Abstracts

Rationale: Electrical stimulation mapping is commonly used in epilepsy surgery to identify eloquent cortices and assist in seizure localization.  It is assumed that resection of aura elicited sites (AES) is a predictor of good surgical outcome.  We investigated this hypothesis by comparing these sites to known electrophysiological biomarkers:  the seizure onset zone (SOZ), early spread zone, and both interictal and ictal high-frequency oscillations (HFOs).   Methods: We included patients undergoing phase II evaluation between 2014-2016 using stereotactic depth arrays at Columbia University Medical Center in whom electrically stimulated auras were induced.  Seizure onset and early spread zones were identified by clinical consensus.   HFOs were defined using visual criteria after high-order FIR filtering (80 – 150 Hz; sampling rate 500 Hz).  Interictal HFOs were identified in association with interictal epileptiform discharges, to avoid including non-pathological events.  Ictal HFOs were characterized as early-onset, i.e. seen within the first 4-5 seconds after seizure onset, or delayed-onset, appearing after the first 4 seconds and sustained for at least 10 seconds. The AES and HFO information was included in the surgical case conference and contributed to the surgical decision-making. Nonparametric tests were conducted for statistical comparisons. Results: Fourteen patients met study criteria. AES corresponded to SOZ in four (29%) and early spread in four (29%) patients. For HFOs, AES corresponded to early ictal sites in five (36%), delayed ictal sites in seven (50%), and to interictal sites in four (29%) patients. AES corresponded to more than one biomarker in five patients and AES which did not correspond to any HFO sites in five patients. Correlation with delayed ictal HFOs was more common than with other types of HFOs, but this was not significant. Seven patients (50%) underwent resective surgery that included AES, of whom one remained seizure free after one year. In this patient, AES corresponded with delayed-onset HFOs and both were included in the resection.  The remaining patients underwent RNS implant or are awaiting surgery. Conclusions: This small study suggests that mapping-induced early seizure symptoms or signs may have limited utility in defining resections.  Furthermore, the comparison with electrophysiological biomarkers that have been shown to be predictive of surgical outcome revealed mixed findings.  We concluded that AES may be located in spread areas and may have only limited value for identifying the onset location. AES correlated most commonly with delayed-onset ictal HFOs, which are the most specific type for identifying regions of ictal invasion.  A larger study is needed to confirm these findings and establish the optimal strategy for employing the various biomarkers in epilepsy surgery procedures. Funding: None