Evaluation of alternative methods for coronary calcium scoring in multi-detector-row computed tomograhy

Background: Determining coronary artery calcifications is one of the methods of coronary heart disease screening. However, the traditional Agatston Calcium Score (CS) shows low interexamination reproducibility. The aim of this study was to evaluate the interscan variability coefficients of calcium measures based on three modifications of the original Agatston equation. Material/Methods: Fifty adults (37 men and 13 women; mean age 46.2 +/- 9.2 years) were included in the study. Each patient was examined with two consecutive, prospectively electrocardiographically triggered, multi-detector-row CT acquisitions to detect and quantify coronary artery calcifications. CS was calculated according to the method by Agatston et al. Alternative scores were calculated using a continuous weighting factor (CS-CM), the average lesion attenuation value (CS-SA), or both (CS-CA). The mean and median interscan percent variabilities of the methods were evaluated using nonparametric analysis of variance. Results: In the 50 patients, 1315 calcified lesions were found. The alternative scores correlated well with CS (for CS vs. CS-SA, CS-CM, and CS-CA, r = 0.990, 0.840, and 0.946, respectively, P < 0.0001). The mean and median percent variabilities did not differ significantly among the methods tested (P = 0.370). For CS, CS-SA, CS-CM, and CS-CA the mean variabilities were 13.24%, 13.36%, 16.00%, and 13.62%, respectively. Except for CS-CM, the methods showed similar distributions of variability vs. score on Bland and Altman plots. Conclusion: None of the tested modifications of the Agatston method brought improvement in the interscan reproducibility of coronary calcium scoring. In our opinion, a significant reduction in variability may be achieved by a standardization of image acquisition and reconstruction.