Noninvasive Quantification of [¹⁸F]SynVesT-2 PET in Healthy Human Brains Using Simplified Reference Tissue Models

PET imaging of synaptic vesicle glycoprotein 2A (SV2A) has proven to be a powerful research tool for neurologic disorders. Dynamic SV2A PET scans provide data related to cerebral blood flow and SV2A density, which have been shown to be altered in neurologic disorders such as Alzheimer disease. [¹⁸F]SynVesT-2, an SV2A PET tracer, has demonstrated fast brain kinetics and high specific binding in human brains. To improve clinical feasibility, we evaluated the performance of 3 simplified reference tissue models (SRTMs) in the quantification of [¹⁸F]SynVesT-2 PET data and the minimum scan times required for reliable estimation of relative cerebral blood flow and SV2A density. Methods: Data were pooled from 14 [¹⁸F]SynVesT-2 scans acquired from 9 healthy volunteers. An SRTM, SRTM with a fitted regionally coupled k' ₂ (SRTMC), and SRTM with a population-based k' ₂ (SRTM2) with the centrum semiovale and cerebellum as reference regions were used to calculate nondisplaceable binding potential (BP_ND) and distribution volume ratio (DVR), respectively, as well as the relative tracer delivery rate (R ₁). Test-retest variability (TRV), absolute TRV, and the minimum scan duration for the reliable estimation of R ₁, BP_ND, and DVR were additionally evaluated. Results: Despite time-activity curves being well-described by all 3 models, SRTM generated unreliable BP_ND and DVR values in 9% and 12% of the regions of interest, respectively. SRTMC and SRTM2 resulted in BP_ND and DVR values consistent with those generated from the 1-tissue compartment model. On the basis of the time stability analysis, BP_ND and DVR estimated using SRTM2 converged after 40 min. Using SRTM2, the TRV and absolute TRV estimated from 40-min dynamic scans were -1.0 ± 11.5% and 9.9 ± 5.8% for BP_ND and 1.7 ± 4.0% and 3.6 ± 2.5% for DVR. Conclusion: The parameters of relative cerebral blood flow (R ₁) and specific binding (BP_ND and DVR) can be reliably estimated from a 40-min dynamic [¹⁸F]SynVesT-2 PET scan by SRTM2, which is 30 min shorter than that required for [¹¹C]UCB-J and [¹⁸F]SynVesT-1. The shortened scan time enables the clinical application of dynamic SV2A PET scans to maximize the physiologically relevant information attainable from a single scan.