Abstracts

Rationale: Greater than 90% of children with Tuberous Sclerosis Complex (TSC) have epilepsy, many medically refractory. The intracranial features of TSC are cortical and subcortical tubers, ‘empty-gyrus', subependymal nodules/giant cell astrocytomas, and white matter radial migration lines. The presence of multiple MRI lesions creates multiple potential targets for surgical resection. Recently high frequency oscillations (HFOs) (>80Hz) have been shown to have a tight relationship with seizure onset zone (SOZ); it has been shown that resecting HFO-generating cortices is critical to achieving seizure freedom in several different etiologies. The aim of this study was to investigate whether ictal HFOs recorded by intracranial EEG are specific to SOZ in children with TSC. Methods: Children with diagnosed intractable epilepsy and TSC who were selected as surgical candidate and underwent intracranial EEG before respective surgery which were performed between January 2008 and June 2011 were retrospectively investigated. The intracranial EEG was recorded with 2000 Hz sampling rate. The seizure onsets were determined with conventional visual inspection at 1-70Hz bandwidth filter first then analyzed with 80Hz high-pass filter along with time-frequency analysis using short time fast Fourier transform. Each electrode was characterized by presence of specific findings in MR image of underlying cortex (i.e. calcification, cystic component), , SOZ, presence/characteristic of HFOs and inclusion in surgical margin. The surgical outcome was also determined and evaluated for association of the above findings using statistical analysis. Results: Thirteen patients were identified and analyzed. HFOs were present in SOZ for all patients with at least 1 year follow-up duration. The electrodes with HFOs in SOZ were categorized as lesional vs non-lesional and further categorized based on specific imaging features. Five patients were seizure-free (ILAE class 1, 38%). Among those patients, HFOs were identified in both lesional and non-lesional categorized electrodes; however, the ratios varied from patient to patient with different types of lesions. Six patients had poor surgical outcome (ILAE 4/5, 46%), even though 2 patients were seizure free for at least 1 year after surgery. These latter patients had a high discordance ratio between non-lesional vs lesional HFOs and SOZ. HFO characteristics in terms of the frequency band were also not specifically associated with either lesional or non-lesional categories. Conclusions: The regions with HFOs were significantly associated with SOZ in patients with TSC. However HFO regions were not always correlated with the tubers, rather dependent on the lesion characteristics (e.g., ‘empty-gyrus' vs calcifications). HFOs may be a biomarker for epileptogenicity of cortex in TSC. Further studies would be helpful to determine any other factors which predict better surgical outcome.