PET/CT Imaging and PET ROI Analysis

At post-natal days 55 and 90 (P55 and P90), age- and sex-matched male GF mice and HG mice were imaged under isoflurane gas anesthesia on a small animal PET/CT scanner (Siemens Inveon Hybrid). CT was acquired for attenuation correction and anatomic localization. Dynamic PET imaging was carried out under anesthesia for 45 min. 5.5 MBq ± 0.4 (range 4.7–6.0) of ¹⁸F-SynVesT-1 was administered intravenously through the lateral tail vein. The 45-min ¹⁸F-SynVesT-1 PET acquisition was reconstructed into 1-min-per-frame images, starting at t = 60 s using 3-dimensional ordered-subset expectation maximization (OSEM, 2 iterations, 16 subsets) followed by a maximum a posteriori probability (MAP) algorithm. The PET and CT images were reconstructed at 128 × 128 × 159 (0.78 × 0.78 × 0.8 mm) and 480 × 480 × 635 (0.21 × 0.21 × 0.21 mm), respectively. PET images were converted to standardized uptake values (SUV). 3D regions of interest (ROIs) were manually placed over the whole brain and SUV_mean (herein referred to as SUV) at each frame was recorded using MIM (MIM Software Inc., Cleveland OH). One GF mouse was excluded from the ¹⁸F-SynVesT-1 analysis due to substantial motion. SUVs at each time point were averaged within each mouse cohort and ¹⁸F-SynVesT-1 time-activity curves were generated for each cohort. The area under the curve (AUC) for GF and HG cohorts were calculated at the peak stabilized interval (4.5–10.5 min).

After 24 h, the same mice were administered 10.7 MBq ± 0.2 (range 10.4–11.1) of ¹⁸F-flurodeoxyglucose (FDG) through the lateral tail vein. Sixty minutes post injection, PET imaging was performed acquiring approximately 50 million counts per mouse. As with the ¹⁸F-SynVesT-1 PET data, ¹⁸F-FDG PET images were reconstructed using 3-dimensional ordered-subset expectation maximization (OSEM, 2 iterations, 16 subsets) followed by a maximum a posteriori probability (MAP) algorithm. CT attenuation, scatter and decay correction were applied to all datasets. The PET and CT images were also reconstructed at the spatial resolutions of 128 × 128 × 159 (0.78 × 0.78 × 0.8 mm) and 480 × 480 × 635 (0.21 × 0.21 × 0.21 mm), respectively.

The Waxholm MR atlas (23) was used to identify the hippocampus, neocortex, amygdala, ventral thalamus, lateral thalamus, globus pallidus, caudate putamen, hypothalamus, and accumbens. After upsampling the PET images to match the resolution of the CT images, brain masks were semi-automatically generated by thresholding the CT image in MATLAB to include brain tissue and were refined using subsequent image dilation and filing steps as needed. All brain masks were visually confirmed for each subject. The Waxholm MR atlas was individually registered to the CT brain masks of each subject using affine registration in MATLAB (imregtform). Then, the registered Waxholm atlases were used to compute SUV values for each brain region.