Abstract
Immune phenotype-guided signature reveals tumor immune microenvironment, intratumoral heterogeneity, TCR Shannon diversity, genome instability and immunotherapy sensitivity of glioma
Department of Thoracic Surgery, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, 330006, China
Departments of Neurosurgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
Departments of Neurosurgery, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, 330006, China
Folia Neuropathol 2026; 64 (3): 240-274
Introduction
The immunosuppressive mechanism of the glioma tumor microenvironment (TME) and TME immunogenomic features have not been completely elucidated.
Material and methods
Through multiomics data on glioma from large-scale datasets and computational algorithms, we determined the glioma TME features. Then, TME clusters and a TME risk signature were constructed and analyzed in combination with these features. Validation analyses confirmed their value.
Results
Based on 2973 glioma samples, the phenotypes of the immune-inflamed, immune-desert, and immune-excluded groups and clusters of TME cluster A, B, C, D within individual tumors were demonstrated to be closely associated with the inflammation stage, tumor stroma and immune activity, genetic variation and individual prognosis. Patients with a high TME risk had unfavorable overall survival (hazard ratio [HR] = 0.15, 95% confidence interval [CI]: 0.12-0.19) and progression-free interval (HR = 0.23, 95% CI: 0.18-0.29). Interestingly, corticosteroid use was linked to poor survival
(HR = 0.65, 95% CI: 0.42-0.99), as well as unfavorable progression-free intervals (HR = 0.63, 95% CI: 0.46-0.88). The prognostic value of the TME risk signature was validated in pan-The Cancer Genome Atlas (TCGA) central nervous system (CNS) tumor cohorts besides glioma (HR = 0.23, 95% CI: 0.16-0.34). The TME high-risk group exhibited inflamed immune TME characteristics, a higher level of T cell receptor (TCR) richness, TCR Shannon diversity, considerable intratumoral heterogeneity and a more aggressive nature compared to the low-risk group. The role of the TME risk signature in predicting immunotherapeutic outcomes was further confirmed in several immunotherapy cohorts. Expectedly, the TME risk signature offered high diagnostic accuracy (near 100%) for distinguishing TME high- and low-risk groups through AdaBoost, GBDT, XGBoost and random forest machine learning models.
Conclusions
The immune phenotypes and TME clusters display the landscape of glioma TME , and the novel TME risk signature acts as a robust prognostic marker for predicting the immunotherapeutic response and as a diagnostic tool.
Keywords
immunogenomic analysis, machine learning, microenvironment, immune phenotype, glioma
Integrated with
