For the assessment of myocardial perfusion with cardiac magnetic resonance imaging, different semiquantitative parameters of the first pass signal intensity time curves can be calculated and myocardial perfusion reserve indices can be determined. In this study we evaluated the feasibility of different perfusion parameters and their perfusion reserve indices for the detection of significant coronary artery stenosis. The signal intensity time curves of the first pass of a gadolinium-DTPA bolus injected via a central vein catheter before and after dipyridamole infusion were investigated in 15 patients with single vessel (stenosis > or = 75% area reduction) and five patients without significant coronary artery disease. For the distinction of ischemic and nonischemic myocardial segments, semiquantitative parameters, such as maximal signal intensity, contrast appearance time, time to maximal signal intensity and the steepness of the signal intensity curve's upslope determined by a linear fit, were assessed after correction for the input function. For each parameter a myocardial perfusion reserve index was calculated and cut off values for the detection of significant coronary stenosis were defined. The diagnostic accuracy of each parameter was then examined prospectively in 36 patients with coronary artery disease and compared with coronary angiography. Where as a distinction of ischemic and normal myocardium was possible with myocardial perfusion reserve indices, semiquantitative parameters at rest or after vasodilation alone did not allow such a distinction. The perfusion reserve index calculated from the upslope showed the most significant difference between ischemic and nonischemic myocardial segments (1.19 +/- 0.4 and 2.38 +/- 0.45, p < 0.001) followed by maximum signal intensity, time to maximum signal intensity and contrast apperance time. Sensitivity, specificity and diagnostic accuracy was 87, 82 and 85% for the detection of hypoperfusion induced by significant coronary artery stenoses using the perfusion reserve index calculated from the upslope. The steepness of the first pass signal intensity curve's upslope, determined by a linear fit, is a feasible parameter for the detection of significant coronary artery disease with MR. Based on a myocardial perfusion reserve index of this parameter, ischemic myocardium can be identified with high diagnostic accuracy.