Prediction of interval breast cancer risk using genetic variants

项目概要/摘要 这是加州大学的内科医生Yiwey Shieh博士的K08奖申请书， 旧金山弗朗西斯科（UCSF）。Shieh博士的长期职业目标是成为一名领先的临床研究者， 精准医疗和癌症筛查/预防的交叉点。K08奖将提供博士。 在谢先生的支持下，实现了以下培训目标：1）培养他对 乳腺癌异质性和基因表达谱的使用来描述肿瘤; 2）加深他的熟悉程度 与遗传关联技术; 3）学习和实施预测建模中的先进技术; 4） 熟练操作和分析高维数据集; 5）建立一个独立的重新 搜索职业为了帮助他实现这些目标，Shieh博士组建了一个模范导师团队， 补充专业知识。Elad Ziv博士，一位领先的癌症遗传流行病学家，将担任主要的男性- 监督研究目标和培训目标的完成。共同导师将包括：劳拉范特维尔博士， 乳腺癌基因表达分析专家; Mi-Ok Kim博士，生物统计方法专家， 研究设计;和Laura Esserman博士，乳腺外科医生和精准肿瘤学的领先创新者。的 拟议的工作将受益于现有的专业知识和丰富的制度环境的加州大学旧金山分校，特别是 海伦·迪勒家庭综合癌症中心，全国最好的癌症中心之一。谢博士 将利用上述资源解决癌症控制的一个重要问题，即能力有限， 目前的乳腺癌筛查策略，以检测生物侵袭性癌症。这些癌症往往 在正常的乳房X光检查后，通常表现为“间隔癌”，并且通常处于晚期。 博士Shieh过去的工作主要集中在使用临床风险模型和多聚乙烯预测整体乳腺癌风险。 代表多个单核苷酸多态性联合效应的基因风险评分（PRS） （SNP）。Shieh博士建议采用这种方法来构建一个基于预处理的间隔癌症风险模型， 初步数据表明，先前发现的SNPs可能对乳腺癌有额外的预测作用， 表型。鉴于基因表达谱能够提供详细的肿瘤表型信息，Dr。 Shieh建议识别与快速发展（攻击性）间隔相关的基因表达特征， 癌症（目标1）。然后，Shieh博士将使用这些基因表达特征来选择与ag相关的SNP， gressiveness（目标2）。这些选择的SNP将用于构建侵袭性癌症的PRS。最后，博士。 Shieh将测试这种PRS预测间隔期癌症的能力，无论是单独还是与乳腺癌联合， 临床危险因素（Aim 3）。拟议的目标将利用Shieh博士对数据的独特访问， 临床试验、肿瘤图谱和乳腺癌筛查队列。其结果将直接导致后续 关于使用间隔癌症风险模型前瞻性地为筛查和预防提供信息的研究，以及 研究新型生物标志物在改善高危女性风险分层中的作用。

PROJECT SUMMARY/ABSTRACT This is an application for a K08 award for Dr. Yiwey Shieh, a general internist at the University of California, San Francisco (UCSF). Dr. Shieh's long-term career goal is to become a leading clinician-investigator working at the intersection of precision medicine and cancer screening/prevention. This K08 award will provide Dr. Shieh with the support necessary to achieve the following training goals: 1) develop his understanding of breast cancer heterogeneity and the use of gene expression profiling to profile tumors; 2) deepen his familiarity with genetic association techniques; 3) learn and implement advanced techniques in prediction modeling; 4) become proficient in manipulating and analyzing high-dimensional datasets; and 5) build an independent re- search career. To help him achieve these goals, Dr. Shieh has assembled an exemplary team of mentors with complementary expertise. Dr. Elad Ziv, a leading cancer genetic epidemiologist, will serve as the primary men- tor and oversee completion of research aims and training goals. Co-mentors will include: Dr. Laura van't Veer, an expert in gene expression analysis of breast cancer; Dr. Mi-Ok Kim, an expert in biostatistical methods and study design; and Dr. Laura Esserman, a breast surgeon and leading innovator in precision oncology. The proposed work will benefit from the available expertise and rich institutional environment of UCSF, specifically the Helen Diller Family Comprehensive Cancer Center, one of the top cancer centers in the country. Dr. Shieh will leverage the above resources to address an important problem in cancer control, namely the limited ability of current breast cancer screening strategies to detect biologically aggressive cancers. These cancers tend to present symptomatically following a normal mammogram as “interval cancers”, and often at advanced stages. Dr. Shieh's past work has focused on predicting overall breast cancer risk using clinical risk models and poly- genic risk scores (PRS) representing the combined effects of multiple single nucleotide polymorphisms (SNPs). Dr. Shieh proposes to adopt this approach to construct a risk model for interval cancers based on pre- liminary data suggesting that previously discovered SNPs may hold additional prediction for breast cancer phenotype. Given the ability of gene expression profiling to provide detailed tumor phenotypic information, Dr. Shieh proposes to identify gene expression signatures correlated with rapidly-developing (aggressive) interval cancers (Aim 1). Dr. Shieh will then use these gene expression signatures to select SNPs associated with ag- gressiveness (Aim 2). These selected SNPs will be used to construct a PRS for aggressive cancer. Finally, Dr. Shieh will test the ability of this PRS to predict interval cancers, both alone and in combination with breast den- sity and clinical risk factors (Aim 3). The proposed aims will leverage Dr. Shieh's unique access to data from clinical trials, tumor atlases, and breast cancer screening cohorts. The results will directly lead to subsequent studies on using an interval cancer risk model to prospectively inform screening and prevention, as well as studies on the role of novel biomarkers in refining risk stratification among high-risk women.