Test-retest reproducibility of [11C]PBR28 binding to TSPO in healthy control subjects

Purpose The PET radioligand [C-11]PBR28 binds to the translocator protein (TSPO), a marker of brain immune activation. We examined the reproducibility of [C-11]PBR28 binding in healthy subjects with quantification on a regional and voxel-by-voxel basis. In addition, we performed a preliminary analysis of diurnal changes in TSPO availability. Methods Twelve subjects were examined using a high-resolution research tomograph and [C-11]PBR28, six in the morning and afternoon of the same day, and six in the morning on two separate days. Regional volumes of distribution (V-T) were derived using a region-of-interest based two-tissue compartmental analysis (2TCM), as well as a parametric approach. Metabolite-corrected arterial plasma was used as input function. Results For the whole sample, the mean absolute variability in V (T) in the grey matter (GM) was 18.3 +/- 12.7 %. Intraclass correlation coefficients in GM regions ranged from 0.90 to 0.94. Reducing the time of analysis from 91 to 63 min yielded a variability of 16.9 +/- 14.9 %. There was a strong correlation between the parametric and 2TCM-derived GM values (r=0.99). A significant increase in GM V-T was observed between the morning and afternoon examinations when using secondary methods of quantification (p=0.028). In the subjects examined at the same time of the day, the absolute variability was 15.9 +/- 12.2 % for the 91-min 2TCM data. Conclusion V-T of [C-11]PBR28 binding showed medium reproducibility and high reliability in GM regions. Our findings support the use of parametric approaches for determining [C-11]PBR28 V-T values, and indicate that the acquisition time could be shortened. Diurnal changes in TSPO binding in the brain may be a potential confounder in clinical studies and should be investigated further.