Intracranial or intraventricular blood pools have been suggested as noninvasive sources of an input function for quantitative PET. These techniques measure the concentration of the tracer in whole blood, but the concentration in plasma depends on the equilibration of the tracer between plasma and erythrocytes. METHODS FDG, 6-[18F]fluoro-L-m-tyrosine (FmT), or its major metabolite, 6-[18F]fluoro-3-hydroxyphenylacetic acid (FHPAA), was added to blood samples obtained from healthy fasting volunteers along with radioiodinated human serum albumin (RIHSA). Samples were incubated at 37 degrees C for times between 10 s and 2 h and then plunged into an ice bath and centrifuged. Whole blood and plasma were counted for 18F and 125I activities. The resulting time courses were fit to successively more complex models, evaluated using an F test. RESULTS All radioactivity associated with RIHSA remained in the plasma, whereas FDG equilibrated instantaneously between plasma and erythrocytes. FmT took about 1 h to equilibrate between plasma and erythrocytes; this time course could be described by a single exponential with a half-life of 10 min. FHPAA equilibrated within the first 5 min of the study. CONCLUSIONS Our results show that, unlike FDG, the partitioning of FmT between plasma and erythrocytes is a relatively slow process. We present an analytic correction that may be applied to the measured time course of radioactivity in whole blood to obtain the time course of the tracer in plasma.