Abstracts

Rationale: Studies have shown favorable results after epilepsy surgery at fixed time intervals of postoperative follow-up or in selected subgroups of patients depending on the surgical site. Long-term longitudinal outcome in all patients, irrespective of resection site, is a clinically meaningful measure of success of epilepsy surgery but remains poorly understood. We investigated this in the current study. Methods: We retrospectively analyzed consecutive patients who underwent epilepsy surgery between January 2008 and July 2012. We excluded patients with malignant tumor or follow-up of less than 6 months. We defined outcome as favorable (Engel I/II) or unfavorable (III/IV) at 6-month and 1/2/3-year periods. We evaluated outcome using univariate and survival analyses. Variables with statistical significance on univariate analysis were further tested using a multivariate Cox regression model.Results: Among 72 operated patients (21-68 years, 39 male), there were 55 (76%) temporal and 17 (24%) extratemporal resections. The follow-up period was 6-63 (median 21) months. Mean epilepsy duration was 16 years. MRI was normal in 14 (19%) and abnormal in 58 (81%); abnormalities included MTS (n=31), mass (n=12), vascular lesions (n=5) and other (n=10). Six patients showed more extensive MRI changes beyond the intended resection (bilateral MTS, widespread tubers, atrophy). Nine patients had intracranial monitoring while 35 had intraoperative electrocorticography. Pathology was normal in 5 patients and abnormal in the rest: MTS or gliosis (n=39), benign tumor (n=9), vascular (n=4), other (n=15). Outcome at last follow-up was favorable in 85% (n=61, temporal 52, extratemporal 9) and unfavorable in 15% (n=11, temporal 3, extratemporal 8). Extratemporal seizure onset was a predictor of unfavorable outcome at all follow-up periods whereas uncertainty in ictal EEG localization was a predictor only at 1/2/3-year follow-up periods. Both were independent predictors using multivariate analysis. Outcome was not associated with age, gender, history of generalized seizures, epilepsy duration, preoperative seizure frequency, Wada memory impairment, interictal EEG, lesional status or pathology. Extensive MRI abnormalities, except bitemporal MTS, were associated with early unfavorable outcome. Survival analysis showed that the chance of seizure recurrence was lower for temporal than extratemporal resections. At 1 year, recurrence risk leveled off for temporal resections but not for extratemporal resections. Conclusions: Extratemporal epilepsy and uncertainty in ictal EEG localization are independent predictors of unfavorable outcome. Presumed reasons for unfavorable outcome include incomplete resection and inadequate seizure focus localization. One-year seizure freedom predicts future seizure freedom after temporal but not extratemporal resections.