Mammographic Density Laws and Inclusion—Time for Change

Over 40 population-based studies—primarily of women of Northern European ancestry—show that high mammographic density is associated with 2- to 6-fold increased breast cancer risk.¹ High density also increases the likelihood that a breast cancer will be missed on mammographic screening and, consequently, not detected until locally advanced.^1,2 While mammographic density predicts breast cancer risk in Northern European populations,¹ mammographic density is a poor predictor of breast cancer risk in individuals¹ and is not validated in women of other racial and ethnic populations.² Latinx and Black and African American women have lower average mammographic density than Asian and non-Hispanic White women.² Furthermore, instead of predicting aggressive breast cancer biology, high mammographic density in women with breast cancer predicts good (not poor) survival.³ Despite the lack of evidence-based guidelines to inform clinician decisions, 38 US states have enacted legislation requiring radiologists to inform all women with high mammographic density that they (1) are at increased breast cancer risk and (2) may benefit from supplemental breast cancer screening (eg, whole-breast screening ultrasonography).² These notification laws have the unintended consequences of (1) inflating breast cancer risk in women of Northern European and Asian ancestry and (2) minimizing risk in Latinx and Black and African American women. Black and African American women have the highest burden of early-onset breast cancer and breast cancer death⁴; notification laws minimize risk in the very women who are in need of additional screening.

Mammographic density is the result of differential penetration of human breast tissue by x-rays and is the sum of many complex variables. X-rays readily penetrate adipose tissue; a high relative volume of adipose tissue correlates with low mammographic density.¹ In contrast, epithelial and stromal tissue are relatively radiographically opaque, and a high ratio of epithelial/stromal to adipose tissue is generally associated with high mammographic density.¹ However, it is likely that other factors (eg, collagen structure) play a role in determining mammographic density; ultimately, the mechanism(s) underlying the association between mammographic density and breast cancer risk are poorly understood.

Mammographic density is highly variable throughout a woman’s lifetime and is associated with age, menopausal status, and body mass index (BMI, calculated as weight in kilograms divided by height in meters squared).² Low mammographic density is associated with increasing age and BMI; high mammographic density is associated with young age and low BMI.² If a woman exercises and loses weight, her mammographic density increases. Clinical mammographic density estimates are not standardized for BMI or age. As a result, as currently assessed, older women and obese women have a decreased breast density relative to younger women and normal-weight women, respectively.² This lack of accounting for age and BMI when estimating breast density has the potential to limit the power of the BI-RADS breast density (Breast Imaging Reporting and Data System) to predict risk in individual women.

There are significant racial and ethnic variations in mammographic density.² On average, Asian and non-Hispanic White women have higher mammographic density than Latinx and Black and African American women.² However, when adjustments are made for age and BMI, the adjusted breast density of Asian women is significantly lower than that of Black and African American women.² In a study by Razzaghi et al,⁵ BMI-corrected mammographic density–associated risk was greatest in patients with a BMI greater than 30. These observations underscore the potential importance of correcting for age and BMI when assessing breast density in women of diverse races and ethnicities.

While population-based mammographic density studies demonstrate a clear association between high mammographic density and cancer risk, these studies have not translated to risk prediction in individuals.¹ In a study by Cecchini et al¹ of 13 409 postmenopausal high-risk women participating in the NSABP Study of Tamoxifen and Raloxifene (STAR), high BI-RADS breast density provided only slight improvement to the Gail score for predicting the incidence of invasive breast cancer. While mammographic density was positively associated with breast cancer risk, the area under the curve (AUC) was 0.55, only slightly above the line of identity (AUC = 0.5).¹ These studies highlight the weak predictive power of mammographic density in evaluating breast cancer risk in individuals.

Given that high mammographic density has the potential to mask or obscure a breast cancer, it might be predicted that high mammographic density in women with breast cancer would predict poor prognosis. This is not the case. There is evidence that low mammographic density, rather than high density, is associated with poor breast cancer survival and death. In the Breast Cancer Surveillance Consortium cohort, Gierach et al³ prospectively investigated the relationship between mammographic breast density and breast cancer death. The investigators observed a significant increased risk of death in women with low breast density (HR, 2.02; 95% CI, 1.4-3.0).³ In a second study, breast cancer death was increased in women with low (<25%) and very low (<10%) breast density, vs women with higher breast density (75.3% vs 90.2%; P = .003), even after adjusting for possible confounders.²

The biological mechanism(s) by which breast density might increase breast cancer risk are not well understood. Recent studies provide evidence that inflammation is a powerful driver of aggressive breast cancer biology and subsequent collagen damage. Studies show that a stiff extracellular matrix predicts poor survival in women with breast cancer.² The highest-risk collagen, TACS-3 (tumor-associated collagen signature 3), is characterized by bundles of straight, stiff collagen fibers aligned perpendicular to the epithelial boundary.² TACS-3 collagen predicts poor survival in White women with breast cancer.² Recent studies show that structurally damaged, parallel collagen fibers can act to traffic breast cancer cells within the mammary gland.² These studies provide evidence that mammographic density, per se, may be an oversimplification of a complex biological process.

Despite the inability of mammographic density to predict breast cancer risk in individuals, 38 states require that women with high density be informed that they are at increased risk for breast cancer.⁶ In many states such as Texas, Illinois, and Connecticut,⁶ high mammographic density (heterogeneously or extremely dense) allows women access to supplemental screening, including screening breast ultrasonography.

Mammographic density notification laws are not supported by research, and as implemented, are unequal. Density is lower on average in Latinx and Black and African American women—this increases access of non-Hispanic White and Asian women to supplemental screening and reduces access for Latinx and Black and African American women. Black and African American women bear the greatest burden of early-onset breast cancer and the greatest mortality. As enacted, mammographic density notification laws may fail women of other racial and ethnic populations by systematically minimizing risk estimates for the very women at highest risk of dying from breast cancer.

We need tools to risk-stratify average-risk women and new imaging techniques that directly image biology⁷; there is also a great need for increased dialogue between physicians and women about both breast cancer risk and the strengths/weaknesses of specific breast imaging techniques. As written, however, the current laws do not facilitate this extremely important dialogue; instead, these laws mandate that physicians provide information about breast cancer risk that is potentially inaccurate and not currently supported by clinical trials.

When passing mammographic density notification laws, lawmakers may believe that they are supporting women’s health; in actuality, they are supporting legislation that has the potential to misinform a woman about her breast cancer risk and reduce access of women of other racial and ethnic populations to supplemental screening. The practice of modern medicine is based on science; unless mammographic density laws are supported by rigorous peer-reviewed clinical trials, these state-mandated clinical practice laws should be eliminated.

Clearly there is a need for change.