Annual Meeting Abstracts: View

  • (Abst. 2.076), 2018
  • Anatomic Early Electrographic Seizure Onset Localization in Neocortical-Responsive Neurostimulation
  • Authors: Allen L. Ho, Stanford University; Emily Mirro, NeuroPace, Inc.; Babak Razavi, Stanford University Medical Center; Kai J. Miller, Stanford University; Dora Hermes, Stanford University; Matthew Markert, Stanford University; Gerald A. Grant, Stanford University Medical Center; Jaimie M. Henderson, Stanford University; and Casey H. Halpern, Stanford University
  • Content:

    Rationale: Brain-responsive neurostimulation (RNS® System, NeuroPace, Inc.) has proven efficacious for the treatment of refractory epilepsy arising from neocortical foci (Jobst, et al, Epilepsia, 2017). Depth leads can be the preferred RNS System lead placement strategy for neocortical seizure onsets when cortical strip leads are not feasible due to dural adhesions/scarring or when the seizure focus is below the cortical surface (as obtained by stereo-EEG). While longitudinal recording of seizure onsets has been enabled by the RNS System, highly precise localization remains difficult given the large neocortical surface area covered. MRI-based cortical reconstruction and volumetric anatomic segmentation were utilized to systematically localize electrodes and correlate them with early electrographic seizure onsets. Methods: Thirteen adult patients with epilepsy arising from neocortical foci and placement of two NeuroPace® Depth Leads were evaluated. In addition to a neocortical focus, some depth leads were also placed in non-neocortical foci. All patients received brain-responsive neurostimulation on both depth leads. The Freesurfer image analysis suite was utilized for cortical reconstruction and volumetric segmentation of pre-operative MRI imaging. Cortical models were then parcellated and labeled via the Destrieux pial surface anatomic atlas, and co-registered to post-op CT imaging to determine precise electrode location. RNS System electrocorticographic (ECoG) data were reviewed to identify electrodes with early electrographic seizure activity. Percent seizure reduction compared to pretreatment was calculated for patients with >6 months follow-up. Results: The patients underwent placement of 33 quadripolar depth leads for treatment. Of the 104 electrodes connected to the neurostimulators, the most common locations were the hippocampus (10%), temporal pole (8%), superior frontal gyrus (5%), paracentral sulcus (4%), precentral gyrus, insula, cingulate gyrus, superior parietal lobule (all 3%), and white matter (10%). Of the electrodes demonstrating the earliest electrographic seizure onsets, 23% were in hippocampus; the remaining were in the superior frontal gyrus (18%), insula (14%), white matter (14%), orbital gyrus (9%), supramarginal gyrus (9%), amygdala (4%), and straight gyrus (4%).Seven of the 13 patients received >6 months of brain-responsive neurostimulation and were followed for an average of 1.5 years (range: 1.0 to 3.4 years). These seven patients had a median seizure frequency reduction of 50% (mean: 48%, range: 0-100%). One patient with depth leads placed in the motor cortex became seizure free. There were no functional neurologic deficits as a result of depth leads placed in the neocortex and there were no side effects caused by neocortical-responsive neurostimulation. Conclusions: Systematic cortical reconstruction and volumetric segmentation can be utilized to precisely localize depth lead electrodes and identify early electrographic seizure onsets in the neocortex. This anatomic specificity is critical as we continue to improved seizure control provide by the RNS System. Funding: None