Sep 29

SS 27 - Advances in Multimodal Therapy for Glioblastoma: From Phase I to Phase III Trials

242 - Intratumoral Heterogeneity of Mutational Drivers Associated with Radiation Response in Glioblastoma

01:20pm - 01:30pm ET

Room 153

Presenter(s)

Taylor Weiskittel, MD, PhD, MS, BS - Mayo Clinic Rochester, Rochester, MN

T. M. Weiskittel¹, E. Harfoush², D. Grief³, M. R. Blomquist⁴, R. O. Kowalchuk¹, A. A. Chaudhuri¹, W. Breen¹, L. Hu⁵, N. Tran³, and S. Fortin Ensign⁶; ¹Department of Radiation Oncology, Mayo Clinic, Rochester, MN, ²Mayo Clinic, Rochester, MN, ³Mayo Clinic, Phoenix, AZ, ⁴University of California San Francisco, San Francisco, CA, ⁵Department of Radiology, Mayo Clinic, Phoenix, AZ, ⁶Department of Medical Oncology, Mayo Clinic, Phoenix, AZ

Purpose/Objective(s): Precision radiotherapy strategies in glioblastoma (GBM) assume that genomic biomarkers from a single surgical specimen reflect the biology of the entire tumor. However, GBM is spatially heterogeneous, and recurrences arise from unresected non-enhancing (NE) margins. We investigated spatial heterogeneity of genomic drivers linked to radiation response using multiregional tumor sampling. We tested whether multiregional profiling reveals clinically relevant heterogeneity in GBM drivers linked to treatment resistance and stress/DNA damage response.

Materials/Methods: Thirty-nine treatment-naïve GBM patients underwent multi-institutional, spatially registered intraoperative sampling from contrast-enhancing (CE) cores and NE margins (mean 5 biopsies/tumor; 1-2 CE, 3-4 NE). Whole-exome sequencing was performed on all samples. An initial radiation relevant driver set five radiation response-associated genes (EGFR, NF1, PTEN, TP53, and CDKN2A) were assessed for pathogenic mutations and copy number alterations. Broader actionable and radiotherapy-resistance gene/pathway annotation is planned. In seven patients, multiregional profiles were compared to matched clinical sequencing derived from a single standard-of-care biopsy from CE only. Spatial heterogeneity was evaluated at both patient and cohort levels.

Results: Multiregional sequencing revealed substantial spatial heterogeneity that was not captured by clinical testing. Among seven patients with matched clinical sequencing, pathogenic TP53 mutations were detected exclusively in spatially distinct regions of two patients and were missed clinically. Across all patients, 18 copy number alterations were identified by multiregional sequencing, yet only 4 were reported on clinical assays. Across the cohort, 25/39 patients had CE-exclusive pathogenic mutations, while 7/39 (17.9%) harbored pathogenic drivers restricted to the NE region. In one extensively sampled tumor (11 biopsies), spatially distinct subclones included EGFR p.R108K (72.7%), EGFR gain (54.5%), NF1 loss (54.5%), and EGFR p.A244V (18.2%), demonstrating intratumoral diversification of radiation-relevant pathways. Greater numbers of NE biopsies correlated with increased detection of heterogeneity and exclusive mutations.

Conclusion: Limited profiling of five radiation response–associated genes revealed substantial spatial heterogeneity in newly diagnosed GBM. Single sample clinical sequencing underestimates genomic diversity, particularly within the unresected NE region. These findings challenge current biomarker paradigms for precision radiotherapy and suggest that radiation sensitivity may be governed by spatially distinct subclones outside the resection cavity. These findings support biologically informed radiotherapy strategies (e.g., escalation/boost concepts) and motivate spatially resolved profiling and predictive mapping to link regional genomics with recurrence and response patterns.