Radiomic profiling for insular diffuse glioma stratification with distinct biologic pathway activities

Area of Science:

• Neuro-oncology
• Radiomics
• Genomics

Background:

• Insular gliomas present surgical challenges, with limited radiogenomic data on prognostic imaging markers.
• Current management relies on surgical complexity and tailored resection strategies.

Purpose of the Study:

• To develop and validate a multiparametric MRI-based radiomic model for predicting prognosis in insular gliomas.
• To uncover the biological mechanisms underlying radiomic prognostic traits through radiogenomic analysis.

Main Methods:

• An 18-feature radiomic risk signature (RRS) was developed using preoperative MRI data from 364 insular glioma patients (295 training, 69 validation).
• Radiogenomic analysis involved paired MRI and RNA-sequencing data (N=39) using Gene Set Enrichment Analysis (GSEA) and Weighted Gene Coexpression Network Analysis (WGCNA).

Main Results:

• The RRS significantly predicted overall survival (OS) in the validation set (p=0.00058, HR=3.595).
• A combined radiomic-pathological-clinical model (R-P-CM) improved prognostic accuracy.
• Radiogenomic analysis identified 322 intersectional pathways, with 13 prognostic radiomic features correlating significantly.

Conclusions:

• The RRS is an independent predictor of insular glioma prognosis, outperforming clinical and pathological profiles.
• Prognostic radiomic features are associated with immune, proliferative, migratory, metabolic, and cellular functions, providing biological insights.