Mammography enables early detection of breast cancer, a critical factor in improving treatment outcomes and breast cancer mortality. Yet, not all women benefit equally, and striking racial disparities in breast cancer mortality persist, with Black women 40% more likely to die from breast cancer compared with non-Hispanic White women. The current issue of Cancer Epidemiology, Biomarkers & Prevention presents three informative reports revealing racial and ethnic variations in mammography's performance in risk stratification, detection, and surveillance. The performance dynamics of mammography across different racial and ethnic groups highlight the urgency for additional research and innovative interventions to ensure equitable breast cancer control. We emphasize a pressing need for a comprehensive evaluation of multilevel influences on the performance and implementation of mammography in racially and ethnically diverse populations, complemented by equally urgent efforts to address factors influencing the risk of aggressive tumor subtypes and timely and effective treatment delivery.

See related articles by Kerlikowske et al., p. 1524, Hubbard et al., p. 1531, Nyante et al., p. 1542

Mammography is a vital tool for breast cancer screening and early detection in the United States, and early detection reduces breast cancer mortality. Yet, mammograms have limitations such as false negatives and false positives, and not all women benefit equally. For example, because the sensitivity of mammography is reduced in dense breasts, women with denser breasts are more likely to receive a false-negative result at the time of screening than those with less dense breasts. False-negative results lead to false reassurance, delayed diagnosis, and potentially more advanced cancer. False-positive results can also occur, leading to further evaluation, psychologic distress, and economic burden. In fact, up to half of women receiving annual mammograms over 10 years will have at least one false-positive finding (1). Furthermore, despite similar overall screening rates, Black and Hispanic women in the United States continue to be diagnosed at later stages and Black women are 40% more likely to die from their breast cancer compared with non-Hispanic White women (2).

This issue of Cancer Epidemiology, Biomarkers & Prevention features three studies based on data from the Breast Cancer Surveillance Consortium (BCSC), a network of regional breast imaging registries that collect individual-level demographic and clinical information from participating mammography facilities and link to cancer and pathology databases. The findings shed light on the performance landscape of mammography, delving into risk stratification based on breast density and utility of mammograms for cancer detection and informing supplemental screening strategies for major racial and ethnic groups in the United States. Kerlikowske and colleagues (3) found that the prevalence of mammographically dense breasts—notification of which may trigger supplemental screening—within different race and ethnic groups is confounded by body mass index (BMI). Nyante and colleagues (4) showed that established performance metrics (e.g., cancer detection rate, sensitivity, false-positive rate) of diagnostic mammography in women presenting with breast lumps differed by race and ethnicity. Finally, using simulated BCSC-derived populations, Hubbard and colleagues (5) reported differences in performance of screening mammography in women with prior breast cancer (surveillance mammography) with racial and ethnic variation in clinical, personal, and social factors. Collectively, these studies underscore the imperative of considering racial and ethnic differences in the assessment of the performance and implementation of mammography for breast cancer screening and detection, alongside the increasingly deployed risk-stratified approaches. In this context, we offer additional considerations of these results to advance equitable breast cancer control and highlight research gaps and future opportunities.

While racial and ethnic differences in screening mammography performance have been documented (6), Nyante and colleagues analyzed for the first time how diagnostic mammography performance varies by race and ethnicity among women reporting a breast lump. Women presenting with breast symptoms have a different risk profile than those attending routine screening and may include those who are unscreened or underscreened and those whose tumors were missed in or progressed rapidly since prior exams. Because breast cancers diagnosed in women presenting with a lump have worse prognosis, understanding performance patterns may inform strategies to reduce racial and ethnic disparities in breast cancer.

Nyante and colleagues found that compared with non-Hispanic White women, Black and Asian/Pacific Islander women had higher cancer detection rates and higher false-positive rates. These differences persisted in fully adjusted models that included breast density, history of breast biopsy, breast ultrasound, and area-level socioeconomic factors. These findings point to the likely significance of unmeasured and yet unknown factors, operating at multiple levels of influence, that differentially impact racial and ethnic groups.

For instance, the findings gleaned from Nyante and colleagues suggest a role for regional registry and mammography facility in racial and ethnic patterns across various performance metrics. First, adjustment for registry attenuated the odds ratio for higher sensitivity and false-positive rates for Asian/Pacific Islander women, while increasing the differences in the same metrics for Black women, both as compared with non-Hispanic White women. Second, adding facility in the fully adjusted model increased the odds of cancer detection and biopsy recommendation rates for Asian/Pacific Islander women, relative to non-Hispanic White women, but had marginal impact on the metrics for Black women. Deeper investigation is necessary to elucidate the specific features of mammography registries and facilities—encompassing geography, health care system, and provider characteristics—that drive the observed racial and ethnic differences in mammography performance. Furthermore, Hubbard and colleagues demonstrated the role of social determinants of health (SDoH), such as individual-level educational attainment and neighborhood income, in surveillance mammography performance by race and ethnicity, although available data did not capture less-easily measurable factors such as placed-based influences and the impact of racism. Ultimately, meaningful progress towards reducing breast cancer disparities demands confronting structural and upstream factors, including those present within health care systems, that disproportionately affect marginalized population groups.

Nyante and colleagues demonstrated that Black women presenting with a lump had higher breast cancer detection rates and their cancers often displayed worse prognostic and aggressive features, such as advanced stage and estrogen receptor–negative tumors, compared with cancers detected in women of other racial and ethnic groups. This finding suggests that improving cancer detection through mammography may not be enough to reduce the higher breast cancer mortality in Black women. Certain risk factors disproportionately experienced by Black women, such as higher parity and lower breastfeeding rates, have been linked to breast cancer aggressiveness (7) that is more prevalent among Black women with breast cancer. Inequities in breast cancer care following screening including delays in diagnosis, limited access to and lower completion of evidence-based treatment, and fewer treatment options for estrogen receptor–negative tumors contribute to these disparities. In addition, as shown by Hubbard and colleagues, disparities in the clinical characteristics and treatment of prior breast cancer may in turn contribute to racial and ethnic differences in surveillance mammography performance. Thus, efforts to improve the performance and implementation of screening, diagnostic, and surveillance mammography in diverse populations must be accompanied by concurrent strategies that foster equity across the breast cancer control continuum, spanning prevention to treatment.

Advocacy for mandatory dense breast notification in at least 38 states in the United States, along with its impending national requirement by the FDA in 2024, originated from concerns over the potential impact of false-negative findings, particularly the “masking” of small tumors on mammograms in women with mammographically dense breasts. In addition, higher breast density—a known risk factor for future breast cancer—is being integrated into risk prediction tools in an effort to improve the implementation of personalized breast cancer screening and risk reduction strategies. Compelling evidence indicates that women from marginalized groups (e.g., Black and Hispanic women, women with less education) have lower awareness of breast density and knowledge of its impact following dense breast notification than more socially advantaged groups (8). Yet, despite decades of research and nearly 15 years of experience with state-level breast density notification laws, there is a major gap in knowledge regarding how integrating breast density into breast cancer screening and risk stratification affect breast cancer outcomes and disparities.

Kerlikowske and colleagues demonstrated that accounting for BMI shifts the population distribution of dense versus non-dense breasts, as assessed by clinical radiologists and used in dense breast notification legislation, across racial and ethnic groups. Specifically, after adjusting for BMI, the prevalence ratio of having dense breasts decreased from 48% to 19% higher for Asian/Pacific Islander women and changed from 18% lower to 8% higher for Black women, compared with the estimated overall prevalence for the U.S. population. Given that dense breast notification legislation does not take into account other characteristics like BMI, this finding suggests a potential for “overnotification” and unnecessary follow-up in Asian/Pacific Islander women and “undernotification” and underutilization of supplemental screening in Black women. The data presented by Nyante and colleagues show that Black women presenting with lumps are less likely to receive supplemental ultrasound than members of other racial and ethnic groups. While the reasons behind the lower rates of ultrasound are not known, prior studies have also reported lower rates of screening MRI among Black and Hispanic women compared with non-Hispanic White women (9). The use of breast density, or other mammogram-derived information, for population stratification approaches based on predicted breast cancer risk versus predicted mammography detection performance merits further consideration. Hubbard and colleagues contribute a valuable perspective, indicating that these distinct approaches yield differing proportions of women recommended for supplemental imaging surveillance, with variation across racial and ethnic groups.

The nationwide adoption of dense breast notification highlights the urgency of equity considerations in the development and implementation of new policies in breast cancer control. Addressing SDoH in the implementation of dense breast notification is necessary to avoid exacerbating breast cancer disparities, which may be partly influenced by how mammography is used for clinical breast density assessment, cancer detection, and recommendation and provision of supplemental imaging.

The findings from the BCSC, along with emerging data from trials of risk-based screening approaches that incorporate breast density, will help to inform the future role of mammography in improving outcomes for all individuals impacted by breast cancer. To move beyond the “one-size-fits-all” approach to screening, accurate breast cancer risk prediction tools in racially and ethnically diverse populations are also needed. Importantly, the most widely used clinical prediction models perform poorly in Black women (10) who experience greater breast cancer burden. Emerging mammogram-based deep learning models offer a new promise for leveraging mammograms to improve risk stratification. Like traditional risk models, however, it will be crucial to test these tools in racial and ethnic minority populations. Equitable roll-out and dissemination of these models, alongside considerations of upstream and downstream factors influencing evidence-based patient care implementation, must be prioritized.

We also recognize the women who are left out of the discussion. While the BCSC includes substantial numbers of women from the largest racial and ethnic groups in the U.S., certain populations, such as Native America/Alaskan Native women, Middle Eastern/North African women, and women of mixed race, remain underrepresented, necessitating dedicated efforts to include these groups in research. Expanded focused research is also needed to understand how urbanicity/rurality, socioeconomic indicators, immigration history, and other contextual factors shape breast cancer disparities both across and within broad racial and ethnic groups (e.g., disaggregated analysis of Asian/Pacific Islander populations).

The articles presented in this issue of Cancer Epidemiology, Biomarkers & Prevention represent important steps toward understanding the nuanced performance of mammography as a tool for breast cancer screening, diagnosis, and surveillance in racially and ethnically diverse populations. Future research must build on these analyses to explicate factors and pathways amenable to interventions and policies that can advance optimal and equitable breast cancer control.

No disclosures were reported.

1.
Hubbard
RA
,
Kerlikowske
K
,
Flowers
CI
,
Yankaskas
BC
,
Zhu
W
,
Miglioretti
DL
.
Cumulative probability of false-positive recall or biopsy recommendation after 10 years of screening mammography: a cohort study
.
Ann Intern Med
2011
;
155
:
481
92
.
2.
Giaquinto
AN
,
Sung
H
,
Miller
KD
,
Kramer
JL
,
Newman
LA
,
Minihan
A
, et al
.
Breast Cancer Statistics, 2022
.
CA Cancer J Clin
2022
;
72
:
524
41
.
3.
Kerlikowske
K
,
Bissell
MCS
,
Sprague
BL
,
Tice
JA
,
Tossas
KY
,
Bowles
EJA
, et al
.
Impact of BMI on prevalence of dense breasts by race and ethnicity
.
Cancer Epidemiol Biomarkers Prev
2023
;
32
:
1524
30
.
4.
Nyante
SJ
,
Abraham
L
,
Aiello Bowles
EJ
,
Lee
CI
,
Kerlikowske
K
,
Miglioretti
DL
, et al
.
Racial and ethnic variation in diagnostic mammography performance among women reporting a breast lump
.
Cancer Epidemiol Biomarkers Prev
2023
;
32
:
1542
51
.
5.
Hubbard
RA
,
Pujol
TA
,
Alhajjar
E
,
Edoh
K
,
Martin
ML
.
Sources of disparities in surveillance mammography performance and risk-guided recommendations for supplemental breast imaging: a simulation study
.
Cancer Epidemiol Biomarkers Prev
2023
;
32
:
1531
41
.
6.
McCarthy
AM
,
Yamartino
P
,
Yang
J
,
Bristol
M
,
Conant
EF
,
Armstrong
K
.
Racial differences in false-positive mammogram rates: results from the ACRIN digital mammographic imaging screening trial (DMIST)
.
Med Care
2015
;
53
:
673
8
.
7.
Palmer
JR
,
Boggs
DA
,
Wise
LA
,
Ambrosone
CB
,
Adams-Campbell
LL
,
Rosenberg
L
.
Parity and lactation in relation to estrogen receptor negative breast cancer in African American women
.
Cancer Epidemiol Biomarkers Prev
2011
;
20
:
1883
91
.
8.
Nickel
B
,
Copp
T
,
Brennan
M
,
Farber
R
,
McCaffery
K
,
Houssami
N
.
The impact of breast density information or notification on women's cognitive, psychological, and behavioral outcomes: a systematic review
.
J Natl Cancer Inst
2021
;
113
:
1299
328
.
9.
Haas
JS
,
Hill
DA
,
Wellman
RD
,
Hubbard
RA
,
Lee
CI
,
Wernli
KJ
, et al
.
Disparities in the use of screening magnetic resonance imaging of the breast in community practice by race, ethnicity, and socioeconomic status
.
Cancer
2016
;
122
:
611
7
.
10.
Palmer
JR
,
Zirpoli
G
,
Bertrand
KA
,
Battaglia
T
,
Bernstein
L
,
Ambrosone
CB
, et al
.
A validated risk prediction model for breast cancer in US black women
.
J Clin Oncol
2021
;
39
:
3866
77
.