Abstract
Background
Quantification of absolute myocardial perfusion using positron emission tomography (PET) augments diagnostic performance and aids risk stratification in patients with coronary artery disease (CAD). Similar to PET, cardiac magnetic resonance imaging (CMR) allows for quantification of absolute myocardial blood flow (MBF). To date, quantitative CMR perfusion has not been compared with [15O]H2O PET, which is considered to be the gold standard for quantification of MBF.
Purpose
The purpose of this study was to compare CMR measurements of absolute myocardial perfusion to those obtained using [15O]H2O PET.
Methods
Fifty-one patients with stable CAD underwent [15O]H2O PET followed by CMR within 7 days. Late gadolinium enhancement was used to rule out presence of scar tissue. CMR perfusion imaging was performed using a single bolus dual sequence technique. Myocardial perfusion was quantified for the three main vascular territories at rest and during vasodilator stress.
Results
Pooled CMR estimates of absolute MBF correlated well with PET (r=0.79; p<0.001) and showed good agreement (intraclass correlation coefficient [ICC]=0.78; p<0.001). However, rest and stress MBF correlated weakly (r=0.31; p<0.001 for rest MBF and r=0.35; p<0.001 for stress MBF, figure) and showed poor agreement between PET and CMR (ICC=0.24; p<0.001 for rest MBF and ICC=0.34; p<0.001 for stress MBF). Similarly, CMR derived coronary flow reserve (CFR) correlated weakly (r=0.28; p<0.001) and showed poor agreement with PET derived CFR (ICC=0.23; p<0.001). Nevertheless, for a cut-off value of 2.3 mL/min/g for stress MBF and 2.5 for CFR, CMR and PET were concordant in 116 (79%) vascular territories for stress MBF and 104 (75%) vascular territories for CFR.
Per vessel – CMR vs PET perfusion
Conclusions
Resting MBF, stress MBF and CFR measurements correlated poorly with [15O]H2O PET. Nevertheless, stress MBF and CFR were concordant between CMR and [15O]H2O PET in 79% and 75% of vascular territories, respectively.
Acknowledgement/Funding
None
Voxel-wise quantification of myocardial blood flow with cardiovascular magnetic resonance: effect of variations in methodology and validation with positron emission tomography
