Visual estimation of coronary calcium on computed tomography for attenuation correction

Abstract

BackgroundThe coronary calcium score (CCS) provides independent diagnostic and prognostic information on top of myocardial perfusion imaging (MPI) in patients suspected for coronary artery disease, but requires an additional computed tomography (CT) scan. ObjectiveWe investigated the accuracy and inter-reader reproducibility of visual estimation of the CCS on the CT used for attenuation correction. Methods250 patients undergoing single photon emission computed tomography MPI and Agatston CCS were included. The CCS was also visually estimated on the CT for attenuation correction by two separate readers blinded to the Agatston CCS, and was categorized into a six-point scale (0, 1–10, 11–100, 101–400, 401–1000 and > 1000). ResultsThe median Agatston CCS was 82 [25th–75th percentile: 0-562], with a range from 0 to 7287. Of the visually estimated CCS, 60% (reader 1) and 65% (reader 2) were classified correctly into the 6 categories. 93% (reader 1) and 88% (reader 2) of the visually estimated CCS did not vary by more than one category from the Agatston CCS. The intraclass correlation coefficient for agreement between the Agatston CCS and the visually estimated CCS was 0.95 for reader 1 and 0.94 for reader 2. The intraclass correlation coefficient for inter-reader reproducibility of the visually estimated CCS was 0.96. ConclusionThe CCS can be accurately estimated on the CT for attenuation correction, as high agreement is demonstrated with the Agatston CCS and inter-reader reproducibility is excellent. If no traditional Agatston CCS is performed, the degree of atherosclerosis should be assessed by means of estimating CCS on the CT for attenuation correction.