Study protocols and patient consent

The protocol for this academic, investigator-initiated, cross-sectional study was approved by the Ethics Committee of the Hospital District of Southwest Finland. All study participants provided signed informed consent and the study was conducted in accordance with the World Medical Association’s declaration of Helsinki.

Study participants

Ten patients with early stage, ten with moderate-stage idiopathic PD and six healthy controls were enrolled for this study. All patients fulfilled the clinically defined criteria for idiopathic Parkinson’s Disease without dyskinesia [18]. PD was considered to be at early stage if time from diagnosis was < 5 years, and at moderate stage if time from diagnosis was between 5 and 15 years. PD cohorts were recruited from the outpatient neurology clinics of Turku University Hospital and through study advertisements published on the Finnish Parkinson Foundation forums. Exclusion criteria included the presence of drug-induced dyskinesia, history of other neurological or psychiatric diseases, or another significant comorbidity. Demographics and clinical features of PD patients and healthy controls are summarized in Table 1. To avoid the possible short-term effects of dopaminergic medications on A2A receptor availability, PET scans for PD patients were performed after 12 h (for standard levodopa) and 24 h (for prolonged release levodopa/carbidopa, MAO-B inhibitors and/or dopamine agonists) cessation of dopaminergic medication (Supplementary Table 1). Moreover, all study subjects were instructed to avoid consuming caffeinated beverages within 24 h before the scanning considering the antagonistic action of caffeine on A2A receptors and half-life of caffeine of 2.3–9.9 h [19,20,21].

Table 1 Demographics and clinical parameters of healthy controls and PD patientsClinical assessment

All patients underwent a thorough medical and neurological examination prior to the PET scan. Definition of the disease duration, Unified Parkinson’s disease rating scale (UPDRS) I–V scoring, Mini Mental Status Examination (MMSE) and calculation of the Levodopa Equivalent Dose (LED) were acquired as part of the examination.

[11C]TMSX radioligand production and administration

Production of [11C]TMSX radioligand was performed according to the methodology described in detail in our earlier study [22]. The radiochemical purity of the produced radioligand was 97.6 (± 0.5) percent. At the start of the dynamic PET scan, a smooth single bolus of [11C]TMSX was injected into the left antecubital vein and was subsequently flushed with saline. The mean (± SD) injected dose (MBq) of [11C]TMSX for controls, and for early-stage and moderate-stage PD patients was 490 (± 25.1), 483 (± 26.6) and 492 (± 21.0), respectively, without significant differences in doses between the groups (P = 0.705). One patient from the early PD group and one from moderate PD group were excluded from the final analyses, since the [11C]TMSX PET data were not reliably quantifiable due to the relatively low injected dose of the radiotracer and high injected mass, respectively.

PET and MR acquisition and image processing

A 60-min dynamic brain PET scan with [11C]TMSX radiotracer was acquired using the ECAT HRRT scanner (Knoxville, USA). Tissue attenuation maps were obtained with a 6-min transmission scan for attenuation correction using 137Cs point source prior to the dynamic scan. Twenty-seven timed frames (6 × 10, 1 × 30, 5 × 60, 5 × 150, and 8 × 300 s) lasting a total of 3600 s were used for PET image reconstruction in accordance with methods described in our earlier study [22].

All subjects underwent brain MRI with 1.5T Nova Dual scanner (Philips, Netherlands). Axial 3D T1 weighted images were used as anatomic reference for PET images. The T1 images were co-registered with motion corrected dynamic PET scans, where the image processing was carried out using Statistical Parametric Mapping software (SPM12, The Wellcome Centre for Human Neuroimaging, UCL, London) as detailed in our earlier study [23]. For the evaluation of region-specific radioligand binding, reference tissue input Logan graphical analysis within 20–60-min interval was used to estimate distribution volume ratios (DVR). This was applied to region of interest (ROI)-based time-activity curves (TAC) using clustered cerebral grey matter as a reference region acquired with a supervised clustering algorithm (modified from the Super PK software) [24]. The supervised clustering algorithm assumes that each dynamic PET data TAC is a combination of four kinetic class TACs that correspond to healthy grey matter, white matter, blood and high specific binding gray matter. In this study, kinetic class TACs corresponding to normal [11C]TMSX binding in grey matter, white matter and blood were first defined from a healthy volunteer group (n = 7). The kinetic class TAC corresponding to high specific [11C]TMSX binding was acquired from thalamus in multiple sclerosis patients (n = 12) and was compared with anterior putamen binding in PD patients (n = 9), as described in [23]. Next, the contribution of each kinetic class to the voxel-level [11C]TMSX TACs was estimated by using non-negative least squares estimation. Lastly, the clustered reference region TAC was calculated from the dynamic [11C]TMSX PET data as a weighted average by using the grey matter coefficients as weighting factors. Parametric maps for voxel-level analysis were generated using non-displaceable binding potential (BPND) images using basis function implementation of simplified reference tissue model (SRTM) with 250 basis functions, and by using the values 0.06 and 0.8 for the lower and upper limit for θ3.

Region of interest and voxel-level analyses of the [11C]TMSX PET data

For the ROI analyses, Freesurfer software (v6.0, http://surfer.nmr.mgh.harvard.edu/) was used for anatomical parcellation of the 3D T1 MR images to extract region-specific DVR values from the co-registered PET scans [25]. Analyses were limited to three bilateral ROIs, i.e., caudate, putamen and pallidum for the primary outcome.

Mean DVR values of the whole dorsal striatum were calculated using a combined mask of caudate and putamen, to check associations with clinical parameters following previous studies [15]. To assess the within and between-group effect of lateralization, contralateral and ipsilateral ROIs were drawn manually for caudate and putamen using in-house software (Carimas v2.9). These unilateral ROIs were defined as contralateral or ipsilateral with respect to each patient’s clinically more affected side, based on the respective unilateral sums of the UPDRS III scores. Comparison of DVRs in ipsilateral or contralateral ROIs of the PD groups with healthy controls was carried out against the combined average DVR of the left and right striatal ROIs of controls.

Parametric [11C]TMSX BPND images were used to evaluate voxel-wise differences between healthy controls and patients with early- and moderate-stage PD utilizing SPM12. Images were smoothed with 3D Gaussian 4 mm FWHM filter to improve signal-to-noise ratio and ensure normal data-distribution before voxel-wise analyses.

Statistical methods

ROI-level statistical analyses were performed with GraphPad Prism software (v9.1, California, USA). Normality of distribution was checked graphically and using the Shapiro–Wilk test. Differences in categorical variables were evaluated using non-parametric Fischer’s exact test. Mann–Whitney U test was used to assess differences in disease duration between patients with early- vs. moderate-stage PD. One-way ANOVA was utilized to evaluate differences in age, molar activity, injected mass, radiochemical purity and injected radioligand dose between groups. To investigate local changes at the ROI level, an independent t-test was used with Bonferroni correction for multiple comparisons applied afterward. Paired t-test was utilized to calculate the within-group differences of ipsilateral and contralateral sides determined by the clinically more affected side of the patient. Cohen’s d was calculated to measure the effect size for statistically significant comparisons. Spearman test was used to evaluate correlations between clinical parameters and striatal and pallidal [11C]TMSX DVR values. P-values < 0.05 from two-tailed tests were considered statistically significant in ROI-level analyses.

For the voxel-level analyses in SPM12, a two-sample t-test was carried out using parametric BPND image with overlaid striatal mask, including pallidum, in MNI-152 space. The cluster-based inference with cluster-defining threshold of P = 0.001 (T = 3.85) was used, and the P < 0.05 family-wise error-corrected critical cluster size was 115 voxels.