Evaluation of the reliability in kinetic analysis for dual tracer injection of FDG and flumazenil PET study

The kinetic analysis for dual tracer injection with 2-input compartment model is challenging in order to assess the two different functions In the same time and same situation. In this study, we investigated the possibility of kinetic analysis with two tracers, ¹⁸F-FDG and ¹¹C-flumazenil (FMZ), by means of the computer simulation. The reliability of estimated parameters was Investigated for various injection protocols and noise levels. Simulated decaying tissue time activity curves were generated for various injection protocols with input function of FDG and FMZ and true k-values by using the 2-input 3-tissue compartment 5-parameter model, i.e. 2-tissue compartment 3-parameter for FDG and 1-tissue compartment 2-parameter for FMZ. The injection Interval of two tracers was changed from 0 to 20 minutes. The noise was generated depending on the total collected count and added each decaying tissue time activity curve. The rate constants for FDG and FMZ were estimated by nonlinear least square method. The reliability of parameter estimates was evaluated by mean absolute difference between true and estimated value of one thousand runs for each injection protocol and noise level. As a result, it was found that parameters were estimated most reliably when FDG was injected 15 minutes later than FMZ injection. In 5% last frame noise, the mean absolute difference between true and estimated value of Ki, reflecting the uptake of FDG, was about 8%, that of DV, distribution volume of FMZ, was 7%. The reliability was independent on the ratio of administration dose of FDG to that of FMZ. In the simulation study, the possibility of kinetic analysis for dual tracer injection was shown