Quantitation of regional myocardial blood flow (MBF) in absolute terms with positron emission tomography (PET) has been difficult to achieve in part because of errors induced by the relatively low spatial resolution of current tomographic instruments. We previously demonstrated that MBF could be accurately measured over a wide range of flows after intravenous administration of H2 15O when the arterial input function and myocardial radiotracer content were measured directly. To extend this quantitative approach for noninvasive estimates of MBF with PET. We recently developed and implemented a novel mathematical approach whereby partial volume and spillover effects were estimated along with flow within the operational one-compartment flow equation. Noninvasive estimates of flow correlated closely with flow measured directly with radiolabeled microspheres. In the present study, with the use of a commercially available cardiac phantom, we assessed our ability to obtain true time-activity curves from observed PET data contaminated by partial volume and spillover effects. Computer simulations demonstrated that the approach developed is relatively insensitive to most potential sources of error, but is sensitive to timing discrepancies between the arterial input function and the tissue time-activity curve. Implementation of this approach provides accurate quantitation of regional MBF in absolute terms and should be useful in noninvasive evaluation of the efficacy of treatments designed to enhance nutritional perfusion in human subjects.