The long-term goal of the proposed research is to develop multi-modality, image-based markers forassessing breast density and parenchyma! structure that may be used alone or together with clinicalmeasures, as well as biomarkers, for use in determining risk of breast cancer. The general hypothesis isthat inclusion of automated analyses of the parenchyma will improve the assessment of breast cancer risk.The specific objectives of the proposed research are (1) to perform mage-based categorization of patientdatabases based on breast density, parenchyma morphology, and parenchyma kinetics [that will beautomatically extracted], (2) perform correlation and modeling of the various descriptors of breast densityand parenchymal characteristics (i.e. image-markers) with known surrogate markers of risk (such asBRCA1 and BRCA2 heterozygotes and presence of cancer on the contralateral breast) to yield new imagebasedmarkers of risk, (3) perform correlation of the various descriptors of breast density and parenchymalcharacteristics (i.e. image-markers) with developing biomarkers and candidate genes to yield betterunderstanding of breast cancer risk, and (4) perform preclinical assessment and translation of the densityand parenchymal characteristics of women at high risk using these new models. This clinical translationalcomponent will involve quantitative comparison with the current method of risk assessment of the Gailmodel and a case control study with databases from other institutions relating the image-based markers toonset of cancer. In the future, it is expected that such image-based markers will be useful for improvedassessment of patients at high risk for breast cancer and for monitoring the response of preventivetreatments. The proposed research is novel in that other correlative research in breast cancer risk withimage-based analyses involves only breast density. However, here we incorporate two additional,potentially complementary, analyses of the breast parenchyma into the correlative and modeling research.The University of Chicago is extremely well-positioned to perform this correlative research on multimodalityimage-based analyses for breast cancer risk because of its 20-year history of developing multimodalitycomputer-aided diagnosis methods for mammography, sonography, and MRI, and its integrationwith the University of Chicago Cancer Risk Clinic.
Showing the most recent 10 out of 203 publications
