Comparison of Visual Scoring of Coronary Artery Calcium (CAC) on Coronary Computed Tomography Angiogram (CCTA) to Agatston Score

Abstract

Purpose: Detection and quantification of coronary artery calcium (CAC) has a prognostic value for future cardiovascular events, beyond that accrued from other cardiovascular risk factors. CAC is conventionally measured as an Agatston score from an EKG-gated non-contrast CT of the heart using special software. The predominant indication for coronary CT angiography (CCTA) in clinical practice is to evaluate obstructive coronary artery disease (CAD) in a low to intermediate pretest probability population. At present there is no commercially available software for CAC measurement on CCTA scans. The purposes of this study were twofold. First, the study examines the correlation of estimated CAC from CCTA, using a novel method of visual scoring on the CCTA exam and comparing with the measured Agatston score from a non-contrast CAC score scan. Second, the study evaluates the performance of such estimation of CAC across radiologists with varying degrees of experience. Methods:Two cardiac radiologists and a chest radiologist, with varying degrees of experience, evaluated 100 coronary CT angiograms for visual scoring of CAC. The three major coronary vascular segments; left anterior descending artery (LAD) including left main coronary artery (LM), left circumflex coronary artery (LCX) and right coronary artery (RCA) were examined and scored for linear extent (0-4) as well as thickness (0-3). The linear and thickness scores for each segment was multiplied to get a vessel score (0-12) and total CAC score was derived by adding all the vessel scores (0-36). Statistical analysis was performed using Kappa statistic and Spearman Rank Correlation for the agreement with the Agatston scores from non contrast EKG gated calcium score CT scans performed at the same time as the CCTA. The scores of the senior cardiac radiologist with 30 years of experience were used for the primary analysis and the readings from the other two radiologists were used to calculate the inter-reader variability.Results: Overall, 59 % and 57 % of patients had CAC detected by Agatston score and visual score respectively. CAC by visual score on CCTA strongly correlated with the Agatston score (Spearman correlation coefficient 0.88, P < 0.001). In 7 cases visual scoring failed to detect minimal CAC with low Agatston CAC score (mean 4.1, range 1-13). Visual score of > 7 identified CAC > 100 with an area under the curve of 0.965, sensitivity of 0.88 and a specificity of 0.92. A visual score of > 12 identified CAC > 300 with an area under the curve of 0.970, sensitivity of 0.95 and specificity of 0.90. The inter-reader agreement with the other cardiac and chest radiologist was high, with Spearman correlation coefficients of 0.90 & 0.91 (p < 0.001). Conclusion:The visual estimation and stratification of CAC on CCTA is feasible and correlates well with the Agatston score and shows good inter-reader correlation. This additional information from CCTA will be useful for cardiovascular risk stratification and management.

Coronary Artery Calcium Data and Reporting System (CAC-DRS): A Primer

“…Since visual score does not require electrocardiogram (ECG) gating, the scoring can be done in any routine chest CT. visual score is a feasible method and correlates well with AS. Sonavane et al 21) showed a good correlation between visual score and AS with a good inter-reader correlation. Einstein et al 22) reported a high degree of association between visual score and AS on ungated low-dose CT scans used for attenuation correction on positron emission tomography/CT and single-photon emission CT/CT.…”