The role of radiomics and deep learning in glioma MRI: advances, challenges, and LMIC-tailored opportunities for precision medicine

Abstract

Glioma is the most common primary malignant central nervous system tumour in adults, with limited access to advanced molecular diagnostics in low- and middle-income countries, like Pakistan. Radiomics and deep learning offer cost-effective, non-invasive alternatives for glioma magnetic resonance imaging characterisation, but their clinical translation remains unoptimised for resource-constrained settings. A structured literature search from January 2018 to October 2025 was conducted across PubMed, Embase, Web of Science, Cochrane Library and Chinese biomedical databases with focus on radiomics and deep learning for glioma magnetic resonance imaging endpoints. Of the 2,132 initial records, 41(1.92%) studies were analysed. The narrative review synthesised advances in radiomics and deep learning for glioma magnetic resonance imaging segmentation, like convolutional neural networks, generative adversarial networks and transformer architectures, preoperative grading (area under the curve: 0.85-0.90), molecular typing, survival stratification and post-radiotherapy response differentiation. Key limitations included small cohorts, lack of external validation, and non-standardised imaging protocols. Radiomics and deep learning enhance glioma precision medicine, but face barriers in terms of standardisation and workflow integration. Prioritising lightweight models and multicentre validation from low- and middle-income countries will accelerate adoption in resource-limited settings, like Pakistan.

Key Words: Glioma, Magnetic resonance imaging, Radiomics, Deep learning, Low- and middle-income countries.