STUDY OF EXPOSURES, BEHAVIORS, AND BIOMARKERS IN CANCER EPIDEMIOLOGY

描述（由申请人提供）：现在人们普遍认为环境暴露和遗传因素都是导致大多数疾病的原因。癌症流行病学研究通常通过调查工具（例如，问卷调查）确定暴露和行为因素（一般称为暴露因素），并通常检查候选基因或途径中的暴露变异或血液暴露水平或暴露标记（生物标记），以确定疾病或疾病前体的风险因素。问卷项目的回答可能是相关的。同样，生物标志物可能由于共同的潜在生物学功能或连锁不平衡而相关，并且可能以低频率发生。这使得数据分析成为一项具有挑战性的任务。尽管已经开发了分析方法来应对这一挑战，但仍然需要研究有效分析暴露因素和生物标志物的新方法，并为后续研究开发最佳研究设计。本建议的长期目标是通过以下四个具体目标来满足这些需求。在目标1中，我们将研究如何对相关暴露因素和生物标志物进行有效分析。在目标2中，我们将研究如何以有效的方式评估暴露途径和潜在生物学特性的影响。由于多种原因，包括基因暴露相互作用，危险因素的病因效应可能是复杂的。目标3将检查用于评估相互作用的简约分析方法。在Aim 4中，我们将开发一种新的研究设计，当由于后勤或成本考虑而无法对整个队列进行评估时，从已建立的队列中抽取信息丰富的亚人群来评估额外的风险因素。我们的提议是由几项合作研究推动的。我们将进行模拟来研究在每个目标下提出的方法的特性，并得出它们在实际环境中使用的指导方针。我们将使用合作研究的数据对方法和指南进行实证评估。我们的建议有两个主要目标。首先，我们打算发展和改进评估癌症流行病学研究中相关危险因素的研究设计和方法。其次，我们打算通过使用来自各种研究的数据进行实证调查来评估这些方法的实际相关性。我们预计，从这些调查中获得的见解将对我们对分析和研究设计的理解产生重大影响，以有效和节省的方式确定与公共卫生相关的风险因素。公共卫生相关性：人们普遍认为，环境暴露和遗传因素都会导致大多数疾病。癌症流行病学研究通常通过调查工具（例如，问卷调查）确定暴露和行为因素，并检查候选基因或血液暴露水平的易感性变异或暴露标记，以确定疾病特征的危险因素。问卷项目的回答可能是相关的。同样，遗传变异可能由于其共同的潜在生物学功能而表现出相关性，并且可能以低频率发生。这使得数据分析成为一项具有挑战性的任务。本提案的长期目标是探索有效分析相关风险因素的新方法，并为后续研究制定最佳研究设计。

DESCRIPTION (provided by applicant): It is now widely accepted that both environmental exposures and genetic factors contribute to most diseases. Cancer epidemiology studies often ascertain exposures and behavioral factors (referred to generally as exposure factors) through survey instruments (for example, questionnaires), and typically examine variants or blood levels of exposures or markers of exposures (biomarkers) in candidate genes or pathways to identify risk factors for disease or for disease precursors. The responses to questionnaire items may be correlated. Similarly, the biomarkers may be correlated due to common underlying biological function or linkage disequilibrium and may occur at a low frequency. These render data analysis a challenging task. Although analytic methods have been developed to address this challenge, there continues to be a need to investigate novel approaches for efficient analyses of exposure factors and biomarkers, and to develop optimal study designs for follow-up research. The long-term goals of this proposal are to address these needs through the following four specific aims. In Aim 1 we will examine how to conduct efficient analysis of correlated exposure factors and biomarkers. In Aim 2 we will examine how to evaluate the effects of routes of exposures and underlying biological characteristics in an efficient manner. The etiologic effects of the risk factors are likely to be complex due to a number of reasons, including gene-exposure interactions. Aim 3 will examine parsimonious analytic methods for evaluating interactions. In Aim 4 we will develop a novel study design for sampling an informative sub-population from an established cohort to evaluate additional risk factors, when evaluation of the full cohort is not feasible due to logistic or cost considerations. Our proposal is motivated by several collaborative studies. We will conduct simulations to investigate the properties of the methods proposed under each aim, and derive guidelines for their use in practical settings. We will conduct an empirical evaluation of the methods and guidelines using data from the collaborative studies. Our proposal has two major objectives. First, we intend to develop and improve research designs and methodologies for evaluating correlated risk factors in cancer epidemiology studies. Second, we intend to assess the practical relevance of these methods by conducting empirical investigations using data from various studies. We anticipate that the insights gained from these investigations will have a significant impact on our understanding of analysis and study designs for identifying risk factors of public health relevance in an efficient and parsimonious manner. PUBLIC HEALTH RELEVANCE: It is widely accepted that both environmental exposures and genetic factors contribute to most diseases. Cancer epidemiology studies often ascertain exposures and behavioral factors through survey instruments (for example, questionnaires), and examine susceptibility variants in candidate genes or blood levels of exposures or markers of exposures to identify risk factors for disease traits. The responses to questionnaire items may be correlated. Similarly, the genetic variants may exhibit correlation due to their common underlying biological function and may occur at a low frequency. These render data analysis a challenging task. The long-term goals of this proposal are to investigate novel methods for efficient analyses of correlated risk factors, and to develop optimal study designs for follow-up research.