Glioma is characterized by high heterogeneity and poor prognosis. Attempts have been made to understand its diversity in both genetic expressions and radiomic characteristics, while few integrated the two omics in predicting survival of glioma. This study was intended to investigate the connection between glioma imaging and genome, and examine its predictive value in glioma mortality risk and tumor immune microenvironment (TIME). Clinical, transcriptomics and radiomics data were obtained from public datasets and patients in our center. Correlation analysis between gene expression and radiomic feature (RF) was performed, followed by survival analysis to select RF-related genes (RFRGs) and gene expression-related RFs (GRRFs). After that, RFRGs and GRRFs were used to construct mortality risk prediction model of all glioma and isocitrate dehydrogenase (IDH) wild type (WT) glioma. The association between RFRGs and TIME was explored. Six cohorts composed of 1,754 glioma patients were included. Thirty-five genes and eighty-two RFs demonstrated high correlation with each other. Gene score based on RFRGs was independent predictor of both glioma (P < 0.05) and IDH-WT glioma (P < 0.05). Same score based on GRRFs was also able to stratify risk of both glioma (P < 0.0001) and IDH-WT glioma (P < 0.0001), with nomograms constructed separately. The TIME of gliomas predicted with RFRGs' score found mismatched risk of death with immune response. RFRGs and GRRFs were able to predict glioma mortality risk and TIME. Further studies could validate our results and explore this genome-imaging interactions.
Keywords: Glioma; Imaging genomics; Immune microenvironment; Isocitrate dehydrogenase; Survival.
© 2025. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.