[11C]UCB-J PET white matter ROI optimization for Parkinson’s Disease imaging of synaptic density

Introduction: [¹¹C]UCB-J, a radioligand for synaptic vesicle glycoprotein 2A (SV2A), is a potential in vivo biomarker for synaptic density (1, 2). [¹¹C]UCB-J can be used to investigate changes in synaptic density in several neurological diseases, including Parkinson’s disease (PD) (3). PD is characterized by depletion of central and peripheral dopamine, a neurotransmitter generated in the substantia nigra (SN), leading to progressively worsening difficulties with movement, balance, and cognition (4). To fully evaluate the specific binding of [¹¹C]UCB-J in the SN, a reference region with low specific binding and with minimal noise is needed. The centrum semiovale (CS) has been used as a reference region, but its relatively high noise due to its small size is detrimental. In this study, we generated, optimized, and evaluated new white matter (WM) regions of interest (ROIs) derived from individual MR T1-weighted images (T1Wis) for more reliable synaptic density measurements in the SN of healthy controls (HC) and PD subjects.

Methods: Nine HCs (age: 59.2±7.8 y, M:F = 3:6) and seven PD subjects (age: 60.6±10.5 y, M:F=2:5) participated in this study. All subjects underwent at least 60 min ¹¹C-UCB-J PET acquisition with arterial blood sampling and metabolite assay. The volume of distribution (V_T) was estimated by one-tissue compartment modeling. The ROIs of SN and CS were defined in the automated anatomic labeling (AAL) space. Those ROIs were applied to the individual V_T images by using non-linear transformation from AAL space to individual T1WIs and linear transformation from individual T1WIs to individual PET image space. The individual T1WIs were segmented by FreeSurfer (FS) 6.0 and cerebral WM ROI was smoothed and thresholded to generate a sequence of FSWM ROIs ranging in size from 0.5 to 200 mL (Fig A). The distribution volume ratio (DVR) within the SN was calculated using CS or FSWM ROIs of different sizes as a reference region. Partial volume correction was applied to the V_T in CS and FSWM ROIs using iterative Yang algorithm with FS segmentation. The DVRs in SN were compared between HC and PD groups using a t-test without correction for multiple comparisons.

Results: Fig B shows the V_T per CS and FSWM volumes for both groups. The percent differences of V_T in CS and each FSWM ROI between groups ranged from 2.4 – 6.9%, with increasing difference with increasing ROI size. This indicates that due to their size, larger FSWM ROIs capture increasing spillover from the specific binding in gray matter regions that could be different between PD and HC subject groups (3). Note that there was no statistically significant group differences in V_T for CS or other sizes of FSWM ROIs.

Fig C shows the mean DVR in the SN for both groups using the CS and each FSWM ROI as the reference region. The DVR in the SN was significantly different between HC and PD groups (p<0.05) using each reference region, with the 8 mL reference region exhibiting the highest percent differences (27.4%) with the highest effect size (Cohen’s d =2.80).

Conclusions: Compared to the CS, V_T using the 8 mL FSWM ROI showed the lowest difference in PD vs HC groups. Moreover, compared to CS-derived results, the SN DVR calculated using the 8 mL FSWM ROI as a reference region exhibited lower variability, as well as the largest and most significant difference between the PD and HC groups. Taken together, these results suggest that the 8 mL FSWM ROI may be an optimal reference region for studying synaptic density changes in PD and other neurodegenerative diseases using [¹¹C]UCB-J.

• Download figure
• Open in new tab
• Download powerpoint