A.7 In vivo hippocampal mGluR5 abnormalities predict MTLE post-surgical outcome

Background: PET imaging of [ 11 C]ABP688 shows reduced hippocampal mGluR5 availability in mesial temporal lobe epilepsy (MTLE) patients, however the relation with post-surgical outcomes is unclear. Here, we tested whether [ 11 C]ABP688 binding in hippocampal subfields vulnerable to glutamate excitotoxicity is related to post-surgical outcome. Methods: [ 11 C]ABP688-PET was obtained from 31 unilateral MTLE patients and 30 controls. Hippocampal subfields were automatically segmented into 1) CA1-3, 2) CA4/dentate gyrus (DG), and 3) Subiculum and manually corrected. Partial volume corrected [ 11 C]ABP688 non-displaceable binding potential (BP ND ) was calculated in the subfields and compared between seizure-free and non-seizure-free patients. Results: [ 11 C]ABP688 BP ND was significantly reduced in ipsilateral CA1-3 & CA4/DG (p<0.001) compared to controls. No difference was seen in Subiculum. Ipsilateral CA1-3 [ 11 C]ABP688 BP ND was lower in seizure-free (p=0.012; Engel Ia, n=13) vs non-seizure-free (Engel Ic-III, n=10) patients, and this effect was independent of subfield volume. In a subset of patients with [ 18 F]FDG-PET, CA1-3 [ 11 C]ABP688 BP ND was significantly lower in seizure-free patients (p=0.03), while no difference was found for [ 18 F]FDG uptake. Conclusions: Reduced CA1-3 mGluR5 availability was associated with post-surgical seizure-freedom independent of atrophy and hypometabolism. Thus, [ 11 C]ABP688-PET may offer a potential biomarker for surgical outcomes and may be particularly relevant for pre-surgical workup in MRI- and [ 18 F]FDG-negative MTLE patients.