This study compared positron emission tomography (PET) using oxygen-15-labeled water for measurement of coronary reserve with intracoronary Doppler in patients with left anterior descending artery stenosis and patients with no coronary lesion and a coronary reserve 3 as assessed by the invasive technique. To determine whether PET measurement of coronary reserve is altered by partial volume effect, patients with left ventricular dysfunction due to idiopathic cardiomyopathy were studied with both techniques. Direct ultrasonic measurement of coronary reserve was performed the day prior to the PET study: a Doppler catheter was placed in the proximal left anterior descending artery; mean velocity was recorded at baseline and after dipyridamole administration. Using a time-of-flight PET system, patients underwent: (1) an intravenous bolus of oxygen-15-labeled water at baseline and 4 to 6 min after intravenous infusion of dipyridamole using the same protocol as for Doppler study and (2) a 18F-fluorodeoxyglucose (FDG) myocardial imaging. Oxygen-15 time-activity curves were recorded in myocardial regions of interest (ROIs) drawn on a static FDG image. Using the left ventricular time-activity curve as an input function, a standard model with a single-tissue compartment was fitted to the PET data; myocardial blood flow was estimated as the blood-to-tissue transfer rate constant. Coronary reserve measured by PET was well correlated with the measured by intracoronary Doppler (r = 0.98, p < 0.001 for global population). This PET method is an accurate and reliable tool to noninvasively measure coronary reserve in patients, even in those with left ventricular dysfunction.