Abstract
Purpose
Evidence on the associations between categorical mammographic density and breast cancer risk by tumor invasiveness remains limited in Asian women. This large, population-based cohort study investigated the distribution of mammographic density by age and menopausal status, as well as its association with the risks of invasive breast cancer and ductal carcinoma in situ (DCIS) in South Korean women.
Methods
Mammographic screening was performed on 6,365,522 women between 2009 and 2014 through the Korean National Cancer Screening Program. Mammographic parenchymal composition was classified using the fourth edition of Breast Imaging Reporting and Data System. We computed multivariable-adjusted hazard ratios (HRs) and 95 % confidence intervals (CIs) using Cox proportional hazards regression models for the association between mammographic parenchymal composition and the risk of invasive cancer and DCIS.
RESULTS
Overall, 40.6% of women had dense breasts, with the proportion decreasing with increasing age. A total of 44,468 incident breast cancer cases (0.7%) were documented. Compared with almost entirely fatty breasts, increasing mammographic density was associated with a higher risk of breast cancer (HR, 1.55; 95% CI 1.51-1.60 for scattered fibroglandular densities; HR, 2.14; 95% CI 2.08-2.21 for heterogeneously dense breasts; and HR, 2.59; 95% CI 2.50-2.69 for extremely dense breasts). Associations between mammographic density and breast cancer risk were similar for invasive cancer and DCIS, and did not vary significantly by menopausal status.
CONCLUSIONS
Mammographic density may be a significant risk factor for both invasive cancer and DCIS, regardless of menopausal status. It should be incorporated into breast cancer risk stratification and screening strategies.
Keywords: mammography, density, age, menopausal status, breast cancer risk, Korean
Introduction
Breast cancer is the most prevalent cancer among women in most Asian countries, with a steadily increasing incidence (1,2). In South Korea, breast cancer has consistently increased since 2000 and the incidence was 134.5 cases per 100,000 people as of 2021. The 2021 crude incidence rate of invasive breast cancer was 111.6, an increase of over 4.4 times compared to 2000. For ductal carcinoma in situ (DCIS), the incidence rate was 22.9, a significant increase of 14.3 times (3), showing distinct differences in the rates of increase between invasive cancer and DCIS.
Breast cancer epidemiology in South Korean women presents unique characteristics compared to Western women. Between 2000 and 2021, the peak incidence age of breast cancer in South Korea shifted upward from 46.9 to 53.4 years. Unlike the continuous age-related increase observed in Western countries, in South Korea, after peaking around the age of 53, incidences of breast cancer tend to decline. In addition, though breast cancer remains more common in postmenopausal women, the prevalence of premenopausal breast cancer is relatively higher in South Korea than in Western populations (3).
Mammographic density is considered an independent risk factor for breast cancer (4,5). Women exhibiting 75% or greater breast density have a 4–6 times higher risk of developing breast cancer compared to those with 10% or less density (6,7). However, there are doubts about whether mammographic density is a risk factor for breast cancer in Asian women (8,9), because, despite higher mammographic density compared to Western women (10-13), the incidence of breast cancer in Asian women remains paradoxically lower as the incidence of breast cancer in Asian women paradoxically remains lower (2,14,15). Most studies documenting the association between mammographic density and breast cancer have been conducted in Western cohorts (6,16-18). Only a few studies have investigated this association in South Korean women, and these studies had limitations such as relatively small sample sizes (19-21) and short follow-up durations (20). Recent large-scale studies involving 2 to 8 million South Korean women have shown that an increase in mammographic density over successive biennial mammographic screenings was associated with an elevated breast cancer risk (22,23), and that certain breast cancer risk factors were associated with dense breasts (24). However, the proportion of mammographic density categories according to age and menopausal status remains unclear, as does the associations between categorical mammographic density and risks for breast cancer by tumor invasiveness (22-24).
To address the gaps in our understanding of the importance of mammographic density and its relationship to breast cancer incidence in South Korean women, we conducted a large population-based cohort study of over 6 million South Korean women who underwent biennial mammographic screening. To our knowledge, this is the largest cohort study to investigate the association between the Breast Imaging Reporting and Data System (BI-RADS) breast density on mammography and the risks of both invasive breast cancer and DCIS by menopausal status in Korean women.
Our study aimed to estimate the proportion of mammographically dense breasts by age and menopausal status, and to examine the associations between mammographic density and the risks of both invasive breast cancer and DCIS, using the BI-RADS breast density classification.
Patients and Methods
Study Population
South Korea adopted the Korean National Cancer Screening Program (KNCSP) in 2002 to provide breast cancer screening via biennial mammography for all South Korean women aged 40 years and older. The KNCSP provides screening services to Medical Aid beneficiaries and National Health Insurance Service (NHIS) enrollees. The NHIS program is a single compulsory government-managed social insurance system covering the entire South Korean population (25,26). A total of 6,628,034 women underwent screening mammography through the KNCSP between 2009 and 2014 and completed questionnaires without missing data. After excluding 213,389 women whose mammographic examinations did not match their screening results within 30 days and 49,123 women with a prior diagnosis of breast cancer, a final total of 6,365,522 women were included in the study. These women were followed until the end of 2016, with a median follow-up of 5.5 years (interquartile range: 3.4–6.4 years). This study was conducted with the approval of the Institutional Review Board (IRB) of Seoul St. Mary’s Hospital, Korea (IRB No: KC17EESI0189) and informed consent was waived.
Assessment of Mammographic Density
Mammography was performed using film screen, computed radiography, or digital mammography. Breast density was assessed using BI–RADS, classifying parenchymal compositions into four categories: almost entirely fatty (<25% glandular), scattered fibroglandular densities (25%–50% glandular), heterogeneously dense (51–75% dense), and extremely dense (>75% dense). Radiologists interpreted mammograms and rated them using the four-point KNCSP scale, which was modified using BI–RADS categories: 1, negative (category 1); 2, benign (category 2); 3, incomplete, additional evaluation needed (categories 3 and 0); 4, suspected breast cancer (categories 4 and 5). Cases with a KNCSP scale of 3-4 were recalled. The KNCSP scale is the standard reporting system for mammographic assessments in the National Cancer Screening Program, modified from BI-RADS categories to align with national screening protocols. BI-RADS is still widely used in hospitals and private screening centers across Korea. When multiple serial mammograms were available for the same participant, the analyses used the initial mammogram.
Ascertainment of Breast Cancer and Covariates
With permission from the South Korean Ministry of Health and Welfare, we used de-identified data from the NHIS system, including eligibility (age, sex, and socioeconomic variables), medical treatment, and health examination databases (27,28). The NHIS requires all medical care providers to contract with the NHIS and submit claims data using the International Classification of Disease, 10th revision (ICD-10) codes (C50 for invasive breast cancer and D05 for DCIS). The NHIS maintains the registration system for cancer patients and the government covers 90% of cancer care costs. Patients are typically registered for financial support of cancer-related medical services after their diagnosis. Thus, breast cancer cases were unlikely to be missed or misclassified, ensuring a high validity of the classification, even though they were based on claims data rather than registry data (25,29).
Trained nurses measured participants’ height and weight during health examinations. Data on health behaviors, family history of breast cancer, and reproductive and hormonal factors were collected through self-administered standardized questionnaires, which included information on smoking history, alcohol consumption, exercise habits, family history of breast cancer, menopausal status, hormone therapy use, parity, breastfeeding, and oral contraceptive use.
Statistical Analysis
Categorical variables were expressed as numbers (percentage) and continuous variables as mean ± standard deviation according to parenchymal composition. The incidence rate of breast cancer was calculated by dividing the number of incident cases by the total follow-up duration (person-years). The cumulative incidence of breast cancer by mammographic density was illustrated using Kaplan-Meier curves, and the log-rank test was used to assess differences between groups. We computed hazard ratios (HR) and 95% confidence intervals (CI) for the associations between mammographic density and breast cancer risk using multivariable Cox proportional hazard models. Analyses were stratified by menopausal status and conducted separately for invasive breast cancer and DCIS as the outcome. The proportional hazards assumption was evaluated by Schoenfeld residuals with the logarithm of the cumulative hazards function based on Kaplan–Meier estimates for mammographic density. Potential confounders were identified a priori based on a literature review and presumed causal relationships among the covariates (30). The following confounders were included in the multivariable-adjusted models: age (continuous), baseline menopausal status (binary), height (continuous), body mass index (BMI, continuous), hormone therapy (never, <2 years, <5 and ≥2 years, ≥5 years, unknown), parity (0, 1, ≥2 births), duration of breastfeeding (none, <6 months, <12 and ≥6 months, ≥12 months, unknown), use of oral contraceptives (never, <1 year, ≥1 year, unknown), age at menarche (<14, ≥14 years), alcohol consumption (never drinker, mild to moderate, heavy), smoking status (never, current, or past), and regular exercise (binary) (Suppl Fig 1). Both height and BMI were included as covariates, as they independently influence breast cancer risk, height reflecting early-life growth and hormonal factors, and BMI capturing adiposity-related effects (31,32).
Statistical significance was evaluated with two-sided tests, with a significance level of 0.05. All analyses were performed using SAS 9.4 software (SAS Institute Inc., Cary, NC).
Results
Baseline Characteristics of Study Participants
Characteristics of participants by mammographic density are shown in Table 1. Overall, 40.6% had dense breasts (either heterogeneously or extremely dense). Women with higher mammographic density were younger and were less likely to be post-menopausal at baseline.
Table 1. Demographic data of screening population according to mammographic density.
| Mammographic parenchymal composition | |||||
|---|---|---|---|---|---|
| Number (%) | 25%< | 25-50% | 51-75% | 76-100% | |
| 1933840 (30.4) | 1848284 (29.0) | 1780992 (28.4) | 802406 (12.6) | ||
| Age, mean (SD), years | 57.1±10.62 | 64.14±9.72 | 58.27±9.42 | 52.23±8.4 | 48.23±7.17 |
| Height, mean (SD), cm | 154.87±5.87 | 152.72±5.87 | 154.62±5.64 | 156.35±5.43 | 157.37±5.34 |
| Body mass index, mean (SD), kg/m2 | 23.92±3.23 | 24.75±3.36 | 24.23±3.15 | 23.38±2.97 | 22.41±2.85 |
| Family history of breast cancera | 126727 (5.6) | 27250 (3.9) | 36728 (5.4) | 42788 (6.7) | 19961 (8.4) |
| Menarche under age of 14 | 578457 (9.1) | 87764 (4.5) | 143624 (7.8) | 221546 (12.4) | 125523 (15.6) |
| Oral contraceptive use | |||||
| Never | 5180243 (81.4) | 1566367 (81.0) | 1480368 (80.1) | 1461349 (82.1) | 672159 (83.8) |
| <1 year | 568537(8.9) | 161930(8.4) | 171559(9.3) | 163230(9.2) | 71818(9.0) |
| ≥ 1 year | 329204(5.2) | 113931(5.9) | 107835(5.8) | 79354(4.5) | 28084(3.5) |
| Unknown | 287538 (4.5) | 91612 (4.7) | 88522 (4.8) | 77059 (4.3) | 30345 (3.8) |
| Number of live birth | |||||
| 1 | 597637 (9.4) | 88454 (4.6) | 153666 (8.3) | 226986 (12.7) | 128531 (16.0) |
| ≥ 2 | 5562864 (87.4) | 1817911 (94.0) | 1651215 (89.3) | 1478121 (83.0) | 615617 (76.7) |
| 0 | 205021 (3.2) | 27475 (1.4) | 43403 (2.4) | 75885 (4.3) | 58258 (7.3) |
| History of breast feeding | |||||
| <6 months | 801949 (12.6) | 110729 (5.7) | 195615 (10.6) | 313527 (17.6) | 182078 (22.7) |
| <12 months | 1234480 (19.4) | 281781 (14.6) | 362578 (19.6) | 406231 (22.8) | 183890 (22.9) |
| ≥ 1 year | 3609909 (56.7) | 1441175 (74.5) | 1115659 (60.4) | 784674 (44.1) | 268401 (33.5) |
| Never | 719184 (11.3) | 100155 (5.2) | 174432 (9.4) | 276560 (15.5) | 168037 (20.9) |
| Menopausal status at baseline | |||||
| premenopausal | 1632601 (25.7) | 133834 (6.9) | 302196 (16.4) | 700600 (39.3) | 495971 (61.8) |
| postmenopausal | 4732921 (74.4) | 1800006 (93.1) | 1546088 (83.7) | 1080392 (60.7) | 306435 (38.2) |
| History of hormone therapy | |||||
| Never | 5315701 (83.5) | 1668930 (86.3) | 1508473 (81.6) | 1458931 (81.9) | 679367 (84.7) |
| < 2 years | 432850 (6.8) | 117469 (6.1) | 149290 (8.1) | 126284 (7.1) | 39807 (5.0) |
| 2 -5 years | 163704 (2.6) | 40159 (2.1) | 57280 (3.1) | 50641 (2.8) | 15624 (2.0) |
| ≥ 5 years | 129973 (2.0) | 31601 (1.6) | 46609 (2.5) | 40360 (2.3) | 11403 (1.4) |
| Unknown | 323294 (5.1) | 75681 (3.9) | 86632 (4.7) | 104776 (5.9) | 56205 (7.0) |
| History of smoking | |||||
| Never | 5994050(95.3) | 1853986(96.2) | 1741802(95.3) | 1659308(94.7) | 738954(94.2) |
| Ex-smoker | 88807(1.4) | 20177(1.1) | 25005(1.4) | 29249(1.7) | 14376(1.8) |
| Current smoker | 209230(3.3) | 52410(2.7) | 61568(3.4) | 64492(3.7) | 30760(3.9) |
| Alcohol consumption | |||||
| Never | 5234819 (82.2) | 1717269 (88.8) | 1545462 (83.6) | 1380490 (77.5) | 591598 (73.7) |
| Mild, Moderate | 1090768 (17.1) | 208573 (10.8) | 291272 (15.8) | 386629 (21.7) | 204294 (25.5) |
| Heavy | 39935 (0.6) | 7998 (0.4) | 11550 (0.6) | 13873 (0.8) | 6514 (0.8) |
| Regular exercise | 2700958 (42.4) | 685783 (35.5) | 787226 (42.6) | 839403 (47.1) | 388546 (48.4) |
| History of diabetes mellitus | 701198 (11.0) | 305288 (15.8) | 225407 (12.2) | 133873 (7.5) | 36630 (4.6) |
variables with missing value more than 25%
Distribution of Mammographic Density by Age and Menopausal Status
The proportion of dense breasts progressively declined with age (P <.0001). Among women with heterogeneously dense breasts, the proportion was 42.7% in their 40s, 34.1% in their 50s, 17.4% in their 60s, and 11.5% at age 70 or older. For extremely dense breasts, the proportion was 30.6% in their 40s, 11.6% in their 50s, 3.3% in their 60s, and 1.7% at age 70 or older (Figure 1A). Mammographic density distribution also differed by menopausal status. The proportion of heterogeneously and extremely dense breasts was higher in premenopausal than in postmenopausal women at 42.9% and 30.4% in premenopausal women compared to 22.8% and 6.5% in postmenopausal women, respectively (P <.0001). Within the same age group, this trend persisted: among women aged 40–49 years, the proportion of heterogeneously and extremely dense breasts was 43.1% and 32.3% in premenopausal women compared to 40.4% and 19.3% in postmenopausal women; in women aged 50–59 years, the corresponding proportions were 42.8% and 21.5% in premenopausal women versus 32.9% and 10.3% in postmenopausal women (each P <.0001) (Figure 1B).
Figure 1. Distribution of mammographic parenchymal composition by age group (A) and by age and menopausal status (B).
A. The proportion of heterogeneously and extremely dense breasts declines with increasing age (P <.0001). B. The proportion of dense breasts is higher in premenopausal women than in postmenopausal women, and remains higher in premenopausal women within the same age group (P <.0001).
Association between Mammographic Density and Risk of Breast Cancer
During a median follow-up of 5.5 years, 44,468 incident cases of breast cancer were documented (0.7%). The incidence rate of breast cancer was approximately three times higher in women with extremely dense breasts compared to those with almost entirely fatty breasts (Table 2, Figure 2, Suppl Figure 2). Compared with almost entirely fatty breasts, higher percent density was associated with an increased risk of breast cancer (HR 1.55; 95% CI 1.51-1.60 for scattered fibroglandular densities; HR 2.14; 95% CI 2.08-2.21 for heterogeneously dense breasts; and HR 2.59; 95% CI(2.50-2.69 for extremely dense breasts) after adjusting for potential confounders. Associations tended to be stronger in postmenopausal women, although there was no significant interaction by menopausal status at baseline (Table 2, Suppl Figure 2).
Table 2. Association between mammographic density and breast cancer risk, overall and by menopausal status.
| Mammographic parenchymal composition |
Subjects N |
Breast cancer event N |
IR per 1000 PY |
Crude HR (95% CI) |
Adjusted HR (95% CI) |
|
|---|---|---|---|---|---|---|
| Overall | 25%< | 1933840 | 7626 | 0.76 | 1(reference) | 1(reference) |
| 25-50% | 1848284 | 11546 | 1.27 | 1.66(1.62,1.71) | 1.55(1.51,1.60) | |
| 51-75% | 1780992 | 16130 | 1.88 | 2.45(2.38,2.51) | 2.14(2.08,2.21) | |
| 76-100% | 802406 | 9166 | 2.33 | 3.05(2.95,3.14) | 2.59(2.50,2.69) | |
| Premenopausal | 25%< | 133834 | 858 | 1.17 | 1(reference) | 1(reference) |
| 25-50% | 302196 | 2216 | 1.37 | 1.17(1.08,1.26) | 1.19(1.10,1.28) | |
| 51-75% | 700600 | 7377 | 2.05 | 1.75(1.63,1.87) | 1.77(1.65,1.90) | |
| 76-100% | 495971 | 6278 | 2.50 | 2.13(1.99,2.29) | 2.17(2.02,2.34) | |
| Postmenopausal | 25%< | 1800006 | 6768 | 0.73 | 1(reference) | 1(reference) |
| 25-50% | 1546088 | 9330 | 1.25 | 1.71(1.65,1.76) | 1.61(1.56,1.66) | |
| 51-75% | 1080392 | 8753 | 1.75 | 2.39(2.31,2.46) | 2.16(2.09,2.24) | |
| 76-100% | 306435 | 2888 | 2.04 | 2.77(2.66,2.90) | 2.51(2.40,2.63) |
Abbreviation: IR, incidence rate; PY, person-years; HR (95% CI), hazard ratio (95% confidence interval)
Adjusted for age (continuous), height (continuous), BMI (continuous), hormone therapy (never, <2 years, <5 and ≥2 years, ≥5 years, unknown), parity (0, 1, ≥2 births), duration of breast feeding (none, <6 months, <12 and ≥6 months, ≥12 months, unknown), use of oral contraceptives (never, <1 year, ≥1 year, unknown), age at menarche (<14, ≥14 years), alcohol consumption (never drinker, mild to moderate, heavy), smoking status (never, current, or past), regular exercise (binary), and baseline menopausal status.
Figure 2. Cumulative incidence of breast cancer stratified by BI-RADS mammographic density categories.
The cumulative incidence of breast cancer increases over time as mammographic density increases. The y-axis indicates the cumulative incidence probability, and the x-axis represents years of follow-up.
Association between Mammographic Density and Risk of Invasive Breast Cancer and DCIS
The association between mammographic density and breast cancer risk by tumor invasiveness is shown in Table 3 and Suppl Figure 3. The overall associations between mammographic density and DCIS (HR 1.65; 95% CI 1.47-1.84 for scattered fibroglandular densities; HR 2.20; 95% CI 1.97-2.47 for heterogeneously dense breasts; and HR 2.62; 95% CI 2.31-2.98for extremely dense breasts) were like those for invasive cancer, with hazard ratios increasing with density category (HR 1.55; 95% CI 1.50-1.59 for scattered fibroglandular densities; HR 2.14; 95% CI 2.07-2.21 for heterogeneously dense breasts; and HR 2.59; 95% CI 2.50-2.69 for extremely dense breasts). Although associations with DCIS also tended to be stronger in postmenopausal women, no significant interaction with menopausal status was found.
Table 3. Association between mammographic density and invasive breast cancer and ductal carcinoma in situ, overall and by menopausal status.
| Mammographic parenchymal composition |
Number | Breast cancer | IR per 1000 Women/y |
Adjusted HR(95% CI) | ||
|---|---|---|---|---|---|---|
| Invasive cancer | Overall | 25%< | 1933840 | 7119 | 0.71 | 1(reference) |
| 25-50% | 1848284 | 10670 | 1.18 | 1.55(1.50,1.59) | ||
| 51-75% | 1780992 | 14813 | 1.72 | 2.14(2.07,2.21) | ||
| 76-100% | 802406 | 8364 | 2.13 | 2.59(2.50,2.69) | ||
| Premenopausal | 25%< | 133834 | 790 | 1.07 | 1(reference) | |
| 25-50% | 302196 | 2029 | 1.25 | 1.18(1.09,1.28) | ||
| 51-75% | 700600 | 6699 | 1.86 | 1.76(1.63,1.90) | ||
| 76-100% | 495971 | 5673 | 2.26 | 2.16(2.01,2.34) | ||
| Postmenopausal | 25%< | 1800006 | 6329 | 0.68 | 1(reference) | |
| 25-50% | 1546088 | 8641 | 1.16 | 1.60(1.54,1.65) | ||
| 51-75% | 1080392 | 8114 | 1.62 | 2.16(2.08,2.24) | ||
| 76-100% | 306435 | 2691 | 1.90 | 2.53(2.40,2.65) | ||
| Ductal carcinoma in situ | Overall | 25%< | 1926721 | 507 | 0.05 | 1(reference) |
| 25-50% | 1837614 | 876 | 0.09 | 1.65(1.47,1.84) | ||
| 51-75% | 1766179 | 1317 | 0.15 | 2.20(1.97,2.47) | ||
| 76-100% | 794042 | 802 | 0.20 | 2.62(2.31,2.98) | ||
| Premenopausal | 25%< | 133044 | 68 | 0.09 | 1(reference) | |
| 25-50% | 300167 | 187 | 0.11 | 1.23(0.93,1.63) | ||
| 51-75% | 693901 | 678 | 0.18 | 1.90(1.48,2.45) | ||
| 76-100% | 490298 | 605 | 0.24 | 2.31(1.79,2.98) | ||
| Postmenopausal | 25%< | 1793677 | 439 | 0.04 | 1(reference) | |
| 25-50% | 1537447 | 689 | 0.09 | 1.75(1.55,1.98) | ||
| 51-75% | 1072278 | 639 | 0.12 | 2.24(1.96,2.56) | ||
| 76-100% | 303744 | 197 | 0.13 | 2.37(1.98,2.84) |
Abbreviation: IR – incidence rate, y – year, HR (95% CI) - hazard ratio (95% confidence interval)
Adjusted for age (continuous), height (continuous), BMI (continuous), hormone therapy (never, <2 years, <5 and ≥2 years, ≥5 years, unknown), parity (0, 1, ≥2 births), duration of breast feeding (none, <6 months, <12 and ≥6 months, ≥12 months, unknown), use of oral contraceptives (never, <1 year, ≥1 year, unknown), age at menarche (<14, ≥14 years), alcohol consumption (never drinker, mild to moderate, heavy), smoking status (never, current, or past), regular exercise (binary), and baseline menopausal status.
Discussion
In this large, nationwide, population-based cohort study, we observed that dense breasts were associated with an increased risk of breast cancer over a 5.5-year follow-up period. South Korean women with heterogeneously dense breasts had 2.1 times the risk of breast cancer, while those with extremely dense breasts had 2.6 times the risk of breast cancer compared to women with almost fatty breasts.
The proportion of dense breasts among South Korean women aged 40 years and older was 40.3 %, comparable to (33,34) or even lower than that reported for women in the p-for United States (35,36). This finding contrasts with previous reports suggesting that Asian women had higher mammographic density compared to Western women (10-13). In our study, mammographic density declined with increasing age, and premenopausal women consistently showed a higher proportion of dense breasts than postmenopausal women, both overall and within the same age groups, aligning with findings from Western studies (35,37). However, the decline in dense breast proportion was more pronounced among South Korean women over 60, unlike the more gradual decrease observed in Western women. Among South Korean women in their 50s, the proportion of dense breasts was 45% compared to 38%-43% (35) and 56% (37) in Western studies. By their 60s, dense breast proportion dropped to 20% among South Korean women, versus 31–33% (35) and 44% (37) in Western women, and in their 70s, it fell further to 8%, compared to 26-28% (35) and 35% (37) in Western cohorts .
In our study, women with dense breasts had a higher risk of developing breast cancer (2.1 times for heterogeneously dense and 2.6 times higher for extremely dense breasts) compared to women with fatty breasts. This association was similar to the result of a cohort study of 74,249 South Korean women, where hazard ratios for heterogeneously dense and extremely dense were 1.96 and 2.86, respectively (21). A meta-analysis reported pooled odds ratios of 1.08 and 1.25 for heterogeneously dense and extremely dense breasts, respectively, compared with fatty breasts in Chinese women, which were weaker than our results (38). Our association was not stronger than the findings among women in the United States (2.0–2.9 times for heterogeneously dense breasts and 2.4–4.0 times for extremely dense breasts) (16,33).
When stratified by menopausal status, the association between dense breasts and breast cancer risk tended to be slightly stronger in postmenopausal than in premenopausal women, although there was no significant interaction. Previous findings were not consistent in their association by menopausal status. There was no difference in association with menopausal status in a study including both Western and Asian women (18). In contrast, a study of Asian women reported a two-fold higher risk of breast cancer for premenopausal women compared to postmenopausal women (39).
In this study, dense breasts were associated with both invasive cancer and DCIS, with no significant difference in association based on tumor invasiveness. Few studies have examined the association between mammographic density and breast cancer risk by tumor invasiveness. A nested case-control study within the Multiethnic Cohort (40) and a pooled case-control study (41) in the United States reported a slightly stronger but generally similar positive association between mammographic density and invasive cancer compared with DCIS; however, these studies used a different mammographic density scale. The associations tended to vary by age: invasive cancer was associated with mammographic density across all age groups, whereas for DCIS, this association was observed only in the age group under 55 years old (41).
In our study, the associations with invasive cancer and DCIS were consistent in both premenopausal and postmenopausal women. By contrast, a case-control study in South Korea using the same BI-RADS mammographic parenchymal composition reported a stronger association for DCIS than for invasive cancer (29). Although our findings on tumor invasiveness differed somewhat from previous studies, our results suggest that DCIS and invasive cancer might share common etiologic factors that interact with mammographic density, potentially influencing the progression of DCIS to invasive cancer (40).
The strengths of this study include a prospective design in a large general population using a well-established longitudinal national database. This large-scale study offers significant clinical implications and epidemiologic evidence regarding mammographic density in Asian women.
Our study had several limitations. First, disease classification relied on medical claims data rather than a cancer registry, which may affect validity. Nonetheless, all medical care providers in South Korea must submit claim data in the form of ICD-10 codes and patients must be registered to receive cancer-related medical service with the government’s financial support, reducing the likelihood of missed or misclassified breast cancer cases. Second, there is some interobserver variability in classifying mammographic density, especially for dense breasts, as observed in a multicenter study (4). However, interobserver agreement was reported to be moderate to substantial, ranging from 0.4 to 0.7 (4,5,42). Even with this variability, the transition from fatty to extremely dense breasts has been well-reflected, and we consistently observe an increased incidence of breast cancer in dense breasts. Future studies should incorporate quantitative measurements of mammographic density to enhance accuracy. Third, our assessment of menopausal status as an effect modifier was based on interaction testing within the Cox model, which may have been underpowered. Alternative methods, such as evaluating effect modification on the additive scale, could provide further insights. Additionally, visualizing confidence interval functions within menopausal strata may help better interpret associations. Fourth, while we adjusted for BMI as a continuous variable, adjustment for BMI modeled as fine categories or alternative adiposity measures, such as waist-to-hip ratio, may better capture adiposity-related confounding. Finally, missing data were handled using the missing indicator method in regression models, which may introduce residual confounding. Future studies should consider multiple imputation to improve data completeness and reduce potential bias.
In conclusion, this study adds to the evidence that mammographic density is an independent risk factor for breast cancer in South Korean women, though the association was not as strong as that observed in Western women. The proportion of dense breasts among South Korean women was similar to or lower than that of Western countries, which may partly account for the lower incidence of breast cancer in Asian women compared to Western women. When developing screening guidelines and personalized screening strategies in Asian countries, it is essential to consider the unique epidemiologic features that distinguish this population, such as the age-specific distribution of dense breasts.
Supplementary Material
Funding
This work was supported, in part, by the Intramural Research Program of the National Institutes of Health, the National Institute of Environmental Health Sciences [Z01-ES044005] (DPS), and by a grant from the Arkansas Breast Cancer Research Program (YMP).
Footnotes
Competing Interests
The authors have no relevant financial or non-financial interests to disclose.
Ethics approval
This study complies with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Seoul St. Mary's Hospital, The Catholic University of Korean (2017/KC17EESI0189).
Consent to participate
Informed consent was waived
Data Availability
The data supproting this study are accessible through the National Health Insurance Sharing Service (https://nhiss.nhis.or.kr/). We accessed the database after submitting the study protocol, the IRB approval document, and the reviewed request form by the committee.
References
- 1.Shin HR, Joubert C, Boniol M, et al. Recent trends and patterns in breast cancer incidence among Eastern and Southeastern Asian women. Cancer Causes Control 2010;21:1777–1785. [DOI] [PubMed] [Google Scholar]
- 2.Youlden DR, Cramb SM, Yip CH, Baade PD. Incidence and mortality of female breast cancer in the Asia-Pacific region. Cancer Biol Med 2014;11:101–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Korean Breast Cancer Society. Breast Cnacer Facts & Figures 2024, Seoul: : Korean Breast Cancer Society, 2024. [Google Scholar]
- 4.Sprague BL, Conant EF, Onega T, et al. Variation in Mammographic Breast Density Assessments Among Radiologists in Clinical Practice: A Multicenter Observational Study. Ann Intern Med 2016;165:457–464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Freer PE. Mammographic breast density: impact on breast cancer risk and implications for screening. Radiographics 2015;35:302–315. [DOI] [PubMed] [Google Scholar]
- 6.McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2006;15:1159–1169. [DOI] [PubMed] [Google Scholar]
- 7.Yaghjyan L, Colditz GA, Rosner B, Tamimi RM. Mammographic breast density and breast cancer risk: interactions of percent density, absolute dense, and non-dense areas with breast cancer risk factors. Breast Cancer Res Treat 2015;150:181–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chen Z, Wu AH, Gauderman WJ, et al. Does mammographic density reflect ethnic differences in breast cancer incidence rates? Am J Epidemiol 2004;159:140–147. [DOI] [PubMed] [Google Scholar]
- 9.Kawahara M, Sato S, Ida Y, et al. Factors influencing breast density in Japanese women aged 40-49 in breast cancer screening mammography. Acta Med Okayama 2013;67:213–217. [DOI] [PubMed] [Google Scholar]
- 10.Rajaram N, Mariapun S, Eriksson M, et al. Differences in mammographic density between Asian and Caucasian populations: a comparative analysis. Breast Cancer Res Treat 2017;161:353–362. [DOI] [PubMed] [Google Scholar]
- 11.Habel LA, Capra AM, Oestreicher N, et al. Mammographic density in a multiethnic cohort. Menopause 2007;14:891–899. [DOI] [PubMed] [Google Scholar]
- 12.del Carmen MG, Hughes KS, Halpern E, et al. Racial differences in mammographic breast density. Cancer 2003;98:590–596. [DOI] [PubMed] [Google Scholar]
- 13.El-Bastawissi AY, White E, Mandelson MT, Taplin S. Variation in mammographic breast density by race. Ann Epidemiol 2001;11:257–263. [DOI] [PubMed] [Google Scholar]
- 14.Chia KS, Reilly M, Tan CS, et al. Profound changes in breast cancer incidence may reflect changes into a Westernized lifestyle: a comparative population-based study in Singapore and Sweden. Int J Cancer 2005;113:302–306. [DOI] [PubMed] [Google Scholar]
- 15.Sung H, Rosenberg PS, Chen WQ, et al. Female breast cancer incidence among Asian and Western populations: more similar than expected. J Natl Cancer Inst 2015;107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Razzaghi H, Troester MA, Gierach GL, Olshan AF, Yankaskas BC, Millikan RC. Mammographic density and breast cancer risk in White and African American Women. Breast Cancer Res Treat 2012;135:571–580. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Boyd NF, Lockwood GA, Byng JW, Tritchler DL, Yaffe MJ. Mammographic densities and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1998;7:1133–1144. [PubMed] [Google Scholar]
- 18.Pettersson A, Graff RE, Ursin G, et al. Mammographic density phenotypes and risk of breast cancer: a meta-analysis. J Natl Cancer Inst 2014;106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Shin J, Lee JE, Ko HY, et al. Association between mammographic density and tumor marker-defined breast cancer subtypes: a case-control study. Eur J Cancer Prev 2018;27:239–247. [DOI] [PubMed] [Google Scholar]
- 20.Kim BK, Choi YH, Nguyen TL, et al. Mammographic density and risk of breast cancer in Korean women. Eur J Cancer Prev 2015;24:422–429. [DOI] [PubMed] [Google Scholar]
- 21.Kim EY, Chang Y, Ahn J, et al. Mammographic breast density, its changes, and breast cancer risk in premenopausal and postmenopausal women. Cancer 2020;126:4687–4696. [DOI] [PubMed] [Google Scholar]
- 22.Kim S, Park B. Association between changes in mammographic density category and the risk of breast cancer: A nationwide cohort study in East-Asian women. Int J Cancer 2021;148:2674–2684. [DOI] [PubMed] [Google Scholar]
- 23.Tran TXM, Kim S, Song H, Lee E, Park B. Association of Longitudinal Mammographic Breast Density Changes with Subsequent Breast Cancer Risk. Radiology 2023;306:e220291. [DOI] [PubMed] [Google Scholar]
- 24.Kim S, Mai Tran TX, Kim MK, et al. Associations between breast cancer risk factors and mammographic breast density in a large cross-section of Korean women. Eur J Cancer Prev 2024;33:407–413. [DOI] [PubMed] [Google Scholar]
- 25.Kang JI, Sung NY, Park SJ, Lee CG, Lee BO. The epidemiology of psychiatric disorders among women with breast cancer in South Korea: analysis of national registry data. Psychooncology 2014;23:35–39. [DOI] [PubMed] [Google Scholar]
- 26.Suh M, Choi KS, Park B, et al. Trends in Cancer Screening Rates among Korean Men and Women: Results of the Korean National Cancer Screening Survey, 2004-2013. Cancer Res Treat 2016;48:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lee J, Lee JS, Park SH, Shin SA, Kim K. Cohort Profile: The National Health Insurance Service-National Sample Cohort (NHIS-NSC), South Korea. Int J Epidemiol 2017;46:e15. [DOI] [PubMed] [Google Scholar]
- 28.Yang HK, Han K, Kwon HS, et al. Obesity, metabolic health, and mortality in adults: a nationwide population-based study in Korea. Sci Rep 2016;6:30329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Park B, Cho HM, Lee EH, et al. Does breast density measured through population-based screening independently increase breast cancer risk in Asian females? Clin Epidemiol 2018;10:61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology 1999;10:37–48. [PubMed] [Google Scholar]
- 31.Gremke N, Griewing S, Kalder M, Kostev K. Positive association between body height and breast cancer prevalence: a retrospective study with 135,741 women in Germany. Breast Cancer Res Treat 2022;196:349–354. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kast K, John EM, Hopper JL, et al. Associations of height, body mass index, and weight gain with breast cancer risk in carriers of a pathogenic variant in BRCA1 or BRCA2: the BRCA1 and BRCA2 Cohort Consortium. Breast Cancer Res 2023;25:72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Vacek PM, Geller BM. A prospective study of breast cancer risk using routine mammographic breast density measurements. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2004;13:715–722. [PubMed] [Google Scholar]
- 34.Bahl M, Baker JA, Bhargavan-Chatfield M, Brandt EK, Ghate SV. Impact of Breast Density Notification Legislation on Radiologists' Practices of Reporting Breast Density: A Multi-State Study. Radiology 2016;280:701–706. [DOI] [PubMed] [Google Scholar]
- 35.Sprague BL, Gangnon RE, Burt V, et al. Prevalence of mammographically dense breasts in the United States. J Natl Cancer Inst 2014;106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kerlikowske K, Zhu W, Hubbard RA, et al. Outcomes of screening mammography by frequency, breast density, and postmenopausal hormone therapy. JAMA Intern Med 2013;173:807–816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Heller SL, Hudson S, Wilkinson LS. Breast density across a regional screening population: effects of age, ethnicity and deprivation. The British journal of radiology 2015;88:20150242–20150242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Bai S, Song D, Chen M, Lai X, Xu J, Dong F. The association between mammographic density and breast cancer risk in Chinese women: a systematic review and meta-analysis. BMC Womens Health 2024;24:131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bae JM, Kim EH. Breast Density and Risk of Breast Cancer in Asian Women: A Meta-analysis of Observational Studies. J Prev Med Public Health 2016;49:367–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Gill JK, Maskarinec G, Pagano I, Kolonel LN. The association of mammographic density with ductal carcinoma in situ of the breast: the Multiethnic Cohort. Breast cancer research : BCR 2006;8:R30–R30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Bertrand KA, Tamimi RM, Scott CG, et al. Mammographic density and risk of breast cancer by age and tumor characteristics. Breast Cancer Res 2013;15:R104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Nicholson BT, LoRusso AP, Smolkin M, Bovbjerg VE, Petroni GR, Harvey JA. Accuracy of assigned BI-RADS breast density category definitions. Acad Radiol 2006;13:1143–1149. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supproting this study are accessible through the National Health Insurance Sharing Service (https://nhiss.nhis.or.kr/). We accessed the database after submitting the study protocol, the IRB approval document, and the reviewed request form by the committee.