Racial survival disparities in head and neck cancer clinical trials

Abstract

Background: Survival disparities between Black and White head and neck cancer patients are well documented, with access to care and socioeconomic status as major contributors. We set out to assess the role of self-reported race in head and neck cancer by evaluating treatment outcomes of patients enrolled in clinical trials, where access to care and socioeconomic status confounders are minimized.

Methods: Clinical trial data from the Radiation Therapy Oncology Group studies were obtained. Studies were included if they were therapeutic trials that employed survival as an endpoint. Studies that did not report survival as an endpoint were excluded; 7 Radiation Therapy Oncology Group Studies were included for study. For each Black patient enrolled in a clinical trial, a study arm-matched White patient was used as a control.

Results: A total of 468 Black participants were identified and matched with 468 White study arm-specific controls. White participants had better outcomes than Black participants in 60% of matched pairs (P < .001). Black participants were consistently more likely to have worse outcomes. When outcomes were measured by progression-free survival or disease-free survival, the failure rate was statistically significantly higher in Black participants (hazard ratio [HR] = 1.50, P < .001). Failure was largely due to locoregional failure, and Black participants were at higher risk (subdistribution HR =1.51, P = .002). The development of distant metastasis within the paired cohorts was not statistically significantly different.

Conclusion: In this study of clinical trial participants using self-reported race, Black participants consistently had worse outcomes in comparison to study arm-specific White controls. Further study is needed to confirm these findings and to explore causes underlying this disparity.

Figure 1. Aggregate survival outcomes by race using composite variable for Disease Free Survival (DFS) or Progression Free Survival (PFS). Probability of DFS or PFS by race estimated using Kaplan Meier curves in the full matched sample. The hazard ratio (HR) is 1.50, 95% Confidence Interval 1.30-1.73, as estimated by a Cox regression. (P < .001).

Figure 2.

Probability of locoregional failure by race as estimated using cumulative incidence curves in the full matched sample. The subdistribution hazard ratio (sHR) is 1.51, 95% Confidence Interval 1.25-1.83, as estimated by a competing risk regression. (P < .001).

Figure 3.

Overall Survival by race. Probability of all-cause survival by race estimated using Kaplan Meier curves in the full matched sample. The hazard ratio is 1.54, 95% Confidence Interval 1.32-1.80, as estimated by a Cox regression. (P <.001).