Introduction: 

Following conventional non-invasive presurgical evaluation (CNPE) for pharmacoresistant partial epilepsy, approximately half of patients are considered poor surgical candidates due to unclear localization of the epileptogenic zone. This study assesses the clinical added value of magnetoencephalography (MEG) in the presurgical evaluation of these patients.

Patients and methods: 

Thirty-two patients (mean age: 35 y;18 M) with refractory epilepsy and non-localizing CNPE results were included in this study. CNPE included clinical and neurological examination, neuropsychological evaluation, one week of video-electroencephalography (EEG) monitoring, optimized 3 Tesla structural brain magnetic resonance imaging (MRI) and [¹⁸F]-fluorodeoxyglucose positron emission tomography (FDG-PET). Results of CNPE were considered non-localizing when no or multiple ictal onset zone(s) were identified during video-EEG monitoring (14 patients), when no or multiple lesions were identified on 3T MRI (5 patients) or when both investigations were non-localizing (13 patients). In 8/32 patients no interictal epileptiform discharges (IED) were observed during the video-EEG monitoring. All patients underwent 1-hour MEG recording using the whole-head 306-channel Elekta Neuromag® system installed into a lightweight magnetically shielded room (Maxshield, Elekta Neuromag Oy). Interpretable MEG data were visually screened for the presence of IED by 2 independent investigators (EC, XDT). Equivalent current dipoles (ECD, g/% >80%) were fitted in the patients' spherical head model and then superimposed on their co-registered MRI. ECD localisation was classified as either clustered or scattered. The added value of MEG was then evaluated by comparing the suitability of the patient for epilepsy surgery before and after adding the MEG results to the conventional investigations.

Results: 

Good signal-to-noise ratio was obtained in 27/32 patients (84%). IEDs were identified in 21/27 patients (78%) with interpretable MEG data. In 4/8 patients without IEDs on video-EEG monitoring, IEDs were observed in MEG data. ECDs were clustered in 16/21 patients and scattered in 5/21. MEG results changed patients' management in 7/21 patients (33%) either by preventing immediate RS without invasive video-EEG monitoring (IVEM, 1 patient), leading to a hypothesis for IVEM (2 patients) or by adjusting the electrode implantation scheme for the planned IVEM (4 patients).

Conclusion: 

This study highlights the clinical added value of MEG in the presurgical evaluation of refractory epilepsy patients when CNPE does not result in unambiguous localisation of the epileptogenic zone. Indeed, in 33% of patients with abnormal MEG and non-localizing CNPE, adding MEG to the presurgical evaluation changes the surgical management either by leading to IVEM or by adjusting the electrode implantation for the planned IVEM.