STRATEGIC STUDIES IN BREAST CANCER SURVEILLANCE

这项建议的长期目标是增强新汉普郡乳房X光摄影网络的研究能力，方法是扩大数据收集范围，以确定发生乳房X光检查异常且随后的乳房结果未在我们的登记中记录的妇女的结果，确定经历假阴性乳房X光检查的妇女获得后续护理的动机，以及收集或合并补充数据以使我们能够进行五个特别研究项目。这些专题研究将包括：1)比较筛查和未筛查的NH妇女的风险和行为特征；2)确定乳腺良性病变特征与乳房X光摄影表现之间的关系；3)探讨月经周期对乳腺密度和乳腺癌筛查性能的影响；4)建立浸润性和非浸润性乳腺癌的风险预测模型(包括将乳腺密度作为乳腺癌风险的标志)；以及5)使用回归建模技术建立由筛查依从性、乳房X光检查结果、随访和单个受试者的疾病结果定义的乳房X光检查/健康状态的纵向模型，并确定这些状态之间转换的预测因素，包括对所有原因死亡率和乳腺癌死亡率的乳房X光检查相关预测因素的评估。我们建议的项目将在两个方面有利于乳腺癌监测联盟的工作。首先，确定不充分后续行动的来源(无论是患者因素还是数据收集问题)将确定基于人群的乳房X光检查登记的捕获率，并可能为那些发现异常但拒绝寻求治疗的人确定可能的干预点。其次，我们计划合并来自三个注册处(新罕布夏州、佛蒙特州和北卡罗来纳州)的数据，以实现我们的三个研究目标。通过汇集这三个登记处的数据，用于分析的癌症病例数量将会增加，数据中所代表的人群的特征也将更加多样化。我们的具体目标可以很容易地通过来自三个拟议站点的数据来解决；然而，任何验证研究都可以将这项拟议工作中的成熟方法应用于未来从该联盟获得的更大的数据集。我们的目标将使我们能够利用当地的分析专业知识，同时为监视联盟提供新的知识。

The long term objective of this proposal is to enhance the New Hampshire Mammography Network's research capacity by expanding data collection to include determining outcomes in women who experienced mammographic abnormalities and whose subsequent breast outcomes are not recorded in our registry, determining what motivated women who experienced a false negative mammogram to obtain follow-up care, and collecting or merging supplemental data to allow us to conduct five special research projects. These special studies will include: 1) comparing risk and behavioral characteristics of screened and unscreened NH women; 2) determining the relationships between benign breast pathology characteristics and mammography performance; 3) exploring the influence of menstrual cycle phase on breast density and breast cancer screening performance; 4) developing risk prediction models for invasive and non-invasive breast cancer (including using breast density as a marker for breast cancer risk); and 5) using regression modeling techniques to develop a longitudinal model of mammography/health states defined by screening compliance, mammography outcomes, follow-up and disease outcomes in individual subjects and to determine predictors for transitions between these states, including an evaluation of mammography-related predictors for all cause mortality and breast-cancer mortality. Our proposed projects will benefit the work of the Breast Cancer Surveillance Consortium in two ways. First, determining sources of inadequate follow-up (whether it is a patient factor or data collection issue) will identify the capture rate of population-based mammography registries and may identify possible intervention points for those with abnormal findings who are resistant to seeking care. Second, we plan to merge data from three registries (New Hampshire, Vermont, and North Carolina) to conduct three of our study aims. By pooling data from these three registries, the number of cancer cases for analysis will be increased and the characteristics of the population represented in the data will be more diverse. Our specific aims can easily be addressed with data from the three proposed sites; however, any validation studies could apply well established methods from this proposed work to a larger dataset accessed from the Consortium in the future. Our aims will allow us to employ local analytic expertise while contributing new knowledge to the Surveillance Consortium.