Clinical Value of Radiomics Analysis of a 2cm Pericavitary Edema Habitat in Predicting Recurrence in Postoperative Glioma Patients

Motivation: Accurate prognosis prediction for postoperative glioma patients is challenging due to tumor heterogeneity in the pericavitary edema zone. Goal(s): To develop model integrating habitat imaging and radiomic features of the 2cm pericavitary edema zone to predict prognosis in glioma patients. Approach: Multimodal MRI was used to perform habitat imaging analysis on the pericavitary edema zone. K-means clustering divided the region into habitat subregions. Radiomic features from these subregions were extracted and combined with clinical factors to build predictive models. Results: The high-risk habitat model showed superior predictive performance over traditional radiomic and clinical models, with higher efficiency in both training and validation groups. Impact: This study introduces a novel method to visualize and quantify tumor heterogeneity in the pericavitary edema zone, improving prognosis prediction for postoperative glioma patients and aiding clinicians in precisely delineating radiotherapy target volumes.