Related Experiment Video
Updated: Mar 6, 2026
06:32
Evaluation of Biomarkers in Glioma by Immunohistochemistry on Paraffin-Embedded 3D Glioma Neurosphere Cultures
Published on: January 9, 2019
8.3K
Machine Learning-Based Preoperative Predicting TERT Promoter Mutation and EGFR Gene Amplification Phenotype in IDH
Yan Su1,2, Wei Guo1,2, Yuwei Pan1
1From the Department of Radiology (Y.S., W.G., Y.P., P.H., D.S.), The First Affiliated Hospital, Fujian Medical University, Fuzhou, Fujian, P.R. China.
AJNR. American Journal of Neuroradiology
|March 4, 2026
Summary
Advanced MRI tumor habitat imaging accurately predicts TERT promoter mutations and EGFR amplification in IDH-wildtype glioblastoma. This model aids in identifying aggressive tumor subtypes and guiding prognosis for patients with glioblastoma.
Area of Science:
- Neuro-oncology
- Radiology
- Genetics
Background:
- Telomerase reverse transcriptase (TERT) promoter mutations are key indicators of poor prognosis in isocitrate dehydrogenase (IDH) wild-type glioblastoma.
- Epidermal growth factor receptor (EGFR) gene amplification is a potential prognostic factor in glioblastoma.
Purpose of the Study:
- To evaluate the utility of a tumor habitat imaging model using advanced MRI.
- To predict TERT promoter mutation and EGFR gene amplification status in IDH wild-type glioblastoma.
Main Methods:
- Utilized advanced MRI (conventional, DWI, DSC-PWI) on 179 patients.
- Applied k-means clustering to ADC and CBV maps to define tumor habitats.
- Developed random forest models for predicting TERT and EGFR status, validated on test and validation sets.
Main Results:
- The TERT promoter prediction model achieved AUCs of 0.877 (training), 0.783 (test), and 0.796 (validation).
- The EGFR amplification prediction model achieved AUCs of 0.877 (training), 0.784 (test), and 0.878 (validation).
- Prediction probabilities closely matched actual outcomes, demonstrating model reliability.
Conclusions:
- The advanced MRI tumor habitat imaging model accurately predicts TERT promoter mutation and EGFR amplification status.
- This imaging model is valuable for characterizing IDH wild-type glioblastoma.
