Pancreatic Cancer Risk Prediction: Integrating Individual-Level Clinical and Genetic Data

胰腺导管腺癌(PDAC)是一种高致命性的肿瘤(5年生存率为10%)。 部分原因是大多数病例是在癌症无法手术的晚期被诊断出来的 医疗治疗是有限的。特别是，在黑人中发现的PDAC是在后期阶段发现的 具有较高的发病率和死亡率。到目前为止，PDAC筛查主要针对有以下症状的个人 遗传性倾向，忽略了90%是零星的。普及型PDAC筛查 没有症状的成年人被认为是不可行的，这促使了对新筛查的探索 针对高危人群的方法。对PDAC的易感性由基因决定 因素和临床暴露，但目前用于识别这些个人的风险预测工具是 不充分，因为它们没有纳入来自跨祖先群体的基因数据，并且没有 全面整合临床和遗传风险因素。为了解决这个问题，我们建议 通过利用来自当地、国家和国际的数据来改进对PDAC风险的预测 生物库/数据集，包括退伍军人企业数据仓库、退伍军人百万计划、 宾夕法尼亚医学生物库、英国生物库和胰腺癌联盟。我们假设 改进的基因型别和表型数据的结合将改进高危人群的识别 退伍军人使用PDAC而不是传统的基于风险因素的方法。我们的具体目标是1) 基于电子病历的PDAC表型算法的研究 非结构化数据，2)通过以下方式识别普通人群中的退伍军人 将遗传学整合到临床预测模型中，以及3)在全基因组范围内执行跨祖先 百万退伍军人计划中PDAC的关联研究(GWAS)，与现有的Meta分析 PDAC数据，以及第一个针对非洲血统个人的全球气候变化数据。成功完成这项工作 该项目不仅将使早期发现PDAC成为可能，而且在多个血统中普遍存在 退伍军人群体，还能提高遗传学的预测能力，最终帮助 弥合黑人在PDAC相关发病率和死亡率方面的差异。这些结果将 作为在PDAC和帮助中制定个性化筛查策略的基础 在PDAC中实现精准医学提高生存率的承诺。这份提案详细说明 年促进王露易斯博士作为内科科学家独立职业生涯的五年计划 胰腺癌早期发现的遗传风险预测。王医生是三年级的学生 胃肠病学研究员，将在毕业前完成临床流行病学的正式培训 在资助期内，她将继续接受遗传学和生物信息学方面的培训，以增强她的技术 精准医学方面的技能。她的培训将通过全面的指导来促进 计划包括以下内容：1)每周至每月与她的指导团队举行会议；2)正式 可视化生物医学和临床数据的程序设计课程，统计方法 分析遗传数据，高级预测建模，将遗传学整合到临床 通过临床流行病学和生物统计中心(CCEB)、基因组学和 计算生物学(GCB)研究生组和生物统计、流行病学和 宾夕法尼亚大学信息学(DBEI)，3)结构化研究研讨会和国家 会议，以及4)在授权期的后期发展未来的赠款。 她的长期目标是通过利用 电子健康记录链接的生物库，整合遗传和临床风险因素。

Pancreatic ductal adenocarcinoma (PDAC) is highly deadly (5-year survival rates <10%), in large part because most cases are diagnosed at advanced stages when the cancer is inoperable and medical therapies are limited. In particular, PDAC among Blacks are discovered at later stages with higher incidence and mortality. To date, PDAC screening has focused on individuals with hereditary predispositions, neglecting the 90% that is sporadic. Universal PDAC screening in asymptomatic adults has been deemed infeasible, motivating the quest for new screening approaches that target high risk individuals. Susceptibility to PDAC is determined by genetic factors and clinical exposures, but current risk prediction tools to identify these individuals are inadequate as they do not incorporate genetic data from a trans-ancestry population and fail to comprehensively integrate clinical and genetic risk factors. To address this problem, we propose to improve the prediction of PDAC risk by leveraging data from local, national, and international biobanks/datasets, including the VA Corporate Data Warehouse, VA Million Veteran Program, Penn Medicine BioBank, UK Biobank, and pancreatic cancer consortia. We hypothesize that the combination of refined genotypic and phenotypic data will improve identification of high-risk Veterans for PDAC over traditional risk factor-based approaches. Our specific aims are 1) develop an electronic health record-based phenotype algorithm for PDAC using structured and unstructured data, 2) identify Veterans in the general population at high risk for PDAC by integrating genetics into a clinical prediction model and 3) perform a trans-ancestry genome-wide association study (GWAS) for PDAC in the Million Veteran Program, meta-analysis with extant PDAC data, and the first GWAS for individuals of African ancestry. Successful completion of this project will not only make early detection of PDAC possible in the multi-ancestry general population of Veterans, but also improve the predictive capability of genetics to ultimately help bridge the disparities in PDAC-related morbidity and mortality seen in Blacks. These results will serve as the foundation for developing personalized strategies for screening in PDAC and help actualize the promise of precision medicine for improving survival in PDAC. This proposal details a 5-year plan to promote the independent career of Dr. Louise Wang as a physician scientist in genetic risk prediction for early detection of pancreatic cancer. Dr. Wang is a third-year gastroenterology fellow who will complete formal training in clinical epidemiology prior to the funding period and will continue training in genetics and bioinformatics to augment her technical skills in precision medicine. Her training will be facilitated through a comprehensive mentorship plan consisting of the following: 1) weekly to monthly meetings with her mentorship team, 2) formal coursework in programming to visualize biomedical and clinical data, statistical methods for analyzing genetic data, and advanced prediction modeling to integrate genetics into clinical prediction through the Center for Clinical Epidemiology and Biostatistics (CCEB), Genomics & Computational Biology (GCB) graduate group, and Department of Biostatistics, Epidemiology and Informatics (DBEI) at the University of Pennsylvania, 3) structured research seminars and national conferences, and 4) development of future grants during the latter portion of the award period. Her long-term goal is to identify Veterans at risk for gastrointestinal malignancies by leveraging electronic health record-linked biobanks to integrate genetic and clinical risk factors.