Authors :
Presenting Author: Roberto Santalucia, MD – Institute of Neurosciences, UCLouvain, Brussels, Belgium
Carlotta Carosio, BS – University Children's Hospital, Department of Neuropaediatrics, Zurich, Switzerland
Antonio Giulio Gennari, MD – University Children's Hospital, Department of Neuropaediatrics, Zurich, Switzerland
Enrique Germany Morrison, PhD – Université catholique de Louvain
Amir Ghasemi Baroumand, PhD – Medical Image and Signal Processing, Ghent University, Ghent, Belgium
Pascal Vrielynck, MD – Centre Hospitalier Neurologique William Lennox, Ottignies, Belgium
Alexane Fierain, MD – Cliniques universitaires Saint-Luc, Brussels, Belgium
Antoine Guilmot, MD – Centre Hospitalier Neurologique William Lennox, Ottignies, Belgium
Vincent Joris, MD – Institute of Neurosciences, UCLouvain
Susana Ferrao-Santos, MD, PhD – Cliniques Universitaires Saint Luc
Pieter van Mierlo, Ph.D. – Clouds of Care, NV, Ghent, Belgium
Georgia Ramantani, MD, PhD – Department of Neuropediatrics, University Children's Hospital and University of Zurich, Zurich, Switzerland
Riëm El Tahry, MD, PhD – Institute of Neurosciences, UCLouvain, Brussels, Belgium
Rationale:
To evaluate the feasibility of low-density (LD) interictal (IIC) and ictal (IC) electrical source imaging (ESI), and to assess their respective and combined diagnostic accuracy and predictive value in children with drug-resistant epilepsy (DRE) who underwent resective surgery before the age of 7.Methods:
De-identified EEG and MRI data were retrospectively collected and (semi)-automatically processed, blinded to clinical information, to compute both IIC and IC-ESI. Concordance of ESI localizations with the resection zone (RZ) at sublobar level and association with surgical outcome were assessed.
Results:
Thirty-two children were included in the study. IIC and IC-ESI showed an accuracy of 66% (CI 95% 47-81%) and 72% (CI 95% 53-86%) and a diagnostic odds ratio (DOR) of 3.0 (CI 95% 0,66 -13,69; p = 0,15) and 5.0 (CI 95% 0,91 -27,47; p = 0,06), respectively. The combined approach increased diagnostic performance, achieving an overall accuracy of 75% and a DOR of 11.4 (CI 95% 1.08-120,35; p = 0,042). In multivariate logistic regression analysis, the combined IIC/IC ESI result was the strongest predictor of postsurgical freedom (OR: 222,28; p = 0.0262; AUC: 0.87).
Conclusions:
Combined (semi)-automated LD-IIC and IC-ESI is feasible and can accurately localize the EZ and predict seizure outcome in resective surgery before the age of 7. ESI may help increase referral of young children for surgery, shorten the time from epilepsy onset to surgical treatment, and ultimately improve postsurgical outcome.
Funding: The research work was supported by Fond de Recherche Clinique (Cliniques Universitaires Saint-Luc, Brussels, Belgium)