Radiomic profiles in diffuse glioma reveal distinct subtypes with prognostic value

Area of Science:

• Neuro-oncology
• Medical imaging analysis
• Computational pathology

Background:

• Diffuse gliomas are a heterogeneous group of brain tumors with variable prognoses.
• Accurate stratification is crucial for effective treatment planning and patient management.
• Current methods may not fully capture the complex biological and clinical heterogeneity.

Purpose of the Study:

• To assess a radiomic approach for stratifying diffuse gliomas based on prognosis.
• To correlate radiomic features with clinicopathological and molecular characteristics.
• To explore the potential of radiomics in predicting patient outcomes.

Main Methods:

• Retrospective analysis of multi-channel MRI data from 166 diffuse gliomas.
• Extraction and survival analysis of 704 radiomic features.
• Consensus clustering to identify prognostic glioma subtypes.
• Integration of multi-layered molecular data and immune cell infiltration analysis.

Main Results:

• 318 radiomic features significantly correlated with patient overall survival.
• Two distinct prognostic subgroups were identified through radiomic feature clustering.
• Significant differences in histological grade, IDH status, and MGMT methylation between subtypes.
• Inferior prognosis subtype showed potential for increased immunotherapy response.

Conclusions:

• A radiomic model from multi-parameter MRI successfully stratified diffuse glioma patients by risk.
• Radiomics offers a valuable alternative for survival estimation in gliomas.
• This approach may improve clinical decision-making and personalized treatment strategies.