Identifying women at risk for breast cancer is not part of the current clinical paradigm for women's health even though strong risk factors, such as breast density, have been identified. The relationship between the spatial distribution of mammographic structures and breast cancer risk is not known. This knowledge gap represents an important opportunity to further our understanding of the relationship between breast structure and breast cancer. The long-term goal of this proposed research is to determine the best local measures of breast features for risk assessment of high-mortality cancers. The objective of this application is to describe the relationship of specific local textural measures to cancer risk in invasive cancers as well as DCIS cases and to discover how the spatial distribution of textural features varies based on known cancer risk factors. The central hypothesis is that subregional measurements of textural measures are stronger local and global breast cancer risk factors than global textural measures. Our secondary hypothesis is that cancer risk directly affects breast morphology. This hypothesis has been formulated based on preliminary data measured in the applicant's laboratory (Approach) showing stronger risk association than that found for mammographic density for certain subregional feature measures. The hypothesis will be tested by pursuing the following two specific aims: 1) Identify textures and subregions of the breast that are associated with breast cancer risk and 2) Identify breast texture topographies that are associated with biomarkers known to impact breast density, risk, and function. Under the first aim, we will start with a predefined set of textural features (Approach), many already shown by the applicants to be independent risk factors to breast density and use our established SFMR mammography cohort to compare different subregional textural features. Under the second aim, each feature will be estimated within grid regions and classified according to feature topography for women with different clinical risk factors. Our approach is unique in that it measures textural features on local regions of the breast and in that we will control for breast density using a calibrated volumetric measure (SXA) that separates textural variations from breast density, unlike mammographic density. The rationale for the proposed research is that understanding the local risk association of breast structure has the potential to translate into stronger clinical risk classification and allow identification of breast structures associated with cancer. This contribution will be significant because it will fundamentally change our understanding of the relationship between breast texture and risk by identifying what structures drive cancer in the breast. The knowledge gained from this research has the potential to reduce breast cancer mortality by allowing better targeting of women at risk for breast cancers.
The proposed research is relevant to public health because it will attempt to significantly improve the ability to quantify a woman's personal risk of breast cancer using measures taken from her screening mammography images. This is relevant to the NIH's mission because improved measures of breast cancer risk will advance significantly the Nation's capacity to protect and improve health.
