Career Center AES Connect

Abstracts

Non-invasive EEG Source Imaging of Irritative Hubs Predicts Post-surgical Outcome and Reveals Fingerprints of Epileptogenic Pathologies

Abstract number : 1.324
Submission category : 9. Surgery / 9B. Pediatrics
Year : 2023
Submission ID : 36
Source : www.aesnet.org
Presentation date : 12/2/2023 12:00:00 AM
Published date :

Authors :
Presenting Author: Lorenzo Ricci, MD – Department of Medicine and Surgery, Università Campus Bio-Medico di Roma, Via Alvaro del Portillo, 21 - 00128 Roma, Italy . Fondazione Policlinico Universitario Campus Bio-Medico, Via Alvaro del Portillo, 200 - 00128 Roma, Italy

Marilisa Boscarino, MD – Neurorehabilitation Department, IRCCS Salvatore Maugeri Foundation, Institute of Milan, Milan, Italy; Navaneethakrishna Makaram, PhD – Fetal-Neonatal Neuroimaging and Developmental Science Center, Division of Newborn Medicine, Department of Medicine, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Jeffrey Bolton, MD – Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Scellig Stone, MD, PhD – Division of Epilepsy Surgery, Department of Neurosurgery, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Christos Papadelis, PhD – Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, Fort Worth, TX 76104, USA; Patrician Grant, MD – Fetal-Neonatal Neuroimaging and Developmental Science Center, Division of Newborn Medicine, Department of Medicine, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Alexander Rotenberg, MD, PhD – Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Phillip Pearl, MD – Division of Epilepsy and Clinical Neurophysiology, Department of Neurology, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA; Giovanni Assenza, MD, PhD – Department of Medicine and Surgery, Università Campus Bio-Medico di Roma, Via Alvaro del Portillo, 21 - 00128 Roma, Italy. Fondazione Policlinico Universitario Campus Bio-Medico, Via Alvaro del Portillo, 200 - 00128 Roma, Italy; Eleonora Tamilia, PhD – Fetal-Neonatal Neuroimaging and Developmental Science Center, Division of Newborn Medicine, Department of Medicine, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA.

Rationale:
Delineation of the epileptogenic zone (EZ) is required in children with drug-refractory epilepsy (DRE) undergoing resective surgery. Scalp EEG is a well-established and widely available diagnostic tool for the non-invasive estimation of the EZ. A traditional approach to estimate the EZ is to localize the irritative-zone via electrical source imaging (ESI) of interictal spikes. Yet, the irritative-zone may lack specificity to the EZ, particularly in cases where the EZ is not well-circumscribed and is instead an anatomically distributed network. Functional Connectivity (FC) analysis can identify hyperconnected seizure networks and complement the interictal spike ESI, but may also not fully describe the EZ boundaries. We therefore developed the concept of irritative-hubs (IH) defined by the combination of spike ESI and FC. We now test whether the novel IH-index, as overlayed on the surgical resection cavity, predicts postoperative seizure outcome. We also asked whether some IH -index values are associated with specific epileptogenic pathologies.

Methods:
We analyzed presurgical EEG (19-24 channels, 5-min) from 45 children with DRE who had epilepsy surgery with known Engel outcome (29 had Engel 1 outcome). For each patient, we placed ~1,200 virtual sensors (VSs) across the whole cortex, reconstructed their activity (ESI; Beamformer method) during spikes and computed FC between them (Figure 1A). We defined the IH-index for each VS (Figure 1B) as the product of FC and spike-ESI. To assess the presurgical value of our IH-index, we estimated and compared (by Wilcoxon sign-rank test) values inside and outside the resection, and used them to predict postsurgical seizure-freedom (SF, Engel 1) via stepwise binomial logistic regression and ROC curves. Generalized mixed effect models were used to test the association between IH- index and epileptogenic pathology (Figure 1C), which was defined based on histopathological findings from the resected brain tissue.

Results:
Spike ESI values did not differ between inside and outside resection in good (p=0.5) or poor outcomes (p=0.81). The IH-index instead was higher inside than outside the resection in SF patients but not in poor outcomes (Figure 2A). The values of IH-index inside and outside resection were able to predict outcome with 86.7% accuracy, outperforming traditional spike-ESI (AUC: 0.91 vs 0.70, Fig 2B). When studying pathologies, the IH-index was higher inside than outside the resection in focal cortical dysplasia, whereas it showed an opposite pattern in glial tumors (Figure 2C).

Conclusions:
We report a novel method to analyze interictal scalp EEG that combines traditional spike localization with FC to localize irritative hubs (rather than irritative zones). Our IH-index outperforms traditional spike ESI, and has the potential for revealing fingerprints of epileptogenic pathologies. Our findings suggest that integrating traditional ESI with FC analysis based on five minutes of interictal EEG enables an EZ mapping tool that does not require seizure capture.

Funding:
R03NS127044 by the NINDS of NIH

Surgery