Abstracts

Rationale: MEG is an imaging technique that has been validated for use in epilepsy pre-surgical evaluation, but the use has remained limited to particular epilepsy centers. Of the many studies validating MEG as a clinical tool, only a few have examined MEG use in the pediatric population. The pediatric epileptic population has unique challenges including more diverse locations of epileptogenic foci and disparate etiologies. The data for MEG as a tool to predict resective surgery outcome is limited in pediatrics. We therefore aimed to determine the predictive value of MEG for pediatric resective epilepsy surgery outcomes. To do so, we first determined how MEG is used for pre-surgical evaluation at a pediatric epilepsy surgery center and evaluated the correlation between MEG interpretation and surgical outcome. Methods: We reviewed 700 subjects who had MEG performed at the Children’s Hospital of Philadelphia from 2010-2016, and then included those that underwent a focal cortical resection surgery for epilepsy and had consented for use in research studies. Data including scalp EEG, seizure history, AED history, MRI imaging and MEG reports was extracted and analyzed. Whole-brain SAMg2 MEG images of kurtosis (i.e., source waveform level ‘spikiness’) were evaluated for the number of independent peaks found and quantified into discreet foci or cortical regions involved. Subjects were excluded if there was an absence of EEG data, MRI imaging, or they were not followed at CHOP post-operatively. Subjects who underwent alternative epilepsy surgical interventions including corpus callosectomy and VNS placement were not included in this study. Surgical outcomes were divided into good outcomes (Engel class I and II) and poor outcomes (Engel class III, IV) at one year after surgery. Statistical analysis was preformed using logistical regression models with Matlab software and Fisher’s exact t-test. Results: 74 subjects met study criteria. The median age was 9 years and 48.6% of subjects were male. 56.7% had a diagnosis of developmental delay. At one year, 66.2% of subjects had an Engel class of I or II and 33.7% an Engel class of III or IV. Based on chart extraction, those with good Engel class had a decrease in seizure frequency per month (p=0.009) and reduced number of AEDs post-surgery (p=0.002). MEG localization as measured by number or regions or number of foci was significantly correlated with surgical outcome. There was significant difference in the number of regions noted in those subjects with high Engel class compared with those with low Engel class (p=0.0032). Conclusions: Fewer MEG SAMg2 detected regions were associated with a better surgical outcome after resective epilepsy surgery. MEG may be a useful tool in predicting not only those pediatric patients who will benefit from resective surgery but also those who will have a significant improvement in seizure burden. MEG should be utilized in evaluation of pediatric patients for possible resective surgery. Funding: N/A