The integration of risk assessment into clinical breast screening holds promise in decreasing breast cancer morbidity and mortality and increasing health care efficiency. Currently, clinical recommendations regarding risk counseling, screening, and chemoprevention are being made based on a woman’s personal risk. One of the criteria that can be used to categorize a woman as “increased risk” is her projected 5-year risk for invasive breast cancer as determined by one of various risk models which are heavily based on family history. We hypothesized that the frequency of screening mammography patients at our medical institution who would be considered at “increased risk” would be different according to different risk models and according to race/ethnicity, thus impacting the volume of patients who are targeted for risk-reducing intervention. Risk scores were calculated for 307 White, Hispanic, and Asian screening mammography patients according to the Gail, BCSC, and Tyrer-Cuzick models. Scores were compared within and between race/ethnicities, according to the different models, individually and in combination. As expected, White women had higher risk scores than Hispanic and Asian women according to all models tested (p<0.05), and a higher percent of White women were categorized as “increased risk” (p<0.0001). However, the correlations between models were moderate, resulting in inconsistencies of increased risk status for many women. Depending on the volume of patients undergoing risk assessment and the resources of staff and services providing the risk counseling and other downstream services, a prevention program may opt to use a combination of risk models suitable for their patient population instead of just one risk model.

Abstract: The USPSTF recommends that women who are at increased risk for breast cancer and at low risk for adverse medication effects should be offered risk-reducing medications, such as tamoxifen or raloxifene, by their clinicians. The NCCN also recommends risk counseling for women with a 5-year risk of ≥1.7% as calculated by the NCI-developed Breast Cancer Risk Assessment Tool (BCRAT, based on the Gail model) or other risk model. The integration of risk assessment into clinical breast screening holds promise in reducing breast cancer risk for many women; however, the criteria that contribute to elevated risk status are different according to different risk models. The University of California, Irvine (UCI) Athena Breast Health Network has integrated a computerized risk assessment largely based on the BCRAT into the screening mammography process at UCI mammography centers. Women identified to be at elevated risk are provided with personalized risk counseling by a Breast Health Specialist. In recent years, breast density has been increasingly recognized as a risk factor for breast cancer; however, it is not part of the BCRAT. In this pilot analysis, we sought to determine if incorporating breast density into the risk assessment program would affect the % of women in our cohort who have a 5-year risk of ≥1.7%. We hypothesized that an increased % of Asian women would reach this threshold since breast density is known to be higher in Asians. We used the BCSC model, which includes density as a variable, to calculate risk scores for a subset of our screening population (n=309), age-matched for three race/ethnic groups, and compared them to their scores according to the BCRAT. Our results showed that while Asians did exhibit significantly higher breast density (30.3% had BIRADS breast density classification 4, extremely dense) than White and Hispanic women (8.7% and 5.6%, respectively), their BCRAT and BCSC scores were significantly lower than in White women but higher than in Hispanic women, with 14.1% of Asians having a BCSC score ≥1.7% compared to 43.7% of Whites and 10.3% in Hispanics, and 17.2% of Asians having a BCRAT score of ≥1.7% compared to 34.0% of Whites and 6.5% of Hispanics. Interestingly, when differences were explored between BCRAT and BCSC scores within the same sub-group of women, the scores were not statistically different among White or Hispanic women but were approaching statistical significance in Asian women. Contrary to our hypothesis, in Asian women, BCSC-calculated risk was actually lower than according to the BCRAT. These findings suggest that, despite higher breast density, Asian women still exhibit lower breast cancer risk scores than White women; thus, a smaller proportion of Asian women in our cohort will be recommended for high risk counseling and chemoprevention. Citation Format: Hannah Lui Park, Stephanie M. Tran, Jennifer Lee, Deborah Goodman, Argyrios Ziogas, Richard Kelly, Kathryn M. Larsen, Andrea Alvarez, Chris Tannous, Julie Strope, Wendy Lynch, Hoda Anton-Culver. A pilot study comparing breast cancer risk scores using models with and without breast density among women of different race/ethnicities undergoing breast screening in the University of California, Irvine Athena Breast Health Network cohort. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3243. doi:10.1158/1538-7445.AM2014-3243

Background

Breast density is an important risk factor for breast cancer and is known to be associated with characteristics such as age, race, and hormone levels; however, it is unclear what factors contribute to changes in breast density in postmenopausal women over time. Understanding factors associated with density changes may enable a better understanding of breast cancer risk and facilitate potential strategies for prevention.

Methods

This study investigated potential associations between personal factors and changes in mammographic density in a cohort of 3,392 postmenopausal women with no personal history of breast cancer between 2011 and 2017. Self-reported information on demographics, breast and reproductive history, and lifestyle factors, including body mass index (BMI), alcohol intake, smoking, and physical activity, was collected by an electronic intake form, and breast imaging reporting and database system (BI-RADS) mammographic density scores were obtained from electronic medical records. Factors associated with a longitudinal increase or decrease in mammographic density were identified using Fisher's exact test and multivariate conditional logistic regression.

Results

7.9% of women exhibited a longitudinal decrease in mammographic density, 6.7% exhibited an increase, and 85.4% exhibited no change. Longitudinal changes in mammographic density were correlated with age, race/ethnicity, and age at menopause in the univariate analysis. In the multivariate analysis, Asian women were more likely to exhibit a longitudinal increase in mammographic density and less likely to exhibit a decrease compared to White women. On the other hand, obese women were less likely to exhibit an increase and more likely to exhibit a decrease compared to normal weight women. Women who underwent menopause at age 55 years or older were less likely to exhibit a decrease in mammographic density compared to women who underwent menopause at a younger age. Besides obesity, lifestyle factors (alcohol intake, smoking, and physical activity) were not associated with longitudinal changes in mammographic density.

Conclusions

The associations we observed between Asian race/obesity and longitudinal changes in BI-RADS density in postmenopausal women are paradoxical in that breast cancer risk is lower in Asian women and higher in obese women. However, the association between later age at menopause and a decreased likelihood of decreasing in BI-RADS density over time is consistent with later age at menopause being a risk factor for breast cancer and suggests a potential relationship between greater cumulative lifetime estrogen exposure and relative stability in breast density after menopause. Our findings support the complexity of the relationships between breast density, BMI, hormone exposure, and breast cancer risk.