Meeting Banner
Abstract #3199

Genomic Mapping and Survival Prediction in Glioblastoma: Role of Tumor Blood Volume Versus Molecular Sub-Classification - A TCGA Glioma Phenotype Research Group Project

Rajan Jain1, Laila Poisson, Jayant Narang, David Gutman2, Adam Flanders3, Carl Jaffe4, Brat Daniel2, Tom Mikkelsen

1Henry Ford Health System, Detroit, MI, United States; 2Emory University; 3Thomas Jefferson University Hospital; 4Boston University

The purpose of this study was to assess the utility of tumor blood volume measured using DSC T2* MR perfusion versus molecular sub-classification of glioblastoma for patient survival prediction. Our results show that rCBV measures do better than molecular mapping and sub-classes in predicting survival in this group of highly malignant and uniformly fatal tumors. Even though genomic mapping helps in better understanding of the molecular basis of tumor cell origin, aggressiveness and heterogeneity of glioblastoma, still non-invasive imaging biomarkers can be an important adjunct for patient prognosis.

Keywords

abnormalities accounting additionally adjusting affect appears applied association atlas attributed available background becomes best better biologic biology blood brat cancer cell characterization characterized class classes classical classification classified classifier clinically closely comprehensive confounding conjunction considered correction correlated curves deaths defines degree delineate derived diagnosis disease dropped earlier enhancing expression five ford future gene generated genome grade greater hazard health henry hospital human identifies important included increasing independent individualized institutes integrative interestingly leakage lesion lesser long looked mapping marginally materials measured measures median model models molecular nature neither neoplasms network neural overall parsimony part pathways patient patients pattern perfusion phenotype play poor potential potentially predicted prediction predictor prognosis progression project proliferative prominent rank regimens relate relevant remains represented respectively risk role sake sample samples scanner scheme software sole square stages still studies subclass subclasses subsequent survival surviving system table targeting therapy though trans treatment tumor type understanding underwent versus volume worst years