Abstracts

Rationale: Magnetoencephalography (MEG) is used in the pre-surgical evaluation of people with refractory epilepsy. It is currently unclear, however, if MEG has an added diagnostic value in localizing the epileptogenic zone (EZ) if compared to 1) other diagnostic tools, 2) invasive EEG (iEEG) and 3) the resection site and if co-localisation of MEG with the resection site has an influence on post-operative outcome. Methods: In this retrospective study at the University Medical Centre in Utrecht and the National Hospital of Neurology and Neurosurgery in London 209 candidates for epilepsy surgery who had MEG evaluation between June 2007 and June 2015 were included. The hypothesis on the location of the EZ from all diagnostic methods including MEG was extracted from the individual’s record, as was resection site and seizure outcome. The hypothesis based on the MEG findings was compared to 1) the hypothesis based on all diagnostics before the MEG, 2) the hypothesis based on the invasive EEG – if performed – and 3) the resection site. The hypothesis on the location of the EZ based on the diagnostics before MEG and the resection site was compared to the MEG hypothesis on a lobar level. Concordance with iEEG was defined as concordant if the iEEG hypothesis was within the area of the MEG source. Seizure outcome was compared between patients whose MEG localisation was lobar concordant with the resection site and patients in whom this was not the case. Results: The hypothesis before MEG and the MEG generated-hypothesis were lobe concordant in 131 individuals (63%) and lobe discordant in 25 (12%). The MEG revealed no source (“negative”) in 53 people (25%). When only considering those with a MEG source the concordance between the pre-MEG and the MEG hypothesis was higher (84%).  Invasive EEG was performed in 98 candidates (47%). In this group, the MEG hypothesis was concordant with the iEEG hypothesis in 60 (61%), discordant in 17 (17%) and the MEG was negative in 21 (22%).  Ninety-three (45%) underwent resection, 67 (72%) of whom after iEEG. The MEG hypothesis was concordant with the resection site in 57 (61%), discordant in 11 (12%) and the MEG was negative in 25 individuals (27%). When only considering those with a MEG source this concordance was higher (84%). In those 77 individuals who had a resection with at least one-year follow-up, 37 (47%) had an Engel 1 score at last follow-up (mean last follow up = 3 years). Seizure outcome did not differ significantly between those in whom the MEG was concordant with the resection site (Engel 1: n= 21, 41,%) and those in whom this was discordant (Engel 1: n= 3, 30%). Those who were MEG negative had a significantly better seizure outcome (Engel 1: n=12, 70,6%) than those who had a localized MEG source (Engel 1: n= 24, 39,3%) (X2=6,3 p Conclusions: In 39% of individuals, resective surgery was performed despite disconcordant or negative MEG results, suggesting that MEG not significantly impacts decision-making in a considerable proportion of people. In the other 61%, MEG confirmed what was already suspected from prior investigations in most individuals. Concordance of MEG and resection site did not seem to contribute to a better seizure outcome; interestingly MEG being negative seemed to relate to a better seizure outcome. Funding: None