? DESCRIPTION (provided by applicant): Breast cancer still kills 45,000 women a year in the US alone and over 270,000 women are given a diagnosis of either invasive or in situ disease. Screening is our major public health intervention. And yet we likely overdiagnose as many or more women than we save with screening and it does not impact the outcomes of the most aggressive cancers. We have assembled an extraordinary set of resources that include datasets with long term follow-up as well as a unique prospective trial that will include comprehensive host risk and tumor annotation to address the underlying biology (from both the tumor and host perspective) of indolent (IDLE) and interval cancers. Our goal is to identify better ways to screen for and treat the most aggressive cancers and avoid overdiagnosis and overtreatment as well as the inadvertent labeling of indolent lesions as cancers. Testable Hypotheses 1. Commercially available assays can identify ultralow risk breast tumors (MammaPrint Ultralow Risk for invasive cancer, Oncotype-DCIS-category 1 for DCIS). 2. The combined use of commercially available assays plus additional genomic, pathology, and immune based assays along with mode of detection can further differentiate IDLE from ultralow breast lesions. 3. Among the malignant features differentiating screen-detected from interval breast cancers are the degrees of cellular and molecular heterogeneity and type/extent of immune microenvironment and host response. 4. Since interval cancers are often biologically distinct from screen detected cancers, we hypothesize that genetic risk factors will be useful to distinguish the risk of interval from screen detected cancers. Specific Aims: 1. Stratify low risk invasive tumors into low vs. ultralow vs. IDLE and high risk into interval vs. screen- detected using gene expression profiling, pathology features, immune profiling in fully annotated invasive cancer data sets and validate the best predictors in a prospective California-wide screening trial. 2. Develop adjunctive assays to stratify DCIS lesions into IDLE, ultralow, moderate and high-risk DCIS breast lesions using gene expression profiling, and measures of tumor immune micro-environment in established data sets and validate using a prospective registry of 300 DCIS cases 3. Develop a model using known germline breast cancer risk variants to predict women predisposed toward ultralow and IDLE screen detected tumors, and those predisposed to interval detected breast cancers, using data from a fully annotated California-wide screening trial that includes germline and tumor profiling." |