Researchers are grappling with the question whether it’s possible to separate race from our understanding of cancer.
Bridgette Hempstead was resting on her couch one evening in 1996, when she heard a loud voice urging her to get a mammogram. Whether it was intuition or a voice "from the Lord," as Hempstead tells it, she made an appointment with her doctor. At the visit, Hempstead recalled, her doctor asked her why she wanted the test. The doctor then listed the reasons why Hempstead didn’t need it, she said. “The very last one was the most disturbing,” Hempstead said. “She said Black women don’t need to get mammograms because they don’t need to worry about it.”
At the time, the American Cancer Society urged women to get their first mammograms at 35 years old. That mammogram would then serve as a baseline when they began to get the test regularly at age 40. It’s not clear how many women under 40 received this screening, and other relevant health organizations such as the U.S. Preventive Services Task Force and the Centers for Disease Control and Prevention maintained that screening should start at 50 years. Still, a 2013 New York Times article describes the earlier screening as a “relatively common practice at the time.” Hempstead, however, had to argue with her doctor for the test.
Her hunch may have saved her life: The mammogram led to a diagnosis of early-stage breast cancer on her 35th birthday. As she drove home from the doctor’s office after learning the news, Hempstead sang a hymn to stay strong. Over the next year, she underwent a mastectomy, followed by radiation therapy and reconstructive surgery. Throughout the process, she said she encountered race-based assumptions, much like those her first doctor held, that affected her care as a Black woman with breast cancer.
Hempstead’s physician was dated, but not alone in her thinking. Over the past 100 years, misguided assumptions about race have skewed scientists’ beliefs about cancer risk. In the early 20th century, many in the scientific community thought of cancer in general — and breast cancer in particular — as a White person’s disease.
“In U.S. medicine more than elsewhere, there’s this assumption that humans are different in many ways, and that race is a relevant axis of difference,” said medical historian David Jones of Harvard University.
She said Black women don’t need to get mammograms because they don’t need to worry about it.
Even now, researchers include race as a factor when trying to understand why some patients have different responses to treatments or seem more susceptible to certain forms of breast cancer. Although researchers and physicians periodically decry this use of race, it persists both in laboratory studies and in clinics. “We’ve sort of struggled with it, and then it died down, and then something else will happen that brings it back to the foreground,” said Erica Warner, a cancer epidemiologist at Harvard Medical School. Both the use of race in cancer research and overlooking its effects can spur racial inequalities, she added.
Biologists largely concur that race is a social construct, but sociocultural forces do influence health — and the research conducted to understand it — in many ways. Living in a food desert, redlining, or lack of access to preventive health care can all worsen a person’s health. In the U.S., factors like these are deeply intertwined with a person’s race. Similar sociocultural forces have driven scientific research. Over the decades, research efforts to identify new drugs or breast cancer’s genetic roots have overwhelmingly focused on people of northern European ancestry, which means that any findings might not hold as strongly for people whose genetic roots lie in other regions.
Researchers and clinicians continue to rely on race as a vague stand-in representing these intertwined strands of social influences and genetic ancestry. But as they learn more about the myriad ways social stress and ancestral genetics can drive health and disease, it’s becoming increasingly clear that race is a poor proxy for these influences. Today’s researchers are grappling with the question of whether it’s possible to separate race from our understanding of cancer.
“What we want to do is remove bias — unfounded, stereotypical treatment or a generalization of a very heterogenous group,” said Melissa Davis, a molecular geneticist at Weill Cornell Medicine. Still, she said, “In research, race could be important, depending on the question that you’re asking.”
In the early 1900s, cancer in the United States was thought to occur only rarely in people of color “because, on average, African Americans and so-called primitive people did not live long enough to reach the so-called cancer age in great numbers,” writes historian Keith Wailoo of Princeton University in his 2010 book, “How Cancer Crossed the Color Line.” “Early 20th-century epidemiologists and health statisticians were certain that race played an important role in cancer disparities.”
Back then, records of cancer cases were gathered primarily by life insurance companies and relied on just two racial categories: Black and White. It’s difficult to ascertain the many reasons why cancer seemed less frequent in people categorized as Black. Many factors might have contributed to it. Researchers pointed to reasons such as their reduced access to diagnostics and care, allowing cancer cases to go undiagnosed. And their lifespans may have been shortened by other diseases, making them less likely to develop a disease like cancer, which disproportionately affects older people. But even in the 1920s, some scholars suggested that the disparity was not because of racial differences in the biology of cancer, but simply bad recordkeeping: a failure to chronicle the health of Black populations in the American South.
Mammograms began to enter clinics in the late 1960s. In 1990, researchers evaluated whether a 1989 push for greater use of the technology via media coverage and national campaigns had paid off. A clear divide emerged: Women of color, including Black women, were less likely to get a mammogram than White women. Richer and more educated women were also more likely to receive the test.
By the turn of the century, researchers had begun to record an anomaly: White women were still being diagnosed with breast cancer at higher rates, but Black women were dying of the disease far more often. “Explaining these epidemiological mysteries required speculation beyond the numbers and statistics themselves,” Wailoo writes. “It required theories of difference and of social change, and an active racial imagination.”
Data on breast cancer cases from 2014 to 2018 revealed that for every 100,000 women in the population, 139.2 non-Hispanic White women and 125.8 Black women developed breast cancer. But for every 100,000, there were 27.7 deaths among Black women and 20 among non-Hispanic White women. According to the most recent available data, Black women are about 40 percent more likely than White women to die after a diagnosis of breast cancer when adjusted for age. But that wasn’t always the case. Before 1980, death rates were slightly lower for Black women. The flip bothered Ismail Jatoi, a surgical oncologist at University of Texas Health Science Center at San Antonio.
Jatoi and his colleagues tracked trends in death rates and found that disparities emerged when two key medical interventions became widely adopted in clinics: mammograms and tamoxifen, a drug that treats breast cancers that are sensitive to the hormone estrogen. "These same two interventions triggered the disparity," Jatoi said. Both saved lives and lowered death rates among White women. But these advances did not benefit Black women’s mortality rates equally, leading to the divide evident in present-day data.
A number of factors likely contributed to the mortality difference, said Davis. Early on, Black patients who were eligible for tamoxifen might not have had the right insurance coverage, and they may have faced other barriers to care. Additionally, people of color were largely left out of clinical trials to test the medication. As a result, scientists only learned later that Black women are more likely to experience side effects from the drug and to develop tumors that don’t respond to it.
Molecular geneticist Melissa Davis in her lab at Weill Cornell Medicine. Davis studies triple negative breast cancers, which tend to start earlier in life, grow faster, and disproportionately affect Black women.
In 1994, scientists reached another milestone — the discovery of genetic variants that predicted a woman’s risk of breast cancer. That year, researchers at the National Institutes of Health met three unrelated Ashkenazi Jewish families with a history of breast cancer. Studies of their genes revealed that certain versions of two genes, named BRCA1 and BRCA2, were associated with a significantly higher risk of breast and ovarian tumors. A 1996 comparison found that these cancer-linked mutations were more common among them than in people of European ancestry who were not of Ashkenazi descent.
The researchers recommended testing for BRCA1 and BRCA2 mutations as “an appropriate clinical test for Jewish women at high risk of cancer,” but made no mention of women who did not share the genetic ancestry of Ashkenazi Jewish families.
Together with these genetic advances, newer drugs, tests, and treatments reframed cancer as a molecular disease with a clear course of clinical care. Prior to a diagnosis, women now receive annual screening mammograms starting at age 40 if they are of average risk. Those at high risk because of a family history of breast cancer or known genetic risk factors might undergo additional screening tests or receive preventive doses of tamoxifen. If diagnosed with breast cancer, women may receive chemotherapy or other drugs before surgery. After a tumor is biopsied or surgically removed, doctors can now conduct genetic tests to estimate the risk of a cancer’s return and offer additional drugs or radiation therapy.
These genetic tests reveal a cancer’s molecular signature and the best ways to treat it. The vast majority of breast tumors carry one or more of three kinds of proteins on the surface of cancer cells, and clinicians can use specific drugs to target each of these molecules. Only about 10 to 20 percent of tumors express none of these three markers. These tumors are referred to as triple negative breast cancers, or TNBC, because of the absence of these hallmarks. TNBC tumors tend to grow faster and are more prevalent among Black women.
This streamlined care is based on several genetic and molecular advances that occurred in the decades after 1980, when the racial disparity in deaths first became visible. And while these more recent advances reduced inequities, that racial gap still exists “because the benefits of those advances are not evenly distributed across the population,” said Kent Hoskins, an oncologist at the University of Illinois Hospital. People of color may experience unequal access to care, and they may face implicit bias from their health care providers. Additionally, they are often underrepresented in the clinical research that leads to new treatments. “We have to start considering the fact that the people who are receiving these treatments may have important differences from the people who were in the original studies, and that has to be accounted for,” said Hoskins.
But that accounting can be tricky. One advance in cancer care, for instance, was the use of mathematical tools to gauge a person’s risk of disease. These calculators relied on details such as a person’s family history of cancer, their age when their first child was born, and other known risk factors to make calculated clinical choices. For breast cancer, one of the oldest of these tools was developed in 1989 by researchers at the National Cancer Institute, or NCI.
The original equation was created based on data from a nationwide cancer registry and included race corrections to factor in differences in the incidence of breast cancer among women of different races. The NCI algorithm produces different estimates based on race: Plug in information for two women, otherwise identical, and the model will suggest a Black or Hispanic woman is less likely to develop cancer than a White woman. Used as a risk factor in this way, race is meant to serve as a proxy for a jumble of contributors — unknown genetic variants, social factors, environmental exposures, and more — that result in different rates of disease in women of color.
The models provide rough estimates that clinicians use along with their own experience and judgement to determine what a patient might qualify for in terms of enhanced screening and prevention measures, according to Hoskins. “I never try to get too precise about an exact number for patients,” he said. “I try to think more in terms of categories. The only problem that creates, of course, is that some of the interventions are based on an actual number.”
The people who are receiving these treatments may have important differences from the people who were in the original studies, and that has to be accounted for.
To decide who gets a particular test or prophylactic drug, the Food and Drug Administration and professional groups set clinical guidelines with exact cutoff values. If the NCI model estimates a woman’s lifetime risk of cancer to be greater than 20 percent, current clinical guidelines recommend breast MRI screening in addition to an annual mammogram. If her 5-year-risk score on the same model crosses 1.7 percent, the FDA recommends she take a preventive dose of tamoxifen to avoid developing cancer.
Factoring in race can result in different treatment recommendations for Black and White women. Using the NCI’s calculator to estimate risk for two identical 40-year-old women, for instance, the White woman might qualify for prophylactic tamoxifen while the Black woman does not. “The traditional risk models have a large impact on access to care,” said radiologist Connie Lehman, who leads the breast imaging program at Massachusetts General Hospital. “These traditional risk models carry a lot of weight.”
Similar algorithms are used to guide treatment and screening decisions for other cancers and various diseases. When medical students and educators began to call out these problematic uses of race in 2017 — and demand the removal of the race correction in one algorithm for kidney function tests — they were met with resistance, stemming in part from concerns that the calculator would be more imprecise and might cause harm. “Even though it was incredibly easy for these things to filter into medicine, it's been very hard to get them out of medicine,” Jones, the Harvard historian, said. Policymakers and clinicians evaluated the request and national panels of experts spent a year weighing the pros and cons. “No one, as far as I can tell, required that level of scrutiny to insert a race correction,” he added. Clinicians seemed “so primed to expect and accept that race matters.”
The BRCA mutations provide another recent example of that expectation. Initially, “there was sort of an overall story” that these mutations were only in women of Ashkenazi descent, said Allison Kurian, a cancer geneticist at Stanford University. Twenty years ago, she said, patients would come into her office and tell her they’d had other doctors who’d refused to test them for the mutations. The doctors said, “We’re not going to test you because you’re Asian and Asians don’t have these,” she recalled the patients telling her.
But in late 2007, more than a decade after the first reports linking certain BRCA gene mutations to cancer, researchers at Stanford studied the rates of these mutations in women who self-identified as being of one of four racial or ethnic categories. They found that these gene variants were prevalent across all four groups, although women of Ashkenazi Jewish descent still had the highest rate. The study was the first to suggest a need for more testing for all women. More recently, in 2021, two large studies of people with breast cancer established that 12 gene variants linked to cancer risk, including the BRCA variants, are equally common across all racial groups.
Still, not all variants are distributed so uniformly across groups. And because people of northern European ancestry are overrepresented in genetic databases, genetic risk tests are less useful for people with certain ancestries. People of African, Middle Eastern, and Native American descent are more likely to find they have a mutation of unknown significance than people of European ancestry, Davis said, referring to mutations that have not been closely studied. Clinicians know how to monitor, guide, and treat people with a known cancer-linked mutation, but they’re at sea when it comes to variants of unknown significance. “Literally there’s nothing that the clinician can particularly do,” Davis said. “Even if those mutations may be pathogenic, we just don’t know.”
Similar unknowns have emerged in studies of how people responded to treatments. For several decades, researchers observed that Black women benefit less than White women from tamoxifen for hormone receptor positive cancers, and experienced more nerve damage, a problematic side effect that leads to stopping use of the medication. Black women were also more likely to die of these cancers, which are considered the most treatable of all breast cancers. Veronica Jones, a breast surgeon at City of Hope, a cancer research and treatment center, said she was struck by this disparity “between Black women and really every other race in terms of outcomes.”
When she and her colleagues examined responses to hormone-based therapy for tumors by race using information from the National Cancer Database, they found that Black women were more likely to respond when they began the treatment at earlier stages of disease. But if they were diagnosed later — as is common for people of color — their cancers were more likely to grow resistant to the treatment.
Other researchers have found that people with cancer who are obese or have diabetes, impaired metabolism, or chronic health conditions often benefit less from endocrine therapy.
Those chronic conditions occur more frequently among minority populations in the U.S., including Black Americans, in part because systemic racism gives them less access to fresh produce or exposes them to more air pollution. Some of the disparities evident in breast cancer mortality could be a result of these differences, Davis said, but that doesn’t mean that all Black patients will fail to benefit from therapy because of these problems.
After tamoxifen or other treatments, people with breast cancer may undergo surgery to remove the remains of tumors. To decide whether a patient needs further chemo or radiation therapy, doctors rely on genetic tests such as Oncotype, which uses variants in 21 genes to predict the odds of a tumor’s return. In a recent analysis, Hoskins and his colleagues discovered that the test’s predictions are less accurate among Black women than among White women because the gene variants were studied only in people of northern European ancestry.
In recent years, researchers and patients have sought to reduce the factors that likely contribute to racial disparities in cancer outcomes. These efforts have focused, in part, on enrolling more people of color in clinical trials.
In 2021, the FDA data revealed that 84 percent of women in trials for new cancer drugs were White, and only 2 percent of participants were Black. The reasons range from a mistrust of medical research to doctors who fail to recommend or suggest that Black patients participate in clinical research. Based on their own experiences, Hempstead and others have launched efforts to fix the problem.
In October 2013, 18 years after she pushed her doctor to order a mammogram for her, Hempstead headed to an urgent care facility because of a severe cough. After looking at X-rays of her lungs, one physician urged her to go see her doctors the next day, Hempstead recalled. What had seemed like “just a shadow” on her lungs, as one of the other urgent care physicians said, turned out to be advanced metastatic tumors — a reoccurrence of her cancer. This time, her oncologist gave her one year to live and told her she’d never sing again, she said.
But in the years since her initial diagnosis, Hempstead had launched Cierra Sisters, an organization to improve awareness of breast cancer and help Black women advocate for themselves. Thanks to her work, she knew of a clinical trial that seemed promising for cases like her own. Hempstead asked her doctor about entering the trial. “‘I will not allow you to go into a clinical trial,’” Hempstead recalls her doctor telling her. “And I said, ‘How dare you tell me what you will and will not allow?’” She switched doctors, entered the trial, and — nine years later — continues to work, receive cancer treatments, and sing.
We’re trying to equip our patients just with the chutzpah to stand up for themselves and ask for care.
Black women are so underrepresented in trials that even routine chemotherapy can feel like “a faith walk” because of the many unknowns, said Ricki Fairley, the CEO of Touch, The Black Breast Cancer Alliance, and a former marketing professional who was diagnosed with triple negative breast cancer in 2011. In January of this year, Fairley launched an effort to engage more Black women in clinical trials. She surveyed nearly 300 self-identified Black women with breast cancer, their family members, and caregivers two years ago and found that doctors often didn’t invite Black women to participate in clinical trials. When people use the term “underserved” to describe Black communities’ challenges with access to care, it “sounds like you’re having a dinner party for 10 people but you only have enough chicken for four,” Fairley said. “But frankly the Black people don’t even know there’s a dinner party.”
Fairley found that patients often initiated the conversation about clinical trials with their doctor. People unfamiliar with how science works may be hesitant to ask, she pointed out. The fear of the unknown — of receiving a sugar pill and dying, or of being a guinea pig — is a concern for many. For those who do ask, doctors can explain that clinical cancer trials typically do not use placebos, for example, but they may “still walk away feeling ill-equipped to talk about it and explain it to their families,” Fairley said. When they try, she said, even those wanting to participate might hear their parents say things like, “Wait, you’re going to be a guinea pig? No, wait, you’re not doing that.”
Without reliable sources of information, many find themselves unable to get the information they need. “It’s breeding health inequity. It’s breeding it because we don’t know what to ask for,” she said. “We’re trying to equip our patients just with the chutzpah to stand up for themselves and ask for care.”
Simultaneously, some researchers are working to change hospital protocols so women of all racial and ethnic identities get the same care without needing to ask. In 2020, Lehman and her colleagues at Massachusetts General Hospital wanted to address one source of disparities: delays after abnormal mammograms. If the test reveals something suspect, clinical guidelines recommend a biopsy within a month of the mammogram. Black women are less likely to undergo the procedure within this recommended time frame. When the clinic limited appointments due to the Covid-19 pandemic in 2020, Lehman and her colleagues saw an opportunity to study whether a new workflow, put in place to encourage same-day readings, might also reduce disparities.
Radiologist Connie Lehman at Massachusetts General Hospital where she leads the breast imaging program. After implementing a new workflow, Lehman and her team were able to eliminate racial disparities in follow-up procedures to an abnormal mammogram.
The new workflow was as follows: A dedicated radiologist would wait in a quiet room and read screening mammograms as the scans were completed. People coming in for the test would wait for the radiologist’s verdict. The technician would immediately relay whether the results required more tests, either a second scan or being referred for a biopsy. Before the program was implemented, patients who were not White had significantly lower odds of having these exams performed on the same day. “When we made the change, those racial disparities went away,” Lehman said.
When speaking with colleagues who find the new workflow daunting, Lehman begins by asking whether they read any screening mammograms immediately. The answer is usually “Of course! There will be a patient that requests that, or a provider or a doctor that calls down and says, ‘Hey, my patient’s coming down. Would you mind, while she’s there, just going ahead and reading her mammogram?’” Lehman said.
Offering this same-day option only when it’s explicitly requested results in “favoring wealthier White patients,” Lehman said. “It’s not that anyone is consciously doing that. It’s just, who thinks it’s an option, thinks to speak up?”
Quantifying and removing these systemic barriers to equal care is key to addressing such sources of inequity, Lehman said. “If we continue to do that and engage more people in doing that, I don’t think there’s any question that we can have a significant impact on the disparities.”
Researchers also want to better understand how genetic ancestry influences cancer risk and response to treatment, Davis said, but this is complicated because certain genetic ancestries do overlap with the social groups we call races.
For example, women in the racial group “White” tend to be of northern European ancestry. Similarly West African, Afro-Caribbean, sub-Saharan and many other ancestries are grouped under the racial category of “Black” in the U.S. Research into the genetics of cancer could reinforce flawed ideas that race has biological roots, Davis said.
In her work, Davis homed in on triple negative breast cancer, the most aggressive form of the disease. TNBC occurs three times more often in Black Americans than in other races and tends to start earlier in life than other kinds of breast cancer. This aggressive growth hints at some underlying biological cause, Davis said. Instead of focusing on racial groups in the U.S., however, Davis and her colleagues looked across a range of genetic ancestries that all fall under the same racial category of Black, including people who self-identify as Black in the U.S., Europe, and different parts of Africa.
In 2019, they found that a genetic variant present in people of West African ancestry — but not East African — correlated with TNBC risk. The data take researchers a step closer to understanding why they see more cases of TNBC amongst Americans who self-identify or are categorized as Black. Just as researchers now know that the BRCA1 risk variants occur in everyone regardless of their race, Davis said, identifying genes linked to TNBC risk could remove race from one more aspect of understanding cancer. With such a variant in hand, clinicians could simply use a genetic test to determine one’s risk of developing TNBC.
Someone whose genetic ancestry is 80 percent from Africa can still have the European variation, Davis said. “This is why it’s important not to generalize based on self-reported race and not even the predominance of African ancestry.”
In Davis' lab, triple negative breast tumors are imaged in a way that allows the researchers to identify how immune cells infiltrate the area. In this image, cancer cells appear cyan and pink, immune cells appear yellow and lime, and cell nuclei appear blue.
The variants the team has identified so far don’t fully explain why TNBC is so aggressive, but the work serves as a proof of principle of why it’s essential to include diverse populations in genomics research, she added. Similar efforts could help identify genetic explanations for other conundrums currently linked to race, such as the prevalence of stomach cancer in people of Asian descent.
Studying the genetic lineages of the people encompassed by terms such as “minoritized populations” or “people of color” doesn’t reinforce biological essentialism, Davis said. “It doesn’t have to be a polarizing issue,” she added. “Identifying differences among us doesn't divide us. It makes us more rigorous in our science to be certain that all of the clinical tools work for everyone.”
Nonetheless, studies of genetic ancestry cannot eliminate the question of race from science — nor should they, she added. This year, Davis and her colleagues studied the genes associated with TNBC in several populations across Africa, Europe, and the U.S. Again, they found evidence that despite many of the participants self-identifying as Black, they represented a range of genetic backgrounds that couldn’t be encompassed by a single racial category.
But the researchers also compared how genes were influenced by people’s environments. Here, they found that race — or more precisely, experiences of racism — mattered. They identified a subset of genes linked to inflammation, immune response, and chronic conditions such as diabetes, obesity, and heart disease that were differently expressed among people who self-identified as Black. In the U.S., researchers have found that experiences of systemic racism spike the risk of these conditions in people of color. Racism that results in exposure to environmental toxins or food insecurity “can cause a bias in literally the way your genes are expressed,” Davis said. “Until racism doesn't exist, race is always going to be a variable that will be an important factor to consider.”