AI Pathomics Model for Treg Infiltration and Radiotherapy Response in Glioblastoma

An AI-based pathomics model predicts regulatory T cell infiltration and patient outcomes in IDH-wild-type glioblastoma. The model was developed using features from H&E-stained biopsies across multi-center cohorts involving over 300 patients. Author Shaoli Peng published the supporting data under a CC-BY-4.0 license in May 2026.

Use Cases

• Predicting regulatory T cell infiltration levels in glioblastoma based on pathomics features.
• Stratifying patient prognosis and overall survival using the derived pathomics score.
• Investigating associations between pathomics signatures and immune-evasive pathways like Notch and IL-6/JAK/STAT3 signaling.
• Exploring potential biomarkers for radiotherapy response, such as DNA repair gene RAD50 expression.

Strengths

• Model validated across multi-center cohorts with a sample size exceeding 300 patients.
• Demonstrated significant hazard ratios for overall survival (e.g., HR=2.16 in TCGA).
• Data is openly shared under a permissive CC-BY-4.0 license.

Limitations

• Row count and column-level documentation are unknown, limiting suitability assessment.
• The dataset is very small (2.9 KB), suggesting it may contain only model outputs or summary statistics, not the underlying patient-level data.

Provenance

Source

figshare, author Shaali Peng.

Collection Method

Features were extracted from hematoxylin and eosin-stained biopsies, with model construction via gradient boosting after mRMR and Relief feature selection.

Freshness

Last updated 2026-05-21 05:35:22.