Simplified ¹¹C-UCB-J PET quantification evaluated in Alzheimer’s disease, epilepsy, and healthy individuals

Objectives: Human PET studies with ¹¹C-UCB-J, a synaptic vesicle protein 2A (SV2A) radiotracer, demonstrated high brain uptake and fast kinetics with excellent test-retest reproducibility (3-9%) for distribution volume (V_T) estimation using the one-tissue compartment (1TC) model (Finnema et al, JCBFM, 2017). However, the 1TC model requires arterial blood sampling and a PET scan of 60-90 min. To simplify the scan protocol, we evaluated the tissue-to-plasma (TTP) and tissue-to-reference (TTR) ratio method for quantification against the gold standard distribution volume (V_T) and binding potential (BP_ND) in Alzheimer’s disease (AD), epilepsy (EP), and healthy controls (HC).

Methods: All PET scans were acquired for 90 min on the HRRT scanner. ¹¹C-UCB-J was administered as a bolus injection over 1 min. Arterial blood sampling and metabolite analysis were conducted to obtain the input function to compute 1TC model parameters. Static TTP and TTR values were calculated at 60-90 min after injection in scans of three datasets: A: 26 HC scans (20 regions of interest (ROIs), including neocortices, putamen, caudate, cerebellum, and hippocampus); B: 19 scans of HCs and ADs (HC: n = 9, AD: n = 10, hippocampus); and C: 5 scans of EPs (ipsilateral and contralateral hippocampus). Regional TTR and BP_ND values were determined using centrum semiovale (CS) as a reference region. For dataset A, % differences of TTP and TTR-1 were computed for comparison with V_T and BP_ND, respectively. In the patient groups, additional analyses were performed to compare the disease-specific changes using modeling and TTR. For dataset B, between-group differences were compared using t-test. For dataset C, asymmetry indices were calculated as 100 x (ipsilateral - contralateral) / contralateral.

Results: In dataset A, regional TTP and TTR-1 correlated well with V_T (R²=0.81, P < 0.0001) and BP_ND (R²=0.93, P < 0.0001), respectively. Since plasma and tissue were not at equilibrium, TTP overestimated V_T (TTP = 1.64 × V_T + 1.94), with similar % difference between TTP and V_T among the examined regions (CS: 79 ± 19%; gray matter: 77 ± 20%). Because there was similar overestimation between the target and reference regions, TTR-1 was very similar to BP_ND (TTR-1 = 0.96 × BP_ND + 0.03) and the % difference between TTR-1 and BP_ND was very close to zero (-2% ± 7%). In dataset B, % differences of TTP and TTR-1 were similar to those of the dataset A (TTP: 75 ± 21%; TTR-1: 3 ± 9%). Hippocampus TTR-1 matched well with BP_ND (TTR-1 = 1.00 × BP_ND + 0.03, R²=0.98, P < 0.0001). TTR-1 in the AD group slightly underestimated BP_ND (TTR-1: 1.48 ± 0.29 for HC and 0.90 ± 0.60 for AD, 1TC BP_ND: 1.49 ± 0.34 for HC and 0.94 ± 0.60 for AD). The HC-AD group difference was significant using both BP_ND and TTR-1 (TTR-1: P = 0.028, 1TC BP_ND: P = 0.018). In dataset C, the % overestimation of TTP over V_T was larger (92 ± 12%) than that of the dataset A, but the % difference between TTR-1 and BP_ND was equally small (0 ± 8%). Hippocampus TTR-1 matched well with BP_ND (TTR-1 = 1.18 × BP_ND - 0.26, R²=0.94, P < 0.0001). Asymmetry indices were -14 ± 29% with 1TC BP_ND and -17 ± 25% with TTR-1. Simulations based on the fitted parameters suggested that the overall excellent agreement between TTR-1 and BP_ND is time-dependent, i.e., TTR-1 would overestimate BP_ND by ~36% (TTR-1 = 1.36 × BP_ND + 0.36) at later times (~5 hours post-injection) once transient equilibrium is reached [2].

Conclusions: TTR-1 from 60-90 min matched extremely well with 1TC BP_ND in multiple subject groups (HCs, ADs, EPs), suggesting that a short scan after tracer injection may be sufficient for accurate quantification of ¹¹C&#8209;UCB&#8209;J specific binding. Since the agreement between TTR-1 and 1TC BP_ND is expected to be time-dependent, careful validation of simplified methods should be performed in each patient cohort. Research Support: 1R01NS094253-01, 1R01AG052560-01A1 References: [1] Finnema et al, J Cereb Blood Flow Metab, 2017, epub. [2] Carson et al, J Cereb Blood Flow Metab, 13:24-42, 1993