[¹¹C]-Carfentanil is a high affinity opiate agonist that can be used to localize mu opiate receptors in humans by positron emission tomography (PET). A four-compartment model was used to obtain quantitative estimates of rate constants for receptor association and dissociation. PET studies were performed in five normal subjects in the absence and presence of 1 mg/kg naloxone. Arterial plasma concentration of [¹¹C]-carfentanil and its labeled metabolites were determined during each PET study. The value of k₃/k₄ = B_max/k_d was determined for each subject in the presence and absence of naloxone. There was a significant reduction in the value of k₃/k₄ from 3.4 ± 0.92 to 0.26 ± 0.13 in the thalamus (p < 0.01) and from 1.8 ± 0.33 to 0.16 ± 0.065 in the frontal cortex (p < 0.001). Mean values of frontal cortex/occipital cortex and thalamus/occipital cortex ratios were determined for the interval 35–70 min after injection when receptor binding is high relative to nonspecific binding. The relationship between the measured region/occipital cortex values and the corresponding values of k₃/k₄ in the presence and absence of naloxone was: regions/occipital cortex = 0.95 + 0.74 (k₃/k₄) with r = 0.98 (n = 20). Simulation studies also demonstrated a linear relationship between the thalamus/occipital cortex or frontal cortex/occipital cortex ratio and k₃/k₄ for less than twofold increases or decreases in k₃/k₄. Simulation studies in which thalamic blood flow was varied demonstrated no significant effect on the region/occipital cortex ratio at 35–70 min for a twofold increase or fourfold decrease in blood flow. Therefore, the region/occipital cortex ratio can be used to quantitate changes in k₃/k₄ when tracer kinetic modeling is not feasible.