Abstracts

Rationale: Only small series from the tertiary care Pediatric Epilepsy centers inform us on the utility of epilepsy surgery in TSC. Using a large cohort of TSC patients enrolled in the TSC Natural History Database (TSCNHD), we studied pre-surgical evaluation, epilepsy surgery types, and seizure outcome after surgery. Methods: Relevant data were queried from the TSCNHD, a database of phenotype, genotype, and longitudinal clinical course on each patient enrolled via a national TSC clinic network. Ordinal logistic regression analysis was done to study the association of surgery type with seizure frequency after surgery. Poisson regression analysis was done to test the association between number of antiepileptic drugs (AEDs) after surgery and surgery type. Results: Of 2059 patients in the TSCNHD, 1759 (85%) had epilepsy of which 202 (11%) had epilepsy surgery. 106 (53%) were male. Mean age at surgery was 8.4 years (13 days-54.9 years, SD=9.7 years). Besides brain MRI and VEEG, only minority had additional tests, FDG-PET in 46 (23%), MEG in 25 (12%), and SPECT in 20 (10%).  117 (58%) underwent straight surgery. Subdural grids +/- depth electrodes for invasive long term VEEG were done in 85 (42%) patients. Of 202, 145 (72%) subjects had one, 41 (20%) had two, and 16 (8%) had three or more surgeries. Surgeries were tuberectomy (91, 45%), Lobectomy/multi-lobar resection of a large tuber complex (70, 35%), corpus callosotomy (CC) (17, 8%), and hemispherectomy (6, 3%). In 18 (9%) ‘other’ procedures were noted, disconnections, laser surgery, and undescribed. Post-operative seizure outcome was noted after the last surgery using ILAE outcome classification (Wieser et al, 2001) at the last follow up (median 5.8 years). Seizure outcome data was not available in 52 (26%, loss of follow up/unknown). In 150 seizure outcome showed a good (class 1+2) outcome in 61 (41%), moderate (class 3+4) in 33 (22%), and poor (class 5+6) in 56 (37%). Palliative CC was used as a reference to compare the seizure outcome between the curative epilepsy surgeries. Using ordinal logistic regression model and after adjusting for age and gender, in comparison to CC, tuberectomy was more likely to be associated with post-operative seizure freedom (OR=5.9, p-value=0.005) followed by lobectomy/multi-lobar resection (OR=4.1, p-value=0.026). Poisson regression used to calculate AEDs after potentially curative surgeries in comparison to the CC showed most AEDs reduction after surgery in subjects after tuberectomy (p-value=0.0007, AEDs burden 0.57 of the AEDs after CC) and after lobectomy/multilobar resection (p-value=0.023, AEDs burden 0.68 of the AEDs after CC). Conclusions: In TSCNHD, most epilepsy surgeries were performed without using the modern imaging tools and MEG. Good (ILAE 1 +2) epilepsy surgery outcomes were noted in 35%, a lower frequency compared to the published series from the established epilepsy surgery centers. Best seizure outcome was reported after resection of a single tuber and lobar/multi-lobar resections with a high likelihood in AEDs reduction after epilepsy surgery. Funding: The authors thank Jo Anne Nakagawa, the Tuberous Sclerosis Alliance, and all contributors to the TSC Natural History Database.