Objectives In PET studies for neuroreceptors, tracer kinetics are described by the two-tissue compartment model (2-TCM), and binding parameters, including the total distribution volume (V T), non-displaceable distribution volume (V ND), and binding potential (BPND), can be determined from model parameters estimated by kinetic analysis. The stability of binding parameter estimates depends on the kinetic characteristics of radioligands. To describe these kinetic characteristics, we previously developed a two-phase graphic plot analysis in which V ND and V T can be estimated from the x-intercept of regression lines for early and delayed phases, respectively. In this study, we applied this graphic plot analysis to visual evaluation of the kinetic characteristics of radioligands for neuroreceptors, and investigated a relationship between the shape of these graphic plots and the stability of binding parameters estimated by the kinetic analysis with 2-TCM in simulated brain tissue time-activity curves (TACs) with various binding parameters. Methods 90-min TACs were generated with the arterial input function and assumed kinetic parameters according to 2-TCM. Graphic plot analysis was applied to these simulated TACs, and the curvature of the plot for each TAC was evaluated visually. TACs with several noise levels were also generated with various kinetic parameters, and the bias and variation of binding parameters estimated by kinetic analysis were calculated in each TAC. These bias and variation were compared with the shape of graphic plots. Results The graphic plots showed larger curvature for TACs with higher specific binding and slower dissociation of specific binding. The quartile deviations of V ND and BPND determined by kinetic analysis were smaller for radioligands with slow dissociation. Conclusions The larger curvature of graphic plots for radioligands with slow dissociation might indicate a stable determination of V ND and BPND by kinetic analysis. For investigation of the kinetics of radioligands, such kinetic characteristics should be considered.
Read full abstract