Synaptic PET imaging using [11C]UCB‐J in frontotemporal dementia

Abstract

Background: A number of pathophysiological mechanisms are associated with neurodegeneration, including synaptic dysfunction and/or loss. Novel radiotracers which enable the quantification of synaptic proteins in vivo have not previously been explored in frontotemporal dementia (FTD). We used [11 C]UCB‐J PET to measure the density of synaptic vesicle protein 2A (SV2A) in FTD. Method: Six participants with behavioural variant FTD (mean age 61.8, SD 5.4) and 17 cognitively normal participants (63.7, 12.5) underwent 90‐minute dynamic acquisition PET scans following injection of [11 C]UCB‐J, with metabolite corrected arterial input function. Thirteen regions of interest (ROIs) were defined on individual MR images using the CIC anatomical atlas. Regional density was evaluated using the regional V T normalised to the V T in the centrum semiovale (DVR‐1). Groups were compared using Mann‐Whitney tests with Bonferroni correction for multiple comparisons. W‐scores were computed from the 17 controls, with age and gender as covariates, to compare the magnitude of group differences in regional volumes versus [11 C]UCB‐J distribution. Volumes were expressed as a percentage of total intracranial volume. A w‐score of <‐1.65 was considered abnormal. Result: We found significant reductions in regional density of SV2A in FTD compared to controls in multiple cortical and subcortical regions: frontal, temporal, parietal, insula, anterior cingulate and posterior cingulate cortices (all 0.001), as well as the hippocampus, amygdala, thalamus and cerebellum (all 0.001). The w‐score analysis revealed significant reductions in SV2A density in FTD in all 13 ROIs, with the greatest reduction in the anterior cingulate (‐4.2), and in volume in 8 ROIs, with the greatest reduction in the temporal lobe (‐6.4). The magnitude of SV2A density loss was greater than the magnitude of volume loss in most ROIs (9/13), and there was more variability in the reduction in volume than there was in SV2A density (SD range [11 C]UCB‐J 0.7‐2.1, volume 1.4‐3.9). Conclusion: Significant reductions in [11 C]UCB‐J distribution indicate reduced synaptic density in disease‐relevant regions in FTD, which surpass reductions in volume. This suggests that [11 C]UCB‐J PET imaging could be a sensitive biomarker of synaptic dysfunction in FTD, with future studies focused on its relationship with symptoms and whether it becomes abnormal before symptom onset.