A comparison of cardiovascular magnetic resonance and single photon emission computed tomography (SPECT) perfusion imaging in left main stem or equivalent coronary artery disease: a CE-MARC substudy.

Background: Assessment of left main stem (LMS) stenosis has prognostic and therapeutic implications. Data on assessment of LMS disease by cardiovascular magnetic resonance (CMR) and single photon emission computed tomography (SPECT) are limited. CE-MARC is the largest prospective comparison of CMR and SPECT against quantitative invasive coronary angiography (QCA) for detection of coronary artery disease (CAD), and provided the framework for this evaluation. The aims of this study were to compare diagnostic accuracy of visual and quantitative perfusion CMR to SPECT in patients with LMS stable CAD.
Methods: Fifty-four patients from the CE-MARC study were included: 27 (4%) with significant LMS or LMS-equivalent disease on QCA, and 27 age/sex-matched patients with no flow-limiting CAD. All patients underwent multi-parametric CMR, SPECT and QCA. Performance of visual and quantitative perfusion CMR by Fermi-constrained deconvolution to detect LMS disease was compared with SPECT.
Results: Of 27 patients in the LMS group, 22 (81%) had abnormal CMR and 16 (59%) had abnormal SPECT. All patients with abnormal CMR had abnormal perfusion by visual analysis. CMR demonstrated significantly higher area under the curve (AUC) for detection of disease (0.95; 0.85-0.99) over SPECT (0.63; 0.49-0.76) (p = 0.0001). Global mean stress myocardial blood flow (MBF) by CMR in LMS patients was significantly lower than controls (1.77 ± 0.72 ml/g/min vs. 3.28 ± 1.20 ml/g/min, p < 0.001). MBF of <2.08 ml/g/min had sensitivity of 78% and specificity of 85% for diagnosis of LMS disease, with an AUC (0.87; 0.75-0.94) not significantly different to visual CMR analysis (p = 0.18), and more accurate than SPECT (p = 0.003).
Conclusion: Visual stress perfusion CMR had higher diagnostic accuracy than SPECT to detect LMS disease. Quantitative perfusion CMR had similar performance to visual CMR perfusion analysis.