Kinetic analysis of neuroreceptor binding using PET

Abstract Most quantitative dynamic PET studies of radioligand binding to neuroreceptor sites in brain are analysed using models derived from a standard three-tissue compartment model. The full model has the advantage that the individual rate constants (microparameters) can be readily interpreted in terms of physiological and pharmacological properties of the system such as blood flow, transport across the blood–brain barrier, receptor concentration and affinity, or binding potential, etc. In practice, because of factors such as the signal to noise ratio, condensed models with a reduced number of compartments are often used to analyse PET data. These analyses then rely on the estimation of macroparameters, such as the total volume of distribution of the radioligand in target and reference tissues, to estimate binding potential. Whilst robust parameter estimates are obtained using these models, there is less transparency in the analysis and it may be difficult, for example, to ascertain the mechanism or cause of an observed difference in the apparent binding potential between study conditions or between regions of brain. The relationship of the micro- and macroparameters to the underlying physiology is discussed, and an approach to the estimation of microparameters based on the simultaneous estimation of multiple regions sharing common parameters is introduced and applied to a PET study of the NK1 receptor site in man.