New Risk Assessment Paradigm to Predict Screening Detection, Failures and False Alarms

项目摘要--项目1 2016年美国预防服务工作组和2015年美国癌症协会指南呼吁降低筛查强度，并要求提供者与女性讨论她们对筛查的偏好。这种对个性化决策的推动引发了基于风险的筛查利用的实质性变化，这种筛查利用不仅仅依赖于女性的年龄来决定何时开始和停止筛查或筛查频率。然而，将风险评估明确纳入普通人群筛查指南还处于早期阶段。风险预测是基于风险的筛查的一个关键方面；现有的模型预测乳腺癌的总体风险，但没有考虑乳腺癌生物学的异质性或筛查方式检测乳腺癌的能力。项目1旨在推进一种新的基于风险的筛查范例，在明确考虑筛查间隔和使用的方式的同时，确定妇女筛查检测、失败和错误警报的绝对累积风险。 目的1)利用现有和扩大的乳腺癌监测联盟(BCSC)基础设施，项目调查人员将开发新的风险预测模型，以确定以下高危妇女：1)早期筛查发现的癌症(I/IIa期浸润性癌)，2)筛查失败(间期浸润性癌或筛查发现的IIb期或更高)，以及3)假警报(假阳性测试和良性活检)。利用BCSC关于100,000多名40-79岁接受数字乳房X光检查或断层合成的女性以及13,000多名已确诊的浸润性乳腺癌的数据，项目组将确定与影响筛查结果的乳腺癌风险相关的临床和成像因素，为每年、两年和三年筛查方案的每个筛查结果生成六年风险估计。 目的2)项目团队将使用妇女和临床医生调查和Delphi小组来确定筛查失败和错误警报的临床意义风险的可操作水平，以及相关的长期筛查结果。 这项建议将回答：1)哪些妇女筛查失败的风险很高，需要两年一次的乳房X光检查的替代策略，哪些妇女风险较低，可以较少进行筛查？2)筛查结果不佳的风险水平如何影响妇女和临床医生对筛查策略的偏好？ 该项目将提供证据，指导妇女、医疗保健提供者和政策制定者根据筛查失败和错误警报的短期风险制定筛查战略，以期在最大限度减少危害的同时改善对侵袭性肿瘤的早期发现。

PROJECT SUMMARY – Project 1 The 2016 US Preventive Services Task Force and 2015 American Cancer Society guidelines call for reduced screening intensity and for providers to discuss with women their preferences for screening. This push for individualized decision-making has triggered substantial movement toward risk-based screening utilization that does not rely solely on a woman's age to determine when to start and stop screening or screening frequency. Yet, the explicit incorporation of risk assessment into screening guidelines for the general population is in its early stages. Risk prediction is a key aspect of risk-based screening; available models predict breast cancer risk overall, but do not take into account the heterogeneity of breast cancer biology or the ability of screening modalities to detect breast cancer. Project 1 aims to advance a new risk-based screening paradigm that identifies women's absolute cumulative risk of screening detection, failures, and false alarms while explicitly considering screening interval and modality used. Aim 1) Using the existing and expanded Breast Cancer Surveillance Consortium (BCSC) infrastructure, project investigators will develop new risk prediction models to identify women at high risk of: 1) early-stage screen-detected cancer (stage I/IIa invasive cancer), 2) screening failure (interval invasive cancer or screen-detected stage IIb or higher), and 3) false alarms (false-positive tests and benign biopsies). Using BCSC data on over 1,000,000 women aged 40-79 years undergoing digital mammography or tomosynthesis, and over 13,000 invasive breast cancers identified, the project team will identify clinical and imaging factors associated with breast cancer risk that impact screening outcomes, generating six-year risk estimates for each screening outcome for annual, biennial, and triennial screening regimens. Aim 2) The project team will identify actionable levels of clinically-meaningful risk of screening failure and false alarm, and associated long-term screening outcomes using women and clinician surveys and Delphi panels. This proposal will answer: 1) Which women are at high risk of screening failures and require an alternative strategy to biennial mammography, and which women are at low risk and can be screened less often? 2) What risk levels of poor screening outcomes impact women and clinician preferences for screening strategies? This project will provide evidence to guide women, health care providers, and policymakers on screening strategies based on short-term risk of screening failures and false alarms with the goal to improve early detection of aggressive tumors while minimizing harms.