Contrast-free whole heart myocardial blood flow quantification from magnetic resonance imaging-determination of coronary sinus flow
Abstract Background Non-invasive reference quantification of whole heart myocardial blood flow (MBF) requires radioisotopes for PET and gadolinium contrast for cardiovascular magnetic resonance imaging (CMR) which in some cases is contraindicated. MBF may be determined from CMR quantification of flow in the coronary sinus (CSBF), the large vein draining the majority of the myocardium. Comparative studies of CSBF from CMR and MBF as determined from a reference technique have not yet been published in any larger cohorts. Purpose Our objective was to evaluate to what extent CMR CSBF measurements can be used to determine MBF as determined from gadolinium-contrast CMR in a cohort of normal subjects and patients with type 2 diabetes mellitus (T2DM) already demonstrated to display a wide range of MBFs. Methods 147 patients with T2DM and 25 age-matched controls were recruited to a cohort study on cardiovascular changes in DM. MBF was quantified from gadolinium-contrast perfusion sequences based on Fermi-constrained deconvolution. Myocardial segments with late gadolinium hyperenhancement or visually significant perfusion defects were included. CSBF was determined with CMR flow-sequences applied across the coronary sinus (VENC 0.5–1.0 m/s). Patients were studied during rest and maximal coronary artery dilatation by adenosine infusion (140 mg/kg/min). Blood flow stress-reserves were the ratio of stress to rest values. Co-variation of MBF and CSBF were determined from Bland-Altman plots with lines of agreement. Repeatability of CSBF was determined during the same experiment and calculated from single rater random intra class and repeatability coefficients. Results In normal subjects and patients with T2DM, MBF and CSBF increased during adenosine-stress (Fig. 1) with mean absolute increments of 172 and 163 mL/min/100g, and with mean stress-reserves of 3.35 and 3.24, respectively. Bland-Altman plots showed that MBF and CSBF covaried with a small bias, but in some cases with relatively large limits of agreement (Fig. 2). Overall, the mean bias of increase from rest-to-stress was 6 (CI: −1; 14) mL/min/100g with corresponding limits of agreement of 93 (CI 81; 105) and −80.2 (CI −92.6; 67.8) mL/min/100g. The mean bias of stress-reserve was 0.106 (CI: −0.0209; 0.234) with corresponding limits of agreement of −1.43 (CI: −1.65; −1.21) and 1.64 (CI: 1.42; 1.86). Intra-class and repeatability coefficients for coronary sinus flow were 0.95 (CI: 0.90; 0.95) and 5 mL/min/100g, respectively. Conclusion Myocardial blood flow can reproducibly and with a small bias be determined from the non-contrast technique of applying magnetic resonance imaging flow-sequences across the coronary sinus. Determination of MBF from coronary sinus blood flow may be useful in patient-groups where contrast is contraindicated, but limits of agreement with MBF must be taken into account. FUNDunding Acknowledgement Type of funding sources: None. Figure 1 Figure 2
