Study Design

This study was performed at the National Institutes of Health (NIH National Institutes of Health) and Charité–Universitätsmedizin Berlin as a two-center prospective substudy of the CORE320 Combined Noninvasive Coronary Angiography and Myocardial Perfusion Imaging Using 320-Detector Row Computed Tomography multicenter trial to investigate the hypothesis that the diagnostic performance of CT perfusion with whole-heart coverage and single-beat acquisition is similar to that of dynamic myocardial MR perfusion imaging. The sponsor of the main study, Toshiba Medical Systems (Otawara, Japan), was not involved in this subanalysis of the CORE320 Combined Noninvasive Coronary Angiography and Myocardial Perfusion Imaging Using 320-Detector Row Computed Tomography study at any stage of the study design, data acquisition, data analysis, or manuscript preparation, nor was it involved in the decision to submit this substudy for publication. No financial support was provided for this substudy, and all costs related to it were the responsibility of the two respective centers (NIH National Institutes of Health and Charité–Universitätsmedizin Berlin). The CORE320 Combined Noninvasive Coronary Angiography and Myocardial Perfusion Imaging Using 320-Detector Row Computed Tomography protocol has been published by Vavere et al (17). In agreement with the CORE320 Combined Noninvasive Coronary Angiography and Myocardial Perfusion Imaging Using 320-Detector Row Computed Tomography protocol, this substudy used the two reference standards that were prospectively defined by the study steering committee: QCA quantitative coronary angiography and SPECT were used in combination to reflect anatomic and functional disease, while QCA quantitative coronary angiography alone was used to reflect anatomic disease (17). Additional analysis with SPECT alone as a reference standard was available on request during the review process. At QCA quantitative coronary angiography, detection of coronary artery stenosis of at least 50% was considered to represent CAD coronary artery disease for the purpose of this analysis. For combined QCA quantitative coronary angiography and SPECT, alignment between coronary artery stenosis and myocardial perfusion territories was performed as defined in the CORE320 Combined Noninvasive Coronary Angiography and Myocardial Perfusion Imaging Using 320-Detector Row Computed Tomography protocol by Cerci et al (18). This myocardial territory segmentation model consists of six segments distributed in the basal myocardium (segments 1–6), another six segments in the apical myocardium (segments 7–12), and one segment in the apex (segment 13). The numbers are given in ascending order for the anterior, anteroseptal, inferoseptal, inferior, inferolateral, and anterolateral segments. The apex (segment 13) was excluded in our analysis, as it is not part of a standard MR perfusion protocol. Vascular territory maps were defined for the following vessels: left main, proximal left anterior descending artery, middistal anterior left descending artery, left circumflex artery, right coronary artery, and Ramus branch of the left circumflex artery. Primary territories were aligned to the respective coronary artery under the assumption of the most common right dominant anatomic coronary pattern. Secondary territories were assumed to have a possible blood supply from the respective coronary artery in normal anatomic variations, and tertiary territories were assumed not to be supplied by the coronary artery (18). The myocardial segments and the vascular territories were also used for MR image reading to achieve consistency with the reading technique used for CT perfusion and SPECT images. All criteria of the Standards for Reporting of Diagnostic Accuracy statement can be found in this report. The study protocol and this substudy on the comparison of myocardial CT perfusion with MR perfusion imaging were approved by the institutional review boards of the two institutions and the German Federal Office for Radiation Protection. Each patient gave written informed consent. The primary analysis included patients who underwent each imaging procedure, including SPECT, MR perfusion imaging, CT perfusion, and ICA invasive coronary angiography.