Beyond gross total resection (GTR): Deep peritumoral radiomics for predicting overall survival (OS) and O6-methylguanine-DNA-methyltransferase promoter methylation (MGMTpm) status in glioblastoma multiforme.
Authors
Affiliations (9)
Affiliations (9)
- Department of Computer Science, International Institute of Information Technology, Bhubaneswar, Odisha, India.
- Department of Neuro Imaging and Interventional Radiology, National Institute of Mental Health and Neuro Sciences, Bengaluru, Karnataka, India.
- Department of Biotechnology and Chemical Engineering, School of Engineering, Faculty of Science, Technology and Architecture, Manipal University Jaipur, Jaipur, Rajasthan, India.
- Infinity Insights Biotechnology Co., Ltd., Taipei, Taiwan, Taipei, Taiwan.
- Faculty of Health Sciences, Shinawatra University, Pathum Thani, Thailand.
- Department of Medicine Research, Taipei Medical University Hospital, Taipei, Taiwan. [email protected].
- Graduate Institute of Data Science, College of Management, Taipei Medical University, Taipei, Taiwan. [email protected].
- Department of Information Technology Office, Taipei Medical University Hospital, Taipei, Taiwan. [email protected].
- Department of Computer Science & Engineering, Indian Institute of Information Technology Vadodara, Gandhinagar, Gujarat, India. [email protected].
Abstract
Glioblastoma Multiforme (GBM) is an aggressive and highly heterogeneous brain tumor with poor survival outcomes. While conventional radiomic analyses focus on tumor-centric regions, emerging surgical strategies, such as GTR and supratotal resection (SupTR), highlight the importance of the peritumoral zone. Moreover, due to intratumoral heterogeneity, MGMTpm, a key molecular biomarker guiding chemotherapeutic decisions, is assessed invasively and remains prone to sampling bias. Therefore, this study examined the prognostic and predictive utility of deep radiomic features from tumor and peritumoral regions in preoperative MRI to improve OS prediction and enable non-invasive MGMTpm classification. Multi-parametric structural MRI scans (T1, T1-Gd, T2, and T2-FLAIR) from 520 to 200 GBM patients were analyzed for OS and MGMTpm prediction, respectively. Tumor and peritumoral masks were segmented and expanded using morphological dilation from 2 to 12 mm. ~11,000 deep features per patient were extracted using ResNet50 and ViT-B16 models, capturing patterns that may reflect tumor infiltration and microenvironmental changes. Hybrid feature selection using variance thresholding and Recursive Feature Elimination (RFE) was applied, including age and gender. Support Vector Machine (SVM) classifiers were trained using 10-fold cross-validation. In OS prediction, the inclusion of an 8 mm peritumoral margin improved AUC from 0.74 (95% CI: 0.69-0.80) to 0.76 (95% CI: 0.71-0.80). For MGMTpm prediction, combining tumor with a 10 mm peritumoral margin achieved the highest AUC of 0.81 (95% CI: 0.77-0.86), surpassing tumor-only models (AUC 0.71, 95% CI: 0.65-0.77). Peritumoral regions encode prognostic and molecular information. Integrating these regions enhances radiomics-based prediction of survival and MGMTpm status, supporting their role in non-invasive precision GBM management.