In vivo hippocampal cornu ammonis 1–3 glutamatergic abnormalities are associated with temporal lobe epilepsy surgery outcomes

OBJECTIVE: Previous positron emission tomography (PET) studies using [¹¹ C]ABP688 show reduced metabotropic glutamate receptor type 5 (mGluR5) allosteric binding site availability in the epileptogenic hippocampus of mesial temporal lobe epilepsy (MTLE) patients. However, the link between mGluR5 abnormalities and postsurgical outcomes remains unclear. Here, we test whether reduced PET [¹¹ C]ABP688 binding in cornu ammonis (CA) sectors more vulnerable to glutamatergic excitotoxicity relates to surgical outcomes. METHODS: We obtained magnetic resonance imaging (MRI) and [¹¹ C]ABP688-PET from 31 unilateral MTLE patients and 30 healthy controls. MRI hippocampal subfields were segmented using FreeSurfer. To respect the lower PET special resolution, MRI-derived anatomical subfields were combined into CA1-3, CA4/dentate gyrus, and Subiculum. Partial volume corrected [¹¹ C]ABP688 nondisplaceable binding potential (BP_ND ) values were averaged across each subfield, and Z-scores were calculated. Subfield [¹¹ C]ABP688-BP_ND was compared between seizure-free and non-seizure-free patients. In addition, we also assessed subfield volumes and [¹⁸ F]fluorodeoxyglucose (FDG) uptake in each clinical group. RESULTS: MTLE [¹¹ C]ABP688-BP_ND was reduced in ipsilateral (epileptogenic) CA1-3 and CA4/dentate-gyrus (p < .001, 95% confidence interval [CI] = .29-.51) compared to controls, with no difference in Subiculum. [¹¹ C]ABP688-BP_ND and subfield volumes were compared between seizure-free (Engel IA, n = 13) and non-seizure-free patients (Engel IC-III, n = 10). In ipsilateral CA1-3 only, [¹¹ C]ABP688-BP_ND was lower in seizure-free patients than in non-seizure-free patients (p = .012, 95% CI = 1.46-11.0) independently of volume. A subset analysis of 12 patients with [¹¹ C]ABP688-PET+[¹⁸ F]FDG-PET showed no between-group significant difference in [¹⁸ F]FDG uptake, whereas CA1-3 [¹¹ C]ABP688-BP_ND remained significantly lower in the seven of 12 seizure-free patients (p = .03, 95% CI = -3.13 to -.21). SIGNIFICANCE: Reduced mGluR5 allosteric site availability in hippocampal CA1-3, measured in vivo by [¹¹ C]ABP688-PET, is associated with postsurgery seizure freedom independent of atrophy or hypometabolism. Information derived from hippocampal CA1-3 [¹¹ C]ABP688-PET is a promising imaging biomarker potentially impactful in surgical decisions for MRI-negative/PET-negative MTLE patients.