Abstracts

Rationale: The surgical approach to treating refractory insular epilepsy is often challenging owing to its deep-seated location within the lateral sulcus, penetrating branches off of middle cerebral artery, and its contiguity to highly eloquent brain areas. The often misleading seizure semiology and scalp electroencephalogram findings, along with the extensive connectivions with potentially epileptogenic structures can contribute to failed epilepsy surgeries. Operculoinsular resections pose difficulties because the perisylvian area is highly vascular and functional; rates for transient and permanent deficits are higher than for epilepsy surgery in other extratemporal lobes. The responsive neurostimulation system (RNS, Neuropace) is approved for adults with refractory focal epilepsy with 2 or fewer seizure foci. There is little published data on the efficacy of RNS specifically in patients with insular epilepsy. Methods: Patients who had at least 1 depth electrode implanted in the insula and connected to the RNS generator between April 2014 and June 2019 were identified from our center's surgical database. RNS electrode placement was guided by frameless stereotaxy (BrainLab) and targeting was based on seizure localization on prior intracranial EEG and intraoperative electrocorticography performed at the time of RNS implantation. A minimum of 3-4 leads were implanted during each procedure. Insular depth electrodes were implanted obliquely with either posterior-anterior or superior-inferior trajectory orientations, with occipital or frontal approaches respectively, and positioning was confirmed on post-operative high-resolution head CT. Longitudinal outcome assessment was determined by chart review. Medication adjustments were made during the follow up period as clinically indicated. Seizure frequency was compared to pre-implantation baseline and classified as unchanged, modest improvement (25-50% improvement), or significant improvement (>50%). Assessments of 'improvement' were based on the treating clinician's appraisal of clinically disabling seizures and overall electrographic seizure frequency.  Results: Twenty-one patients were implanted with at least 1 insular electrode. There were no perioperative complications. Of these, 17 (81%) had failed prior surgical resections. Three patients had exclusive insular onsets, 4 frontal-insular, 4 temporal-insular, 1 parietal-insular, and 10 had multifocal or poorly localized onsets. Electrode configuration of insular depth arrays were either monopolar or stacked bipolar. Stimulation parameters in these electrodes ranged from an estimated charge density (ECD) of 0.2-3.2 µC/cm2, pulse width 80-160 µs, duration 80-2000 ms, frequency 5-200 Hertz. The follow up period ranged from 9-57 months. One patient's device was removed after 15 months due to an indolent infection. One patient reported intermittent sensory symptoms on the contralateral limbs during stimulation that improved when electrode configuration was modified. Eighteen patients had stimulation enabled for greater than 6 months and of these, 11 (61%) had significant improvement, 5 (28%) had modest improvement, 2 (11%) were unchanged. Patients with multifocal or poorly localized epilepsy demonstrated meaningful improvement in baseline seizure frequency, with 7 (70%) experiencing at modest to significant improvement in seizures.  Conclusions: RNS therapy of insular epilepsy is relatively safe and well-tolerated, and is a viable alternative to resection or thermal ablation of the insular cortex. Almost 90% of our cohort responded favorably to neurostimulation, and with relatively low ECD. Notably, patients with multifocal or poorly localized epilepsy often improved, possibly due to the extensive connections between the insula and neighboring epileptogenic brain regions and neuromodulation of these networks. Funding: No funding