Myocardial flow reserve (MFR) is not routinely assessed in myocardial perfusion imaging (MPI) studies but has been hypothesized to affect test accuracy when assessing disease severity by coronary vessel lumenography. Magnetic resonance imaging (MRI) is an emerging diagnostic technique that can both perform MPI and assess MFR. We studied women (n = 184) enrolled in the Women's Ischemia Syndrome Evaluation (WISE) study with symptoms suggesting ischemic heart disease. Tests performed were coronary angiography and MPI by both MR and gated radionuclide single photon emission computed tomography (gated-SPECT). The MFR index was calculated using the MR data acquired at baseline and under vasodilation (dipyridamole) conditions. The study was structured with a pilot and an implementation phase. During the pilot phase (n = 46) data were unmasked and an MFR threshold was defined to divide patients into those with an adequate (AMFRI) or inadequate (IMFRI) MFR index. During the implementation phase, the MFR index threshold was prospectively applied to patients (n = 138). In the implementation phase, MPI ischemia detection accuracy compared to severe (> or = 70%) coronary artery diameter narrowing by angiography was higher in the AMFRI vs. the IMFRI group for MRI (86% vs. 70%, p < 0.05) and gated-SPECT (89% vs. 67%, p < 0.01). The IMFRI group (n = 55, 30% of study population) had a higher resting rate-pressure product compared with the AMFRI group (10,599 +/- 2871 vs. 9378 +/- 2447 bpm mm Hg, p < 0.01), consistent with higher resting myocardial flow. When compared with each other, MRI and gated-SPECT MPI showed no difference in accuracy among MFR groups. Myocardial perfusion patterns in the IMFRI group may have resulted in atypical perfusion patterns, which either masked or mimicked epicardial coronary artery disease.