Abstracts

Rationale: To determine the incidence and risk factors of hippocampal atrophy contralateral to the side of seizure onset in temporal lobe epilepsy. Methods: Subjects were 42 consecutive patients admitted to the MGH epilepsy-monitoring unit with an electroclinical diagnosis of temporal lobe epilepsy and 21 control subjects with idiopathic generalized epilepsy or psychogenic non-epileptic seizures. All subjects had a 3T epilepsy protocol MRI with T1 images processed through the FreeSurfer v6 including a probabilistic hippocampal subfield volumetry. Hippocampal atrophy was determined by comparison to control subjects. Risk factors for contralateral hippocampal atrophy were evaluated with cross-validation elastic-net generalized linear model. Results: The strongest predictor of contralateral hippocampal atrophy was atrophy ipsilateral to the side of seizure. Nine out of 42 subjects (0.21 95%-CI 0.12-0.36) with TLE had contralateral hippocampal atrophy. Elastic-net cross-validation model found ipsilateral hippocampal atrophy to be the best predictor of contralateral atrophy. Fisher exact test comparing incidence of contralateral hippocampal atrophy when ipsilateral atrophy was present was statistically significant (p=0.02 OR 7.4, 95%-CI 1.5-37). Conclusions: Bilateral hippocampal atrophy is common in apparent unilateral temporal lobe epilepsy. Atrophy ipsilateral to the seizure onset zone is the greatest risk factor for contralateral hippocampal atrophy. The correlation of right and left hippocampal atrophy in seemingly unilateral temporal lobe epilepsy supports the concept of contralateral homologous involvement of seizure network and potentially epileptogenesis in temporal lobe epilepsy. Funding: None