A new look at breast density and breast cancer risk

Abstract

Breast density, as visible on mammograms, comprises connective and epithelial tissue and can be seen to represent the glandular target tissue for breast cancer, whereas the non-dense tissue mainly comprises fat. High percentages of density are established to be one of the strongest risk factors of breast cancer risk. A hypothesis that the actual number of target cells is directly related to breast cancer risk has not yet been established as such, which may be partly due to the predominant use of relative density, rather than absolute amounts of dense tissue. In a series of studies we used computerized measurement of mammographic tissues from women who participated in the DOM breast cancer screening project (Diagnostisch Onderzoek Mammacarcinoom, Utrecht, The Netherlands) to investigate the biology of breast density and the effect that using a relative measure has had on current knowledge. We found that there is a large genetic influence on the overall appearance of the breast on a mammogram, but a possible modifying ability of other factors, such as parity. A model that combines monozygotic twins with dizygotic twins or sisters yielded an estimate of the heritability of over 100%, which indicated that dominant gene effects, genetic interactions or gene–environment effects could be involved. The last effect was seen for parity as heritability estimates ranged from 90% in sisters that were both nulliparous, to 2% in sister pairs discordant for nulliparity. In a population of 418 postmenopausal women we found that relative density was influenced by a combination of the factors that influence dense and/or non-dense tissue. These influences were in the same direction as those on dense tissue, but the magnitudes were the resultant of the effects on dense and non-dense tissues. The influence of BMI on relative density was completely due to an effect on non-dense tissue. These results showed that, although relative density is a relevant prognostic factor, inferences about the aetiology of breast density should better be made on the basis of absolute measures. We did one such study into the relation between the Dutch famine and breast density in 1035 women with ascending levels of famine-exposure; unexposed, moderately exposed and severely exposed. Only the amount of non-dense tissue was found to be significantly reduced after high exposure before age 10 (p-trend=0.03). This group also appeared to have smaller breast with more absolute and relative density, but these results were not statistically significant. In our main study, all newly diagnosed invasive breast cancers in 15.5 years of follow-up were determined (361) and follow up experience was obtained for a random cohort sample of 600 women. The highest risk for breast cancer was found in the women that had above median amounts of both dense and non-dense tissues (HR 2.20; 95%CI 1.37-3.52) compared to women with below median amounts of both tissues. This indicated that non-dense tissue plays a role in the development of breast cancer, whereas thus far the highest risk had always been associated with the highest percent density.