Genetic Testing in Diverse Populations: Title and subTitle BreakAre Researchers Doing Enough to Get Out the Correct Message?

More than 10 years after BRCA1 and BRCA2 were discovered as major breast cancer susceptibility genes, the medical community is just beginning to get a glimpse of how germline mutations in these genes might be distributed among racial/ethnic minority populations in the United States. While a growing body of evidence documents the benefits of preventive measures having minimal risk for women with identifiable BRCA1 and BRCA2 mutations, genetic testing services remain underutilized by minority women. In this issue of JAMA, John and colleagues¹ provide information regarding the prevalence of pathogenic BRCA1 mutations in a population-based study among 5 US racial/ethnic groups in Northern California. Prevalence was particularly high among young African Americans, a group that has long been recognized as having a disproportionate burden of aggressive young-onset breast cancer.

With more than 200 000 women estimated to be diagnosed with breast cancer and more than 40 000 deaths from the disease each year,² a renewed emphasis on prevention for control of breast cancer is warranted. Women with germline mutations in BRCA1 and BRCA2 genes face dramatically increased risks of breast and ovarian cancers. The decision to pursue genetic testing for BRCA1 and BRCA2, with its associated costs, depends on 2 prerequisites: valid mutation detection methods for the test and effective cancer prevention strategies for mutation carriers. Regarding the first prerequisite, the specificity of genetic testing for BRCA mutations is believed to be 100% and sensitivity is estimated at 85%,³ although recent studies have shown that analysis of genomic rearrangements, which was not performed in the United States until 2006, may significantly increase sensitivity.^{4 - 5} For the second prerequisite, mutation carriers now have the option of choosing from a spectrum of evidence-based prevention strategies, including intensive breast cancer surveillance using magnetic resonance imaging,^{6 - 8} chemoprevention,⁹ risk-reducing prophylactic mastectomy,¹⁰ and salpingo-oophorectomy.^{11 - 14}

What are the reasons women are not offered genetic counseling and testing as part of a comprehensive risk assessment program? Are minority patients less likely to accept genetic counseling, or are there barriers to physicians offering the test to minority women? A recent case-control study found that African American women were 78% less likely to use genetic counseling and BRCA genetic testing than white women.^{15 - 16} Data on BRCA testing from Myriad Genetics Laboratories showed that less than 10% of individuals tested were from minority populations, such as Hispanics, African Americans, Asian Americans, and Native Americans.¹⁷

Given that most literature on genetic testing has focused on Ashkenazi Jewish and non-Hispanic white women, it is conceivable that clinicians are not aware of the clinical usefulness of BRCA testing among US minority populations. To compound the problem, most of the available risk assessment tools were developed using empirical data collected mainly in non-Hispanic white populations.^{17 - 20} Their applicability in other populations is uncertain. Other models were developed based on mendelian principles and the Bayes theorem.^{21 - 22} Of these, the BRCAPRO model has been widely used in the genetic counseling setting, and its performance has been evaluated mostly in white populations.^{3 ,23 - 25} However, as an essential parameter of the BRCAPRO model, the prevalence of mutation carriers is available only for the Ashkenazi Jewish and non-Hispanic white populations.

The study by John et al¹ provides important new information that can be used to refine available risk assessment tools. Since it is impractical to examine directly mutation prevalence using simple random sampling from the general population, the authors used a clever design: a 2-stage stratified sampling of female patients with breast cancer younger than 65 years in the population-based Greater Bay Area Cancer Registry. In the first stage, a brief telephone interview was administered to all patients. According to family history of cancer and age at diagnosis, patients were divided into 2 categories. Category A included women whose cancers are likely to be hereditary, and category B included those whose cancers are likely to be sporadic. Patients in category A were oversampled for BRCA1 mutation testing in the second stage. The current study differs from clinic-based studies in that the proportion of patients whose cancers were likely to be hereditary was known. Factoring in sampling weights, the authors estimated the BRCA1 mutation prevalence to be 3.5%, 1.3%, 0.5%, 8.3%, and 2.2% in Hispanic, African American, Asian American, Ashkenazi Jewish, and other non-Hispanic white patients with breast cancer, respectively.

While the study by John et al¹ provides much-needed information on BRCA1 mutation prevalence among US minority populations, the data should be interpreted with caution. Although patients with breast cancer were randomly selected to undergo the BRCA1 test, the process for the selected women to have completed the test was not random. Only 57% of eligible patients in category A and 50% of eligible patients in category B completed BRCA1 testing. In addition, these proportions varied by racial/ethnic groups, with Asian Americans having the lowest testing proportion (43% in category A). Although eligible patients who were tested and those not tested had similar age at diagnosis, whether they have similar profile of family history is unknown. This raises a concern that the cross-ethnic variation in mutation carrier prevalence may not be as large as reported. For example, one population-based study²⁶ documented a prevalence of 1.1% in Chinese patients with breast cancer, which is higher than the 0.5% reported in the current study. Two studies conducted in China and Singapore found that BRCA1 mutation carrier prevalences were between 5% and 9% in patients with early onset cancers,^{27 - 28} which are also higher than the 2.4% reported in the current study for the group younger than 35 years.

John et al¹ have provided a good starting point for narrowing the knowledge gap in characterizing the BRCA1 gene. Further studies are needed to estimate the prevalence of BRCA2 in racial/ethnic minority populations as well as to study the penetrance of both BRCA1 and BRCA2 in diverse populations. Relative risk and penetrance (absolute risk of developing a cancer in mutation carriers) in minorities are assumed to be the same as those obtained predominantly from Ashkenazi Jews and non-Hispanic whites. However, genetic and environmental factors modify the risk in mutation carriers, and the prevalence of these modifiers may vary across populations. Research studies are warranted to examine factors related to genetic testing uptake in minority populations and to identify the barriers to genetic service and preventive care in general. It also is important to design and evaluate interventions for improving genetic testing uptake in underserved populations, so that genetic testing can achieve full potential as a tool for effective cancer control and prevention.

As documented by John et al,¹ more than half of BRCA1 mutation carriers would be detected in female patients with breast cancer whose cancers are likely to be hereditary based on age at diagnosis and family history. The differences in BRCA1 mutation prevalence across populations should be used to more accurately calculate the pretest probability of having a mutation, rather than as evidence against genetic testing in minority populations. While there has been great debate about the role of race/ethnicity in health research, clinicians interested in providing patients with personalized assessment of cancer risk must understand the contributions of BRCA1 and BRCA2 mutations in diverse populations, because potential modifying factors particular to patients' race/ethnicity, family history, ancestral country of origin, and environmental factors may work in concert to influence outcomes.