[11C]Ro15-4513 has been introduced as a PET radioligand to image the GABAAα5 receptor
subtype thought to be important in learning, memory and addiction. However, the in vivo
selectivity of the ligand remains unknown and a full assessment of different analysis
approaches has yet to be performed. Using human heterologous competition data, with
[11C]Ro15-4513 and the highly selective GABAAα5 selective negative allosteric modulator
Basmisanil (RG1662), we quantify the GABAAα5 selectivity of [11C]Ro15-4513, assess the
validity of reference tissues and evaluate the performance of four different kinetic analysis
methods. The results show that [11C]Ro15-4513 has high but not complete selectivity for
GABAAα5, with α5 representing around 60-70% of the specific binding in α5 rich regions.
Competition data indicate that the cerebellum and pons are essentially devoid of α5 signal
and might be used as reference regions under certain conditions. Off-target non-selective
binding to other GABAA subtypes means that the choice of analysis method and the
interpretation of outcome measures must be considered carefully. We discuss the merits of
two tissue compartmental model analyses to derive both and
, band-pass spectral
analysis for estimation of and the simplifified reference tissue model for estimation of BPnd.