Abstract
Despite improvements in overall cancer survival, there are racial and ethnic disparities in mortality rates among cancer survivors. Further investigation of both socioeconomic factors and genetic ancestry is needed to better understand contributors to disparities in cancer mortality. The aim of this study was to compare self-identified race-ethnicity (SIRE) with genetic ancestry as predictors of all-cause mortality among cancer survivors enrolled in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Multivariable Cox-proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) of all-cause mortality and to assess the effect of SIRE (non-Hispanic White [NHW], non-Hispanic Black [NHB], Asian & Pacific Islander [API], Native Indian [NI], and Hispanic) and genetic ancestry (continuous percentages of African, Asian, and European ancestry summing to 1, obtained from genotype array data and SNP weights) among all cancer survivors and by primary cancer site (>350 survivors). Models were (1) age-adjusted and (2) fully adjusted for: age at cancer diagnosis, year of cancer diagnosis, sex, stage, study arm, education, income, occupation, and percent urbanicity. Likelihood-ratio tests (LRT) were used to independently assess the effect of SIRE or genetic ancestry on overall survival. Participants were followed from cancer diagnosis and censored at first occurrence of death, last contact, or December 31, 2017. Among 27,249 cancer survivors, mean age was 71 years (SD 7.1), and mean follow-up time was 7.1 years (SD 6.1). SIRE and genetic ancestry were both predictors of all-cause mortality in age-adjusted models for all cancer survivors (p-LRT<0.001), and by prostate, colorectal, ovarian, and breast cancers. In the fully adjusted model, SIRE remained significant overall (p-LRT<0.001) and for prostate cancer survivors (p-LRT=0.005). Compared to NHW, API had a lower mortality (HR: 0.76, CI: 0.68-0.84) overall, which was similar for prostate cancer survivors. In comparison, genetic ancestry within the fully adjusted model remained a significant predictor overall (p-LRT<0.001), and for prostate (p-LRT=0.008), colorectal (p-LRT=0.017), and breast cancer survivors (p-LRT=0.014). Compared to European ancestry, African ancestry had an elevated mortality (HR: 1.11, CI: 1.004-1.23) overall but not significant by cancer site. Asian ancestry had a lower mortality (HR: 0.74, CI: 0.66-0.82) overall, as well as for prostate, breast, and colorectal cancer survivors when compared to European ancestry. Our analyses indicate both SIRE and genetic ancestry are important albeit different predictors of all-cause mortality among cancer survivors, with SIRE representing a social construct and genetic ancestry representing a measure of biological variation. Lack of treatment data on all participants was a limitation for this study, and future studies should consider examining disparities by treatment access. Selection of SIRE or genetic ancestry should be carefully considered when evaluating disparities in mortality.
Citation Format: Derek W. Brown, Jacqueline B Vo, Ian D. Buller, Naoise C. Synnott, Rena R. Jones, Maria Teresa Landi, Wen-Yi Huang, Mitchell J. Machiela, Amy Berrington de Gonzalez, Neal D. Freedman. Comparing the association of self-reported race-ethnicity and genetic ancestry with all-cause mortality: A pan-cancer survivor analysis in the PLCO Screening Trial [abstract]. In: Proceedings of the AACR Virtual Conference: 14th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2021 Oct 6-8. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr PO-192.