Abstracts

Rationale: We evaluated the utility of MEG as a pre-surgical tool in pediatric drug-refractory focal epilepsy, compared its efficacy with other pre-surgical modalities and analyzed it among children with seizure freedom vs. those with surgical failure.  Methods: This prospective study included 228 (age at onset:4.98±4.66 years; M:F=135:93) children with drug-resistant focal epilepsy who underwent MEG (306-channel Neuromag). Fifty three patients (age at surgery:10.86±4.83 years) underwent resective or disconnection (except corpus callosotomy) surgeries. EEG, MRI, video-EEG, NPA were performed (100%) and PET was done in some. Congruence of epileptogenic zone (EZ) delineated by MEG was compared with the EZ derived from other evaluations after pre-surgical conference. Inter-rater agreement was calculated by Cohen’s kappa Results: In MRI positive children, MEG showed a definitive cluster in 148(81.31%). Agreement by Cohen’s kappa between MEG and presumed EZ was 0.59(p<0.001) while it was 0.71(p<0.001) in extra-temporal epilepsies. In children with MRI positivity, concordance between MEG and presumed EZ was lobar-specific:92(50.54%), lobar-specific with additional sources:36(19.68%), hemispheric-specific but not lobar-specific: 34(18.68%) and discordant: 20(10.98%). Definitive clusters in MEG was noted in 28/46(60.86%) children with normal MRI [kappa-0.32(p<0.001)]. Concordance between MEG and presumed EZ was lobar/hemispheric- specific: 24(52.17%), hemispheric-specific but not lobar-specific: 14(30.43%) and discordant: 8(17.39%) in MRI negative group. Post-surgically (follow-up:19.52±11.27 months), kappa was 0.67 each in those with good (Engel I=37/53) and poor outcome (Engel=II=16/53). 27 children (50.94%) with good outcome had concordant and region-specific localization with respect to EZs determined by MEG and other tests. In children with poor outcome, lobe specific concordance was seen in 68.75%. Post-op MEG (n=14) revealed localizing clusters in 6, and 3 of them with dense clusters had poor outcome Conclusions: There was statistically significant agreement and concordance of EZ based determined by MEG and other tests; therefore MEG served as an important and additional pre-surgical tool in this cohort Funding: None