Systolic versus diastolic myocardial blood flow in patients with suspected coronary artery disease - a cardiovascular magnetic resonance study

Abstract

Summary: This study has shown that in patients with suspected and confirmed CAD, estimates of MBF by perfusion-CMR are significantly higher in diastole than systole during maximal hyperemic stress. Background: Differences in myocardial blood flow (MBF) between systole and diastole have been reported in healthy volunteers but the impact of cardiac phase on detecting coronary artery disease (CAD) is unknown [1]. This study aimed to compare MBF estimates from cardiovascular magnetic resonance (perfusion-CMR) imaging in systole and diastole in patients with suspected CAD and determine if either phase has greater diagnostic accuracy. Methods: Following invasive coronary angiography, 40 patients (68% men, 64 ± 8 yrs) underwent stress perfusion-CMR (1.5T Philips) acquired at mid-systole and end-diastole simultaneously [1]. Based on angiographic stenosis >70%, patients were grouped as having 'CAD' or 'no CAD'. In patients with CAD, myocardial segments were classified as 'stenosis-dependent' (downstream of a significant stenosis) or 'remote'. For each segment, MBF (Fermi-constrained deconvolution) and myocardial perfusion reserve (MPR) were calculated. The diagnostic accuracy of each phase was determined with receiver operator characteristic analysis. Results: Twenty-one patients (53%) had CAD. A typical example of a patient with ischemia is shown in Figure 1. Resting MBF was similar in the two cardiac phases for both normal and CAD patients (all p values > 0.05). MBF at stress was greater in diastole than systole in normal, remote and stenosis-dependent segments (3.75 ± 1.5 vs. 3.15 ± 1.1 ml/g/min; 2.75 ± 1.20 vs. 2.38 ± 0.99 ml/g/min; 2.49 ± 1.07 vs. 2.23 ± 0.90 ml/g/min; all p values < 0.01). MPR was also greater in diastole than systole in all three segment groups (all p values <0.05) (Figure 2). On receiver operator characteristic analysis, the optimal MPR cut-off for the detection of CAD was 1.95 for systole and 2.04 for diastole (area under curve 0.82 vs. 0.79; p=0.30). Conclusions: Estimates of stress MBF and MPR by perfusion-CMR in this study were greater in diastole than systole in normal and CAD patients. Although the diagnostic accuracy of both phases was similar, the MPR cut-off values were different. These observations are particularly (Figure presented) important in the emerging field of 3D perfusion-CMR where the acquisition phase may be specifically chosen. Different estimates of MBF and different MPR cut-off values between phases mean a universal standard needs to be agreed for 3D acquisitions.