Pre-operative MRI-Based Radiomics for Predicting Telomerase Reverse Transcriptase Promoter Mutation Status in Glioma...

Area of Science:

• Neuro-oncology
• Medical imaging
• Genetics

Background:

• TERT promoter (TERTp) mutations are crucial for glioma prognosis and treatment.
• Tissue testing for TERTp status faces limitations like sampling errors and surgical inaccessibility.
• Magnetic Resonance Imaging (MRI)-based radiomics presents a non-invasive alternative for TERTp status prediction.

Purpose of the Study:

• To assess the diagnostic accuracy of pre-operative MRI radiomics for predicting TERTp mutations in glioma.
• To compare the performance of radiomics-only, clinical-only, and combined radiomics-clinical models.
• To identify factors influencing model performance and heterogeneity.

Main Methods:

• A systematic review and meta-analysis following PRISMA-DTA guidelines.
• Searched PubMed, Embase, Web of Science, and Scopus up to October 13, 2025.
• Included 14 retrospective studies (2,863 patients) evaluating MRI radiomics models against molecular reference standards.
• Utilized bivariate random-effects models to pool sensitivity, specificity, and AUC, assessing risk of bias with QUADAS-AI.

Main Results:

• MRI-only radiomics models showed moderate discriminative performance (AUC 0.79) with substantial heterogeneity.
• Clinical-only models had lower pooled performance (AUC 0.73).
• Combined radiomics-clinical models demonstrated numerically higher performance (AUC 0.82), though not definitively superior.
• Subgroup analyses indicated classifier type, validation strategy, and software influence performance.

Conclusions:

• Pre-operative MRI radiomics offers moderate accuracy for predicting TERTp mutation status in glioma.
• Combined models show potential but require further validation due to study limitations (retrospective design, heterogeneity).
• Current models should be adjunctive; prospective, multicenter validation with standardized workflows is essential for clinical implementation.