Positron emission tomography with`50-labeled water (H2150) can be used to delineate abnormal regional myocardial blood flow in experimental animals. To determine the feasibility of this method in humans, we studied 33 subjects (9 normal volunteers and 24 patients with angiographically documented coronary artery disease) at rest and after myocardial hyperemia induced with intravenous infusion of dipyridamole. At rest, the myocardial region demonstrating the lowest relative H2`50 activity exhibited 71 + 8% of activity in the region with peak activity in control subjects and 62 + 17% in patients (p = NS). After the dipyridamole infusion, differences between the two groups were accentuated. In control subjects, activity in the region with lowest relative radioactivity averaged 77 5% of that in the region with peak activity. In patients, it averaged 55 ±+22% of activity in the region with peak activity (p<0.01). Results in patients with ischemia with or without a history of remote myocardial infarction were not significantly different. In 22 of the 24 patients, the region with lowest relative perfusion corresponded anatomically to the region of myocardium distal to a stenosis. Thus, delineation of regional myocardial perfusion in patients with coronary artery disease is possible with positron emission tomography and H215O. Further studies will be necessary to prospectively determine sensitivity and specificity. (Circulation 1988;78:612- Received July 27, 1987; revision accepted May 5, 1988.`3 N-labeled ammonia2-4 and 52Rb5-7 have been successfully used to estimate regional myocardial blood flow in experimental animals and in patients. 82Rb is particularly attractive in the clinical setting because it is generator produced, which obviates the need for an on-site cyclotron. However, the use of these tracers for quantification of flow entails some limitations. Single-pass extraction and myocardial retention of`3N-labeled ammonia are influenced by flow23,8 as well as by cell membrane integrity9 and the metabolic state of the myocardium.8 9 Extraction of 82Rb also varies with flow and metabolism. 5,6,10 An ideal perfusion tracer should be either freely diffusible or completely extractable and retainable by myocardium on a single pass through the coronary circulation without having its kinetics altered throughout the range of flow studied or by the metabolic status of the myocardium. 150-labeled water (H2`50) is essentially a freely diffusible tracer in the heart with uptake that does not vary despite wide variations in flow rate11 or changes in the metabolic state of myocardium associated with prolonged ischemia or reperfusion. 11-14 150 is a useful