Date of Award

January 2020

Document Type

Open Access Thesis

Degree Name

Master of Public Health (MPH)


School of Public Health

First Advisor

Melinda Irwin

Second Advisor

Heather Eliassen


Background and Aims. Long-term use of menopausal hormone therapy, typically in the form of estrogen or estrogen + progesterone, can increase the amount of dense breast tissue in women, which is associated with breast cancer. These changes are commonly measured using mammograms to calculate percent density; higher percent density is also associated with breast cancer. Newer methods of mammogram analysis look at spatial texture feature variation; these approaches are more sensitive than percent density, however the association between texture features and hormone therapy use is not well-documented. This study aims to analyze a specific texture feature, the V metric, which is a measure of the standard deviation of greyscale pixel intensity values from a mammogram image. We evaluated the V metric and its associations with both estrogen and estrogen + progesterone menopausal hormone therapy for durations of less than or equal to 5 years as well as more than 5 years. Our goal was to further establish V as a sensitive predictive measure of breast cancer.

Methods. This was a retrospective cohort study that used participant data from the Nurses’ Health Study I and II. The study included 1,986 postmenopausal women that had available mammograms and menopausal hormone therapy use data. From both low and high-resolution mammograms, the V metric is computed. To evaluate the outcome variable V, linear mixed models were fit, adjusting for age, BMI, race, menopause, family history of breast cancer, personal history of benign breast disease, total breastfeeding, age at first birth, parity, age at menarche, age at menopause, physical activity, and alcohol intake. Models were then stratified by menopause type and percent mammogram density. Sensitivity analyses were conducted with the subset of mammograms that were high-resolution.

Results. Adjusting for covariates, participants that used estrogen + progesterone for more than 5 years had significantly higher mean V measures than those that used estrogen or never used menopausal hormone therapy. Participants using any form of menopausal hormone therapy for any duration had significantly higher mean V measures than never users.

Conclusions. Menopausal hormone therapy is significantly associated with higher mean V measures. Various durations of therapy can affect breast tissue physiology, leading to increases in dense tissue. This can increase risk of breast cancer. Notably, the V metric may quantify some changes that percent mammogram density does not register. Particularly in women that used estrogen + progesterone therapy for longer durations, the V metric may be a more accurate predictor of breast cancer risk. Additional evaluation of related physiological factors and their associations with the V metric will help increase its validity.


This is an Open Access Thesis.

Open Access

This Article is Open Access