Sensitivity of kinetic macro parameters to changes in dopamine synthesis, storage, and metabolism: A simulation study for [¹⁸F]FDOPA PET by a model with detailed dopamine pathway

Quantitative interpretation of brain [¹⁸F]FDOPA PET data has been made possible by several kinetic modeling approaches, which are based on different assumptions about complex [¹⁸F]FDOPA metabolic pathways in brain tissue. Simple kinetic macro parameters are often utilized to quantitatively evaluate metabolic and physiological processes of interest, which may include DDC activity, vesicular storage, and catabolism from ¹⁸F-labeled dopamine to DOPAC and HVA. A macro parameter most sensitive to the changes of these processes would be potentially beneficial to identify impaired processes in a neurodegenerative disorder such as Parkinson's disease. The purpose of this study is a systematic comparison of several [¹⁸F]FDOPA macro parameters in terms of sensitivities to process-specific changes in simulated time-activity curve (TAC) data of [¹⁸F]FDOPA PET. We introduced a multiple-compartment kinetic model to simulate PET TACs with physiological changes in the dopamine pathway. TACs in the alteration of dopamine synthesis, storage, and metabolism were simulated with a plasma input function obtained by a non-human primate [¹⁸F]FDOPA PET study. Kinetic macro parameters were calculated using three conventional linear approaches (Gjedde-Patlak, Logan, and Kumakura methods). For simulated changes in dopamine storage and metabolism, the slow clearance rate (k_loss) as calculated by the Kumakura method showed the highest sensitivity to these changes. Although k_loss performed well at typical ROI noise levels, there was large bias at high noise level. In contrast, for simulated changes in DDC activity it was found that K_i and V_T, estimated by Gjedde-Patlak and Logan method respectively, have better performance than k_loss.