Tumor-associated epilepsy at diagnosis in glioblastoma patients reveals an inflammatory molecular signature and is associated with better overall survival

Dimova L, Stritzelberger J, Dimov D, Ort J, Hamou HA, Clusmann H, Schnell O, Hamer H, Putz F, Corradini S, Singer L, Dörfler A, Ricklefs F, Drexler R, Simon M, Heiland DH, Delev D (2026)

Publication Type: Journal article

Publication year: 2026

Journal

Neuro-Oncology Oxford University Press

Book Volume: 28

Pages Range: 2461-2471

Journal Issue: 10

DOI: 10.1093/neuonc/noag134

Abstract

Background: Glioblastoma (GB) is the most aggressive primary brain tumor in adults. Tumor-associated epilepsy (TAE) at diagnosis is common, yet its prognostic significance remains unclear. Methods: We analyzed a retrospective multicenter test cohort of 855 GB patients (Aachen, Hamburg, Bielefeld) and validated findings in a prospectively collected cohort of 344 patients (Erlangen). Survival was assessed using multivariable Cox regression, propensity score matching, and interaction modeling of TAE and extent of resection (EOR). Molecular profiling included methylation-based classification, epigenetic deconvolution, and spatial transcriptomics. Results: Tumor-associated epilepsy was independently associated with improved survival (hazard ratio [HR] 0.81, 95% CI 0.69-0.99, P = .036, absolute survival advantage ∼4-5 months). This effect was validated in the independent cohort (C-index 0.68, 95% CI 0.62–0.74) and persisted in propensity-matched analyses (HR 0.74, 95% CI 0.56-0.96, P = .027). Interaction modeling revealed that gross total resection (GTR) improved survival in both groups but particularly in patients with TAE (EOR interaction HR 0.69, 95% CI 0.49-0.99, P = .041). In this subgroup, partial resection provided no significant advantage over biopsy, whereas patients without seizures benefited incrementally from both partial resection and GTR. Molecular analysis demonstrated enrichment of the RTK II subtype, differentiated cell states, and an inflammatory microenvironment in GB with TAE; tumors without seizures displayed neuronal and stem-like features. Functional validation using electrogenomics showed that GB cortical slices with increased inflammatory score exhibited synchronization of action potentials characteristic for seizure-like epileptiform activity. Conclusions: Tumor-associated epilepsy at diagnosis is a favorable prognostic marker in GB, defining a biologically distinct subgroup. Seizure status modifies the prognostic effect of surgical resection, underscoring the importance of GTR particularly in patients presenting with TAE.

Authors with CRIS profile

Jenny Stritzelberger Department of Radiation Oncology Oliver Schnell Lehrstuhl für Neurochirurgie Hajo Hamer Professur für Epileptologie Florian Putz Department of Radiation Oncology Stefanie Corradini Department of Radiation Oncology Ludwig Singer Neuroradiologisches Institut Arnd Dörfler Department of Neuroradiology Dieter Henrik Heiland Department of Neurosurgery

Involved external institutions

Rheinisch-Westfälische Technische Hochschule (RWTH) Aachen DE Germany (DE) Universität Bielefeld DE Germany (DE) Universitätsklinikum Hamburg-Eppendorf (UKE) DE Germany (DE)

How to cite

APA:

Dimova, L., Stritzelberger, J., Dimov, D., Ort, J., Hamou, H.A., Clusmann, H.,... Delev, D. (2026). Tumor-associated epilepsy at diagnosis in glioblastoma patients reveals an inflammatory molecular signature and is associated with better overall survival. Neuro-Oncology, 28(10), 2461-2471. https://doi.org/10.1093/neuonc/noag134

MLA:

Dimova, Lilyana, et al. "Tumor-associated epilepsy at diagnosis in glioblastoma patients reveals an inflammatory molecular signature and is associated with better overall survival." Neuro-Oncology 28.10 (2026): 2461-2471.

BibTeX: Download