Development of [18F]-Flumazenil-PET for localisation of the epileptogenic zone in patients with medically refractory focal epilepsy.

Introduction: Studies of GABAA/central benzodiazepine receptor (GABAA/cBZR) distribution in the CNS using [11C]-flumazenil-PET (FMZ-PET) have enabled localisation of the epileptogenic zone (EZ) in patients with medically refractory epilepsy. [11C]-FMZ-PET images show a more restricted region of abnormality with increased sensitivity when compared with FDG-PET. However, use of [11C]-FMZ in routine clinical practice has been hindered by practical limitations of [11C]. The aim of the current study is to develop an [18F]-radiolabelled FMZ tracer with high specificity and improved imaging quality for EZ localisation in routine clinical practice. Methods: Preclinical studies: Presaturation, displacementand uptake PET scans were performed to define the kinetics of [18F]-FMZ binding in non-epileptic rats (n=8). Bmax (receptor density) and KD (binding affinity) were then quantified in the brains of epileptic (n=9; kainic acid-induced model of temporal lobe epilepsy) vs. non-epileptic control animals (n=10). Clinical study: 4 patient groups have been studied; healthy controls (n=20), patients with well-localised TLE from MRI scans (n=10), patients with lateralised TLE(n=7), and patients with other focal epilepsies (n=4). Each participant underwent a single 60 minute dynamic [18F]-FMZ-PET scan. Patients also underwent an FDG-PET scan. Blinded visual assessment of images to locate the EZ was performed. Parametric images of binding potential (BP) were generated. Datasets were processed using ROI analysis and SPM to assess differences in BP between patients and controls and localisation of the EZ in patients. Results:[18F]-FMZ was shown to be a suitable PET radiotracer for imaging GABAA/cBZR in vivo, with reversible and competitive binding and low non-specific binding. [18F]-FMZ-PET reliably detected decreased Bmax in the hippocampi of epilepticrats (left 16.3, right 15.9) compared with controls (left 20.9, right 19.8, p=0.022, p=0.049), with no change in KD (left 8.24 vs 8.46, p=0.82, right 7.43vs 8.07 p=0.56). There were no changes in whole brain Bmax or KD. To date the visual assessment of the clinical data has shown [18F]-FMZ-PET to have high sensitivity (100%) and positive predictive value (100%) for the EZ in patients, with a more restricted localisation of the EZ compared to FDG-PET. Quantitative analysis is ongoing. Conclusions: The pre-clinical studies have demonstrated that [18F]-FMZ-PET is a reliable radiotracer for quantification of CNSGABAA/cBZR expression in vivo. Preliminary analysis in our current clinical study indicates that [18F]-FMZ-PET also has excellent imaging characteristics in humans, and shows promise as a new clinical tool for localising the EZ in TLE patients