Test–retest reproducibility of cerebral adenosine A 2A receptor quantification using [ 11 C]preladenant

Objective To evaluate the reproducibility of cerebral adenosine A 2A receptor (A 2A R) quantification using [ 11 C]preladenant ([ 11 C]PLN) and PET in a test–retest study. Methods Eight healthy male volunteers were enrolled. Dynamic 90 min PET scans were performed twice at the same time of the day to avoid the effect of diurnal variation. Subjects refrained from caffeine from 12 h prior to scanning, and serum caffeine was measured before radioligand injection. Arterial blood was sampled repeatedly during scanning and the fraction of the parent compound in plasma was determined. Total distribution volume ( V T ) was estimated using 1- and 2-tissue compartment models (1-TCM and 2-TCM, respectively) and Logan graphical analysis (Logan plot) ( t * = 30 min). Plasma-free fraction ( f P ) of [ 11 C]PLN was measured and used for correction of V T values. Distribution volume ratio (DVR) was calculated from V T of target and reference regions and obtained by noninvasive Logan graphical reference tissue model (LGAR) ( t * = 30 min). Absolute test–retest variability (aTRV), and intra-class correlation coefficient (ICC) of V T and DVR were calculated as indexes of repeatability. Correlation between DVR and serum concentration of caffeine (a nonselective A 2A R blocker) was analyzed by Pearson’s correlation analysis. Results Regional time–activity curves were well described by 2-TCM models. Estimation of V T by 2-TCM produced some erroneous values; therefore, the more robust Logan plot was selected as the appropriate model. Global mean aTRV was 20% for V T and 14% for V T / f P (ICC, 0.72 for V T and 0.87 for V T / f P ). Global mean aTRV of DVR was 13% for Logan plot and 10% for LGAR (ICC, 0.70 for Logan plot and 0.81 for LGAR). DVR estimates using LGAR and Logan plot were in good agreement ( r 2 = 0.96). Coefficients of variation for V T , V T / f P , DVR (Logan plot), and DVR (LGAR) were 47%, 47%, 27%, and 18%, respectively. Despite low serum caffeine levels, significant concentration-dependent effects on [ 11 C]PLN binding to target regions were observed ( p < 0.01). Conclusions In this study, moderate test–retest reproducibility and large inter-subject differences were observed with [ 11 C]PLN PET, possibly attributable to competition by baseline amount of caffeine. Analysis of plasma caffeine concentration is recommended during [ 11 C]PLN PET studies. Trial registration UMIN000030040.