The hypothesis for the proposed study is that women diagnosed with breast cancer have higher mammographic density patterns in prediagnostic mammograms than women without a diagnosis of breast cancer. In addition, we postulate that despite differences in mammographic density patterns among ethnic groups, the relation between mammographic densities and breast cancer risk is similar in Japanese, Native Hawaiian, and Caucasian women and that dietary factors contribute to the variation in mammographic densities among ethnic groups. A large multiethnic cohort in Hawaii established between 1993 and 1996 with well documented differences in breast cancer risk offers a unique opportunity to research these hypotheses. The specific aims of this project are to investigate the association between mammographic density patterns and breast cancer risk in women of Japanese, Native Hawaiian, and Caucasian ancestry, to examine the possible effect of dietary fat, fruit and vegetable intake, soy consumption, and other dietary components on mammographic densities, and to elucidate the relative importance of percent densities vs. the absolute size of the dense areas in relation to breast cancer risk in women of different ethnicity.

We are proposing a nested case-control study using mammograms from women who are part of the multiethnic cohort. We will include 3 ethnic groups who represent the largest sub-groups: Caucasian, Japanese, and Native Hawaiian. Mammograms performed before the diagnosis of breast cancer will be obtained for women diagnosed with breast cancer between 1994 and 2000. For an equal number of healthy controls, matched by age and ethnicity, we will identify mammograms taken within 2 years of the time when the respective cases selected for Inclusion in the study received their mammogram. Women with the following characteristics will be excluded from the study: a diagnosis of breast cancer before entry into the cohort, bilateral breast cancer, no previous mammogram, and a previous history of breast augmentation, reduction surgery, or an extensive surgical biopsy with a diagnosis of benign breast disease. Based on breast cancer incidence in the cohort, mammography rates, and estimated participation rates, we plan to enroll approximately 1,000 cases and 1,000 controls with 500 Japanese, 300, Caucasian, and 200 Native Hawaiian per group. Potential study subjects will be contacted by mail and asked for permission to retrieve a suitable mammogram from their mammography clinic. Women will also complete a short questionnaire related to breast health. The statistical analysis will include logistic regression to estimate the risk of developing breast cancer in relation to mammographic density patterns and multiple linear regression to examine the association between diet and density patterns.

R01CA085265

2000-06-01

2005-05-31