ATLANTA -- The first breast cancer risk-prediction model based entirely on data from Hispanic women, including whether a woman was born in or outside of the United States, provided a more accurate assessment of Hispanic women's risk of developing breast cancer compared with existing models based on data from non-Hispanic women, according to a study presented at the Eighth American Association for Cancer Research (AACR) Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved, held Nov. 13-16.
"Currently, there is no breast cancer risk-prediction model for Hispanic women," said Matthew P. Banegas, PhD, MPH, investigator with Kaiser Permanente Center for Health Research in Portland, Oregon, and primary author of the study. "We developed a model based on data on ethnicity, nativity, and breast cancer risk factors, as well as incidence and mortality rates in Hispanic women, which allowed us to create a more specific tool to predict their risk of developing invasive breast cancer."
Presently, physicians and researchers use the Breast Cancer Risk Assessment Tool (BRCAT) from the National Cancer Institute (NCI) to estimate risk, counsel patients, and design breast cancer prevention trials. However, since it is based, in part, on data from women of other races and ethnicities, it does not accurately reflect the risk of breast cancer in Hispanic women, and tends to underestimate their risk, Banegas said.
"The goal of our work is to enable Hispanic women to better understand their risk of developing invasive breast cancer. They will be able to discuss this information with their physician and what it means for them specifically," added Banegas.
Factors that are incorporated into the new prediction model include:
- A woman's age at first full-term pregnancy: Women who have children at younger ages tend to have a lower risk of breast cancer. Studies show that Hispanic women born outside the United States tend to start having children at a younger age than Hispanic women born in the United States.
- A woman's age at first menstrual period: The younger a woman is when she starts menstruating, the greater her lifetime exposure to estrogen, which has been shown to increase breast cancer risk. Prior research has shown that Hispanic women born outside the United States may be older when they start menstruating than Hispanic women born in the United States.
- Having had a biopsy for benign breast disease: Breast cancer risk is increased among women with benign breast disease. In the risk-prediction model, the risk associated with this factor was slightly greater for Hispanic women born outside the United States than for Hispanic women born in the United States.
- Family history of breast cancer in first-degree relatives: Women with a family history of breast cancer have higher risk of developing breast cancer. Prior studies show that Hispanic women born outside the United States are less likely to have a family history of breast cancer compared with Hispanic women born in the United States.
Banegas and colleagues used data from the San Francisco Bay Area Breast Cancer Study, focusing on 1,086 Hispanic women with breast cancer and 1,411 without breast cancer, to develop a Hispanic-specific breast cancer risk-prediction model. They separated the women into two groups: those who were born in the United States and those who were born outside the United States, then estimated risks for both groups, applying their estimates to incidence and mortality data from the California Cancer Registry and the NCI's Surveillance, Epidemiology, and End Results (SEER) program. The researchers validated their prediction model, in part, against data from Hispanic women in the Women's Health Initiative, and found that the model was well calibrated for Hispanic women born in the United States, but overestimated the risk in foreign-born Hispanic women. Prior research has shown that foreign-born Hispanic women have about half the breast cancer risk of U.S.-born Hispanic women.
Since the model was developed using data from women in the San Francisco Bay area, it will be most applicable to women in that region, Banegas said. As researchers gather more data from Hispanic women in other parts of the United States and from those born outside the United States, those data should be incorporated into the model to increase the accuracy for those populations.
###
To interview Matthew P. Banegas, contact Julia Gunther at julia.gunther@aacr.org or 215-446-6896.
This study was supported by the NCI. Banegas declares no conflicts of interest.
Follow the meeting on Twitter: #AACRdisp15
Follow us: Cancer Research Catalyst http://blog.aacr.org; Twitter @AACR; and Facebook http://www.facebook.com/aacr.org
About the American Association for Cancer Research
Founded in 1907, the American Association for Cancer Research (AACR) is the world's oldest and largest professional organization dedicated to advancing cancer research and its mission to prevent and cure cancer. AACR membership includes more than 35,000 laboratory, translational, and clinical researchers; population scientists; other health care professionals; and patient advocates residing in 101 countries. The AACR marshals the full spectrum of expertise of the cancer community to accelerate progress in the prevention, biology, diagnosis, and treatment of cancer by annually convening more than 30 conferences and educational workshops, the largest of which is the AACR Annual Meeting with nearly 19,300 attendees. In addition, the AACR publishes eight prestigious, peer-reviewed scientific journals and a magazine for cancer survivors, patients, and their caregivers. The AACR funds meritorious research directly as well as in cooperation with numerous cancer organizations. As the Scientific Partner of Stand Up To Cancer, the AACR provides expert peer review, grants administration, and scientific oversight of team science and individual investigator grants in cancer research that have the potential for near-term patient benefit. The AACR actively communicates with legislators and other policymakers about the value of cancer research and related biomedical science in saving lives from cancer. For more information about the AACR, visit http://www.AACR.org.
Abstract: A09
Title: Projecting Individualized Absolute Invasive Breast Cancer Risk in Hispanic Women
Authors: Matthew P. Banegas1, Esther M. John2, Marty Slattery3, Scarlett L. Gomez2, Mandi Yu4, Andrea LaCroix5, David Pee6, Mitchell H. Gail4. 1Kaiser Permanente, Center for Health Research, Portland, OR, 2Cancer Prevention Institute of California, Fremont, Ca, 3University of Utah, School of Medicine, Salt Lake City, UT, 4National Cancer Institute, Bethesda, MD, 5University of California San Diego, School of Medicine, San Diego, CA, 6Information Management Services Inc., Rockville, MD.
Background: The Breast Cancer Risk Assessment Tool (BCRAT) of the National Cancer Institute (NCI) is widely used for estimating absolute risk of invasive breast cancer, counseling women about breast cancer risk and designing breast cancer prevention trials. However, BCRAT absolute risk estimates for Hispanic women rely in part on data from non-Hispanic white women. We developed a model for projecting absolute invasive breast cancer risk in Hispanic women, based on nativity-specific (United States (US)-born versus foreign-born) case-control data, incidence and competing mortality rates in Hispanic women.
Methods: The Breast Cancer Health Disparities (BCHD) Study includes participants from two US population-based case-control studies: the San Francisco Bay Area Breast Cancer Study (SFBCS) and the 4-Corners Breast Cancer Study (4-CBCS). Data from 1086 Hispanic women with breast cancer (cases: US-born=533; foreign-born=553) and 1411 women without breast cancer (controls: US-born=464; foreign-born=947) in the SFBCS were used to estimate relative risks (RRs) and attributable risk (AR), separately for US-born and foreign-born women, for age at first full term pregnancy (<20 years (yrs) (ref); 20-29 yrs; ?30 yrs/nulliparous), age at menarche (?14 yrs (ref); 12-13 yrs; <12 yrs), history of breast cancer in 1st-degree female relatives (No (ref); Yes), and previous biopsy for benign breast disease (No (ref); Yes). Nativity-specific absolute risks were estimated by combining RR and AR information from the SFBCS with nativity-specific invasive breast cancer incidence and competing mortality rates from the California Cancer Registry and NCI's Surveillance, Epidemiology and End Results (SEER) program, respectively. The resulting model includes separate calculations for US-born and foreign-born Hispanic women. We evaluated the RR feature by comparing the model RRs to those estimated for Hispanic women in data from the 4-CBCS (cases=731, controls=836). Data from Hispanic women in the Women's Health Initiative (WHI, n=6220) were used to assess model calibration and discriminatory accuracy, by estimating the ratio of observed to expected number of breast cancers (O/E ratio) and area under the curve (AUC), respectively.
Results: Risk factors for the US-born Hispanic model included age at first full term pregnancy (RR=1.26), biopsy for benign breast disease (RR=1.10) and family history of breast cancer (RR=1.18); risk factors for the foreign-born Hispanic model included age at first full term pregnancy (RR=1.60), age at menarche (RR=1.30), biopsy for benign breast disease (RR=1.62) and family history of breast cancer (RR=2.48). For most US-born Hispanic women, the model estimated lower risk projections than the BCRAT; for foreign-born Hispanic women, the model estimated higher risk projections than the BCRAT in most instances. There were no significant differences between RRs estimated from 4-CBCS and those used in the US-born Hispanic model; however, the RRs used in the foreign-born Hispanic model were significantly higher than those from 4-CBCS (p<0.05). The model was well calibrated for US-born women (O/E ratio = 1.07; 95% Confidence Interval (CI): 0.82-1.40), but overestimated risk in foreign-born women (O/E ratio=0.66; 95% CI: 0.41-1.06), in independent data from the WHI. However, overestimation of risk in WHI foreign-born Hispanic women (with only 17 cases) was not statistically significant. The discriminatory accuracy of the model was modest for both US-born Hispanics (AUC = 0.564) and foreign-born Hispanics (AUC=0.625).
Conclusion: This is the first absolute risk model for Hispanic women that is based entirely on data from Hispanic women. A unique feature of the model is the incorporation of nativity-specific SEER incidence rates, which are lower in foreign-born than in US-born Hispanic women. Although the model has been evaluated in independent datasets, further studies in Hispanic women are warranted to evaluate its validity.