Saturation analysis in PET--analysis of errors due to imperfect reference regions.

In the determination of specific binding in receptor binding techniques in vitro as well as in vivo, determination of the nonspecific binding as well as the free component is of crucial importance. If a low proportion of specific binding is included when determining the nonspecific binding, relatively large errors may be obtained. In the present study, benzodiazepine (BZ) receptor binding in the human brain was determined in vivo using position emission tomography (PET) by applying a saturation procedure using [11C]flumazenil as an example of this problem. Analysis of the errors in Bmax and KD obtained using Scatchard analysis in PET was performed using a priori information from in vitro [3H]flumazenil binding in the pons, used normally as a reference region in BZ receptor binding studies. Even if the density of BZ receptors in the reference region pons is only 2% compared to that in the frontal cortex, this small proportion of specific binding sites will result in a 10% error in the Bmax and KD values. Simulation of a number of Scatchard plots was performed at varying ratios between the nonspecific and the specific binding.