Test-retest reproducibility of binding parameters in humans with 11C-LY2795050, an antagonist PET radiotracer for the κ opioid receptor

(11)C-LY2795050 is a new antagonist PET radioligand for the κ opioid receptor (KOR). In this study, we assessed the reproducibility of the binding parameters of (11)C-LY2795050 in healthy human subjects. Sixteen healthy subjects (11 men and 5 women) underwent 2 separate 90-min PET scans with arterial input function and plasma free fraction (fP) measurements. The 2-tissue-compartment model and multilinear analysis-1 were applied to calculate 5 outcome measures in 14 brain regions: distribution volume (VT), VT normalized by fP (VT/fP), and 3 binding potentials (nondisplaceable binding potential, binding potential relative to total plasma concentration, and binding potential relative to free plasma concentration: BPND, BPP, BPF, respectively). Since KOR is distributed ubiquitously throughout the brain, there are no suitable reference regions. We used a fixed fraction of individual cerebellar VT value (VT,CER) as the nondisplaceable VT (VND) (VND = VT,CER/1.17). The relative and absolute test-retest variability and intraclass correlation coefficient were evaluated for the outcome measures of (11)C-LY2795050. The test-retest variability of (11)C-LY2795050 for VT was no more than 10% in any region and was 12% in the amygdala. For binding potential (BPND and BPP), the test-retest variability was good in regions of moderate and high KOR density (BPND > 0.4) and poor in regions of low density. Correction by fP (VT/fP or BPF) did not improve the test-retest performance. Our results suggest that quantification of (11)C-LY2795050 imaging is reproducible and reliable in regions with moderate and high KOR density. Therefore, we conclude that this first antagonist radiotracer is highly useful for PET studies of KOR.