Despite years of research, optimal breast cancer screening strategies remain elusive, especially for women between the age of 40 and 49. Academic societies and agencies differ in their recommendations regarding the age to begin mammography and the screening intervals. One potential solution is risk-adapted screening, where decisions around the starting age, stopping age, frequency, and modality of screening are based on individual risk to maximize the early detection of aggressive cancers and minimize the harms of unnecessary screening. In addition to demographics, family history, breast density, and other risk factors, single nucleotide polymorphism (SNP) profiling of germ line DNA has been incorporated into breast cancer prediction models that can further guide our clinical recommendations for screening. Of relevance to every woman, the ~100 low penetrant single nucleotide polymorphism (SNP) confers a small risk of breast cancer development but affects many women due to the high risk allele frequency. High to median penetrant mutations of cancer susceptible genes, such as BRCA and Lynch syndrome genes, are associated with a higher risk of breast cancer development but affect only a minority of women who are carriers. Women Veterans in the Million Veteran Program (MVP) represent a cohort of women for whom comprehensive genetic information and clinical covariates have been obtained, providing an exceptional opportunity to develop, optimize and/or validate a risk adapted breast cancer screening strategy. Women predicted to have an elevated risk of developing breast cancer by prediction models may benefit from screening beginning at a younger age and more frequent breast imaging including the incorporation of breast MRI. Women predicted to have a low(er) risk for breast cancer may do well with less intense screening. Because women Veterans in MVP may have unique military and environmental exposures, it is unknown whether previously developed breast cancer risk prediction models can be applied to this population. Moreover, since 28% of women Veterans in the current MVP cohort are of African American descent, while the genetic markers that contribute to the construction of genetic prediction models are developed from studies involving Caucasians, it is not clear if these instruments can be applied to women that are of diverse ethnic backgrounds. Our study will determine if breast cancer prediction models built on currently available SNPs can be validated in women Veterans in the MVP. Moreover, we will determine whether mutant alleles of cancer susceptibility genes with median to high penetrance will confer the same (breast) cancer risks as previously established. A higher cancer risk incurred by mutation in these cancer susceptible genes may make universal testing cost effective, which can further facilitate and motivate the adoption of genetic profiling to build breast cancer prediction models for every woman. We propose to build breast cancer risk prediction models in this two-year pilot project with the ultimate goal to apply and validate these models in the entire MVP population. Our work, focusing on Veteran women, together with and complemented by a prospective trial that is being launched will greatly enhance our ability to optimize breast cancer screening in a personalized manner. In sum, we will build a molecularly full characterized women Veteran cohort in the MVP that we can continue to follow longitudinally. We will focus on building and validating breast cancer risk prediction models with the potential to extend to other cancer or disease types. Our work will significantly enhance our abilities for early detection and optimize and individualize breast cancer screening for all women Veterans and women in general.

Breast cancer is a leading cancer problem for women veterans. Population wide breast cancer screening with mammography has witnessed a decrease in breast cancer death. Recent efforts on breast cancer screening has focused on, instead of solely relying on age, finding genetic and clinical markers that can accurately assign breast cancer risk to individual women. This will allow a more personalized and risk- adapted screening strategy. Women felt to be at a higher risk for breast cancer may need additional imaging such as MRI. Breast cancer risk prediction models have been built primarily based on studies from civilian, Caucasian women. It is not clear if military experience or combat exposure would significantly modify women's risk for breast cancer. Nor is it clear if the prediction models will perform well in African American women who now take up more than 25% of the women veteran population. Leveraging the wealth of information now already captured and the ability to monitor thousands of women in the MVP prospectively, we seek to answer this important question.