Abstracts

Rationale: Excisional surgery achieves seizure-freedom in a large proportion of children with MRI negative epilepsy, but the resections based on functional data are often extensive. We explored the possibility that more restricted resections within a carefully selected MRI negative cohort could achieve comparable success. Methods: We report a subset of 25 children with MRI-negative, intractable partial epilepsy who underwent focal corticectomy at our institution between the years of 2005-2011. The epileptogenic region was identified by integrating multimodal functional data including 3-D EEG source localization, SPECT, PET, and invasive EEG data using co-registration software. Corticectomies were minimized by resecting regions that showed convergence of multimodal abnormality. Results: The mean age of seizure onset was 5.6 years (range, 1 day to 14 years) and mean age at surgery was 11.8 years (range, 6 months to 19.9 years). Corticectomies were convergent with scalp EEG in all cases and 3-D EEG source localization of scalp interictal data in 11 of 12 analyzed. Convergent, focal hyperperfusion was seen in 10 of 15 patients who had ictal SPECT scans and, of the 17 patients that underwent PET scan, 12 had convergent regions of hypometabolism and 2 had highly localized, convergent hypermetabolic areas. Twenty-three children had convergent interictal and/or ictal discharges on intraoperative ECoG and 24 had ictal onset in corresponding regions during extraoperative subdural monitoring. Resections were considered complete in 7 and incomplete in 18. Histopathology revealed definite cortical dysplasias in 19 patients, 7 of which were Palmini type I and 12 of which were type II. The remaining 6 showed mMCD. At 1 year follow-up, 13/25 (52%) were seizure-free, 5 (20%) experienced persistent seizures with an >90% reduction in frequency, and the remaining 7 (28%) had no change in seizure burden. The rate of seizure freedom in children with type I FCD, type II FCD, and mMCD was 3/7 (43%), 7/12 (58%), and 3/6 (50%), respectively [not significant]. Outcomes were unrelated to completeness of resection and all three histopathologic classes were represented in each Engel outcome class. The majority of children with Engel class IV outcomes had data convergence of fewer functional modalities than that seen in almost all children in more favorable outcome classes. Conclusions: Our findings demonstrate that functional lesionectomy is successful in a carefully selected subset of a MRI negative cohort. Outcomes were not related to completeness of resection and did not vary significantly between children with different histopathologies, but more favorable outcomes were associated with a greater number of convergent functional modalities. Minimizing resection based on convergent mulitmodal functional data helps avoid unnecessarily large resections and may have important implications not only for optimizing seizure control but also for improvement in functional and developmental outcomes.