Authors :
Presenting Author: Daniel Zhou, MD – University of Pennsylvania, Philadelphia, PA
Marc Jaskir, BS – University of Pennsylvania
Yombe Fonkeu, MD, MS – University of Pennsylvania
julian Gal, MD, MS – University of Pennsylvania
Jose Garcia, MS – University of Pennsylvania
Christos Davatzikos, PhD – University of Pennsylvania
MacLean Nasrallah, MD, PhD – University of Pennsylvania
Stephen Bagley, MD, MS – Division of Hematology/Oncology, Department of Medicine
Arati Desai, MD – University of Pennsylvania
Colin Ellis, MD – University of Pennsylvania
James Gugger, MD, PharmD – University of Rochester
Joel M Stein, MD, PhD – University of Pennsylvania
Nishant Sinha, PhD – University of Pennsylvania
Kathryn Davis, MD – Center for Neuroengineering and Therapeutics and Penn Epilepsy Center, Department of Neurology, University of Pennsylvania
Rationale:
Approximately half of patients with glioblastoma (GBM) develop tumor-related epilepsy (TRE). However, the relationship between the topographic features of GBM—including peritumoral edema—and seizure risk remains poorly understood. Advances in neuroimaging have enabled more precise delineation of tumor and edema boundaries. We aimed to evaluate how the spatial distribution of the GBM and surrounding edema correlates with TRE.
Methods:
We conducted a retrospective cohort study of patients with pathology-confirmed IDH-wildtype GBM who received brain MRI at initial presentation at the University of Pennsylvania between 11/2020 and 4/2023. The contrast-enhancing tumor, non-enhancing tumor, and peritumoral edema were segmented using a validated deep learning algorithm (Fig. 1). The brain was parcellated into regions of interest (ROIs) using the Harvard-Oxford atlas. Demographic and clinical variables were extracted by chart review. Logistic regression was performed to evaluate associations between tumor and edema volume and outcomes, adjusting for age and sex.
Results:
Among 100 GBM patients (40% female; mean age 67.2 ± 10.1 years), mean tumor volume was highest in the temporal lobe (4.26 ± 6.49 cm3), while edema volume was highest in the frontal lobe (6.19 ± 9.48 cm3). Forty-six patients developed TRE: 35 initially presented with seizure leading to discovery of brain tumor, with the first clinical event occurring at median (IQR) of 14 (3.5-48) days before MRI. Of the 28 TRE cases with seizures after tumor resection, 15 (54%) were associated with tumor progression. In addition to TRE, 6 other patients developed seizures in the setting of intracranial hemorrhage during the acute postoperative period. Patients with an initial seizure had lower subcortical enhancing tumor volume (OR 0.77, 95% CI 0.61-0.97, p=0.03), higher mesial temporal edema volume (OR 1.32, 95% CI 1.01-1.73, p=0.04), and lower parietal edema volume (OR 0.87, 95% CI 0.79-0.96, p=0.004). By ROI, they had greater edema in the parahippocampal gyrus and fusiform gyrus and lower edema in the precentral gyrus, postcentral gyrus, superior parietal lobule, supramarginal gyrus, and parietal operculum (each p< 0.05). Compared to those without TRE, patients with TRE had lower edema in the parietal lobe (OR 0.92, 95% CI 0.87-0.99, p=0.02), and by ROI, in the supramarginal gyrus, angular gyrus, planum temporale, and lateral occipital cortex (each p< 0.05). No significant associations were found between tumor or edema volumes and postoperative TRE seizures.