Social determinants of health, genetic ancestry, and mortality in ECOG-ACRIN E5103.

Abstract

6527 Background: Social determinants of health (SDH) and genetic ancestry have been independently implicated in breast cancer presentation, treatment and mortality. However, little is known about the relationship between SDH and genetic ancestry on clinical trial outcomes. The objective of this study is to assess the association between SDH, genetic ancestry and clinical outcomes in patients enrolled in an adjuvant breast cancer clinical trial. Methods: ECOG-ACRIN (EA) 5103 randomized patients to receive AC + taxane + bevacizumab or placebo. SDH were operationalized as insurance status at trial registration (individual SES) and neighborhood socioeconomic status (nSES). Insurance categories included: (1) Private, 2) Medicare including private/Medicare, military, 3) Medicaid including Medicaid/Medicare, uninsured, 4) self-pay). The nSES index was calculated using zip codes linked to county level data on occupation, income, poverty, wealth, education and crowding. Genome-wide single-nucleotide polymorphism arrays were used to define African ancestry (AA), European ancestry (EA) and other (OA). Multivariable regression and Cox-Proportional Hazard models (odds ratios (OR) and hazard ratios (HR) with corresponding 95% confidence intervals (CI)) were used to assess associations with chemotherapy completion and overall mortality. Estimates were adjusted for the following clinical covariates: age, tumor size, nodal status, hormone receptor status, and primary surgery at randomization. Results: The study cohort included 2453 EA (79.2%), 381 AA (12.2%) and 265 OA (8.6%). Medicaid patients (OR 0.76(0.59-0.99); ref private) and those with AA (OR 0.62(0.49-0.78); ref EA) were less likely to complete chemotherapy. Regarding overall mortality, Medicaid insurance (HR 1.42(1.05-1.92) was associated with a higher mortality than private insurance. Conversely, there was no significant difference in mortality by ancestry (AA HR 1.27 (0.97-1.66); OA HR 0.90 (0.63-1.29): ref EA). Neighborhood socioeconomic status did not appear to be associated with chemotherapy completion or mortality. Conclusions: SDH reflective of individual SES, such as insurance, appear to be stronger drivers of trial completion and mortality compared to nSES among patients enrolled in E5103. Moreover, study results suggest an interplay between ancestry and individual proxies for SDH in trial completion. Nevertheless, the relationship between ancestry and lower rates of chemotherapy completion do not appear to translate into higher mortality rates among patients of AA.