Radiogenomic Mapping of Angiogenesis in Glioblastoma

Angiogenesis is one of the critical molecular events in pathogenesis of glioblastoma. Radiogenomics, an emerging field links genotypes with MRI phenotypes. To find the imaging surrogate of angiogenesis, comprehensive radiogenomic analysis using multiparametric MRI phenotypes and large-scale gene-expression profiling in glioblastoma was presented. Based on The Cancer Genome Atlas (TCGA), glioblastomas with EGFR overexpression were enrolled. Tumors with and without altered expression of target angiogenesis genes were further categorized. Qualitative and quantitative MRI phenotype data in these cases were created for analysis. Kaplan Meier survival statistics, MRI-angiogenesis module correlation analyses, and survival analysis with target MRI traits were performed to gain further insight into the radiogenomic signatures. Results: In the first year project, we have recruited 52 cases in this study. Significant worse survival (p=0.03), higher degree of contrast enhancement (p=0.04), higher contrast-necrosis ratio (p=0.04), larger tumor volume (0.02) and larger surface area (0.01) is noted in augmented angiogenic group. Multiple radiomic features including intensity-based, shape/size-based, and textural features were further extracted. Our preliminary results showed that overall survival of patients is correlated with several quantitative features including maximum 3-dimensional diameter (r = -0.33; p = 0.02), spherical disproportion (r = -0.38;p = 0.01), RLN (r = -0.33; p = 0.02), LRHGLE (r = -0.33; p = 0.02), and high gray-level run emphasis (HGLRE, r = -0.32; p = 0.02). Further investigation of their relationship with different angiogenesis and immune-regulation related genetic modules will be executed and large-scale screening of the VEGF genetic profiles will be proceeded in this second year project. Conclusion: Construction of an MR imaging and mRNA radiogenomic association map has led to identification of MR traits that are associated with dysregulated angiogenesis in glioblastoma. These imaging biomarkers can also serve as prognosis factors in our cohort. Our findings also have potential therapeutic significance since successful molecular inhibition of angiogenesis is one of the current standard therapy of glioblastoma.