Presurgical Tumor and Habitat MRI-based Radiomics Features for Predicting Post-Surgical Recurrence of Glioma

Motivation: Accurate preoperative prediction of glioma post-surgery recurrence holds pivotal clinical importance for guiding clinical therapy. Goal(s): Explore the presurgical tumor and habitat MRI-based radiomics features to identify key biomarkers for the surgical recurrence of glioma. Approach: 103 glioma patients who underwent preoperative T1-CE and T2WI MRI were prospectively included. Three habitats were identified and radiomics features were extracted, prediction models for recurrence were constructed, model performance was evaluated. Results: The habitat 2 with a higher volume ratio in recurrence gliomas exhibits a stronger relation to post-surgical recurrence. The combined model of three habitats shows more excellent predictive performance than the whole tumor model. Impact: The presurgical habitat-based radiomics model may facilitate clinicians in deciding whether to undergo the surgical resection of glioma, thus giving a patient-tailored decision.