Abstract 1879: Serum biomarker candidates for predicting radiotherapy response in glioblastoma

Abstract Introduction: Glioblastoma remains an incurable disease with limited treatment options. Imaging remains the primary diagnostic tool, but it has limitations in distinguish tumor progression from benign events like pseudoprogression. Currently, no biomarkers exist to differentiate these conditions. This study aimed to identify predictive biomarkers of glioblastoma progression by profiling serum proteins, seeking candidates for future validation. Methods: Blood samples were collected from patients pre-therapy, and during the first and fourth weeks of chemoradiotherapy. We selected 11 samples for non-progression (NP), 12 for progression (P), and 6 for pseudoprogression (PS) based on availability and clinical classification. Proteomic analysis was performed using a proximity extension assay platform targeting more than 5400 serum proteins. Differential protein expression was assessed among groups and correlated with clinical outcomes. Results: Pre-therapy analysis (NP vs P vs PS) identified 185 different proteins (Kruskal-Wallis, p<0.05). Enrichment analysis linked these to glutamatergic synapse, cell proliferation, and neuronal activity. Comparing responders (P) vs. non-responders (NP+PS) revealed 86 proteins (Wilcoxon, p<0.05), associated with interferon regulation and calcium homeostasis and clustered samples accordingly. In PS vs. NP+P, 87 proteins (Wilcoxon, p<0.05) were enriched for homologous recombination and neuronal apoptosis and clustered samples accordingly. At the first therapy week, NP vs. P vs. PS comparisons revealed 301 proteins (Kruskal-Wallis, p<0.05), linked to innate immunity and apoptosis. P vs. NP+PS revealed 95 proteins (Wilcoxon, p<0.05) associated with T cell activation. PS vs. NP+P identified 172 proteins that clustered samples accordingly (Wilcoxon, p<0.05), linked to receptor activity and carbohydrate binding. By week four, protein differences decreased. NP vs. P vs. PS comparisons identified 197 proteins (Kruskal-Wallis, p<0.05) related to nervous system development and wound healing. P vs. NP+PS identified 94 proteins (Wilcoxon, p<0.05) linked to cell adhesion. PS vs. NP+P revealed 75 proteins (Wilcoxon, p<0.05) that discriminated and clustered all PS samples on the same group and enriched for defense response and TRAIL production. Conclusion: This study demonstrates that a sensitive proteomic approach can point to candidate biomarkers predictive of glioblastoma progression. These findings offer insights into radiotherapy response and warrant further mechanistic studies and validation in larger cohorts. Citation Format: Paulo Roberto Del Valle, Janette Herr, Miah Elizabeth Nicolls, Kaylie Cullison, Alessandro Valderrama, Gregory A. Azzam, Jessica J. Meshman, Scott Welford, Macarena I. De La Fuente, Eric A. Mellon. Serum biomarker candidates for predicting radiotherapy response in glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1879.