Abstract

Introduction
Glioblastoma (GBM) recurrence is driven by residual disease that infiltrates beyond the resection margin, a region poorly characterised at the molecular level. Fluorescence-guided surgery using 5-aminolevulinic acid (5ALA) identifies metabolically active invasive tumour populations and provides an opportunity to interrogate the biology underpinning local invasion and recurrence.

Methods
Spatially resolved bulk RNA-seq data derived from 5ALA-defined GBM regions were analysed to identify genes enriched at the invasive margin. Differential expression within intra-tumour regions were integrated with protein-protein interaction analysis and patient survival. Protein-coding genes associated with adverse overall and disease-free survival were prioritised for structure-based drug repurposing using binding pocket identification and molecular docking against FDA-approved and experimental libraries. In parallel, 16 invasive-specific lncRNAs were identified and curated, and literature-informed Boolean network modelling is being used to explore their regulatory roles within invasion-associated gene networks.

Results
This integrative pipeline identified an invasive- specific gene signature enriched in 5ALA+ tumour populations. Six genes upregulated in 5ALA+ regions relative to other intra-tumour regions demonstrated invasive-margin specificity and consistent associations with survival and were prioritised for structure-based drug repurposing. Molecular docking identified a focused set of clinically relevant candidate compounds, including fosphenytoin, flecainide, droperidol, lemborexant, and safinamide, which are currently being progressed to wet-lab validation in GBM 2D/3D models. In parallel, 16 lncRNAs uniquely enriched at the invasive margin were identified and selected for Boolean network modelling to explore their potential regulatory roles within invasion-associated gene networks.

Conclusion
This work establishes a multi-layered computational framework linking invasive-margin biology with therapeutic prioritisation and regulatory network modelling. By targeting both protein-coding candidates and lncRNA driven regulatory dynamics, this approach aims to advance strategies directed at residual GBM cells responsible for tumour recurrence.