Evaluating a Risk Prediction Model for Lung Cancer

点击翻译按钮获取中文摘要

 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.