Project 1: Imaging-Based Determination of Breast Cancer Risk
Leader: Maryellen Giger, PhD
Co-Leader: Gillian Newstead, MD
The long-term goal of this project is to further improve multimodality, image-based markers for assessing breast density and parenchymal structure that may be used alone or together with clinical measures and biomarkers to determine the risk of developing breast cancer. The general hypothesis is that inclusion of automated analyses of the parenchyma will improve the assessment of breast cancer risk. In the future, it is expected that the proposed image-based markers will be useful for improved assessment of patients at high risk for breast cancer and for monitoring the response to preventive treatments. The project's aims are to:
Aim 1: Conduct image-based categorization of patient databases using breast density, parenchyma morphology, and automatically extracted parenchyma kinetics.
Aim 2: Yield new image-based markers of risk by correlating and modeling various characteristics of breast density and parenchymal patterns with known surrogate markers of risk, such as BRCA1 and BRCA2 genetic mutations and the presence of cancer in the contralateral breast.
Aim 3: Develop a better understanding of breast cancer risk by correlating various image markers with developing biomarkers and candidate genes.
Aim 4: Use these new models to perform preclinical assessment and translation of the density and parenchymal characteristics of women at high risk.
[Note that this list is limited as it pertains only to aims from the first funding cycle]
- Compiled a large database of breast images of women at high risk for breast cancer and of women at normal risk, and categorized the cases based on quantitative image analysis of breast density, parenchyma morphology, and parenchymal kinetics.
- Conducted correlation studies of these various descriptors of parenchyma (i.e. image-based risk phenotypes) with known indicators of risk (BRCA1 and BRCA2 deleterious mutations) or presence of cancer (using analysis on the contralateral breast) demonstrating the robustness of our quantitative image analysis methods.
- Validated, using these large clinical datasets, our methods for quantitative analyses of mammographic patterns on FFDM, statistically demonstrating again that women at high risk tend to have dense breasts with parenchymal patterns that are coarse and low in contrast.
- Developed an image-based risk assessment interface for ultimately translating the development to clinical use.