Incremental predictive value for obstructive coronary artery disease by combination of Duke Clinical Score and Agatston score

Aims Recent study suggests that algorithms such as the Duke Clinical score (DCS) may overestimate the pretest probability. The Agatston score representing the grade of coronary artery calcification can be simply calculated from low-radiation exposure ECG-gated plain CT. In this study, we investigated whether or not more superior diagnostic performance for obstructive coronary artery disease (CAD) can be obtained by combining DCS with the Agatston score. Methods and results Of 3939 consecutive patients suspected of having stable angina without known CAD who underwent Coronary Computed Tomography Angiography (CCTA) as well as calculation of the DCS and Agatston score at our hospital, 3688 patients were selected as subjects. Obstructive CAD was defined as >50% diameter stenosis on CCTA; we investigated the diagnostic performance based on the area under the curve (AUC) of a receiver operating characteristic (ROC) curve, Net Reclassification Improvement (NRI), and Integrated Discrimination Improvement (IDI). The AUCs of ROCs prepared using the DCS alone and combination of the DCS and Agatston score were 0.7137 and 0.8057, respectively, showing that the diagnostic performance of the combination was significantly superior to DCS alone (P < 0.001). NRI was 0.8132 and IDI was 0.1374, showing that the diagnostic performance was improved by the combination of the DCS and Agatston score compared with DCS alone (P < 0.001). NRI (0.3522) and IDI (0.0287) were improved compared with those of the Agatston score alone (P < 0.001). Conclusion The combination of the DCS and Agatston score improved the diagnostic performance for obstructive CAD compared with DCS alone and Agatston score.