Authors :
Presenting Author: Naoto Kuroda, MD, PhD – Wayne State University
Hiroshi Uda, MD, PhD – Wayne State University
Keiki Inoue, MD – National Hospital Organization Sendai Medical Center
Yu Kitazawa, MD, PhD – Yokomaha City University Graduate School of Medicine
Shin-ichiro Osawa, MD, PhD – Tohoku University
Hitoshi Nemoto, RT – Tohoku University Hospital
Makoto Ishida, PhD – Tohoku University
Kazushi Ukishiro, MD, PhD – Tohoku University Graduate School of Medicine
Aimee Luat, MD – Children’s Hospital of Michigan
Hidenori Endo, MD, PhD – Tohoku University
Nobukazu Nakasato, MD, PhD – Tohoku University
Eishi Asano, MD, PhD – Wayne State University
Rationale:
Pathological brain regions often exhibit pronounced structural and functional abnormalities compared to adjacent, unaffected tissue, whereas healthy brain areas show only mild, intrinsic topographical variability. We hypothesized that the epileptogenic zone responsible for drug-resistant focal seizures is characterized by greater deviations in electrographic activity relative to the surrounding cortex.
Methods:
This international, multicenter observational study included patients with drug-resistant focal epilepsy who underwent curative epilepsy surgery following intracranial EEG recording and had at least one year of postoperative follow-up. We defined high-frequency oscillation (HFO)-spatial volatility as the difference in the interictal HFO occurrence rate between each electrode and its neighboring sites (Figure 1). We evaluated whether higher HFO-spatial volatility, rather than elevated HFO rate at local electrode sites, in resected versus preserved regions independently predicted postoperative seizure freedom, controlling for clinical, imaging, and ictal intracranial EEG findings. We tested whether incorporating either local HFO rate or HFO-spatial volatility into a prediction model improved outcome classification beyond standard presurgical evaluation by comparing the area under the receiver operating characteristic curve (AUROC) using the DeLong test.
Results:
In the derivation cohort (140 patients, 14,933 electrode sites; mean age: 13.1 years old (range: 4-44); female: 68 [48.6%]; seizure freedom: 98 [70.0%]), greater local HFO rate and HFO-spatial volatility in resected areas independently contributed to predicting seizure freedom. Adding HFO-spatial volatility to the standard care prediction model (HFO-spatial volatility model) significantly improved its performance (AUROC increased from 0.745 to 0.801; p = 0.034), whereas adding local HFO rate (local HFO rate model) did not yield a significant improvement (AUROC increased to 0.789; p = 0.098) (Figure 2). In an independent validation cohort (26 patients, 1,626 electrode sites; mean age: 25.7 years old (range: 12-42); female: 14 [53.8%]; seizure freedom: 16 [61.5%]), the HFO-spatial volatility model similarly improved predictive performance (AUROC increased from 0.650 to 0.819; p = 0.020), while the local HFO rate model again failed to show significance (AUROC increased to 0.788; p = 0.053).
Conclusions:
Despite the retrospective design and modest validation sample size, these findings suggest that increased heterogeneity in the local distribution of interictal HFO is a key feature of the epileptogenic zone. Spatial volatility may represent a promising biomarker to enhance the localization of the epileptogenic zone in drug-resistant focal epilepsy.Funding:
This work was supported by the National Institutes of Health (NS064033 to E.A.), the Uehara Memorial Foundation Postdoctoral Fellowship (20210301 to H.U.), and the Japan Society for the Promotion of Science (JP22J23281 and JP22KJ0323 to N.K.).