To compare relative coronary artery vasodilator reserve (rCVR = CVRtarget/CVRreference) to myocardial perfusion stress imaging, 48 patients with coronary artery stenoses (61% +/- 16%; mean, +/- SD; range, 30%-91%) had measurements of target and reference vessel CVR (Doppler-tipped guidewire). rCVR was computed and compared to stress 201thallium or (99m)technetium-sestamibi myocardial tomography. Compared to 24 patients with negative stress imaging studies, 24 patients with positive stress studies had angiographically more severe stenoses (74% +/- 13% vs. 44% +/- 24%; P = 0.0005) with lower CVR(target) (1.68 +/- 0.55 vs. 2.46 +/- 0.74; P = 0.002) and lower rCVR (0.72 +/- 0.22 vs. 1.0 +/- 0.26; P < 0.003). Based on receiver-operator characteristic (ROC) cut points (CVR > 1.9; rCVR > 0.75), compared to CVR, rCVR had similar agreement (Kappa 0.54 vs. 0.50), sensitivity (63% vs. 71%), specificity (88% vs. 83%), and positive predictive value (83% vs. 81%) with myocardial perfusion tomography. A concordant CVRtarget/rCVR only slightly increased sensitivity, specificity, and positive predictive values (77%, 90%, and 87%, respectively). Although rCVR, like CVR, correlates with stress myocardial perfusion imaging results, rCVR did not have significant incremental prognostic value over CVR alone for myocardial perfusion imaging. However, rCVR does provide additional information regarding the status of the microcirculation in patients with coronary artery disease and complements the CVR for lesion assessment.