返回Google圖書搜尋

Integrated Analysis Using Molecular-scale and Imaging-scale Data to Identify Prognostic Subgroups in Glioblastoma

出版	Stanford University, 2015
URL	http://books.google.com.hk/books?id=KHuOAQAACAAJ&hl=&source=gbs_api

註釋Glioblastoma (GBM, World Health Organization [WHO] grade IV) is the most common and most aggressive brain cancer in adults, with a median survival of approximately one year, despite multimodal therapy of microsurgical resection, radiation, and chemotherapy. Previous molecular studies have classified patients into four subtypes based solely on gene expression. However, heterogeneity can be seen within these GBM subtypes, and there is no significant survival difference between the subtypes; thus, there is a pressing need to find better ways of defining disease subtypes based on biological characteristics that impact clinical outcome. We believe that this goal can be accomplished by developing methods to integrate, analyze, and make inferences from multi-scale datasets that span the visible (imaging-based) and molecular (genomic) scales to link imaging phenotype to genomic features. In my thesis, I describe my work developing computational methods to identify imaging-scale markers in combination with molecular-scale markers that better stratify patients into prognostic subgroups than current methods. First, I present imaging informatics methods to quantify two imaging phenotypes, tumor location and cerebral blood perfusion, to stratify patients in GBM. As an extension of perfusion analysis, I also examined inter-vendor and inter-reader variability in perfusion quantification. Second, to relate molecular data to alterations in metabolic activity, I extended an existing pathway analysis framework to incorporate metabolic pathways, enabling interrogation of deregulated energy metabolism in cancer. Third, I developed algorithms for integrating the quantitative imaging phenotypic and biological activity data to identify associations between molecular-scale features and imaging-scale features.