PROJECT 2: Summary/Abstract The ongoing WISDOM Study of risk-based versus annual breast cancer screening is one of the first-ever applications of population-wide genetic testing to inform breast cancer risk. Participants in the risk-based arm receive testing for rare pathogenic variants and common variants (e.g., single nucleotide polymorphisms, or SNPs). The latter are used to calculate polygenic risk scores (PRS), which are used in combination with clinical risk models to inform screening and prevention recommendations. WISDOM is therefore an ideal setting to critically evaluate and accelerate the translation of PRS. Through our firsthand experience implementing PRS- based risk assessment in WISDOM, we have identified several key areas on which current PRSs need to improve: 1) their performance in non-European populations; 2) their ability to predict faster-growing cancers; and 3) their ability to identify women who may benefit from intensive screening. The overall goal of Project 2 is to advance breast cancer risk prediction by addressing these three gaps. In doing so, we aim to enhance the generalizability and clinical utility of breast cancer PRS. We will address the following Specific Aims: 1) Improve the cross-ancestry performance of breast cancer PRS; 2) Develop and validate a PRS for fast-growing breast cancers; and 3) Determine the risk of advanced cancer with annual versus biennial screening across levels of PRS. To bolster the impact and generalizability of our work, we will leverage our unique access to large, diverse datasets from publicly available resources, industry, and institutional cohorts. We have assembled an experienced, highly accomplished research team of experts in genetic epidemiology, biostatistical modeling, and breast cancer risk assessment and control. This project will eventually yield improved PRS’s for overall (Aim 1) and fast-growing breast cancer risk (Aim 2) to be implemented in the WISDOM Study and evaluated in tandem with clinical and mammographic artificial intelligence-based risk scores in Project 3. Together, this work will accelerate the evolution of breast cancer risk stratification to encompass the risk of fast vs slow-growing cancers. Moreover, our results (Aim 3) will directly inform modeling of mammography screening intervals by PRS in Project 4. These models and simulations are expected to improve the accuracy and clinical utility of the current WISDOM risk stratification engine. We expect these advancements will ultimately lead to safer, more effective risk-based screening.