Positron emission tomography (PET) using radioligand [¹⁸F]fluorodopa detects reduced striatal dopamine synthesis capacity in Parkinson’s disease (PD) patients. Demographic factors such as sex and BMI are also associated with dopamine synthesis capacity. The combined contribution of demographic and clinical effects however remains elusive.

For this retrospective register-based study, we used baseline [¹⁸F]fluorodopa PET data acquired at the Turku PET Centre between the years 1988–2016 with three scanners (Ecat 931, GE Advance, HRRT). The data involved 350 adult human subjects, including 132 healthy controls, and 218 PD patients. The primary aim was to simultaneously investigate the effects of PD, age, sex and BMI on regional dopamine synthesis capacity (influx rate constant Ki^ref quantified with Patlak in atlas-based regions of interest) using Bayesian linear regression. Secondary aims were to assess (1) interregional correlations of dopamine synthesis capacity, (2) association between regional presynaptic dopamine synthesis and postsynaptic dopamine type 2 receptor (D₂R) availability in subjects who also had a proximal [¹¹C]raclopride PET scan, and (3) scanner effects and atlas- versus MRI-based quantification approaches. We provide the mean dopamine synthesis brain maps of healthy controls and PD patients in NeuroVault.

Dopamine synthesis capacity was drastically reduced in PD patients, decreased with age, increased with BMI, and higher in females versus males. Across regions, the capacity was positively correlated in both patients and controls. We observed support for positive correlation between the dopamine synthesis capacity and the D₂R in caudate nucleus. Scanner had a substantial influence on Ki^ref estimates. Atlas- and MRI-based normalization methods provide largely comparable Ki^ref estimates for most subjects.

Dopamine synthesis capacity is independently affected by PD and demographic factors and correlated between the striatal and thalamic regions in both controls and PD patients. Adjusting for scanner effects in multi-scanner datasets is recommended. When subject-specific MRI is unavailable, atlas-based normalization may be used with caution to prevent major data loss.