Related Experiment Video
Updated: Jun 9, 2025
05:45
Quantitative Immunohistochemistry of the Cellular Microenvironment in Patient Glioblastoma Resections
Published on: July 31, 2017
9.6K
Machine learning-based new classification for immune infiltration of gliomas
Feng Yuan1, Yingshuai Wang2, Lei Yuan3
1Department of Neurosurgery, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, China.
Plos One
|October 25, 2024
Summary
This study identifies four distinct glioma immune subtypes using machine learning, revealing differences in immune cell infiltration, molecular characteristics, and clinical outcomes to guide immunotherapy strategies.
Area of Science:
- Oncology
- Immunology
- Bioinformatics
Background:
- Glioma is a heterogeneous, poorly immunogenic tumor with limited immunotherapy efficacy.
- The immunosuppressive tumor microenvironment (TME) hinders immunotherapy effectiveness.
- Understanding glioma's immune microenvironment (IME) is crucial for improving treatment.
Purpose of the Study:
- To reveal IME characteristics in glioma.
- To predict distinct glioma immune subtypes using machine learning (ML).
- To provide guidance for glioma immunotherapy.
Main Methods:
- Unsupervised cluster analysis on glioma gene expression data (CGGA and TCGA databases).
- Development and verification of a ML-based classification model for immune subtypes.
- Functional enrichment analysis, immune cell distribution, stemness, phenotype, and molecular/clinical characteristic analysis of identified subtypes.
Main Results:
- Four glioma immune subtypes (IM1-IM4 and IMA-IMD) were identified in both databases.
- A reliable ML model was developed to predict these subtypes.
- Subtypes differed significantly in immune cell infiltration (e.g., Monocytic lineage, NK cells, CD8 T cells), signaling pathways (e.g., IL-8, TNF), stemness, neuronal/mesenchymal phenotypes, and clinical features (GBM proportion, survival, IDH mutation, 1p36/19q13 co-deletion).
Conclusions:
- A reliable ML model for predicting glioma immune subtypes was developed.
- Four distinct glioma subtypes (immunogenic, adaptive immune resistance, mesenchymal, immune tolerance) were defined.
- These subtypes represent different TMEs and tumor development stages, offering insights for targeted immunotherapy.
