The aim of this study was to compare the diagnostic accuracy of perfusion and wall motion (WM) during dipyridamole magnetic resonance in patients with chest pain syndrome. Ninety-three patients with normal baseline left ventricular function were referred for coronary angiography. Additional dipyridamole stress magnetic resonance testing (0.84 mg/kg over 6 minutes; using a Signa Cvi scanner) was performed. Cardiac-gated fast gradient-echo train sequences with a first pass of gadolinium contrast medium were used to assess myocardial perfusion. A perfusion reserve index was calculated as the ratio of dipyridamole to rest upslope. A perfusion reserve index value <1.54 in 2 contiguous myocardial segments was the perfusion positivity criterion. The WM positivity criterion was a segmental score increase of > or =1 grade in > or =2 segments. WM and the perfusion reserve index showed similar diagnostic accuracy for >50% quantitatively assessed coronary diameter reduction (86% for both), with WM having higher specificity (96% vs 66%, p <0.01) and lower sensitivity (82% vs 93%, p <0.05) than the perfusion reserve index. Perfusion had the highest accuracy values for coronary stenoses <75% (cutoff 59%) and WM for coronary stenoses > or =75% (cutoff 84%) (p <0.001). In conclusion, during dipyridamole magnetic resonance stress testing, perfusion and WM abnormalities have similar diagnostic accuracy, with perfusion showing higher sensitivity, particularly in the detection of moderate stenoses, and WM showing higher specificity.