Mammographic Densities and Breast Cancer Prevention

Institution: University of Southern California
Investigator(s): Malcolm Pike, Ph.D. -
Award Cycle: 1996 (Cycle II) Grant #: 2RB-0058 Award: $567,993
Award Type: Research Project Awards
Research Priorities
Detection, Prognosis and Treatment>Imaging, Biomarkers, and Molecular Pathology: improving detection and diagnosis



Initial Award Abstract (1996)
It is well known that mammograms have been found to be useful in identifying early existing breast cancers before they can be felt. Less well known is that mammograms taken in a woman without breast cancer may also be used to convey very useful information as to the woman's future risk of breast cancer. The appearance of the breast in mammograms is determined by the relative amounts of fat, connective tissue, and epithelial tissue. (Breast epithelial tissue is the tissue in which the majority of breast cancer arises.) On mammograms, fat appears as radiologically lucent (clear) areas, whereas connective and epithelial tissue appears slightly darker, resulting in mammographic "densities." Higher levels of mamographic densities have been found to be strongly predictive of breast cancer risk in certain groups of (non-Latina) White women. We propose to study whether mammographic densities can be used to predict the risk of breast cancer in African-American and Asian-American as well as in White (Latina) women, and in younger as well as in older women. We will also study whether the known ethnic differences in breast cancer risk will be reflected in differences in mammographic densities across these ethnic groups. In addition, we will study whether the protective effects of physical exercise on breast cancer risk is reflected in reduced mammographic densities.

The proposed study represents an addition to two ongoing studies in our department, the Los Angeles part of the nationwide Women's CARE study, and a study of breast cancer in Asian Americans. We will determine the extent of mammographic densities in 240 African-American, 240 White, and 150 Asian-American breast cancer cases, and an equal number of African-American, White and Asian-American controls (women without cancer).

If mammographic densities can be shown to provide consistent, wide-ranging predictive power regarding breast cancer risk, and to reflect differences in major breast cancer risk factors, they then could be used to indicate degrees of risk for breast cancer. This could be added to information about other known risk factors, such as family history, being a carrier of certain mutant genes, etc. to help estimate a woman's risk of breast cancer. Indeed, they could be used instead of these and other, as yet unknown, risk factors to provide a more precise measure of risk. Such a "surrogate" measure of risk would be of great value in genetic and other risk factor reduction counseling, and in public health breast cancer prevention efforts, by enabling high-risk target populations to be defined.


Final Report (2000)
The extent of radiodense tissue on a mammogram (mammographic densities) has been strongly associated with breast cancer risk in certain groups of (non-Latina) White women. However, there is little data on whether mammographic densities predict the risk of breast cancer in African American and Asian American as well as in White women, and in younger as well as in older women. We collected mammograms from women aged 35-64 who had participated in two ongoing breast case control studies at our institution.

We obtained mammograms from 1203 women of three different ethnicities: 227 African American cases, 171 African American controls, 304 White cases and 247 White controls and 175 Asian American cases and 79 Asian American controls. Mammographic images were scanned and mammographic density assessed using a computer assisted method we have previously evaluated. Per cent mammographic density was estimated by dividing the area of absolute density with the measured breast size.

Preliminary analyses of mammograms from the first 1090 women suggest that Asian women had the highest, and African American women the lowest percent mammographic density. Cases had statistically significantly higher % density than controls overall. When conducting anlayses by ethnicity, we found statistically significant case-control differences for White women under as well as above age 50, and for Asian American women above age 50.

In all three ethnic groups combined, increased mammographic densities was associated with increasing risk of breast cancer. Compared to women with 0% densities, the odds ratios associated with breast cancer adjusted for age and ethnicity increased for women with increasing mammographic density as follows: 0%--1.0; 1-9%--1.5; 10-24%--1.9; 25-49%--2.4; 50-74%--3.3; and 75% or more densities, 5.8.

We also conducted preliminary crude analyses of the relative risk of breast cancer associated with mammographic density by ethnicity. For these analyses we compared having 60% or more densities to 10% or less densities. These unadjusted relative risks were 3.7 in Whites, 2.3 in Asian Americans, and 1.7 in African Americans (all three figures are within the 95% confidence interval of statistical relevance).

We are conducting further analyses to try and decipher the reasons for the ethnic differences in mammographic densities. Mammographic densities has been suggested as a surrogate endpoint for breast cancer risk. It is therefore crucial that we determine whether it is useful as a surrogate endpoint in African Americans and Asian Americans as well as in Whites.