Visual and automatic grading of coronary artery stenoses with 64-slice CT angiography in reference to invasive angiography

The aim of this study was to assess the performance of a software tool for quantitative coronary artery analysis of computed tomography coronary angiography (CT-QCA) in comparison with invasive coronary angiography with quantitative analysis (CAG-QCA) as standard of reference. Two radiologists reviewed the CT angiography data sets (Siemens Sensation 64) of 25 patients, grading coronary artery stenoses visually and with a software tool (Circulation, Siemens). Twenty-three data sets with sufficient image quality were included in the final analysis. CAG revealed a total of 30 wall irregularities and 28 stenoses, of which 17 were graded as moderate and nine as hemodynamically significant. CT-QCA showed a better agreement to CAG-QCA, with a systematic overestimation of the degree of stenosis of 6.1% and limits of agreement of +36.1% and −23.9; the correlation coefficient was 0.82 (p < 0.0001). Using CT-QCA, sensitivity, specificity, and positive and negative predictive value were 89%, 100%, 89%, and 100%, respectively, for significant area stenoses greater than 75%. The positive predictive value for the visual assessment amounted to 53%. Interobserver variability between CT-QCA and visual assessment showed a kappa value of 0.72. In conclusion, software-supported CT-QCA makes it possible to quantify significant coronary artery stenoses automatically, with good agreement to CAG-QCA. Both Stephanie Busch and Thorsten R. C. Johnson contributed equally to the study.