In about 10 to 30% of patients with typical angina undergoing coronary angiography for suspicion of stenotic coronary artery disease angiographically normal coronary arteries are found. Kemp et al. in 1973 coined the term syndrome X to describe this entity. In a substantial portion of these patients pathologic findings in myocardial scintigraphy are present. Sensitivity and specificity of thallium-201 exercise imaging by visual analysis of images in the presence of significant coronary stenosis is 84 and 88%, respectively. Several investigators have reported abnormal results in radionuclide exercise tests in patients with angiographically normal coronary arteries. Some of these results can be explained by myocardial bridging, vasospasm, left or right bundle branch block, hypertrophic cardiomyopathy, or absorption artifacts. In the majority of cases, however, these abnormalities are not sufficient to explain the scintigraphic findings. Formerly often claimed "false positive", recent studies suggest that endothelial dysfunction might be the reason for the observed perfusion defects. When comparing patients with angiographically unobstructed coronary arteries with and without perfusion defects in stress myocardial perfusion imaging, patients with pathological results show a significantly lower increase of coronary flow after intracoronary injection of the endothelial-dependent vasodilator acetylcholine. Endothelial-independent vasodilation, however, is not impaired in these patients. In addition, intracoronary Doppler measurements reveal that perfusion defects in myocardial scintigraphy only occur if coronary blood flow in this perfusion area is significantly reduced. These results suggest that regional endothelial dysfunction may cause hypoperfusion in myocardial perfusion imaging and underline the important role of the microcirculation in the distribution of radiotracers. Another striking scintigraphic pattern in patients with microvascular angina is the high incidence of reverse redistribution. These perfusion defects, apparent in images obtained 4 hours after exercise stress testing, often cannot be assigned to the perfusion territory of one of the major epicardial vessels. This results in a marked inhomogeneous radionuclide distribution pattern in resting images. The inhomogeneity is associated with a significant reduced resting coronary flow velocity in these patients. As histologically confirmed microvessel disease is often accompanied by slow-flow phenomenon reflecting decreased resting flow velocity, the results suggest that the inhomogeneous perfusion pattern is caused by microvascular dysfunction. Furthermore, the heterogeneity of nuclide distribution supports the hypothesis that endothelial function is not homogeneous in the entire myocardial microcirculation, but varies considerably. In conclusion, microvascular dysfunction by itself seems to cause regional myocardial hypoperfusion, as documented by myocardial scintigraphy. When interpreting pathological scintigraphic results in patients without significant epicardial stenosis, true blood flow and myocardial perfusion abnormalities must be assumed.