The physiologic importance of coronary stenoses can be assessed indirectly by stress myocardial perfusion imaging or directly by translesional pressure and flow measurements. The aims of this study were to compare myocardial fractional flow reserve (FFRmyo), a recently proposed index of lesion significance derived from hyperemic translesional pressure gradients, with directly measured poststenotic flow velocity reserve for the prediction of myocardial perfusion stress imaging results in corresponding vascular beds. Poststenotic coronary flow velocity (0.018-inch guide wire) and translesional pressure gradients (2.7F fluid-filled catheter) were measured at baseline and after intracoronary adenosine (12 to 18 micrograms) in 70 arteries (diameter stenosis: mean 56% +/- 15%, range 14% to 94% by quantitative angiography). Coronary flow reserve was calculated as the ratio of hyperemic to basal mean flow velocity. FFRmyo was calculated during maximal hyperemia as equal to 1-(hyperemic gradient [mean aortic pressure-5]), where 5 is the assumed central venous pressure. Positive and negative predictive values and predictive accuracy for reversible stress myocardial perfusion abnormalities were computed. There was a significant correlation between pressure-derived FFRmyo and distal coronary flow reserve (r = 0.46; p < 0.0001). The strongest predictor of stress myocardial perfusion imaging results was the poststenotic coronary flow reserve (chi square = 33.2; p < 0.0001). The correlation between stress myocardial perfusion imaging and FFRmyo was also significant (chi square = 8.3; p < 0.005). There was no correlation between stress myocardial perfusion imaging and percentage diameter stenosis (chi square = 2.9; p = 0.10) or minimal lumen diameter (chi square = 0.47; p = 0.73). A poststenotic coronary flow reserve of < or = 2 had a positive predictive value of 89% for regionally abnormal myocardial perfusion imaging abnormalities, whereas the positive predictive values of FFRmyo and angiographic percentage diameter stenosis were only 71% and 67% respectively. In conclusion, the predictive value of poststenotic coronary flow velocity reserve for stress-induced myocardial perfusion abnormalities exceeds that of the translesional FFRmyo. These findings should be considered when applying these techniques for clinical decision making in the assessment of coronary stenosis severity.