Tracer kinetic modeling of [11C]AFM, a new PET imaging agent for the serotonin transporter

Objectives: [11C]AFM is a new radioligand for the serotonin transporter which showed higher specific binding than [11C]DASB in nonhuman primate studies. The aim of this study was to determine an appropriate kinetic model to quantify [11C]AFM binding in humans.Methods: Four healthy subjects were involved in a test-retest protocol. Dynamic PET data were acquired for 120 min with an HRRT scanner. Arterial blood sampling and metabolite analysis were performed. A 1-exponential (1e), 2-exponential (2e), and Hill function were fitted to the parent fraction data. Regional time-activity curves (TACs) were obtained and kinetic analysis was performed using one-tissue (1T) or two-tissue (2T) models, and Logan graphical analysis (LGA).Results: Metabolism of AFM was rapid with % parent of 31 plusminus 6.0 and 8.1 plusminus 4.0 at 30 and 90 min, respectively. The 2T model resulted in nonconvergence or unstable estimates. For the 1e and Hill models, means of residuals were far from zero compared with the 2e model. The 1T model with 2e metabolite fitting described the data very well in ROIs with high VT values. Moderate lack of fit was seen in ROIs with low VT in 3/8 scans. VT ranged from 6.5 plusminus 1.0 (cerebellum) to 23.9 plusminus 3.7 (thalamus), with K1 values of 0.33 to 0.43 mL/min/mL. LGA estimates were in excellent agreement with 1T: VT(LGA) = 0.94 VT(1T) + 0.78 (r2=0.99).Conclusions: The 1T model well describes the kinetics of [11C]AFM in most cases. The lack of fit in the remaining cases may be due to uncertainties in the metabolite measurement, and suggest that the ultimate development of a reference region approach is desirable.Research Support: NIMH (MH066624) and Grants-in-Aid for Scientific Research of JSPS (18-6916)2008 SNM Annual Meetin