Evaluating a Risk Prediction Model for Lung Cancer

 描述(由申请者提供)：这个职业发展奖将支持Lori Sakoda博士向独立的肺癌翻译研究人员过渡。她的长期职业目标是通过领导整合复杂、大规模生物医学数据分析的跨学科研究，为肺癌检测和控制提供信息并改进现实世界的战略。风险预测模型可用于优化吸烟者肺癌筛查策略的益害比。然而，为了支持它们在临床实践中的使用，必须有令人信服的证据表明它们的预测能力，以识别患肺癌风险最高的吸烟者和/或区分那些出现恶性和良性肺结节的人。她的指导研究将评估一种新开发的、面向临床的肺癌风险预测模型，无论是建议还是修改，是否可以在肺癌筛查的背景下帮助决策。具体目的是：1)验证模型的预测性能；2)确定将遗传和其他临床预测因子添加到模型中的增量价值；3)检查基线模型和最佳预测扩展模型在符合美国预防服务工作组低剂量计算机断层扫描肺癌筛查资格标准的人群中的预测性能。作为一项探索性目标，这两个模型的预测性能将在偶然诊断为肺结节的筛查合格人群中进行评估。这些目标将通过在凯撒永久北加州(KPNC)基因、环境和健康研究计划中整合大量当代吸烟者的调查、全基因组基因分型和电子健康记录(EHR)数据来实现。该建议建立在候选人之前接受的癌症流行病学培训的基础上，以填补临床领域(肺部病理生理学、肺癌检测和治疗实践以及医疗决策)和与EHR和其他复杂、大规模数据(生物统计学、遗传流行病学和生物医学信息学)的综合分析相关的科学领域的知识空白，这将使她能够更有效地生成科学证据并将其转化为临床实践。培训将在一支高素质的导师和科学顾问团队的指导下，通过课程作业、研讨会、专业学会会议和体验式学习获得。她还将建立临床和科学合作伙伴关系，这对她在目前的环境中取得成功至关重要。KPNC研究部是一个理想的培训环境，因为它在癌症筛查指南方面做出了重要贡献的悠久历史，因为它在科学上的领导地位，以及它可以接触到种族多样化和稳定的成员(目前超过300万成年人)，这些成员的电子健康记录数据被无限期保留。拟议的计划将为候选人提供初步数据，以制定具有竞争力的R01计划，以及成功建立专注于优化肺癌检测和控制策略的跨学科研究计划的专业知识和技能。

 DESCRIPTION (provided by applicant): This Career Development Award will support Lori Sakoda, PhD, in her transition to independence as a translational researcher in lung cancer. Her long-term career goal is to inform and improve real-world strategies for lung cancer detection and control by leading transdisciplinary research that integrates analysis of complex, large-scale biomedical data. Risk prediction models could be valuably employed to optimize the benefit-to-harm ratio of screening strategies for lung cancer in smokers. To support their use in clinical practice, however, there must be convincing evidence of their predictive ability to identify smokers at highest risk for lung cancer and/or to differentiate those presenting with malignant versus benign lung nodules. Her mentored research will evaluate whether a newly developed, clinically-oriented risk prediction model for lung cancer, as proposed or modified, could aid decision-making in the context of lung cancer screening. The specific aims are to 1) validate the predictive performance of the model; 2) determine the incremental value of adding genetic and other clinical predictors to the model; and 3) examine the predictive performance of the baseline model and the best predictive extended model in persons who meet the U.S. Preventative Services Task Force eligibility criteria for lung cancer screening with low-dose computed tomography. As an exploratory aim, the predictive performance of these same two models will be assessed in the subgroup of screening- eligible persons diagnosed incidentally with lung nodules. These aims will be addressed by integrating survey, whole genome genotyping, and electronic health record (EHR) data on a large, contemporary cohort of smokers in the Kaiser Permanente Northern California (KPNC) Research Program on Genes, Environment, and Health. The proposal builds on the candidate's prior training in cancer epidemiology to fill knowledge gaps in clinical domains (lung pathophysiology, lung cancer detection and management practices, and medical decision-making) and scientific domains pertinent to integrated analysis of EHR and other complex, large-scale data (biostatistics, genetic epidemiology, and biomedical informatics), which will allow her to more effectively generate and translate scientific evidence into clinical practice. Training will be acquired from coursework, seminars, professional society meetings, and experiential learning, under the guidance of a highly qualified team of mentors and scientific advisors. She will also build clinical and scientifc partnerships essential to succeed in her current setting. The KPNC Division of Research is an ideal training environment, given its long history of important contributions to cancer screening guidelines, due to both its scientific leadership and its access to an ethnically diverse and stabl membership (currently over three million adults) for whom EHR data are kept indefinitely. The proposed plan will provide the candidate with preliminary data to develop a competitive R01 proposal, along with specialized knowledge and skills to successfully establish a transdisciplinary research program focused on optimizing strategies for lung cancer detection and control.