Abstracts

Rationale: For patients with nonlesional refractory focal epilepsy (NLRFE) drug-resistant focal epilepsy these patients, localization of the epileptogenic zone is more arduous and invasive EEG (icEEG) is frequently required. A number of studies have looked at the added value of MEG in the presurgical evaluation of refractory focal epilepsy. Most of these studies have however indiscriminately mixed patients with lesional (a subpopulation where the location of the epileptic zone can already be fairly assumed) and nonlesional epilepsies. In the present study, we report our experience with MEG in a large series of NLRFE evaluated for epilepsy surgery. Methods: Between April 2006 and May 2012, 61 consecutive patients (mean age 34 years) with NLRFE underwent a MEG study. For the first 37 patients (group A), MEG results were not included in the preoperative consensus for several reasons (pipeline set-up, personnel recruitment and training, learning curve for interpretation of results etc.). For the remaining 24 subjects (group B), MEG findings were presented during the epilepsy surgery conference after initial review of standard presurgical work-up. To determine if MEG results would have changed patient management, anonymized results of the standard presurgical evaluation for all patients were presented randomly to a multidisciplinary epilepsy surgery team at first blinded to MEG results and then with the results. Correlation between MEG and icEEG findings and clinical outcome were also evaluated. Results: Based on the standard presurgical evaluation, 18/61 (30%) were presumed to have nonlesional temporal lobe epilepsy and 43/61 (70%) nonlesional extratemporal lobe epilepsy. MEG showed spikes in 56/61 patients (92%). For group A, MSI would have changed the initial management in 24/37 (65%) had the results been available at the time of decision whether it be by (a) reducing the number of icEEG contacts (42%); (b) increasing the number of icEEG contacts (17%); (c) keeping the same number of icEEG contacts but changing their position (4%); (d) allowing for direct surgery without icEEG (13%); (e) providing enough evidence to now justify an icEEG study in previously rejected candidates (13%) or in patients initially directed to surgery (4%); or (f) rejecting a case previously destined for icEEG as the cluster overlaid an eloquent area (8%). For group B, MEG results modified the treatment plan in 14/27 (52%) by reducing icEEG contacts (50%), or change their position (7%), allow for direct surgery (36%), make eligible to an icEEG someone initially though inoperable (7%). IcEEG recordings available for 25 subjects showed good concordance with MEG results in (64%). In 33% of cases, MEG was localizing but ictal icEEG recordings were non-localizing. Surgery was performed in 25 subjects: 19/21 (90%) subjects with resection of the MEG zone and 4/4 (100%) subjects whose MEG zone was not resected had a poor outcome. Conclusions: MEG affects patient management, icEEG planning and surgical outcome in a significant percentage of patients with NLRFE.