The reasons for differences in breast cancer incidence by ethnicity and geographic location are not well understood. The focus of this proposal is to identify hormonal characteristics that may explain these discrepancies. Insulin-like growth factor I (IGF-I) has been proposed as a link between nutritional factors and breast cancer, in particular among premenopausal women. Prolactin directly affects the growth of breast epithelium and has been associated with breast cancer risk among postmenopausal women. Mammographic densities, a strong predictor of breast cancer risk, appear to be associated with both hormones. The objectives of this project are to compare mammographic densities among four populations at different breast cancer risk, examine the relation of IGF-I, IGFBP-3, and prolactin with mammographic densities, and investigate the relation between these hormones and mammographic density separately by ethnicity and menopausal status. All samples and mammograms were collected as part of previous epidemiologic studies in four locations: Hawaii, Japan, Arizona, and Norway. Combining data from these diverse studies provides the opportunity to examine the consistency and strength of previously observed associations. This design will allow us to include women with a variety of ethnic backgrounds who have experienced different life-styles and exposures.

The final analysis will include 1,520 women of different ethnicities: Caucasian women residing in all four locations, Japanese women living in two locations, Hispanic women from Arizona, and Filipino, Chinese, and Native Hawaiian women from Hawaii. The age of the subjects ranges from 40 to 80 years and approximately 60% of the women are postmenopausal. All subjects were recruited as part of previous projects, were never diagnosed with cancer, have a mammogram free of suspicious lesions, and provided information on reproductive and anthropometric characteristics. Blood samples were taken after an overnight fast, for premenopausal women during the luteal phase, and stored at -70xC. Serum samples will be analyzed in the same laboratory for IGF-I, IGFBP-3, and prolactin using an ELISA assay. After scanning the mammographic images from the four locations into the same computer, one reader will perform computer-assisted quantitative mammographic density assessment. For quality control purposes, a sample of serum samples and mammograms will be analyzed in duplicate. Because serum and mammographic analyses were partially funded by the previous projects, we are only requesting funds for 900 serum analyses and 600 mammographic density assessments. The statistical analysis will include mixed models to explore the relation between hormone levels and breast density while controlling for confounders and for clustering by study location.

R03CA105948

2003-09-27

2006-08-31