Coronary artery calcium and cardiovascular disease prediction by scanner type: the multi-ethnic study of atherosclerosis

Abstract

To evaluate the predictive value of coronary artery calcium (CAC) scoring methods across cardiac computed tomography (CT) scanner types. CAC was measured in participants from the MESA (Multi-Ethnic Study of Atherosclerosis), a prospective cohort study of participants free of baseline cardiovascular disease (CVD), using either EBCT (electron beam CT) or MDCT (multidetector CT). The risks for incident coronary heart disease (CHD) and CVD events were compared for CAC scores per SD using Cox proportional hazards models with multivariable adjustment in 3,362 MESA participants with detectable CAC. Using the Agatston score, the hazard ratio (HR) and 95% confidence interval (CI) for CHD was 1.50 (1.27,1.78) for EBCT and 1.60 (1.37,1.87) for MDCT. Using volume and density scores, the HR for CHD was 2.14 (1.71,2.68) for volume and 0.61 (0.48,0.76) for density on EBCT and 1.73 (1.42,2.11) for volume and 0.88 (0.71,1.10) for density on MDCT. Similar results were seen for CVD risk and in analyses stratified by sex, body mass index (BMI), and age. The volume and density score model demonstrated higher areas under the curve (AUC) for CHD than the Agatston score with EBCT (0.702, 95% CI: 0.666,0.738 versus 0.677, 95% CI: 0.638,0.715, p<0.001) and MDCT (0.669, 95% CI: 0.632,0.705 versus 0.660, 95% CI: 0.622,0.697, p=0.216). The CAC volume and density scores provide better risk prediction than the Agatston score for CHD and CVD events, regardless of CT scanner type. CAC density was strongly and inversely associated with CHD risk. Both density and volume score prediction were stronger for EBCT than MDCT.