Examining the Relationship Between Insulin-Like Growth Factor-I and Breast Density in Post-Menopausal Women
Background: Breast density and insulin-like growth factor-I (IGF-I) have both been linked to breast cancer risk. Prior research has shown an overwhelmingly positive association between these two factors, but almost exclusively in pre-menopausal women. In fact, few studies have focused on this relationship in post-menopausal women. Evidence suggests that reproductive hormones, like estrogen, might play a role in the overarching pathways connecting IGF-I, breast density, and breast cancer. Objective: The objectives of this thesis were to examine the relationship between IGF-I and breast density in post-menopausal women, and then determine if there was any effect modification of this association by estrogen level. Methods: This thesis focused on a cohort of 568 post-menopausal women from a clinical trial that analyzed the efficacy of exemestane in reducing breast cancer risk. Cumulus software was used to calculate percent breast density values from provided mammograms. These values were then dichotomized as being either low (< 25%) or moderately high (≥ 25%). IGF-I concentrations were determined by ELISA from collected serum samples, and then standardized prior to data analysis. The association between IGF-I level (standardized) and percent breast density (dichotomized) was analyzed using multivariable modified Poisson regression. In addition, log-transformed percent breast density values were used to measure the association through multiple linear regression. Effect modification by treatment arm (i.e. exemestane or placebo) was examined to determine if estrogen level plays any part in this relationship. Results: Percent breast density values were lower than anticipated (μ = 5.54%) with few women (13.9%) being classified as having moderately high breast density. Serum IGF-I levels were higher than expected (μ = 178.52 ng/mL), due in large part to exemestane administration. After identifying and controlling for potential covariates, there was no significant association detected between IGF-I and percent breast density in either its dichotomized (p = 0.30) or log-transformed (p = 0.60) form. The inclusion of an interaction term also returned a non-significant result for effect modification of these relationships by estrogen level. Conclusion: There was no observed association between IGF-I and breast density in postmenopausal women. This was seen regardless of estrogen level, as determined by treatment arm.