We aimed at evaluating the adequacy of the commonly employed compartmental model for quantitation of cerebral metabolic rate of oxygen (CMRO2) using 15O-labeled oxygen (15O2) and positron emission tomography (PET). Sequential PET imaging was carried out on monkeys following slow bolus injection of blood samples containing 15O 2-oxyhemoglobin (15O2-Hb), 15O- labeled water (H2 15O), and C15O-labeled hemoglobin (C15O-Hb) into the internal carotid artery (ICA). Clearance slopes were assessed in the middle cerebral artery territory of the injected hemisphere. The time-activity curves were bi-exponential for both 15O2-Hb and H2 15O. Single exponential fitting to the early (5 to 40 seconds) and late (80 to 240 seconds) periods after the peak was performed and the 15O2-Hb and H2 15O results were compared. It was found that a significant difference between the clearance rates of the 15O 2-Hb and H2 15O injections is unlikely, which supports the mathematical model that is widely used to describe the kinetics of 15O2-Hb and H2 15O in cerebral tissues and is the basis of recent approaches to simultaneously assess CMRO 2 and cerebral blood flow in a single PET session. However, it should be noted that more data are necessary to unequivocally confirm the result.