Abstracts

Rationale: Delineation of the seizure onset zone (SOZ) is required in children with drug resistant epilepsy (DRE) undergoing resective surgery. Intracranial EEG (iEEG) serves as the gold standard for this purpose but presents limitations due to its invasiveness and limited spatial sampling. We examine the clinical utility of virtual implantation based on electrical source imaging performed on ictal scalp EEG recordings for mapping the SOZ. We hypothesize that Electrical Source Imaging (ESI) performed on ictal scalp EEG data can delineate the SOZ noninvasively and guide the placement of iEEG electrodes.

Methods: We retrospectively analyzed ictal scalp EEG data (19 channels) from 35 children with DRE who underwent iEEG monitoring and epilepsy surgery. We dichotomized surgical outcome into seizure-free (SF) and non-seizure-free (NSF). We identified discharges at ictal onsets recorded with scalp EEG. Using ESI, we estimated virtual sensors (VSs) at brain locations that matched the iEEG implantation (Fig. 1). We described the seizure onset patterns of virtual EEG and compared them with iEEG. We estimated the agreement between virtual SOZ electrodes and clinically defined SOZ electrodes and built receiver operating characteristic curves (ROC) to test whether it predicted outcome.

Results: Twenty-one patients (60%) were seizure free after surgery. We identified three seizure onset patterns using visual inspection and time-frequency analysis having the following frequencies: (i) low-voltage fast activity (LVFA): 37.1%, (ii) rhythmic activity ≤13 Hz (RA): 34.3%, and (iii) burst of spike-and-waves activity (SW): 29.6% (Fig. 2). Moderate agreement between virtual and iEEG SOZ patterns was found (kappa=0.45, p< 0.001). Virtual SOZ agreement with clinically defined SOZ was higher in seizure-free compared to non seizure-free patients (67.5% vs. 37%, p=0.01). Anatomical concordance of virtual SOZ with clinically defined SOZ predicted seizure freedom (AUC=0.73; 95% CI: 0.6-0.9; sensitivity = 57.1%; specificity = 78.6%; accuracy = 65.7%).

Conclusions: Virtual implantation based on ictal recordings with conventional scalp EEG can delineate non-invasively the SOZ and predict surgical outcome. Our approach promises to optimize the surgical strategy for patients with DRE. Non-invasively mapping the SOZ using virtual intracranial EEG sensors may augment epilepsy surgery planning, tailor the intracranial EEG implantation, and predict surgical outcome in children with DRE undergoing epilepsy surgery.

Funding: Please list any funding that was received in support of this abstract.: No funding was received in support of this abstract.