Discovery of exon, microRNA and clinical prognostic markers of glioblastoma survi

描述(申请人提供)：美国每年有10,000多人被诊断为多形性胶质母细胞瘤，这是最常见的原发脑瘤，中位生存期(不考虑治疗)为一年。这种疾病的多因素性质和先前研究的有限范围导致了对胶质母细胞瘤存活的病因学的不完全了解。这项应用的主要目的是增加关于胶质母细胞瘤生存的预后标志物和决定因素的知识。概念框架包括同时考虑临床和基因组信息，使用包括397个胶质母细胞瘤和匹配的对照样本的大样本，以及使用脆弱模型综合分析胶质母细胞瘤生存和解释变量之间的关系。特异性目的1主要是检测外显子的表达和胶质母细胞瘤生存的临床预后标志物，而特异性目的2则是揭示microRNA表达的标志物以及外显子的表达和临床标志物。将采用以外显子和微RNA为中心和以基因为中心的方法来模拟基因组信息。为了解决拟议研究的高维问题，将实施一系列特征提取策略。考虑到方差的异质性和观察之间的相关性的混合效应比例风险模型将被用来模拟胶质母细胞瘤的存活率。从贝叶斯马尔可夫链蒙特卡罗方法获得的后验密度估计将提供预后标记物与生存之间的关联的完整表征。生存预测方程的校准将使用交叉验证方法，并对胶质母细胞瘤生存的已知标记物进行基准比较。这项申请涉及癌症流行病学研究中的几个高优先级领域，这些领域已经被NCI确定，我们的团队在拟议研究的基础上具有先前的经验。这项研究的结果将是：1)确定胶质母细胞瘤生存的临床和基因组预后标记物，独立、协同或拮抗地起作用；2)表征与这些标记物相关的存活率变化；3)建立更普遍的胶质母细胞瘤存活率预测或预后指数；4)确定与临床和基因组水平的特定组合相对应的胶质母细胞瘤亚类；5)了解广泛的临床和基因组因素之间的相互作用，最终目的是阐明癌症的病因。拟议的对胶质母细胞瘤基因组和临床数据的多因素分析的结果将支持癌症生存功能分子研究的有效设计，促进基础科学与临床和公共卫生实践之间的发现交流，并导致诊断、预防、治疗和治愈癌症的生物医学战略。 公共卫生相关性：项目叙述该项目将揭示胶质母细胞瘤生存的临床和基因组预后标志和决定因素。这项研究的结果包括加强对多种临床和基因组变量之间相互作用的理解，最终目的是阐明癌症的病因和开发可靠的生存预测功能。这些发现将促进功能分子研究，促进基础科学与临床和公共卫生实践之间的发现交流，以及开发诊断、预防、治疗和治愈癌症的生物医学战略。

DESCRIPTION (provided by applicant): Abstract More than 10,000 people per year in the United States are diagnosed with Glioblastoma multiforme, the most common primary brain tumor, and the median survival (irrespective of therapy) is one year. The multifactorial nature of this disease and limited scope of previous studies has resulted in an incomplete understanding of the etiology of glioblastoma survival. The main goal of this application is to augment the knowledge on the prognostic markers and determinants of glioblastoma survival. The conceptual framework encompasses the simultaneously consideration clinical and genomic information, use of a large sample size including 397 glioblastoma and matched control samples, and a comprehensive analysis of the relationship between glioblastoma survival and the explanatory variables using frailty models. Specific Aim 1 centers on the detection of exon expression and clinical prognostic markers of glioblastoma survival meanwhile, Specific Aim 2 entails the uncovering of microRNA expression markers together with exon expression and clinical markers. Exon and microRNA-centric and gene-centric approaches to model the genomic information will be pursued. To address the high-dimensionality of the proposed study, a combination of feature extraction strategies will be implemented. A mixed-effects proportional hazards model that accommodates for heterogeneity of variances and correlation between observations will be used to model glioblastoma survival. The posterior density estimates obtained from a Bayesian Markov chain Monte Carlo approach will offer a complete characterization of the associations between the prognostic markers and survival. Calibration of the survival predictive equation will use cross-validation approaches and benchmarking to known markers of glioblastoma survival. This application addresses several high-priority areas in cancer epidemiology research that have been identified by the NCI and our group has prior experience on the foundations of the proposed study. The outcomes of this study will be 1) the identification of clinical and genomic prognostic markers of glioblastoma survival acting independently, synergistically or antagonistically, 2) the characterization of changes in survival associated with these markers, 3) development of a more universal prediction or prognostic index of glioblastoma survival, 4) identification of glioblastoma subclasses corresponding to particular combinations of clinical and genomic levels, and 5) understanding of the interplay between a wide range of clinical and genomic factors with the ultimate goal of elucidating the etiology of cancer. The results from the proposed multi-factorial analysis of glioblastoma genomic and clinical data will support the effective design of functional molecular studies of cancer survival, facilitate the exchange of discoveries between the basic sciences and clinical and public health practice and lead to biomedical strategies to diagnose, prevent, treat and cure cancer. PUBLIC HEALTH RELEVANCE: Project Narrative This project will uncover clinical and genomic prognostic markers and determinants of glioblastoma survival. The outcomes of this study include the enhanced understanding of the interplay between multiple clinical and genomic variables with the ultimate goals of elucidating the etiology of cancer and development of a reliable survival prediction function. These findings will facilitate functional molecular studies, the exchange of discoveries between the basic sciences and clinical and public health practice and the development of biomedical strategies to diagnose, prevent, treat and cure cancer.