Novel Methods for Missing Subtype Data in Colorectal Cancer

项目总结/摘要 该项目是对PAR-18-021，NCI癌症研究小额赠款计划（NCI）的回应 Omnibus R03）。我们的动力来自于大肠癌遗传学和流行病学方面的资源 财团（GECCO）。我们主要对开发和应用创新的统计方法感兴趣 寻找缺失的癌症亚型数据许多疾病，如结直肠癌（CRC），是异质性的。 肿瘤的分子特征提供了多种肿瘤亚型的证据，这些亚型通过以下途径发展： 激活不同的肿瘤通路。重要的CRC肿瘤亚型包括微卫星不稳定性 (MSI)状态、BRAF和KRAS中的体细胞突变以及CpG岛甲基化表型。比如说， MSI状态与生存结局和治疗反应相关。然而，有些人可能 未知MSI状态和其他肿瘤生物标志物。回归分析可能会遇到挑战 由于研究队列中某些个体的数据缺失。经常需要缺失数据的方法 以解决效果估计和效率中的偏差问题。本提案的具体目标包括： (i)开发和应用方法以考虑多项logistic中缺失的癌症亚型数据 回归分析(ii)开发和应用方法，以调整癌症病例的生存分析，其中 多个肿瘤生物标志物可能缺失。建议中制定的方法适用于 GECCO和其他研究，其中某些研究个体的癌症亚型数据可能未知。的 方法可以应用于其他常见的研究设计，如巢式病例对照设计， 竞争风险的考克斯回归。

Project Summary/Abstract The proposed project is in response to PAR-18-021, NCI Small Grants Program for Cancer Research (NCI Omnibus R03). We are motivated by the resources in the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO). We are primarily interested in developing and applying innovative statistical methods for missing cancer subtype data. Many diseases, such as colorectal cancer (CRC), are heterogeneous. Molecular characterization of tumors has provided evidence of multiple tumor subtypes that develop through activation of diverse neoplastic pathways. Important CRC tumor subtypes include microsatellite instability (MSI) status, somatic mutations in BRAF and KRAS, and CpG island methylator phenotype. For instance, MSI status is associated with survival outcomes and treatment response. However, some individuals may have unknown MSI status, and other tumor biomarkers. Regression analysis may encounter a challenge due to missing data in some individuals of the study cohort. Methodology for missing data is often required to address the issue on bias in effect estimation and efficiency. Specific aims of this proposal include: (i) To develop and apply methods to take into account missing cancer subtype data in multinomial logistic regression. (ii) To develop and apply methods to adjust for survival analysis among cancer cases in which multiple tumor biomarkers may be missing. The methods developed in the proposal are applicable to the GECCO and other studies where cancer subtype data may be unknown among some study individuals. The methods can be applied to other common study designs such as nested case-control designs and Cox regression with competing risks.