Comparison of Two Different Software Systems for Electron-Beam CT-Derived Quantification of Coronary Calcification

Publisher Website
Google Scholar
Abstract
Adamzik M, Schmermund A, Reed JE, Adamzik S, Behrenbeck T, Sheedy PF. Comparison of two different software systems for electron-beam CT-derived quantification of coronary calcification. Invest Radiol 1999;34:767–773. The growing interest in coronary calcium quantification by electron-beam CT (EBCT) has led to the development of various software systems for the analysis of EBCT raw data, but it is unknown whether these software systems yield comparable results. Two sets of EBCT scans were obtained in 73 asymptomatic patients less than 15 minutes apart. Both scans of each patient were analyzed using two different software systems, the Mayo Clinic software and the AccuImage Scoring System. The authors compared the calcium quantities yielded by the two different software systems, analyzed the interscan variability, and calculated the interobserver variability. Finally, they investigated the influence of the CT density factor inherent in the widely used Agatston score for the quantification of coronary calcium on reproducibility. The mean score determined by the Mayo Clinic software was 14% greater than that determined by theAccuImage system. The mean difference between the two systems was 14% ± 25%, and the median difference was 3%. The relative mean and the median difference between the two scans of one patient were 15.3% and 6% determined by the AccuImage system and 17% and 6.5% determined by the Mayo Clinic software. The interobserver reliability calculated by the Mayo Clinic software was better than that of the AccuImage system. There was a trend for better reproducibility using calcium area rather than the Agatson score. Two different scoring systems do not necessarily yield the same result. Calcium quantities were systematically determined to be greater by one system than the other, and there were significant differences with regard to interobserver reliability. Hence, software should be tested with regard to reproducibility data, and the interpretation of calcium quantities should acknowledge which type of software was used.