A phase I/II clinical trial evaluating atezolizumab in combination with standard-of-care treatment for newly diagnosed glioblastoma has demonstrated promising survival outcomes, particularly in patients with immune-enriched tumors. The study, conducted at MD Anderson Cancer Center, provides new insights into biomarkers that could guide patient selection for checkpoint inhibitor therapy in this aggressive brain cancer.
Trial Design and Patient Population
The single-arm prospective trial enrolled 60 patients with newly diagnosed glioblastoma between August 2017 and July 2019. Patients received atezolizumab intravenously every two weeks in combination with radiation and temozolomide during the concurrent stage, followed by atezolizumab with adjuvant temozolomide. The study was unselected for MGMT methylation status and included patients with Karnofsky Performance Scale scores of 60 or higher.
Safety and Efficacy Results
Atezolizumab combined with radiation and temozolomide was generally well tolerated, with 56.7% of patients experiencing grade ≥3 treatment-related adverse events. The most common toxicities included lymphopenia (38.3%), thrombocytopenia (8.3%), and elevated alanine aminotransferase (8.3%). Three patients were removed from the study due to dose-limiting toxicities, including hepatitis and pneumonitis.
The overall cohort achieved a median overall survival of 18.0 months (95% CI [14.2, 25.0]), with IDH1 wild-type patients reaching 16.1 months (95% CI [13.9, 24.6]). Among patients with MGMT methylated tumors, median survival extended to 25.4 months compared to 14.6 months in unmethylated cases. The objective response rate was 23.33%, with 10% achieving complete response and 13.33% partial response.
Immune Infiltration as Key Biomarker
A critical finding emerged when researchers analyzed tumor immune infiltration using the ESTIMATE Immune Score (EIS). Patients with high immune infiltration demonstrated significantly improved survival compared to those with low infiltration (24.8 months versus 14.5 months, p = 0.02, HR = 0.45). Notably, 21% of the high EIS group remained alive at 48.7 months.
To validate this finding's specificity to immunotherapy, researchers analyzed independent glioblastoma cohorts from The Cancer Genome Atlas and the GLASS Consortium, where patients received standard chemoradiation without immunotherapy. In these control populations, immune infiltration showed no association with survival, suggesting that the benefit observed in the atezolizumab trial was specifically related to checkpoint inhibitor treatment.
Genetic Markers Predict Treatment Response
Unsupervised clustering analysis of global gene mutation profiles revealed two distinct patient subgroups with markedly different survival outcomes. All tumors in the better-survival cluster harbored PTEN mutations, while the poorer-survival group was enriched for EGFR mutations. Patients with PTEN-mutant tumors achieved significantly better median overall survival compared to those with EGFR mutations (p = 0.013, HR = 0.325).
In multivariate analysis accounting for known prognostic factors, high immune infiltration (HR = 0.23, p = 0.019) and EGFR mutation status (HR = 6.27, p = 0.029) emerged as the most statistically significant independent predictors of survival in this immunotherapy-treated cohort.
Immune Cell Characterization
Detailed immune profiling revealed that tumors with high immune infiltration were characterized by specific cell populations including activated dendritic cells, M1 macrophages, and CD4+ effector memory T cells. Gene set enrichment analysis showed that patients with longer survival had tumors enriched for interferon gamma and inflammatory response pathways, as well as enhanced antigen presentation programs.
The study also found that 91.7% of mesenchymal subtype glioblastomas had high immune infiltration, compared to only 27.2% of proneural and 31.3% of classical subtypes. Patients with mesenchymal subtype tumors achieved superior median survival (26.5 months) compared to proneural and classical subtypes (15.5 and 15.6 months, respectively).
Microbiome Associations
Analysis of fecal samples from a subset of patients revealed that those with longer overall survival showed enrichment of bacteria from the Bacillota phylum (formerly Firmicutes). Patients with higher immune infiltration also demonstrated greater microbial diversity, consistent with emerging evidence linking gut microbiome composition to immunotherapy efficacy.
Clinical Implications
These findings suggest that immune infiltration status could serve as a biomarker for selecting glioblastoma patients most likely to benefit from anti-PD-L1 therapy. The identification of PTEN and EGFR mutations as predictive markers provides additional tools for treatment stratification. The association between mesenchymal subtype and immune infiltration indicates this historically aggressive glioblastoma variant may be particularly suited for immunotherapy approaches.
The researchers noted that their findings support the hypothesis that timing and duration of anti-PD-L1 treatment may be crucial in the newly diagnosed setting, consistent with recent evidence showing efficacy of neoadjuvant checkpoint inhibition in recurrent glioblastoma. The study provides a foundation for future trials incorporating immune infiltration assessment and genetic profiling to optimize patient selection for checkpoint inhibitor therapy in glioblastoma.
