Black–White Differences in Treatment Receipt and Outcome for Triple-Negative Breast Cancer: The Competing Contributions of Access, Bias, and Biology

Compared with many malignancies, breast cancer is a source of relative optimism, with cumulative gains in treatment efficacy and patient outcomes year after year providing evermore patients an opportunity for cure. However, despite the advent of endocrine therapy for hormone receptor-positive tumors and human epidermal growth factor receptor 2 (HER2)-targeted therapy for tumors that overexpress HER2, triple-negative breast cancer (TNBC) remains an elusive target for therapy and has the worst prognosis of all breast cancer subtypes. TNBC represents 10–15% of breast cancer across the US, but is overrepresented among breast cancer diagnosed in women of African ancestry.¹ An estimated 30% of breast cancer diagnosed in Black women in the US are TNBCs, and, among premenopausal Black women, as many as half of all breast cancers are TNBC.² Black women are also more likely to be diagnosed with late-stage disease,² therefore the combination of aggressive disease and late presentation in the case of TNBC might be expected to portend worse outcomes; but, in fact, there has been conflicting evidence as to whether there are survival differences between Black and White women with TNBC.^{2, 3}

In their analysis of 23,000 women with non-metastatic TNBC in the Surveillance, Epidemiology, and End Results (SEER) database, Cho et al. concluded that Black women were less likely to receive surgery, less likely to receive chemotherapy, and more likely to die than White women.⁴ Furthermore, this increased risk of death persisted after controlling for sociodemographic characteristics and appeared to be at least partly attributable to not receiving surgery and chemotherapy.

Non-receipt of guideline-concordant care represents an important contributor to disparities for all types of malignancy, with multiple studies showing that Black patients with the same disease characteristics and comorbidities are less likely to receive treatment than White patients and less likely to be enrolled in clinical trials.⁵

With regard to the disparities observed in the study by Cho and colleagues, we must ask ourselves what factors led to the non-receipt of treatment they report. Yes, TNBC might be more common among Black patients, but this difference in prevalence does not sufficiently explain the difference in mortality. Furthermore, when focusing on socioeconomic status, the survival difference observed between Black and White women was only seen among women from wealthier, less deprived areas, suggesting that access alone cannot explain this disparity in treatment and, concomitantly, survival.

These findings leave us with at least two potential possibilities for the survival disparity that persists after all other factors are controlled for: (1) the treatment Black patients are receiving is less efficacious for them than for White patients; and (2) the reasons Black patients do not receive comparable treatment reflect a combination of treatment declined and treatment not offered. Unfortunately, both of these contributors can be attributed to structural factors, specifically the systemic racism that has fueled both underrepresentation of Black patients in clinical trials⁶ and the often inadequate extent and quality of shared decision making for Black patients with cancer.⁷ The good news is that both of these historical inadequacies can be remedied through (1) concerted efforts to improve diversity in clinical trials, particularly those focusing on conditions that are more common in people of color (e.g., the IMPACT trial at Penn Medicine⁸), and (2) training of providers and the development of patient-facing materials to better promote treatment initiation and maintenance among diverse patients.⁹ As surgical oncologists, we must commit ourselves to both of these strategies and to redressing the modifiable barriers to both equitable treatment receipt and research development that keep TNBC an elusive therapeutic target … for now.