PhD Scientific Days 2025

Pathological and Oncological Sciences II.

Transcriptomic Signatures Associated with Recurrence and Survival in Glioblastoma

Name of the presenter

Pánczél András

Institute/workplace of the presenter

Department of Bioinformatics / Department of Neurosurgery and Neurointervention

Authors

Pánczél András¹, Fekete János Tibor¹

1: Department of Bioinformatics

Text of the abstract

Introduction: Glioblastoma (GB) is the most aggressive adult glioma subtype, characterized by inevitable recurrence and limited treatment efficacy. Despite initial response to surgery and chemoradiotherapy, most patients experience tumor relapse within months, contributing to poor overall survival.
Aims: Our aim was to identify transcriptomic signatures that distinguish primary from recurrent GBM and to evaluate their association with overall survival.
Methods: Transcriptomic and clinical data were obtained from the Diffuse Glioma dataset via cBioPortal. Differential gene expression analysis was performed using paired t-tests between primary and recurrent GBM samples. Benjamini-Hochberg correction was applied to adjust for multiple testing. Significant genes were subjected to gene ontology (GO) enrichment analysis. A random survival forest machine learning model was used to assess the association of significant genes with overall survival.
Result: The downloaded dataset included 329 patient records and 693 tumor samples. Among these, 111 glioblastoma patients had matched transcriptomic data for both primary and recurrent tumors, along with available overall survival information. The expression dataset contained 35 438 transcripts. Transcripts with zero values in more than 25% of the available samples were excluded, resulting in a filtered set of 18 950 transcripts. T-tests identified 5 115 genes with statistically significant expression differences between primary and recurrent tumors (p adjusted < 0.05). GO enrichment analysis revealed overrepresentation of molecular functions related to DNA binding and processing (GO:0003697, GO:0003688), helicase activity (GO:0017116, GO:0003678), and ATP-dependent DNA-associated mechanisms (GO:0008094). Enrichment was also observed in microtubule-based motor activity (GO:0008574). Random survival forest model identified several genes strongly associated with overall survival, including SLC25A51, ASPN, KIF18A, RP4-730K3.3, and FN1. Patients stratified by RSF-based risk scores showed distinct survival patterns, with significant differences between high- and low-risk groups (p<0.0001).
Conclusion: Transcriptomic profiling of matched primary and recurrent glioblastoma samples revealed molecular differences and survival-associated gene signatures that may inform future research on disease progression.
Funding: No funding