Abstracts

Rationale: Indications for invasive subdural monitoring in pediatric epilepsy surgery are poorly defined.We sought to critically review our experience with pediatric invasive intracranial monitoring to better delineate when invasive EEG is useful. Methods: Patients who had subdural invasive EEG monitoring for epilepsy surgery were identified from an epilepsy surgery database at Miami Children s Hospital Brain Institute. Exclusion criteria were inadequate subdural data and children with infantile spasms/myoclonic epilepsy. Pre-invasive evaluations were retrospectively reviewed including history, interictal/ictal scalp EEG,MRI and functional imaging in selected cases. Indications for subdural implantation were categorized based on pre-invasive data as :1) Mapping of eloquent cortex 2) Poorly localized epileptogenic zone(EZ)3) Discordant data.Subdural EEG reports,surgical diagrams and functional mapping were reviewed for localization of the EZ and compared with pre-invasive data for congruence. Data on surgical resection, completeness, complications and outcome were analyzed. Results: 102 patients met inclusion criteria.12 were implanted for mapping eloquent cortex,76 for non-localized EZ and 14 for divergent data. In the non-localized group 45% had negative MRIs. A focal area for resection was identified in 99%(68% incomplete). Subdural monitoring impacted resection in 84%. Implantation for functional mapping facilitated incomplete resections in 42% due to proximity to eloquent cortex. Of cases implanted solely for language mapping in a presumed dominant hemisphere,5 were redundant as all underwent anterior left temporal lobectomy. Implantations to determine the extent of involvement beyond the mesial/anterior temporal region were rarely useful particularly if there was a focal MRI lesion.90% resulted in anterior temporal lobectomies. Most poorly localized cases were multilobar (paracentral n=21, frontotemporal n=15, posterior quadrant n=9,other multilobar n=15). 1 paracentral case resulted in explantation.The EZ was outside motor cortex in 38% and only partially overlapped motor in several cases resulting in 5 complete resections. Frontotemporal implantations resulted in lobar resections in 73%. Posterior quadrant and other multilobar implantations usually resulted in more restricted resections than anticipated by non-invasive data(79%)and no quadrantectomies. 14 cases were implanted for discordant data with 13 resolving the discordance. With divergent scalp EEG and MRI subdural EEG more often was concordant with MRI than scalp EEG(71% vs 28%). Conclusions: There is limited pediatric data guiding the use of intracranial monitoring to localize an epileptogenic focus for resective surgery. This study provides an overview of current indications at an active pediatric epilepsy surgery centre and examines the utility of those practices. We confirmed the adult experience that subdural monitoring is rarely helpful for MRI-lesional temporal lobe epilepsy. Routine use of functional neuroimaging modalities may obviate the need for many invasive cases previously deemed poorly localized.