Related Experiment Video
Updated: Jul 23, 2025
09:17
Digital Spatial Profiling for Characterization of the Microenvironment in Adult-Type Diffusely Infiltrating Glioma
Published on: September 13, 2022
2.4K
Spatial cellular architecture predicts prognosis in glioblastoma
Yuanning Zheng1, Francisco Carrillo-Perez1,2, Marija Pizurica1,3
1Department of Medicine, Stanford Center for Biomedical Informatics Research (BMIR), Stanford University, Stanford, CA, 94305, USA.
Nature Communications
|July 11, 2023
Summary
Glioblastoma tumor architecture and cell organization significantly impact patient prognosis. Spatial transcriptomics and deep learning reveal how cell subtypes and their arrangement predict glioblastoma treatment resistance and survival outcomes.
Area of Science:
- Oncology
- Computational Biology
- Genomics
Background:
- Intra-tumoral heterogeneity and cell-state plasticity contribute to glioblastoma's therapeutic resistance.
- Understanding glioblastoma's spatial cellular organization is crucial for improving patient prognosis.
Purpose of the Study:
- To investigate the association between spatial cellular organization and glioblastoma prognosis.
- To develop deep learning models for predicting glioblastoma transcriptional subtypes and prognosis from histology images.
Main Methods:
- Leveraged single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics data.
- Developed a deep learning model to predict transcriptional subtypes from histology images, analyzing 40 million tissue spots from 410 patients.
- Validated findings with a separate deep learning model predicting prognosis from histology images and applied it to spatial transcriptomics data.
Main Results:
- Identified consistent associations between tumor architecture and prognosis across two independent cohorts.
- Poor prognosis in glioblastoma patients correlated with higher proportions of tumor cells expressing a hypoxia-induced transcriptional program.
- Clustering of astrocyte-like tumor cells indicated worse prognosis, while their dispersion and connection with other subtypes correlated with decreased risk.
Conclusions:
- Established a critical link between spatial cellular architecture and glioblastoma clinical outcomes.
- Presented a scalable deep learning approach to unravel transcriptional heterogeneity and identify survival-associated gene expression programs.
- Spatial organization and cell-state plasticity are key determinants of glioblastoma therapeutic resistance and prognosis.
