Biom-25. Molecular Profiling Predicts Early And Late Progression In Glioblastoma

Marina Nikolopoulos,Megan Wu,Alexander Bahcheli,Sorcha Kellett,Sten Myrehaug,Arjun Sahgal,Melanie Spears,Shraddha Pai,Jane Bayani,Sunit Das

DOI: https://doi.org/10.1093/neuonc/noae165.0098

2024-11-29

Neuro-Oncology

Abstract:Patients with glioblastoma experience a wide variation in response to standard treatment, with nearly 30% experiencing tumour progression during treatment, and nearly 6% surviving more than 5 years. To date, there are few non-invasive clinical biomarkers to predict response to first – line treatment. Chemical exchange saturation transfer (CEST) MRI may have the potential to fill this gap. CEST MRI is sensitive to treatment-induced changes and changes in tumor metabolism. Our team has obtained CEST data for patients before, during and after the end of standard chemoradiation treatment, and found that CEST provides markers of early response and can identify early, standard and late progressors before treatment initiation. In this study, we aimed to establish molecular profiles of early, standard and late progressors with IDH wild-type glioblastoma. Patients (n=180) with primary, IDH wild-type glioblastoma were imaged with CEST-MRI at multiple time points throughout standard chemoradiation treatment. DNA and RNA were co-extracted from matched normal and tumour pairs and processed for whole genome sequencing, enzymatic methyl-seq and gene expression analysis using Nanostring. Clinical variables such as age, extent of resection, sex, ECOG status and MGMT promoter methylated were also collected. A survival analysis was conducted using the Kaplan-Meier method with log-rank tests. Univariate and multi-variate hazard ratios for clinical variables were calculated by fitting Cox Proportional Hazards Models. Early progressors reported a median progression-free survival (PFS) of 142 days compared to 832 days in late progressors (p<0.0001). Early progressors also harbored distinct and statistically significant differences in gene expression and genomic alterations, namely in DNA damage repair, glucose transport and arginine metabolism pathways. A gene signature was prognostic of PFS and overall survival. Collectively, this data has the potential to serve as a radiogenomic biomarker to assess treatment response before or within early phases of treatment and allow for individual tailoring of the treatment plan.

oncology,clinical neurology