Abstracts

Rationale: Recent evidence shows that substantial number of amygdala enlargement (AE) is found in patients with temporal lobe epilepsy. However, the pathophysiology of AE in human epilepsy is not fully elucidated. We investigated relationship between changes of amygdala volume and seizure rate, using a longitudinal automatic volumetry analysis. Methods: 37 patients (22 male and 17 with AE in the left side) who had AE on initial MRI were followed up for the second MRI (with interval more than 6 months). AE was defined as a definite asymmetry between the two sides, as determined by the visual assessment. The volume of amygdala at each time point was estimated by Freesurfer software. Asymmetry index (AI) was calculated as ratio of the volume of the larger side to the smaller. According to detailed interview at initial visit, two groups were identified and compared to test our hypothesis; (1) 11 subjects with higher seizure frequency (of seizure count >20 in last 3 months); and (2) 26 patients with less seizures ( < 10). Results: Mean age of seizure onset was 43.5 19.3 years. The mean age of patients with higher seizure frequency was older than those with less seizures (p = 0.005). The mean follow-up duration between 2 MRI was 18.1 months. The decrease of enlarged amygdala volume was found in 31 (84%) subjects. Overall rate of volume change was greater in the enlarged side than the other side (p = 0.02). All patients were well controlled with antiepileptic drug medication around the date of second MRI. The AI of initial MRI was greater in group of higher seizure frequency (p < 0.001), but the difference was not significant in AI of the second MRI (p = 0.12). Rate of decline in volume of enlarged amygdala over time was higher in patients with frequent seizures (p < 0.01). The rate of volumetric change of enlarged amygdala was statistically different between two groups after controlling for covariates of age, sex and laterality of the lesion. Conclusions: TLE with AE might not be a homogenous syndrome. Our longitudinal data suggests that AE might be contributed to by higher seizure rate before the treatment. Funding: No funding was received.