Abstracts

Rationale: While patients with non-lesional MR imaging represent a smaller subset of patients undergoing epilepsy surgery, identifying the seizure onset zone (SOZ) non-invasively in these patients remains a major challenge. We sought to determine the clinical utility of MEG in the presurgical evaluation of patients with non-lesional epilepsy by comparing MEG findings to those from intracranial EEG (iEEG).Methods: We identified all MRI-negative patients with drug resistant epilepsy who underwent presurgical evaluations which included MEG and an iEEG between 2009 and 2015 at the Comprehensive Epilepsy Program of the Medical College of Wisconsin in Milwaukee. Patients with incidental findings on the MRI remote from the area of suspected seizure onsets were included. A retrospective review of the electronic medical records was performed to correlate MEG findings in these patients with interictal and ictal iEEG findings, histopathology, and surgical outcome.Results: From a cohort of 72 patients who underwent both MEG and iEEG between 2009 and 2015 we identified 19 non-lesional patients. MEG was non-contributory in 7/19 (36%) of these patients due to absence of epileptiform abnormalities. This was significantly higher (two tailed P<0.05 using Fisher’s exact test) than a rate of 7/53 (13%) observed among the lesional patients in our original cohort. MEG localization of interictal epileptiform abnormalities showed sub-lobar concordance with interictal iEEG findings in 6/19 (31%) patients, and partial-concordance (overlapping regions of abnormality) in another 6/19 (31%). There were no ictal MEG studies in this series. Ictal data were not captured in 3 patients during iEEG recording. Interictal MEG findings showed sub-lobar concordance with the SOZs identified on iEEG in 8/19 (42%) patients, while the SOZ was one of the regions identified by MEG in another 2 (10%) patients with multifocal MEG abnormalities. Of 16 patients in this series who underwent a surgical resection, 6 had histopathological findings other than non-specific astrocytosis (hippocampal sclerosis in 2, cortical dysgenesis in 4). Among the patients who had post-operative follow-up of at least 6 months, Engels class I outcome was seen in 10 patients and class II in 3 patients. Neither histopathology nor surgical outcome correlated with either a positive MEG study or MEG-iEEG concordance.Conclusions: MEG source-modeling of interictal epileptiform discharges can provide useful non-invasive localization of epileptic dysfunction in the cortex that is concordant with iEEG findings in the majority of patients with non-lesional epilepsy. The principal factor limiting the utility of MEG in our series was the absence of interictal epileptiform discharges in about one third of our non-lesional patients which was significantly higher than that observed in our lesional patients. Acquiring MEG data in the context of in-patient AED withdrawal for Video-EEG studies could potentially increase the yield of these studies, as may quantification of novel markers of focal cortical dysfunction such as MEG background abnormalities or connectivity.