Genomic Disparities in Breast Cancer Among Latinas

Abstract

Background

Breast cancer is the most common cancer diagnosed among Latinas in the United States and the leading cause of cancer-related death among this population. Latinas tend to be diagnosed at a later stage and have worse prognostic features than their non-Hispanic white counterparts. Genetic and genomic factors may contribute to observed breast cancer health disparities in Latinas.

Methods

We provide a landscape of our current understanding and the existing gaps that need to be filled across the cancer prevention and control continuum.

Results

We summarize available data on mutations in high and moderate penetrance genes for inherited risk of breast cancer and the associated literature on disparities in awareness of and uptake of genetic counseling and testing in Latina populations. We also discuss common genetic polymorphisms and risk of breast cancer in Latinas. In the treatment setting, we examine tumor genomics and pharmacogenomics in Latina patients with breast cancer.

Conclusions

As the US population continues to diversify, extending genetic and genomic research into this underserved and understudied population is critical. By understanding the risk of breast cancer among ethnically diverse populations, we will be better positioned to make treatment advancements for earlier stages of cancer, identify more effective and ideally less toxic treatment regimens, and increase rates of survival.

Introduction

An estimated 55 million individuals living in the United States identify as being Hispanic or Latino.¹ Latinos are a culturally and genetically diverse group with origins in Mexico, the Caribbean, Central American, and South America. In the United States, 64.0% of Latinos are of Mexican background, 9.6% of Puerto Rican background, 3.8% of Salvadoran background, 3.7% of Cuban background, 3.2% of Dominican background, 2.4% of Guatemalan background, and the remainder are of other origins.² Although the terms Hispanic and Latino/Latina are often interchangeably used, we selected the term Latina for the current manuscript as we feel it extends beyond spoken language to reflect both origin and cultural traditions of women from Latin America.

Breast cancer is the most common cancer diagnosed among Latinas in the United States and is the leading cause of cancer-related death in this population.¹⁰ Although the overall prevalence of breast cancer in Latinas is lower than in non-Hispanic whites, Latinas tend to be diagnosed at a later stage and have worse prognostic features (eg, triple negative disease and HER2–positive disease).³ A myriad of socioeconomic and cultural factors contribute to health disparities in breast cancer among Latinas,^4–6 but biological factors — particularly genomics — remain an important but understudied consideration.

High and Moderate Penetrance Genes

Approximately 10% to 15% of breast cancer cases are attributed to inherited gene mutations.⁷ Although multiple genes confer an inherited risk fir cancer,⁸ BRCA mutations are the most prevalent and penetrant mutations, accounting for the majority of hereditary types of breast cancer.⁹ BRCA mutations result in an increased lifetime risk of breast cancer of up to approximately 60% to 70% and a lifetime ovarian cancer risk of up to 40%.^10–12 Among Latinas, breast cancer is often diagnosed at younger ages and with worse prognostic features, including increased rates of triple-negative disease, than their non-Hispanic white counterparts.^13–16 Triple-negative disease and premenopausal breast cancer are both clinical characteristics associated with a higher probability of having a BRCA1/2 mutation.^17,18

Prevalence of BRCA

The prevalence of BRCA mutations in the general US population is estimated to be 1 in 400, excluding women of Ashkenazi Jewish descent in whom prevalence is 1 in 40.^19–21 However, less is known about the prevalence among racial and ethnic minority groups, including Latinas as a whole or by subethnicity based on country of origin. A review examined the spectrum of BRCA1 and BRCA2 mutations in Latin America and the Caribbean using studies published between the years 1994 and 2015.²² Six of the 33 studies were conducted among Latina living in the United States, with the vast majority of participants drawn from clinic-based samples of patients of Mexican origin with breast cancer residing in California, Arizona, and Texas.²² Prevalence estimates of carrying a BRCA mutation for this US Latina group ranged from 0.7% to 42% and varied based on whether cases were selected or unselected for family history or clinical characteristics (eg, affected vs unaffected, age at diagnosis), cancer site (eg, breast, ovarian), and type of testing (eg, inclusion of large rearrangement testing).²² In the cohorts of unselected patients with breast cancer, the BRCA mutation prevalence was 1.2% to 4.9%, which was consistent with expected rates.²²

BRCA mutations have also been documented in all residents of Latin American countries where these genes have been studied, including Argentina, Brazil, Chile, Colombia, Costa Rica, Cuba, Mexico, Peru, Puerto Rico, Uruguay, and Venezuela.^23–54 Most studies have focused on the spectrum of BRCA mutations.^22,55 In a review of BRCA1 and BRCA2 mutations in persons living in Latin America and the Caribbean, 36% of the 33 studies primarily focused on Mexican or Mexican American patients.²² Of the Mexican study population, the mutation prevalence was between 4.3% and 23.0%.²² For other Latina subethnic groups, the mutation prevalence estimates of each country studied were: Colombia (1.2%–15.6%; 2 studies), Costa Rica (4.5%; 1 study), Cuba (2.6%; 1 study), Peru (4.9%; 1 study), Uruguay (17%; 1 study), and Venezuela (17.2%; 1 study).²² These studies provide insight into areas of future research of BRCA mutation distribution and frequency based on country of origin, the role of specific founder mutations, the contribution of large genomic rearrangements to the spectrum of mutations across various Latina subethnic groups, and the consideration of other non-BRCA genes that increase the risk of breast cancer.

Although recurrent mutations were identified within most studies, the specific mutation varied by study and country.²² BRCA1 185delAG has also been documented in Latinas across Latin America and the United States.^{43,45–47,50,56–58} One of the 3 Jewish founder mutations, BRCA1 185delAG is estimated to have arisen about 800 years ago or earlier and is believed to have been introduced into Latin America about 650 years ago.⁵⁹ When this mutation is identified in Latinos, haplotype analysis supports that this mutation is of the same origin as the Jewish founder mutation, rather than a separate genetic event.^60,61 Pooled mutation estimates performed by Porchia et al⁵⁵ found that BRCA1 185delAG is the second most prevalent BRCA1 mutation and its frequency is not significantly different between Mexico and other Latin American countries (P = .70). However, it is worth noting that not all Central and South American countries were represented in their analysis.⁵⁵

The most common BRCA1 mutation in the same meta-analysis was deletion of exons 9 to 12.⁵⁵ This mutation is estimated to have originated nearly 1,500 years ago near Puebla Mexico.^48,58 However, to date, it has been reported in Mexicans and Mexican Americans alone.^{22,35,55,61,62} The contribution of large genomic rearrangements to BRCA1 in Latin American patients was evaluated in a study of US Latinas and described the prevalence of rearrangements by racial and ethnic groups.⁶³ Large rearrangements were significantly more common in individuals who reported Latin American ancestry, and the prevalence of rearrangements was two-fold higher than in the overall population tested.⁶³ This laboratory-based cohort extracted ethnicity data from genetic testing request forms; therefore, no data about subethnicity was available.⁶³ However, the 2 most frequent BRCA1 rearrangements identified in this study were deletion of exons 9 to 12 and deletion of exons 1 and 2, likely reflecting the underlying US Latino population in whom the majority is of Mexican ancestry.⁶³ In a Puerto Rican study, BRCA1 deletion of exons 1 to 2 was seen in nearly 20% of study patients positive for BRCA1.⁵² Because Puerto Ricans represent the second largest US Latino group after those of Mexican ancestry, these findings support utilizing an assay that includes large rearrangements when testing Latinos.⁵² The study results also highlight the importance of understanding more granular aspects of ethnicity, such as country of origin, to ensure that all mutations that contribute significantly are captured.

Dutil et al²² noted that most Latin American studies they reviewed identified a higher proportion of BRCA1 than BRCA2 mutations, a finding similar to reports in other populations. However, studies from 4 different countries (Costa Rica,⁴⁰ Cuba,³⁸ Puerto Rico,⁵² Uruguay⁶⁴) reported more BRCA2 mutations than those in BRCA1. While these studies may have been limited by sample size and the mutation-detection strategies and technologies,^38,40,52,64 this finding has been also reported in a single US-based clinical site and may warrant further exploration.⁶⁸

The meta-analysis performed by Porchia et al⁵⁵ identified recurrent BRCA2 mutations across all studies with the following pooled prevalence: H372N (0.88%; 95% confidence interval [CI], 0.24–1.92), E49X (0.38%; 95% CI, 0.13–0.75), and 3492insT (0.32%; 95% CI, 0.24–0.53). BRCA2 3492insT has been identified in different regions of Spain with a frequency as high as 2.08%.^65–71 Although it is possible that this mutation was introduced in Latin America by the Spaniards, no haplotype studies of this specific mutation were identified to confirm a shared ancestry rather than a separate mutational event.⁵²

BRCA1 and BRCA2 account for the majority of hereditary breast cancer, but other high- and moderate-risk genes also predispose individuals to breast cancer, including TP53, PTEN, CDH1, STK11, CHEK2, PALB2, ATM, and others.^{25,29,30,72–80} Limited studies have been performed of non-BRCA genes in Latina breast cancer cohorts, leaving much to be learned about the prevalence and spectrum of mutations in these genes among Latinas with breast cancer (Table 1).^{25,29,30,72–80}

Table 1.

Select Non-BRCA Genes Observed in Latina Populations

Study	Country	Cohort		Gene	Analysis	Findings
		No. of Patients	Inclusion Criteria
Assumpção69	Brazil	123	Family history of breast cancer Family history of ovarian cancer Sporadic breast cancer	TP53	Site-specific analysis of TP53 R337H	2.4% of cases carried the mutation (P = .0442)
Carraro22		54	Early-onset breast cancer diagnosis < 30 y	BRCA1 BRCA2 CHEK2 TP53	Coding introns/exons of BRCA1/2, TP53, and site-specific analysis of CHEK2 (c.1100delC)	22% carried mutations mostly in BRCA1/2 2% (n = 1) had TP53 mutation
Felix26		106	HBOC testing	BRCA1/2 CHEK2 TP53	PCR of each exon of BRCA1 Site-specific analyses of BRCA2 (c.5946_5946delT; c.156_157insAlu), CHEK2 (c.1100delC; c.444+1G>A; p.I157T), and TP53 (p.R337H)	2.8% carried mutations (BRCA = 2, TP53 = 1)
Giacomazzi70		874	Family history of cancer (group 1) Consecutive breast cancer (group 2)	TP53	Single-site analysis for p.R337H	p.R337H identified in 3.4% (group 1) and 8.6% (group 2) Higher prevalence when diagnosed ≤ 45 y (12.1%) than ≥ 55 y (5.1%) (P< < .001)
Silva27		120	HBOC testing	ATM BRCA1/2 BRIP1 CDH1 CDKN2A CTNNB1CHEK2 MLH1 MSH6 NBN PALB2 PTEN RAD50 RAD51 TP53	Coding introns/exons of BRCA1/2 Site-specific analysis of CHEK2 (c.1100delC) and TP53 (p. R337H) Array comparative genomic hybridization for CNVs in other 14 genes	26% (n = 31) mutations (BRCA1/2 = 27, CHEK2 = 1, TP53 = 3)
González-Hormazábal72	Chile	137 (BRCA− = 126, BRCA+ = 11)	≥ 2 family members with breast cancer ≥ 2 family members with ovarian cancer Family history of male breast cancer Early-onset breast cancer with no family history	ATM	PCR-based analysis of coding sequence and exon/intron boundaries of ATM Analysis of ATM 5557G>A, IVS38-8T>C, IVS24-9delT	5557G>A, IVS38-8T>C, IVS24-9delT associated with elevated risk of breast cancer if BRCA− Identification of composite genotype that confers 3.19-fold risk for breast cancer
Jara71		143 (BRCA− = 131, BRCA+ = 12)	≥ 2 family members with breast cancer ≥ 2 family members with ovarian cancer Family history of male breast cancer Early-onset breast cancer with no family history	RAD51D	PCR-based analysis of coding sequence and exon-intron boundaries of RAD51D Analysis of RAD51D, c.135G>C	No mutations detected in RAD51D c.135G>C associated with elevated breast cancer risk if BRCA−
Leyton73		436	BRCA+ ≥ 2 family members with breast cancer ≥ 2 family members with ovarian cancer Single case of early-onset diagnosed ≤ 50 y	PALB2	Full gene sequencing in 100 “high-risk” cases Analysis of identified variants	No pathogenic mutations identified 3 variants identified (c.1676A>Ga, c.2993C>Ta, c.1861C>A)
Calderón-Zúñiga74	Mexico	94	Familial breast cancer Early-onset breast cancer	ATM	PCR-FLP of 3 specific mutations (IVS24-9delT, IVS38-8T>C, 5557G>A)	5557G>A (13%) IVS24-9delT (21% vs 8% controls; P = .0122) IVS38-8T>C (< 1%)
Bell76	United States	362b	Early-onset breast cancer	CHEK2	169 cases diagnosed ≤ 40 y had sequencing of coding region of CHEK2 Specific analysis of 1100delC, H143Y, and 8 other CHEK2 variants/mutations	Data not reported by ethnicity, but reported “infrequency” of c.1100delC among Latinas
Bretsky75		101 b	Personal history of breast cancer	ATM	20 specific ATM missense mutations or polymorphisms	L546V had modest but not significant predictor of risk; almost exclusive to African American women (found in n = 2 Latinas)
Damiola77		158 b	Breast cancer diagnosed ≤ 45 y	MRE11, RAD50, NBN	PCR-based analysis of coding sequence and exon/intron boundaries of MRE11, RAD50, NBN	Data not reported by ethnicity MRE11, RAD50, NBN are intermediate-risk genes

^aThese variants play a role in risk of breast cancer.

^bLatinas were part of a larger multiethnic cohort.

CNV = copy number variation, FLP = fragment length polymorphism, HBOC = hereditary breast and ovarian carcinoma, PCR = polymerase chain reaction.

One exception is the Brazilian founder mutation in TP53, R337H. Mutations in TP53 cause Li-Fraumeni syndrome, which is associated with an elevated risk for a wide spectrum of cancers, including adrenal cortical carcinoma, soft-tissue and bone sarcomas, brain tumors, and breast cancer.^81–83 The overall contribution of TP53 mutations to breast cancer is estimated to be less than 1%, unless selecting for early-onset breast cancer.^84,85 In studies of women diagnosed with breast cancer at or before the age of 30 or 35 years, 5% to 8% had TP53 mutations.^86–89

TP53 R337H was first identified in individuals with childhood adrenal cortical carcinomas living in southern Brazil.⁹⁰ This mutation occurs in 2.4% to 8.6% of Brazilian women with breast cancer.^72,73,90,91 In a large study, which included 403 patients with breast cancer diagnosed at 45 years or younger, 12.1% carried the TP53 R337H mutation. Although the mutation was significantly more frequent in younger patients compared with those diagnosed at or above age 55 years (P < .001), 5.1% of the older group carried the mutation.⁷³ To date, no other populations have been identified in whom TP53 makes such a significant contribution to breast cancer. The prevalence of this mutation in southern Brazil has been estimated be approximately 0.3%.^92,93 Additional haplotype analyses support the hypothesis that this recurrent mutation is a founder mutation from a shared ancestor.^73,94

Historically, the genetic assessment for hereditary breast cancer involved the formation of a differential diagnosis followed by a syndrome-by-syndrome evaluation through the sequential testing of genes. However, the rapid integration of next-generation sequencing has enabled simultaneous testing of multiple inherited cancer genes, thereby expanding the use of multigene panels in clinical testing at a reduced cost.⁹⁵ This expansion is reflected in the emerging body of literature on breast cancer focused on multigene panel findings from the research, clinical, and laboratory settings.^96–104

These literature cohorts are predominantly non-Hispanic whites, with Latinas representing less than 1.0% to 7.4% of study participants^{96–99,101–103} — thus highlighting another area where future research is needed. One study of 475 patients undergoing multigene panel testing included 228 Latino patients (47.6% of study population), and it reported that the likelihood of detecting a deleterious mutation was no different among the ethnic and racial groups represented.¹⁰² Of the patients with breast cancer (n = 197), 14.8% (n = 28) carried mutations, and, as expected, BRCA1 and BRCA2 were the most commonly mutated genes; however, 16 mutations were identified in other genes (CDH1 = 4, CHEK2 = 3, MUTYH = 3, PALB2 = 2, TP53 = 1, RAD50 = 1, RAD51D = 1, BARD1 = 1).¹⁰² Of note, the likelihood of identifying more than 1 variant of uncertain significance in Latinos was significantly higher than that of non-Hispanics whites.¹⁰⁵ Thus, a need exists for further research to better classify rare variants, especially given the under-representation of Latinos in laboratory and research databases.

Genetic Counseling and Testing

Patient- and Health Care–Related Factors

An important step toward understanding the role of BRCA and other high- and moderate-risk breast cancer genes in Latinas is to increase the number of individuals who receive genetic counseling and subsequently elect to undergo testing. However, growing evidence identifies disparities in awareness of and access to genetic counseling among Latinas compared with non-Hispanic white women. Data from health interview surveys from 2000, 2005, and 2010 show that Latinas had the lowest level of awareness about genetic testing for inherited cancer risk than all of the other US racial ethnic groups.^106–108 Using telephone surveys, Gammon et al¹⁰⁹ studied 63 Latinas and 84 non-Hispanic whites at increased risk for carrying a BRCA1 or BRCA2 mutation, examining their awareness, cognitions, and psychosocial needs related to genetic counseling and testing. Among those who had not previously undergone genetic counseling (53 of the 120), Latinas were more unaware than their white counterparts of the availability of testing (56.9% vs 34.8%, respectively).¹⁰⁹ Vadaparampil et al¹¹⁰ reported on a sample of Latinas with a personal or family history of breast cancer, all of whom reported an awareness of genetic risk for breast cancer (ie, family history). However, none of the Latinas had a clear understanding of what genetic testing was and had not received physician referral for genetic testing.¹¹⁰ Findings did differ based on country of origin — an important are to consider in future work, given the diversity of Hispanic populations across the United States.¹¹⁰ In another report, Kaplan et al¹¹¹ reported differences in awareness of genetic testing by race and ethnicity such that 19.4% of Latinas had heard of genetic testing compared with 59.4% of whites, 26.1% of Asian Americans, and 31.0% of black women.

In a study of more than 2,400 patients completing a family cancer history form, Mays et al¹¹² found that, overall, despite low levels of initial awareness, 65 patients (2.7%) met criteria for cancer risk assessment; of those, 72.3% expressed interest in receiving genetic counseling. Furthermore, no differences in interest in genetic services were reported across all racial and ethnic groups.¹¹² Among 1,536 women with nonmetastatic breast cancer, Jagsi et al¹¹³ found that Latinas had a greater desire for genetic counseling than other groups (58.8% of Spanish-speaking Latinas; 36.7% of English-speaking Latinas; 27.1% of non-Hispanic whites; and 28.1% of blacks). In addition, Lagos et al¹¹⁴ examined social, cognitive, and cultural variables among Latinas prior to an appointment for genetic counseling. Fifty low-income, underserved Latinas completed the assessment, and the results demonstrated their readiness (having the necessary skills for the genetic-counseling process), low fatalism, and high rate of self-efficacy, and social support.¹¹⁴ However, this study was conducted in women who showed up to their genetic counseling appointments, thus representing a unique group of women.¹¹⁴ Vadaparampil et al¹¹⁵ studied a group of Puerto Rican women (living in Puerto Rico or central Florida) with a family or personal history of breast cancer and found that the vast majority of participants said they would undergo genetic testing within the next 6 months if it was available. Barriers included the cost of testing and potential pain.

Uptake of Services

Given lower levels of patient awareness, physician recommendations may provide a critical approach to increasing utilization of genetic counseling and testing for hereditary risk of breast cancer. However, available studies^113,116,117 suggest a missed clinical opportunity, because both English- and Spanish-speaking Latina survivors of breast cancer were more likely to have unmet needs for discussion with a health care professional about cancer genetic testing than their non-Hispanic white counterparts. For example, Jagsi et al¹¹³ reports that minority patients were the most likely to express an unmet need for a discussion about genetic testing, although they also showed a strong desire for such testing.

Preliminary studies support the uptake of genetic counseling when services are offered.^118–121 One study of predominantly Latina patients (71.4%) offered genetic counseling at the safety-net hospital found that 88.0% kept their appointments.¹¹⁸ Another study of women (69.6% were Latinas) seen in the safety-net hospital setting reported that 96.4% of them underwent BRCA testing when it was recommended to them.¹¹⁹ Once Latinas were referred, Olaya et al¹²⁰ found that they are equally likely as the general population to complete BRCA testing. Overall, 52% completed genetic testing, and no differences by race and ethnicity were observed.¹²⁰ Woodson et al¹²¹ reported on the utilization of group pretest genetic counseling in a community clinic made up of mostly Latinas (62.3%) with breast cancer; the majority (86.7%) underwent BRCA genetic testing when offered.

Overall, these studies have focused on the delivery of cancer genetics services to majority Latina cohorts, demonstrating that genetic counseling and testing is likely well-received by Latinas with breast cancer; however, these studies were all conducted in safety-net hospitals or in community, low-resource settings and were aimed at the provision of service to low-income, uninsured patients.^118–121 Although Latinas continue to be disproportionally uninsured or underinsured, these study findings might not generalize to other health care settings. Thus, further studies are needed across various clinical settings and in a wider representation of Latinas with breast cancer to better understand the utilization of genetic testing as well as the barriers for referrals.

Common Genetic Polymorphisms and Risk of Breast Cancer

Genome-Wide Association Studies

Progress in the discovery of germline genetic polymorphisms associated with breast cancer risk changed pace when technological advances in genotyping made it possible to characterize genome-wide genetic variation at a relatively low cost.¹²² In 2007, the first breast cancer genome-wide association studies were published, and they reported a handful of single nucleotide polymorphisms (SNPs) associated with a modest increase in risk.^123–125 Since then, more than 100 common variants that either increase risk for or are protective against developing breast cancer have been discovered and, including replication efforts, data from more than 120,000 women have been analyzed.^123–146 A small proportion of samples included in these major initiatives are from minority populations in the United States (eg, Latinas, African Americans)^123–146 and the first results of genome-wide association studies of Latinas with breast cancer were published in 2014.¹⁴⁷ This latter study represents important but limited progress, considering that the sample size was one-tenth of that available for genome-wide association studies involving women of European origin.^{123–125,147}

Until the first genome-wide association studies of breast cancer in women of European origin were published, the search for risk predisposing genetic variants was focused on finding polymorphisms within genes that, for known or hypothesized involvement in the biology of the disease, were likely to contribute to breast cancer risk.^123–125 These studies in US Latinas or Latin American women typically consisted of the replication of previously associated polymorphisms reported in Europeans, with few of these studies looking for variation in samples of Latinas before further testing specific polymorphisms for associations in larger samples.^148,149

Compared with the hundreds of genome-wide association studies in non-Hispanic white women, we identified 13 case-control studies or cohorts that include US Latina or Latin American women.^150–163 These studies include populations of women of no more than 100 and up to approximately 5,000 women of Latin American origin; combined, the study populations tally approximately 5,000 Latina women with breast cancer and 11,000 Latina healthy controls.^150–163

Candidate Gene or Pathway Studies

Multiple breast cancer–association studies of candidate genes or pathways have been reported for US Latina and Latin American women during the last 20 years. Genes or pathways studied have included those related to hormone metabolism, hormone receptors, and hormone coactivators or supressors,^{148,164–169} growth factors,^{149,170–176} matrix metalloproteinases,^177,178 inflammation and energy balance,^{158,179–184} metabolism of xenobiotic compounds and oxidative stress,^{159,185–188} DNA repair,^{160,161,163,189} and angiogenesis.¹⁹⁰ Results reported in these publications should be interpreted with caution, given that approximately 60% of the candidate gene analyses included in our report did not adjust for genetic ancestry, which is a known confounder in genetic association studies in admixed populations.^191,192 In addition, no associations in candidate gene or pathways studies, nor any of the interactions with risk factors, genetic ancestry, or tumor characteristics, have been replicated in independent samples of Latinas.

Replication of Identified Single Nucleotide Polymorphisms

Few studies have included Latinas and tested the association between SNPs discovered in genome-wide association studies of breast cancer conducted in samples of European or Asian women.^193–198 The first study genotyped previously reported SNPs in the 2q35 region and FGFR2, TOX3, and MAP3K1, reported statistically significant replications for the polymorphisms in FGFR2 and 2q35.¹⁹³ Two different studies published the results of analyses conducted in the same sample of high-risk families from Chile and healthy controls, testing associations between previously reported variants in FGFR2, MAP3K1, and TOX3 and the 2q35 and 8q24 regions and breast cancer risk.^197,198 They replicated the associations for FGFR2, MAP3K1, TOX3, and 2q35 but not for 8q24.^197,198 An analysis conducted in a pooled sample of Latina cases and controls from the Four-Corners study, San Francisco Bay Area Breast Cancer Study, and a study in Mexico, investigated the association between 10 identified polymorphisms in genome-wide association studies (in region 2q35 and in or near RELN, MRPS30, RNF146, FGFR2, TOX3, LSP1, TLR1, MAP3K1, RAD51L1) and breast cancer risk.¹⁹⁵ They replicated associations for the polymorphisms in RELN, FGFR2, TOX3, and TLR1 and 2q35 and found heterogeneity by ancestry for the RELN, 2q35, and TLR1 SNPs.¹⁹⁵ A follow-up study reported that the heterogeneity by ancestry for the 2q35 polymorphism was likely due to the association between genetic ancestry, use of hormone therapy, and breastfeeding.¹⁹⁴ Another analysis of the FGFR2 polymorphism in the Mexican study reported an interaction between the FGFR2 polymorphism and alcohol intake.¹⁹⁶ The first genome-wide association study of breast cancer in US Latinas also replicated previous associations, with most of the SNPs being concordant in terms of direction and magnitude of association with those reported in European or Asian populations.¹⁴⁷ Twenty-three of the 83 variants tested had probably values below .05.¹⁴⁷

Ancestry

Admixture mapping leverages the demographical history of admixed populations to find genomic regions that may carry trait-associated variants.^199–207 An admixed population results from the combination of 2 or more ancestral groups.²⁰⁰ The principle of admixture mapping is to identify genomic regions in which cases share more of the same genetic ancestry than either population-based controls (case-control analysis) or compared with the average ancestry of the rest of the genome among cases (case-only analysis).²⁰² This approach has identified risk variants or risk regions for multiple complex traits, including obesity, hypertension, and cancer.^{199,201,203–207} The incidence of breast cancer varies across different racial and ethnic groups in the United States, and Latinas have lower incidence rates than non-Hispanic whites but higher rates than American Indian women.²⁰⁸ Genetic ancestry has also been associated with breast cancer risk in US Latinas and Mexican women after adjusting for non-genetic risk factors, suggesting that a genetic component could be responsible for the difference in risk.^158,209,210 An admixture mapping study in Latinas reported a statistically significant association between a region in the long arm of chromosome 6 (6q25) near ESR1 and risk of breast cancer and a suggestive association on chromosome 11.²¹¹ Higher Indigenous American ancestry at chromosome 6q25 was associated with lower risk of breast cancer.²¹¹ This finding was concordant with the previous reports of lower rates of risk of breast cancer among Latinas with high American Indian ancestry compared with women with high European ancestry after adjusting for possible risk factors such as socioeconomic status, number of full-term pregnancies, and breast feeding.^209–211

One included a discovery phase and replication in 3 additional studies.¹⁴⁷ The study reported genome-wide results that were statistically significant for 2 linked SNPs 56kb upstream of ESR1 (rs140068132 and rs147157845).¹⁴⁷ These SNPs have a frequency of between 5% and 23% in Latin American populations and are absent in most all other groups.¹⁴⁷ The minor allele was protective, with an associated odds ratio (OR) of 0.60 (95% CI, 0.53–0.67) per allele and was more protective for estrogen receptor (ER)–negative disease than for ER-positive disease (OR for ER-negative disease 0.34; 95% CI, 0.21–0.54).¹⁴⁷

Treatment

Tumor Genomics

A growing body of evidence suggests differences in the tumor biology of breast carcinoma across various races and ethnicities. Several studies have evaluated the prevalence of phenotypic subtypes of breast cancer in Latinas compared with other population groups.^212–218 Most data have shown a higher proportion of hormone receptor–negative disease types among Latinas when compared with non-Hispanic whites (Table 2).^3,212–216 However, those results have not always been concordant, and the differences seen across these studies could represent small sample sizes, patient age, or unadjusted rates for genetic ancestry. Although studies based on data from the California Cancer Registry indicated a higher proportion of triple-negative tumors,²¹⁷ this finding was not confirmed in a Colorado study.²¹² In a retrospective study performed in Brazil, patients in the southern regions with a higher percentage of European ancestry and higher socioeconomic status presented with the highest proportion of luminal tumors, whereas the more aggressive subtypes were seen in the northern parts of Brazil, an area with a higher African ancestral influence.²¹⁸

Table 2.

Comparison of ER and HER2 Status in Hispanic, Black, and White Women

Study	No. of Patients	Hispanic	White	Black
Chlebowski205	N = 3,800 n = 103 Latinas	ER+: 83.0%	ER+: 87.0%	ER+: 71.0%
Dunnwald206	N = 209,276 n = 5,585 Latinas	ER+: 70.2%	ER+: 78.4%	ER+: 60.5%
Hausauer207	N = 243,906 n = 15,355 Latinas	ER+: 53.2% (unknown ER status: 29.4%)	ER+: 63.5% (unknown ER status: 22.3%)	ER+: 46.4% (unknown ER status: 29.4%)
Hines203	N = 285 n = 69 Latinas	ER+: 63.8% HER2: 31.9% TNBC: 17.4%	ER+: 77.3% HER2: 14.3% TNBC: 15.1%	—
Li204	N = 124,934 n = 7,219 Latinas	ER+: 68.7%	ER+: 78%	ER+: 53.4%
Parise3	N = 143,184 n = 24,078 Latinas	ER+: 74.6 % HER2: 22.2% TNBC: 15.9%	ER+: 82.7% HER2: 17.2% TNBC: 11.2%	ER+: 66.5% HER2: 20.1% TNBC: 24.5%

ER = estrogen receptor, TNBC = triple-negative breast cancer.

Approximately 15% of breast cancer types overexpress erb-b2 receptor tyrosine kinase 2 (ERBB2; also referred to as epidermal growth factor receptor 2 [HER2 or HER2/neu]) protein.²¹⁹ High levels of HER2 expression identify those women who benefit from treatment with HER2-targeted agents, which have been shown to increase survival in the adjuvant and metastatic settings.^220,221 Most studies with HER2-targeted therapies have enrolled majority populations of non-Hispanic whites, although consistent evidence demonstrates that a higher proportion of HER2-positive tumors exist among Latinas, even after adjusting for other tumor characteristics (eg, grade, stage, ER status) and breast cancer risk factors (eg, number of children, alcohol consumption).²¹²

Further tumor characterization has been made possible due to advances in molecular tumor profiling. Oncotype DX (Genomic Health, Redwood City, CA) is a 21-gene breast cancer assay — known as a recurrence score — that provides prognostic and predictive information regarding the benefits of adjuvant chemotherapy in patients with ER-positive tumors. Use of Oncotype DX is part of several guidelines from professional medical organizations, including the National Comprehensive Cancer Network, the American Society of Clinical Oncology, and the European Society for Medical Oncology.^222–224 The characteristics of this assay and the impact of its results on treatment decisions among Latinas with breast cancer are lacking in the medical literature. Kalinsky et al²²⁵ studied 74 Latinas and 145 non-Hispanic white women matched for age, disease stage, and nodal status, and they observed no differences in the overall recurrence score, ER or progesterone receptor status, or HER2 expression by Oncotype DX. However, Latinas had a higher expression of CCNB1 and AURKA, 2 genes that are part of the proliferation score and heavily weighted in the calculation of the recurrence score.

Multiple trials whose study populations were mostly comprised of non-Hispanic white women have shown that use of Oncotype DX affects treatment recommendations and leads to an increase in physician and patient confidence in treatment decisions.^226–229 A small study of 96 patients with breast cancer treated in Mexico showed that use of the Oncotype DX changed treatment decisions for 32% of patients, a finding suggesting that its use has a meaningful impact on recommendations for adjuvant treatment.²³⁰ Results from cost-effectiveness analyses indicated that use of the Oncotype DX assay was projected to improve rates of life expectancy when compared with the current standard of care.²³¹

Pharmacogenomics

Several factors cause variations in the individual response to drugs, including age, body mass index, diet, and genetic variation.^232–234 SNPs in genes related to drug-metabolizing enzymes have been recognized as important determinants of variability to drug response.²³⁵ Most studies have not been sufficiently powered to determine whether specific chemotherapy agents used to treat breast cancer have different rates of effectiveness and toxicities based on race or ethnicity.²³⁶ However, differences in the metabolism of endocrine therapies have been well documented according to race and ethnicity.^237–240

In a study that evaluated clinical data and blood samples from patients with breast cancer undergoing adjuvant tamoxifen therapy, mostly non-Hispanic white women (68%) and Latinas (26%) had significantly higher serum levels of tamoxifen and 4-hydroxytamoxifen, one of tamoxifen metabolites (P = .02 and P = .007, respectively).²³⁷ In 2 other studies, genetic polymorphisms in CYP2D6 associated with lower plasma concentrations of the active metabolites of tamoxifen were described in Mexican, Puerto Rican, and Spanish patients.^238,239 A higher prevalence of this poor metabolizer phenotype has also been observed in non-Hispanic whites.²⁴⁰ In an attempt to clarify whether CYP2D6 allele status influences outcomes from tamoxifen, investigators assessed data from 2 large prospective trials and found that CYP2D6 allele status did not predict clinical benefit of adjuvant tamoxifen in terms of risk of recurrence.^241,242 Therefore, changes in treatment decisions based on CYP2D6 allele status alone are not recommended. Differences in the incidence of polymorphisms of the aromatase gene among different ethnic groups have also been reported and could potentially lead to different outcomes and toxicities among populations.²⁴³ One trial evaluated the benefit of extended hormonal therapy with an aromatase inhibitor after 5 years of tamoxifen treatment in non-Hispanic whites (n = 4,708) and minority women (n = 352; 1.5% were Latinas).²⁴⁴ In general, the researchers found that, compared with non-Hispanic whites, minorities had fewer associated toxicities and no definitive survival benefit with aromatase inhibitors.²⁴⁴ However, these results should be cautiously interpreted, because the minorities participants were less adherent to hormonal therapy and the study was not powered to detect survival benefit in the subgroups.²⁴⁴

Conclusions

Several genetic and genomic factors are related to the health disparities of breast cancer in Latinas. Increasing our knowledge about the contribution of high- and moderate-penetrance mutations to the risk of breast cancer among Latinas overall and for subgroups based on country of origin is an important priority. Although genetic counseling and testing for inherited susceptibility to breast cancer has been clinically available for nearly 20 years, disparities in awareness, referral to services, and access persist. Therefore, interventions to address barriers related to low levels of awareness and lacking physician referrals are critical.^106–111

Even though multiple candidate gene studies have been conducted in this population, only identified variants in genome-wide association studies have been systematically replicated. Much larger sample sizes will be required if we expect to discover similar results in Latinas as would be identified for the European genome in view of their Indigenous American component. We cannot assume that overlapping variants alone between different ancestral genomes will be associated with rate of risk, so our efforts should be focused on reducing research disparities by expanding available resources to include large cohorts and case-control studies of diverse populations in and outside the United States.

Latinas remain systematically underrepresented in pharmacogenomics studies and the current studies were not powered to detect outcome differences. They are also underrepresented in clinical treatment trials and other patient-reported outcomes research. Future research should draw from the few models of success in prior studies that have recruited sufficient numbers of Latinos.^245–248 Lessons can also be learned from successful examples of recruiting other racial and ethnic minority patients who have survived breast cancer.²⁴⁹ Elements that appear to bolster success include partnership with community-based organizations that provide services to Latinas and the provision of language-concurrent clinical care.^250–252 The heterogeneity of the US Latina population must also be considered, because different cultural influences, levels of awareness of, and interest in genetic and genomic services appear to vary by country of origin.¹¹⁰

An urgent need exists to ensure that existing genomic research considers the unique needs of this Latina population. As the US population continues to diversify with up to one-third identifying as Hispanics by 2060,¹ extending genetic and genomic research into this underserved and understudied population will be critical. By understanding the risk of breast cancer among diverse populations, we will be better positioned to make advancements in the number of women diagnosed at earlier stages, identify more effective and less toxic treatment regimens, and increase rates of survival. Meeting these goals will contribute to reducing the current health disparities in these patients with breast cancer.