Assessment of Glioblastoma Multiforme Tumor Heterogeneity via MRI-derived Shape and Intensity Features

Area of Science:

• Medical imaging analysis
• Computational anatomy
• Oncology research

Background:

• Glioblastoma multiforme (GBM) is an aggressive brain tumor with variable patient prognosis.
• Accurate characterization of tumor features, including shape and intensity, is crucial for understanding disease progression and survival.
• Current methods may not fully integrate spatial and intensity information for comprehensive tumor analysis.

Purpose of the Study:

• To develop and validate a novel geometric approach for joint characterization of tumor shape and intensity in glioblastoma.
• To enable objective comparison and statistical analysis of tumor features for prognostic insights.
• To explore the relationship between tumor characteristics, patient survival, and tumor heterogeneity.

Main Methods:

• A geometric representation is proposed that is invariant to translation, scale, rotation, and reparameterization.
• This representation allows tunable emphasis on shape and intensity components for registration, comparison, and statistical summarization (e.g., PCA).
• A composite distance metric integrating shape and intensity from multiple imaging modalities is defined, coupled with distance-based clustering.

Main Results:

• Application to a glioblastoma cohort revealed distinct patient groups with significant differences in median survival.
• Cluster memberships were successfully correlated with tumor heterogeneity.
• The study demonstrates that variations in tumor shape significantly impact disease prognosis.

Conclusions:

• The proposed shape+intensity geometric framework provides an objective and powerful tool for analyzing glioblastoma from MRI.
• This method facilitates the discovery of prognostic subgroups and links tumor characteristics to survival outcomes.
• Tumor shape variation is identified as a critical factor influencing glioblastoma prognosis.