BACKGROUND: To evaluate the diagnostic value of magnetic resonance imaging (MRI) of myocardial perfusion in the assessment of flow-limiting epicardial stenosis in a head-to-head comparison with abnormal thallium-201 ((201)TI) single photon emission tomography (SPECT) studies in patients with predominantly known coronary artery disease (CAD). METHODS AND RESULTS: Twenty-one patients (mean age 65 +/- 10 years) with reversible myocardial perfusion defects on (201)TI-SPECT images during dipyridamole-stimulated hyperemia were recruited for study purpose. Within 5 days of the (201)TI-SPECT study, myocardial perfusion was studied again with MRI during dipyridamole stimulation and at rest. Overall, (201)TI-SPECT identified 30 reversible regional perfusion defects. The sensitivity to detect hypoperfused segments was 70% (21/30) with the GRE-MRI perfusion analysis with (201)TI-SPECT as reference. When patients were subgrouped according to the extent of regional reversible perfusion defects on (201)TI-SPECT, mild- (SDS: 2-4), moderate- (SDS: 5-8), and severe- (SDS > 8) perfusion defects were also identified by GRE-MRI perfusion analysis in 75% (6/8), in 56% (9/16) and 100% (6/6), respectively. CONCLUSIONS: GRE-MRI first-pass stress perfusion imaging may not identify up to 30% of mild-to-moderate perfusion defects in a group of preselected patients with predominantly known CAD and abnormal (201)TI-SPECT studies.