The role of serotonin in CNS function and in many neuropsychiatric diseases (e.g., schizophrenia, affective disorders, degenerative dementias) support the development of a reliable measure of serotonin receptor binding in vivo in human subjects. To this end, the regional distribution and intrasubject test-retest variability of the binding of [18F]altanserin were measured as important steps in the further development of [18F]altanserin as a radiotracer for positron emission tomography (PET) studies of the serotonin 5-HT2A receptor. Two high specific activity [18F]altanserin PET studies were performed in normal control subjects (n = 8) on two separate days (2-16 days apart). Regional specific binding was assessed by distribution volume (DV), estimates that were derived using a conventional four compartment (4C) model, and the Logan graphical analysis method. For both analysis methods, levels of [18F]altanserin binding were highest in cortical areas, lower in the striatum and thalamus, and lowest in the cerebellum. Similar average differences of 13% or less were observed for the 4C model DV determined in regions with high receptor concentrations with greater variability in regions with low concentrations (16-20%). For all regions, the absolute value of the test-retest differences in the Logan DV values averaged 12% or less. The test-retest differences in the DV ratios (regional DV values normalized to the cerebellar DV) determined by both data analysis methods averaged less than 10%. The regional [18F]altanserin DV values using both of these methods were significantly correlated with literature-based values of the regional concentrations of 5-HT2A receptors determined by postmortem autoradiographic studies (r2 = 0.95, P < 0.001 for the 4C model and r2 = 0.96, P < 0.001 for the Logan method). Brain uptake studies in rats demonstrated that two different radiolabeled metabolites of [18F]altanserin (present at levels of 3-25% of the total radioactivity in human plasma 10-120 min postinjection) were able to penetrate the blood-brain barrier. However, neither of these radiolabeled metabolites bound specifically to the 5-HT2A receptor and did not interfere with the interpretation of regional [18F]altanserin-specific binding parameters obtained using either a conventional 4C model or the Logan graphical analysis method. In summary, these results demonstrate that the test-retest variability of [18F]altanserin-specific binding is comparable to that of other PET radiotracers and that the regional specific binding of [18F]altanserin in human brain was correlated with the known regional distribution of 5-HT2A receptors. These findings support the usefulness of [18F]altanserin as a radioligand for PET studies of 5-HT2A receptors.