Abstracts

Rationale: In patients with MRI-negative drug resistant epilepsy, identifying epileptogenic hippocampi is important. We investigated whether structural changes at the subfield level of the hippocampi and ipsilateral amygdala may provide potential biomarkers of secondary hippocampal epileptogenicity.

Methods: We included MRI-negative hippocampi and ipsilateral amygdala from drug-resistant epilepsy patients who underwent stereo-electroencephalography (SEEG) and/or epilepsy surgery and age and sex-matched non-epileptic controls. In epilepsy patients, we grouped the hippocampi into (1) primary, (2) secondarily, and (3) non-epileptogenic based on SEEG data and/or seizure freedom after minimum 12 months of follow-up in the case of spared hippocampi after resection. Primary epileptogenicity was determined by a team consensus at the end of SEEG admission. Secondary epileptogenicity is defined if the hippocampus was implanted, found not to be a part of the primary epileptogenic network but there was no spontaneous hippocampal seizures, stimulation-induced electro-clinical seizures and/or early seizure propagation. We segmented the hippocampi and the ipsilateral amygdala nuclei on (i) T1 (epilepsy and controls) and (ii) T1+ high-resolution T2 (epilepsy only) MRI images using the FreeSurfer segmentation toolbox (version 7.3.2). We also quantified normalised FLAIR signal intensity in the hippocampus using the T1+T2 segmentation results as regions of interest in epilepsy patients. We performed analyses of covariates and mixed-model logistic regressions controlled for the false discovery rate.


Results: 63 patients (male 54.0%, mean age 35.30±10.63 years) were included in the study. T1-only segmentation (63 patients and 68 controls): We collected data on 18, 11, 66, and 136 hippocampi (groups 1, 2, 3, and controls respectively). Central amygdala nuclei volumes were increased ipsilateral to secondarily epileptogenic hippocampi compared to non-epileptogenic and control hippocampi (p=< 0.001 OR 1.219, and p< 0.001 OR=1.128). The basal nuclei were also enlarged in group 2 compared to controls (p=0.001 OR=1.017). T1+T2 segmentation (57 patients): We collected data on 14, 11, and 60 hippocampi (groups 1, 2, and 3 respectively). Secondarily epileptogenic hippocampi exhibited larger hippocampal fissures and CA1 heads compared to the non-epileptogenic hippocampi (p=0.001 OR=1.063 and p=0.001 OR 1.023). They also exhibited larger central amygdala, basal, and accessory basal nuclei compared to non-epileptogenic hippocampi (p< 0.001 OR=1.195, p=0.002 OR =1.035, p< 0.001 OR=1.066). Increased ipsilateral amygdala volumes were also observed in group 2 compared to group 3 (p=0.002 OR=1.009). Normalised FLAIR intensity analysis did not reveal subfield differences between groups.