Abstracts

Rationale: Noninvasive evaluation and management of patients (pts) with intractable focal epilepsy after failed epilepsy surgery is particularly difficult. Yet, reoperation may offer a second chance for seizure-freedom in up to 50%. Unlike scalp EEG, MEG analysis is not influenced by the presence of previous craniotomy or areas of resection. We aimed to assess the usefulness of MEG in the reevaluation of this challenging group of pts.Methods: We included all pts with failed epilepsy surgery who had a positive MEG study as part of their re-evaluation. Pts with no interictal epileptiform discharges (IEDs) during MEG recordings were excluded. MEG was recorded with a 306-sensor array with simultaneous scalp-EEG. MEG source localization of IEDs was obtained using standard ECD methods. We retrospectively reviewed the noninvasive data in all pts; and intracranial EEG and seizure outcome (using Engel classification and minimum 6 months follow-up) in those pts undergoing re-operation.Results: Of 277 consecutive pts undergoing MEG from 2008 to 2011, 35 (12.6%) had history of failed epilepsy surgery and positive MEG: 21 (7.6%) had prior Temporal (T) and, 10 prior Extratemporal (ET) resection. Four pts with hemispherectomy were not included in this analysis. Source analysis of MEG IEDs revealed dipole localizations in an area adjacent to the borders of the previous resection (12 of 21 with T and 7 of 10 with previous ET surgery). MEG implicated a lobe outside the previous resection in the ipsilateral (5 in T and 2 in ET group) or contralateral hemisphere (4 T, and 1 ET). Notably, 2 of the 31 pts had MEG-unique spikes (1 in each group) without any detectable IEDs on scalp EEG. During this study reoperation was accomplished in 9 of 21 (42.8%) pts with previous T (5 achieved Engel I) and 5 of 10 pts with previous ET resection (3 with Engel I). When resection included the area of clustered MEG spike sources (Figure 1) 7 out of 8 pts achieved a favorable outcome. In contrary when the area of resection (guided predominantly by intracranial EEG) did not overlap with the MEG focus (Figure 2), only 1 out of 6 pts became seizure-free (p<0.05). Three additional pts in each group have been offered surgery and are awaiting reoperation. The remaining pts were not considered candidates for reoperation due to: bi-temporal epilepsy (3 in T), discordant data (2 in T), nondisabling seizures (1 in each group), medical issues (2 in T) or unacceptable risk for neurological deficits (2, ET group). Concordant MEG and other noninvasive data were noted in approximately half of the non-operated pts.Conclusions: We present the largest series to-date of pts with persistent seizures following previous T and ET resections studied with MEG. In those pts undergoing reoperation seizure-freedom was accomplished in 57.1%. A localizing MEG in this cohort was associated with a higher chance of seizure-free outcome following resection of the corresponding spike focus. When contemplating reoperation in these challenging pts MEG should be considered an integral part of their re-evaluation.