A computer simulation study on the input function sampling schedules in tracer kinetic modeling with positron emission tomography (PET)

Abstract

Tracer kinetic modeling with positron emission tomography (PET) requires measurements of the time-activity curves in both plasma (PTAC) and tissue (TTAC) to estimate physiological parameters, i.e. to fit the parameters of certain compartmental models using PTAC and TTAC as the model input and output functions, respectively. In this paper, we first explored the optimal blood sampling schedule (OBSS) for the input function, based on the tracer [18F]2-fluoro-2-deoxy-D-glucose (FDG) blood sample experimental data. Then using a 5-parameter FDG model we investigated the effects of the plasma sampling schedule, as well as PTAC measurement noise, on the estimation accuracy and reliability of FDG model macro- and micro-parameters and the physiological parameter local cerebral metabolic rates of glucose (LCMRGlc), using computer simulation. Three different methods were used: (a) estimation of the FDG model parameters ignoring PTAC noise using the traditional PTAC schedule (non-OBSS); (b) estimation of the PTAC model parameters and FDG model parameters simultaneously using both non-OBSS and OBSS; (c) estimation of the PTAC model parameters first, then the FDG model parameters using both non-OBSS and OBSS. The results show that OBSS can provide more reliable estimates and largely simplifies the experiment operations.