Abstracts

Rationale: Studies have shown that the pathologic high frequency oscillations (HFO: >70 Hz)) define a smaller seizure onset zone (SOZ) than the conventional frequency activity (CFA: 1-70 Hz) and that resection of an HFO-defined SOZ results in favorable outcome. However, the efficacy of targeted responsive neurostimulation (RNS) delivered to the ictal or interictal HFO zone for seizure control remains unclear. Methods: We evaluated consecutive patients who underwent RNS implants and had at least 1 month of data available after therapy was enabled between March 2015 and May 2016. We separated them into 2 groups: targeted group consisted of patients with RNS leads located in the brain regions that were previously determined to be the ictal or interictal HFO zones based on extraoperative intracranial monitoring or intraoperative electrocorticography (ECoG); nontargeted group consisted of patients who underwent direct RNS implants without prior determination of the HFO zones. We measured outcome in terms of the rate of postimplant clinical seizures (CS) compared with preimplant baseline and the rate of postimplant user-programmable, electrocorticographic long episodes (LE). Results: Among 8 implanted patients, 4 were in the targeted group (all neocortical, subdural leads), 1 was in the nontargeted group (mesial temporal, depth leads), and 3 did not have data and therefore excluded from analysis. Of the 5 patients, 4 were male, age 23-53 years at implant. Follow-up period was 1-11 months. Overall, 3/5 (60%) benefitted from RNS therapy, with >75% postimplant CS reduction compared with preimplant baseline (responders), 1 showed 25% reduction in CS (nonresponder), and 1 showed no change (nonresponder). Targeted group had 3 responders and 1 nonresponder, whereas the nontargeted group had 1 nonresponder. The programmed LE length range was 15-40 sec. Number of LE per day were 0.2, 3.3, and 5.9 in the targeted group responders and 55.5 in the nonresponder; in the nontargeted group nonresponder, it was 13.9. Responders had bilateral heterotopia (n=2) and MRI-negative epilepsy with prior failed resection (n=1); nonresponders had left temporo-occipital dysplasia (n=1) and left mesial temporal sclerosis (n=1). No adverse effects related to RNS implant were noted. Conclusions: Despite the small number of patients and short duration of follow-up, this preliminary study demonstrates the efficacy of targeting the HFO zones for RNS therapy in seizure reduction. User-programmable long episodes appear to be a reliable marker of the clinical seizure frequency. Optimization of the stimulation parameters to improve efficacy in the nonresponders needs to be further explored. Funding: Abstract submission fees are paid by The Neurology Residency Program at UT Austin Dell Medical School. Author Pradeep Modur receives Investigator funding and travel expenses support from Neuropace company.