Screening mammography frequency following dense breast notification among a predominantly Hispanic/Latina screening cohort.

PURPOSE: Nationally legislated dense breast notification (DBN) informs women of their breast density (BD) and the impact of BD on breast cancer risk and detection, but consequences for screening participation are unclear. We evaluated the association of DBN in New York State (NYS) with subsequent screening mammography in a largely Hispanic/Latina cohort. METHODS: Women aged 40-60 were surveyed in their preferred language (33% English, 67% Spanish) during screening mammography from 2016 to 2018. We used clinical BD classification from mammography records from 2013 (NYS DBN enactment) through enrollment (baseline) to create a 6-category variable capturing prior and new DBN receipt (sent only after clinically dense mammograms). We used this variable to compare the number of subsequent mammograms (0, 1, ≥ 2) from 10 to 30 months after baseline using ordinal logistic regression. RESULTS: In a sample of 728 women (78% foreign-born, 72% Hispanic, 46% high school education or less), first-time screeners and women who received DBN for the first time after prior non-dense mammograms had significantly fewer screening mammograms within 30 months of baseline (Odds Ratios range: 0.33 (95% Confidence Interval (CI) 0.12-0.85) to 0.38 (95% CI 0.17-0.82)) compared to women with prior mammography but no DBN. There were no differences in subsequent mammogram frequency between women with multiple DBN and those who never received DBN. Findings were consistent across age, language, health literacy, and education groups. CONCLUSION: Women receiving their first DBN after previous non-dense mammograms have lower mammography participation within 2.5 years. DBN has limited influence on screening participation of first-time screeners and those with persistent dense mammograms.

Differences in Serum miRNA Profiles by Race, Ethnicity, and Socioeconomic Status: Implications for Developing an Equitable Ovarian Cancer Screening Test.

Serum miRNAs are promising biomarkers for several clinical conditions, including ovarian cancer. To inform equitable implementation of these tests, we investigated the effects of race, ethnicity, and socioeconomic status on serum miRNA profiles. Serum samples from a large institutional biobank were analyzed using a custom panel of 179 miRNA species highly expressed in human serum, measured using the Abcam Fireplex assay via flow cytometry. Data were log-transformed prior to analysis. Differences in miRNA by race and ethnicity were assessed using logistic regression. Pairwise t tests analyzed racial and ethnic differences among eight miRNAs previously associated with ovarian cancer risk. Pearson correlations determined the relationship between mean miRNA expression and the social deprivation index (SDI) for Massachusetts residents. Of 1,586 patients (76.9% white, non-Hispanic), compared with white, non-Hispanic patients, those from other racial and ethnic groups were younger (41.9 years ± 13.2 vs. 51.3 ± 15.1, P < 0.01) and had fewer comorbidities (3.5 comorbidities ± 2.7 vs. 4.6 ± 2.8, P < 0.01). On logistic regression, miRNAs predicted race and ethnicity at an AUC of 0.69 (95% confidence interval, 0.66-0.72), which remained consistent when stratified by most comorbidities. Among eight miRNAs previously associated with ovarian cancer risk, seven significantly varied by race and ethnicity (all P < 0.01). There were no significant differences in SDI for any of these eight miRNAs. miRNA expression is significantly influenced by race and ethnicity, which remained consistent after controlling for confounders. Understanding baseline differences in biomarker test characteristics prior to clinical implementation is essential to ensure instruments perform comparably across diverse populations. PREVENTION RELEVANCE: This study aimed to understand factors affecting miRNA expression, to ensure we create equitable screening tests for ovarian cancer that perform well in diverse populations. The goal is to ensure that we are detecting ovarian cancer cases earlier (secondary prevention) in women of all races, ethnic backgrounds, and socioeconomic means.

Association of census-tract level gentrification and income inequality with 6-year incidence of metabolic syndrome in the Hispanic Community Health Study/Study of Latinos, an epidemiologic cohort study.

BACKGROUND: Metabolic syndrome varies by socio-demographic characteristics, with younger (18-29 years) and older (50-69 years) Hispanic/Latino having higher prevalence compared to other groups. While there is substantial research on neighborhood influences on cardiometabolic health, there are mixed findings regarding the effects of gentrification and few studies have included Hispanic/Latinos. The role of neighborhood income inequality on metabolic health remains poorly understood. OBJECTIVES: Examined associations of neighborhood gentrification and income inequality with metabolic syndrome (MetSyn) using data from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). DESIGN, SETTING AND PARTICIPANTS: The HCHS/SOL is a community-based cohort of adults of Hispanic/Latinos (aged 18-74). Analyses included 6710 adults who did not meet criteria for MetsS at baseline (2008-2011) and completed the visit 2 examination (2014-2017). Poisson regressions estimated odds ratios (IRR) and 95% confidence intervals (CI) for neighborhood gentrification and change in income inequality with MetSyn incidence. MAIN OUTCOME AND EXPOSURE MEASURES: Gentrification was measured with an index that included changes (2000 to 2006-2010) in education, poverty, and income. Change in neighborhood income inequality (2005-2009 to 2012-2016) was measured using the Gini coefficient of income distribution. MetSyn was defined using National Cholesterol Education Program Adult Treatment Panel III criteria. RESULTS: Among 6647 Hispanic/Latino adults, 23% (N = 1530) had incident MetSyn. In models adjusted for socio-demographic, health insurance status, and neighborhood characteristics, gentrification (IRR, 1.00, 95%CI, 0.96-1.03) and income inequality change (IRR, 1.00, 95%CI, 0.99-1.00) were not associated with MetSyn at visit 2. There was no association between cross-sectional income inequality (2005-2009) and MetSyn at visit 2 (IRR, 0.97, 95%CI, 0.82-1.15). CONCLUSION: Neighborhood gentrification and income inequality change were not associated with incidence of MetSyn over 6 years among Hispanic/Latino adults. This study demonstrated that income-based residential changes alone may not be sufficient to explain neighborhood influences on health outcomes among this population.

Scientific uncertainty and perceived mammography benefits in women screened for breast cancer.

PURPOSE: Personal aversion to scientific uncertainty may influence how women perceive the benefits of mammography, a breast cancer screening practice with conflicting scientific opinions and guidelines. Such associations may even exist among women who participate in screening. METHODS: We evaluated the distribution of aversion to ambiguous medical information (AA-Med), using a 6-item scale capturing the level of agreement with statements about obtaining a cancer screening test with conflicting medical recommendations in 665 women (aged 40-60 years; 79.5% Hispanic) recruited during screening mammography appointments in New York City. We assessed the association of AA-Med with perceptions of benefits of mammography (breast cancer mortality reduction, worry reduction, early detection, treatment improvement) using multivariable logistic regression. RESULTS: Over a quarter of participants expressed negative reactions to medical ambiguity about a cancer screening test (e.g., fear, lower trust in experts), but a majority endorsed intention to undergo screening. AA-Med was higher in women who were U.S.-born, non-Hispanic black, and had marginal to adequate health literacy, but there were no differences by clinical factors or screening experiences (e.g., family history, prior breast biopsy). Women with higher AA-Med were more likely to perceive treatment benefits from mammography (OR = 1.37, 95% CI = 0.99-1.90), but AA-Med was not associated with other perceived mammography benefits. CONCLUSIONS: Aversion to uncertainty regarding cancer screening varies by sociodemographic characteristics but has limited associations with perceived mammography benefits in women who already participate in screening.

Breast cancer worry, uncertainty, and perceived risk following breast density notification in a longitudinal mammography screening cohort.

BACKGROUND: Dense breast notification (DBN) legislation aims to increase a woman's awareness of her personal breast density and the implications of having dense breasts for breast cancer detection and risk. This information may adversely affect women's breast cancer worry, perceptions of risk, and uncertainty about screening, which may persist over time or vary by sociodemographic factors. We examined short- and long-term psychological responses to DBN and awareness of breast density (BD). METHODS: In a predominantly Hispanic New York City screening cohort (63% Spanish-speaking), ages 40-60 years, we assessed breast cancer worry, perceived breast cancer risk, and uncertainties about breast cancer risk and screening choices, in short (1-3 months)- and long-term (9-18 months) surveys following the enrollment screening mammogram (between 2016 and 2018). We compared psychological responses by women's dense breast status (as a proxy for DBN receipt) and BD awareness and examined multiplicative interaction by education, health literacy, nativity, and preferred interview language. RESULTS: In multivariable models using short-term surveys, BD awareness was associated with increased perceived risk (odds ratio (OR) 2.27, 95% confidence interval (CI) 0.99, 5.20 for high, OR 2.19, 95% CI 1.34, 3.58 for moderate, vs. low risk) in the overall sample, and with increased uncertainty about risk (OR 1.97 per 1-unit increase, 95% CI 1.15, 3.39) and uncertainty about screening choices (OR 1.73 per 1-unit increase, 95% CI 1.01, 2.9) in Spanish-speaking women. DBN was associated with decreased perceived risk among women with at least some college education (OR 0.32, 95% CI 0.11, 0.89, for high, OR 0.50, 95% CI 0.29, 0.89, for moderate vs. low risk), while those with a high school education or less experienced an increase (OR 3.01, 95% CI 1.05, 8.67 high vs. low risk). There were no associations observed between DBN or BD awareness and short-term breast cancer worry, nor with any psychological outcomes at long-term surveys. CONCLUSIONS: Associations of BD awareness and notification with breast cancer-related psychological outcomes were limited to short-term increases in perceived breast cancer risk dependent on educational attainment, and increases in uncertainty around breast cancer risk and screening choices among Spanish-speaking women.

"It's something I'll do until I die": A qualitative examination into why older women in the U.S. continue screening mammography.

BACKGROUND: Professional guidelines in the U.S. do not recommend routine screening mammography for women ≥75 years with limited life expectancy and/or poor health. Yet, routine mammography remains widely used in older women. We examined older women's experiences, beliefs, and opinions about screening mammography in relation to aging and health. METHODS: We performed thematic analysis of transcribed semi-structured interviews with 19 women who had a recent screening visit at a mammography clinic in New York City (average age: 75 years, 63% Hispanic, 53% ≤high school education). RESULTS: Three main themes emerged: (1) older women typically perceive mammograms as a positive, beneficial, and routine component of care; (2) participation in routine mammography is reinforced by factors at interpersonal, provider, and healthcare system levels; and (3) older women do not endorse discontinuation of screening mammography due to advancing age or poor health, but some may be receptive to reducing screening frequency. Only a few older women reported having discussed mammography cessation or the potential harms of screening with their providers. A few women reported they would insist on receiving mammography even without a provider recommendation. CONCLUSIONS: Older women's positive experiences and views, as well as multilevel and frequently automated cues toward mammography are important drivers of routine screening in older women. These findings suggest a need for synergistic patient, provider, and system level strategies to reduce mammography overuse in older women.

Current regular aspirin use and mammographic breast density: a cross-sectional analysis considering concurrent statin and metformin use.

PURPOSE: The nonsteroidal anti-inflammatory drug aspirin is an agent of interest for breast cancer prevention. However, it is unclear if aspirin affects mammographic breast density (MBD), a marker of elevated breast cancer risk, particularly in the context of concurrent use of medications indicated for common cardiometabolic conditions, which may also be associated with MBD. METHODS: We used data from the New York Mammographic Density Study for 770 women age 40-60 years old with no history of breast cancer. We evaluated the association between current regular aspirin use and MBD, using linear regression for continuous measures of absolute and percent dense areas and absolute non-dense area, adjusted for body mass index (BMI), sociodemographic and reproductive factors, and use of statins and metformin. We assessed effect modification by BMI and reproductive factors. RESULTS: After adjustment for co-medication, current regular aspirin use was only positively associated with non-dense area (ß = 18.1, 95% CI: 6.7, 29.5). Effect modification by BMI and parity showed current aspirin use to only be associated with larger non-dense area among women with a BMI ≥ 30 (ß = 28.2, 95% CI: 10.8, 45.7), and with lower percent density among parous women (ß = -3.3, 95% CI: -6.4, -0.3). CONCLUSIONS: Independent of co-medication use, current regular aspirin users had greater non-dense area with stronger estimates for women with higher BMI. We found limited support for an association between current aspirin use and mammographically dense breast tissue among parous women.

A mixed-methods study of multi-level factors influencing mammography overuse among an older ethnically diverse screening population: implications for de-implementation.

BACKGROUND: There is growing concern that routine mammography screening is overused among older women. Successful and equitable de-implementation of mammography will require a multi-level understanding of the factors contributing to mammography overuse. METHODS: This explanatory, sequential, mixed-methods study collected survey data (n= 52, 73.1% Hispanic, 73.1% Spanish-speaking) from women ≥70 years of age at the time of screening mammography, followed by semi-structured interviews with a subset of older women completing the survey (n=19, 63.2% Hispanic, 63.2% Spanish-speaking) and providers (n=5, 4 primary care, 1 obstetrics and gynecology) to better understand multi-level factors influencing mammography overuse and inform potential de-implementation strategies. We conducted a descriptive analysis of survey data and content analysis of qualitative interview data. Survey and interview data were examined separately, compared, integrated, and organized according to Norton and Chambers Continuum of Factors Influencing De-Implementation Process. RESULTS: Survey findings show that 87.2% of older women believe it is important to plan for an annual mammogram, 80.8% received a provider recommendation, and 78.9% received a reminder in the last 12 months to schedule a mammogram. Per interviews with older women, the majority were unaware of or did not perceive to have experienced overuse and intended to continue mammography screening. Findings from interviews with older women and providers suggest that there are multiple opportunities for older women to obtain a mammogram. Per provider interviews, almost all reported that reducing overuse was not viewed as a priority by the system or other providers. Providers also discussed that variation in mammography screening practices across providers, fear of malpractice, and monetary incentives may contribute to overscreening. Providers identified potential strategies to reduce overscreening including patient and provider education around harms of screening, leveraging the electronic health record to identify women who may receive less health benefit from screening, customizing system-generated reminder letters, and organizing workgroups to develop standard processes of care around mammography screening. CONCLUSIONS: Multi-level factors contributing to mammography overuse are dynamic, interconnected, and reinforced. To ensure equitable de-implementation, there is a need for more refined and empirical testing of theories, models, and frameworks for de-implementation with a strong patient-level component that considers the interplay between multilevel factors and the larger care delivery process.

Breast Density Awareness and Knowledge in a Mammography Screening Cohort of Predominantly Hispanic Women: Does Breast Density Notification Matter?

BACKGROUND: New York State law mandates that women with dense breasts receive a written notification of their breast density (BD) and its implications, but data on the impact of dense breast notification (DBN) on BD awareness and knowledge in diverse populations remain limited. METHODS: Between 2016 and 2018, we collected survey and mammographic data from 666 women undergoing screening mammography in New York City (ages 40-60, 80% Hispanic, 69% Spanish-speaking) to examine the impact of prior DBN on BD awareness by sociodemographic and breast cancer risk factors, and describe BD knowledge by sources of information. RESULTS: Only 24.8% of the overall sample and 34.9% of women receiving DBN had BD awareness. In multivariable models adjusting for DBN, awareness was significantly lower in women who were Spanish-speaking [OR, 0.16; 95% confidence interval (CI), 0.09-0.30 vs. English speakers], were foreign-born (OR, 0.31; 95% CI, 0.16-0.58 vs. U.S.-born), and had lower educational attainment (e.g., high school degree or less; OR, 0.14; 95% CI, 0.08-0.26 vs. college or higher degree). Women receiving DBN were more likely to be aware of BD (OR, 2.61; 95% CI, 1.59-4.27) but not more knowledgeable about the impact of BD on breast cancer risk and detection. However, women reporting additional communication about their BD showed greater knowledge in these areas. CONCLUSIONS: DBN increases BD awareness disproportionately across sociodemographic groups. IMPACT: Efforts to improve communication of DBN must focus on addressing barriers in lower socioeconomic and racially and ethnically diverse women, including educational and language barriers.

Influence of Childhood Adversity and Infection on Timing of Menarche in a Multiethnic Sample of Women.

Childhood adversities (CAs) and infections may affect the timing of reproductive development. We examined the associations of indicators of CAs and exposure to tonsillitis and infectious mononucleosis (mono) with age at menarche. A multiethnic cohort of 400 women (ages 40-64 years) reported exposure to parental maltreatment and maladjustment during childhood and any diagnosis of tonsillitis and/or mono; infections primarily acquired in early life and adolescence, respectively. We used linear and relative risk regression models to examine the associations of indicators of CAs individually and cumulatively, and history of tonsillitis/mono with an average age at menarche and early onset of menarche (<12 years of age). In multivariable models, histories of mental illness in the household (RR = 1.44, 95% CI: 1.01-2.06), and tonsillitis diagnosis (RR = 1.67, 95% CI: 1.20-2.33) were associated with early menarche (<12 years), and with an earlier average age at menarche by 7.1 months (95% CI: -1.15, -0.02) and 8.8 months (95% CI: -1.26, -0.20), respectively. Other adversities indicators, cumulative adversities, and mono were not statistically associated with menarcheal timing. These findings provided some support for the growing evidence that early life experiences may influence the reproductive development in girls.

Independent and joint cross-sectional associations of statin and metformin use with mammographic breast density.

BACKGROUND: Well-tolerated and commonly used medications are increasingly assessed for reducing breast cancer risk. These include metformin and statins, both linked to reduced hormone availability and cell proliferation or growth and sometimes prescribed concurrently. We investigated independent and joint associations of these medications with mammographic breast density (MBD), a useful biomarker for the effect of chemopreventive agents on breast cancer risk. METHODS: Using data from a cross-sectional study of 770 women (78% Hispanic, aged 40-61 years, in a mammography cohort with high cardiometabolic burden), we examined the association of self-reported "ever" use of statins and metformin with MBD measured via clinical Breast Imaging Reporting and Data System (BI-RADS) density classifications (relative risk regression) and continuous semi-automated percent and size of dense area (Cumulus) (linear regression), adjusted for age, body mass index, education, race, menopausal status, age at first birth, and insulin use. RESULTS: We observed high statin (27%), metformin (13%), and combination (9%) use, and most participants were overweight/obese (83%) and parous (87%). Statin use was associated with a lower likelihood of high density BI-RADS (RR = 0.60, 95% CI = 0.45 to 0.80), percent dense area (PD) (ß = - 6.56, 95% CI = - 9.05 to - 4.06), and dense area (DA) (ß = - 9.05, 95% CI = - 14.89 to - 3.22). Metformin use was associated with lower PD and higher non-dense area (NDA), but associations were attenuated by co-medication with statins. Compared to non-use of either medication, statin use alone or with metformin were associated with lower PD and DA (e.g., ß = - 6.86, 95% CI: - 9.67, - 4.05 and ß = - 7.07, 95% CI: - 10.97, - 3.17, respectively, for PD) and higher NDA (ß = 25.05, 95% CI: 14.06, 36.03; ß = 29.76, 95% CI: 14.55, 44.96, respectively). CONCLUSIONS: Statin use was consistently associated with lower MBD, measured both through clinical radiologist assessment and continuous relative and absolute measures, including dense area. Metformin use was associated with lower PD and higher NDA, but this may be driven by co-medication with statins. These results support that statins may lower MBD but need confirmation with prospective and clinical data to distinguish the results of medication use from that of disease.

Associations of Nativity, Age at Migration, and Percent of Life in the U.S. with Midlife Body Mass Index and Waist Size in New York City Latinas.

Migration to the U.S. has been associated with increased body size and obesity risk in Latinas, but results for Caribbean immigrant women are limited and inconclusive. Emerging evidence also suggests that early-life environment associations with women's midlife body mass index (BMI) may be different for larger and smaller women, but this has not been tested within migration life-course history. We examined the associations of nativity and migration timing with midlife body size in a sample of majority Caribbean Latinas and whether these associations varied across the body size distribution. We used interview data from 787 self-identified Latinas (ages 40-65 years) and assessed overall obesity using BMI (kg/m2) and central obesity based on waist circumference (WC, cm). We used linear and quantile regression to examine the association of migration history with BMI and WC and logistic regression for the probability of obesity. Foreign birthplace, later migration age, and lower percent of life in the U.S. were associated with lower BMI and WC means and lower odds of overall and central obesity. Quantile regression showed only inverse associations in the upper quantiles of BMI and WC. For example, relative to U.S.-born women, women living <50% of their lives in the U.S. had lower BMI in the 75th BMI percentile (ß = -4.10, 95% CI: -6.75, -0.81), with minimal differences in the 25th (ß = 0.04, 95% CI: -1.01, 0.96) and 50th BMI percentiles (ß = -1.54, 95% CI: -2.90, 0.30). Our results support that migration to and increasing time in the U.S. are associated with greater body size in midlife Latina women, with stronger influences at higher body size distribution.

"You probably can't feel as safe as normal women": Hispanic women's reactions to breast density notification.

BACKGROUND: Patient advocacy has led to state-level legislative mandates for the release of personal mammographic breast density information to women undergoing screening mammography. More research is needed to understand the impact of this information on women's perceptions and mammography screening behavior. METHODS: Semistructured interviews were conducted in English and Spanish with 24 self-identified Hispanic women who had undergone at least 1 mammogram since breast density notification was enacted in New York State. The women ranged in age from 43 to 63 years. Women were asked about their understanding and perceptions of the communication of New York State-mandated breast density information, and any actions they have taken or would take in response to this information. A content analysis of the qualitative data from the translated and transcribed interviews was conducted. RESULTS: The majority of participants had no prior knowledge of breast density and expressed confusion and apprehension regarding the meaning of dense breasts when presented with the notification information. Many participants understood having dense breasts to be a serious and abnormal condition, and reported feelings of worry and vulnerability. Participants mostly expressed a strong interest in learning about breast density and obtaining additional and more frequent breast cancer screening tests. These behavioral intentions were consistent with participants' overall favorable view of breast cancer screening and a belief that their faith, as well as regular screening, can help to protect them from breast cancer morbidity and mortality. CONCLUSIONS: Hispanic women conveyed proactive breast cancer screening intentions in response to breast density notification, despite inadequate comprehension of this information and negative emotional responses.

Childhood body size and midlife mammographic breast density in foreign-born and U.S.-born women in New York City.

PURPOSE: We investigated whether childhood body size is associated with midlife mammographic density, a strong risk factor for breast cancer. METHODS: We collected interview data, including body size at age 10 years using a pictogram, and measured height and weight from 518 women, recruited at the time of screening mammography in New York City (ages 40-64 years, 71% Hispanic, 68% foreign-born). We used linear regression models to examine childhood body size in relation to percent density and areas of dense and nondense tissue, measured using a computer-assisted method from digital mammograms. RESULTS: In models that adjusted for race/ethnicity, and age and body mass index at mammogram, the heaviest relative to leanest childhood body size was associated with 5.94% lower percent density (95% confidence interval [CI]: -9.20, -2.29), 7.69 cm2 smaller dense area (95% CI: -13.94, -0.63), and 26.17 cm2 larger nondense area (95% CI: 9.42, 43.58). In stratified analysis by menopausal status and nativity, the observed associations were stronger for postmenopausal and U.S.-born women although these differences did not reach statistical significance. CONCLUSIONS: Heavy childhood body size is associated with lower mammographic density, consistent with its associations with breast cancer risk. Suggestive findings by nativity require confirmation in larger samples.

Migration History, Language Acculturation, and Mammographic Breast Density.

Background: Breast cancer incidence is lower in many U.S. ethnic minority and foreign-born population groups. Investigating whether migration and acculturation patterns in risk are reflected in disease biomarkers may help to elucidate the underlying mechanisms.Methods: We compared the distribution of breast cancer risk factors across U.S.-born white, African American and Hispanic women, and foreign-born Hispanic women (n = 477, ages 40-64 years, 287 born in Caribbean countries). We used linear regression models to examine the associations of migration history and linguistic acculturation with mammographic breast density (MBD), measured using computer-assisted methods as percent and area of dense breast tissue.Results: The distribution of most breast cancer risk factors varied by ethnicity, nativity, and age at migration. In age- and body mass index-adjusted models, U.S.-born women did not differ in average MBD according to ethnicity, but foreign-born Hispanic women had lower MBD [e.g., -4.50%; 95% confidence interval (CI), -7.12 to -1.89 lower percent density in foreign- vs. U.S.-born Hispanic women]. Lower linguistic acculturation and lower percent of life spent in the United States were also associated with lower MBD [e.g., monolingual Spanish and bilingual vs. monolingual English speakers, respectively, had 5.09% (95% CI, -8.33 to -1.85) and 3.34% (95% CI, -6.57 to -0.12) lower percent density]. Adjusting for risk factors (e.g., childhood body size, parity) attenuated some of these associations.Conclusions: Hispanic women predominantly born in Caribbean countries have lower MBD than U.S.-born women of diverse ethnic backgrounds, including U.S.-born Hispanic women of Caribbean heritage.Impact: MBD may provide insight into mechanisms driving geographic and migration variations in breast cancer risk. Cancer Epidemiol Biomarkers Prev; 27(5); 566-74. ©2018 AACR.