Inter-scan reproducibility of coronary calcium measurement using Multi Detector-Row Computed Tomography (MDCT)

Abstract

Purpose: To assess inter-scan reproducibility of coronary calcium measurements obtained from Multi Detector-Row CT (MDCT) images and to evaluate whether this reproducibility is affected by different measurement protocols, slice thickness, cardiovascular risk factors and/or technical variables. Design: Cross-sectional study with repeated measurements. Materials and methods: The study population comprised 76 healthy women. Coronary calcium was assessed in these women twice in one session using 16-MDCT (Philips Mx 8000 IDT 16). Images were reconstructed with 1.5 mm slice thickness and 3.0 mm slice thickness. The 76 repeated scans were scored. The Agatston score, a volume measurement and a mass measurement were assessed. Reproducibility was determined by estimation of mean, absolute, relative difference, the weighted kappa value for agreement and the Intra-class correlation coefficient (ICCC). Results: Fifty-five participants (72.4%) had a coronary calcification of more than zero in Agatston (1.5 mm slice thickness). The reproducibility of coronary calcium measurements between scans in terms of ranking was excellent with Intra-class correlation coefficients of >0.98, and kappa values above 0.80. The absolute difference in calcium score between scans increased with increasing calcium levels, indicating that measurement error increases with increasing calcium levels. However, no relation was found between the mean difference in scores and calcium levels, indicating that the increase in measurement error is likely to result in random misclassification in calcium score. Reproducibility results were similar for 1.5 mm slices and for 3.0 mm slices, and equal for Agatston, volume and mass measurements. Conclusion: Inter-scan reproducibilility of measurement of coronary calcium using images from MDCT is excellent, irrespective of slice thickness and type of calcium parameter.