Related Experiment Video
Updated: Oct 19, 2025
Digital Spatial Profiling for Characterization of the Microenvironment in Adult-Type Diffusely Infiltrating Glioma
Published on: September 13, 2022
Deep learning features from diffusion tensor imaging improve glioma stratification and identify risk groups with
Jing Yan1, Yuanshen Zhao2, Yinsheng Chen3
1Department of MRI, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China; Glioma Multidisciplinary Research Group, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China.
A novel deep learning signature (DLS) from diffusion tensor imaging (DTI) effectively predicts overall survival in infiltrative glioma patients. This DLS identifies high-risk groups with specific biological pathway dysregulation, suggesting targeted therapy approaches.
Area of Science:
- Neuroimaging
- Oncology
- Artificial Intelligence
Background:
- Infiltrative gliomas pose a significant challenge in predicting patient survival.
- Diffusion tensor imaging (DTI) offers insights into brain tumor microenvironment.
- Developing advanced predictive models is crucial for personalized glioma treatment.
Purpose of the Study:
- To develop and validate a deep learning signature (DLS) using DTI for predicting overall survival in infiltrative glioma patients.
- To identify and investigate the biological pathways underlying the DLS.
- To assess the DLS's prognostic value and its integration into existing risk assessment tools.
Main Methods:
- A DLS was developed using a deep learning cohort (n=688).
- Key biological pathways were identified using a radiogenomics cohort with paired DTI and RNA-seq data (n=78).
- Prognostic value of pathway genes was validated in public databases (TCGA, CGGA).
Main Results:
- The DTI-derived DLS significantly correlated with patient survival (log-rank P < 0.001) and was an independent predictor.
- A deep learning nomogram incorporating the DLS demonstrated superior survival prediction accuracy compared to existing models.
- Five pathway types, including synaptic transmission and axon guidance, were significantly correlated with the DLS and showed prognostic significance.
Conclusions:
- DTI-derived DLS enhances glioma stratification by identifying risk groups linked to specific survival-associated biological pathways.
- Targeted therapies aimed at inhibiting neuron-to-brain tumor synaptic communication may benefit high-risk glioma patients identified by the DTI-derived DLS.
