calcification are stronger at lower vertebral levels. Correlations were significant (α Conclusion: Providing a granular, segmented assessment of aortic calcification from routine CT scans may provide an approximate measure of CAC. This aortic calcification approach may identify patients at risk for cardiovascular events, who would not have otherwise been identified, or reduce unnecessary radiation exposure by ruling out the need for a CAC scan. This may reduce financial cost of healthcare and the emotional burden on patients and their families.