@article {Rossano74, author = {Samantha Rossano and Richard Carson}, title = {SV2A synaptic density imaging in the neonate rhesus macaque brain: Analysis with partial volume correction}, volume = {61}, number = {supplement 1}, pages = {74--74}, year = {2020}, publisher = {Society of Nuclear Medicine}, abstract = {74Objectives: Synaptic development begins mid-gestation in the nonhuman primate (NHP), where synaptic density increases to exceed the level of the adult at the time of birth, and plateaus at a level over 100\% of the adult by 4 months old1. We have previously used 11C-UCB-J PET longitudinally to image SV2A density in the developing brain, to investigate synaptogenesis. However, the PET uptake levels in the neonate brain relative to the adult did not suggest an overproduction of synapses as suggested by previous literature, as the SUVs did not exceed the adult level, but approached it at 4-6 months old, not at birth. Partial volume effects (PVE) lead to inaccuracies in PET quantification, and are caused by spill-over of activity between neighboring regions due to limited resolution. As the brain grows, regional volumes and PET tracer distribution patterns change, which can contribute to PVE that varies at different ages. Many partial volume correction (PVC) methods have been established. Here, we used the geometric transfer matrix (GTM) method2,3 to correct regional mean values of 11C-UCB-J SUV in the neonate NHP brain to determine whether PVC can reduce the discrepancy between synaptic density levels relative to the adult using in vivo SV2A imaging and ex vivo histology methods. Methods: Six 11C-UCB-J PET studies were completed in one neonate rhesus macaque (Macaca mulatta) performed monthly from 1-6 months old, with a follow up scan at 9 months old. Images were acquired on a Focus220 scanner, and reconstructed using filtered back projection, with image resolution (FWHM) of 2.5mm and 2.7mm in transaxial and axial directions, respectively. Summed PET images (0-30 minutes) were registered using an affine transform to an age-appropriate neonate rhesus MRI template4 (template ages: 2 weeks, 3 months, or 6 months) to quantify PET data in 5 cortical (frontal, temporal, occipital, and parietal cortices, and cerebellum) and 5 subcortical( hippocampus, amygdala, putamen, caudate, subcortical thalamic and hypothalamic structures) gray matter (GM) ROIs. Standardized uptake value (SUV) images from the first 30 minutes of each scan were used for PVC. The GTM method was implemented as follows: Binary ROI maps were transformed from template space into PET space and blurred with a 3D Gaussian filter to PET resolution. The GTM matrix elements were calculated as the total voxel intensity in region i on the smoothed binary image of region j, divided by the number of voxels in region i. The inverse of the GTM matrix was multiplied by the vector of ROI SUVs to produce PVC values and the percent change due to PVC in each region was calculated. Results: PVC increased SUV in all GM ROIs in all scans. Patterns of SUV changes across scans were similar before (observed) and after (corrected) PVC. The observed, corrected, and \% change in values are shown for each scan in Table 1. When corrected, 11C-UCB-J PET uptake increases and plateaus at a level closer to the adult (peak SUV of 5-8 in non-PVC GM regions 5). The observed, corrected, and the \% change in SUV for all 10 regions are shown in Table 2. The greatest PVC effect was observed in the thalamic GM, with an average \% change of 100\%. The lowest \% change was observed in the putamen, with an increase of 29\%. Conclusions: GTM was implemented for PVC of 11C-UCB-J images in the neonate NHP. PVC increased the SUV in all GM ROIs investigated, reducing the differences of synaptic density with respect to adult observed between current in vivo and previous ex vivo studies. As expected, the \% change in SUV due to PVC decreased with age in most regions of interest (e.g., frontal cortex SUV PVC changes was 63\% at 1 month old vs. 56\% at 9 months old), due to brain growth. Ongoing and future PET studies of neonate NHPs will include acquisition of structural MR, which can improve the quality PET image registration to the template. View this table:Table 1: Observed SUV, Corrected SUV, and \% change in cortical and subcortical ROIs View this table:Table 2: Average regional SUV estimates before (obs) and after (corr) PVC, with average \% change}, issn = {0161-5505}, URL = {https://jnm.snmjournals.org/content/61/supplement_1/74}, eprint = {https://jnm.snmjournals.org/content}, journal = {Journal of Nuclear Medicine} }