Minority groups underrepresented in hematologic cancer trials

Abstract

A new study finds that patient participation in clinical trials (CTs) leading to approval for hematologic cancer indications significantly underrepresents certain at-risk demographic groups in comparison with their incidence of these diseases. These disparities are most prominent and most consistent for Black, Hispanic, and Native American patients.

In their article, which appears in the Journal of Clinical Oncology (doi:10.1200/JCO.22.00504), researchers from the Medical College of Georgia at Augusta University in Augusta, Georgia, point out that, as of 2021, hematologic cancers represent only 10% of cancers diagnosed in the United States. This is one reason, they write, that these cancers often are not examined separately from other malignancies when disparities in oncology drug approval CTs are being explored. Filling this gap was a key reason for their new study.

The study focused on CTs leading to drug approval by the US Food and Drug Administration (FDA) for multiple myeloma (MM); myelodysplastic syndrome and myeloproliferative neoplasms; and leukemias, including acute myeloid leukemia, chronic lymphocytic leukemia, chronic myeloid leukemia (CML), and acute lymphoblastic leukemia (ALL).

The researchers identified potentially relevant CTs from the FDA’s Oncology (Cancer)/Hematologic Malignancies Approval Notifications and Novel Drug Approvals databases. Participants’ demographic information and CT geographic site locations were obtained from studies on ClinicalTrials.gov and from articles found in PubMed via the drug name and CT numbers. Population-based incidence and mortality rates, delineated by race, ethnicity, and gender, came from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) 21 registries and US census population estimates for more than 21 locations from 2014 to 2018. The researchers excluded CTs of pediatric populations, lymphomas, and rare hematologic cancers with data insufficient for statistical analysis.

Of the 61 CTs (with 13,731 participants), 67.2% reported data on race, and 48.8% also included ethnicity.

Of the 7287 patients in MM CTs, 80.3% were White, 4.7% were Black, 10.7% were Asian/Pacific Islander (API), <0.1% were Native American, and 2.9% were Hispanic. The corresponding percentages for MM incidence in SEER registries were 68.7%, 27.4%, 3.4%, 0.2%, and 17.5%. All of

these disparities in CT representation were statistically significant.

For the 1629 patients in acute myeloid leukemia trials, the percentages for CT participation and SEER incidence were 79.6% and 77.3%, respectively, for Whites; 3.8% and 12.3%, respectively, for Blacks; 8.8% and 5.6%, respectively, for APIs; 0.1% and 1.5%, respectively, for Native Americans; and 6.3% and 15.3%, respectively, for Hispanics. Again, all of these comparisons were statistically significant.

Varying levels of White and API overrepresentation, in contrast to Black, Native American, and Hispanic underrepresentation, were generally reported for chronic lymphocytic leukemia, CML, and ALL. The exceptions to this pattern were the absence of significant overrepresentation or underrepresentation among Whites in CML and ALL CTs and among APIs in ALL CTs.

The 41 CTs that reported participant race took place in 990 locations, with 30.4% (n = 301) being in the United States. Most of the CT sites were in the Northeast, Southeast, and Midwest. The study authors wrote that they believed that the geographic distribution inadequately reflected the geographic burden of most hematologic cancers, with the notable exception to that conclusion being MM.

“Our analysis demonstrates that pivotal trials of novel agents approved for some hematologic malignancies are being conducted in populations that are not fully representative of the general population that suffers these malignancies,” says study author Jorge E. Cortes, MD, the Cecil F. Whitaker Jr, MD, GRA Eminent Scholar Chair in Cancer at the Georgia Cancer Center in Augusta, Georgia. “There are differences in genetics, including possibly pharmacogenetics, as well as nutritional, social, economic, educational, and other factors that affect how patients may respond to therapy and the safety of a given drug, so we should not expect that results can be generalized to all patients.”

Arif Kamal, MD, MBA, MHS, chief patient officer for the American Cancer Society (based in Chapel Hill, North Carolina), agrees with Dr Cortes. “For blood cancers, particularly regarding multiple myeloma management, the standard of care is changing almost yearly, and the research is moving very quickly.” Dr Kamal says that, because of limited diversity among trial participants, clinicians may not be prescribing the latest, most optimal therapy for minority patients. “The real harm as I see it is when we take a leap of faith based on old data that may not be generalizable to some of our patients.”

Dr Cortes says he believes that participation in CTs is already somewhat biased, as they have very strict eligibility criteria that are probably not met by a large percentage of the population that could benefit from the drug once approved. “This is even worse when you have significant disparities in the racial and sex composition of the population studied compared to the population at risk.” He adds that it is also important to provide study access to all patients. “One cannot expect that patients can just travel to the study sites. … We must bring the studies to the patients, not the patients to the studies.”

In an editorial accompanying the study in the same issue, Namrata S. Chandhok, MD, and Mikkael A. Sekeres, MD, MS, from the Division of Hematology at the Sylvester Comprehensive Cancer Center at the University of Miami in Miami, Florida, wrote that this study confirms that hematologic malignancies in the United States are “inadequately racially and ethnically represented in studies leading to drug approval.” They also wrote that they believe that one driver of this deficiency is a lack of available trials in areas of the country where special populations live.

Dr Chandhok and Dr Sekeres also noted that a lack of understanding of racial differences exists, and this results in lumping together groups of people who have diverse genetic variances and “socioeconomic, cultural, and behavioral practices within ethnic groups.” As a result, they wrote, “obstacles to access for specific populations need to be addressed … such as language barriers, communication, and socioeconomic challenges.”