Measuring the performance of EKG-based image selection for the purpose of cardiac calcification scoring

Coronary artery calcium screening is gaining increased attention as a tool for the noninvasive detection of coronary artery disease. In the case of calcification scoring performed on helical CT images, EKG data collected during image acquisition are used retrospectively to select the image with the least motion artifacts in a heart cycle. Calcification scoring is then only performed on the selected images, which makes appropriate image selection a key factor in the overall success of calcification scoring. In this paper, we examine how well EKG-based image selection matches image selection done by visual inspection. Results are presented for nineteen image sets