Impact of race on prostate cancer selection and cause-specific mortality: A SEER database analysis.

Abstract

248 Background: The objective of this study was to assess relationships between race, treatment, and cause-specific mortality for prostate cancer. Methods: The Surveillance, Epidemiology and End Results (SEER) database yielded 328,151 prostate cancer patients diagnosed between 1990–2007 including 5,129 Japanese-American, 50,717 Black-American, and 282,305 White-American men. Japanese-American men were selected as they have the lowest incidence of prostate cancer and best treatment outcomes. Cox proportional-hazards regression analysis was used to assess association of race with prostate-cancer specific mortality before and after adjusting for age, health, and treatment. To assess impact of race on treatment selection, we compared mean incidence of treatment type [external beam radiation therapy (EBRT), brachytherapy, and surgery] by race using one-way ANOVA and post-testing with Bonferroni correction. Results: Race significantly impacted survival after prostate cancer treatment. Blacks had the highest (HR= 1.198; 95% CI: 1.159 to 1.239, p< .001) and Japanese the lowest (HR= .618; 95% CI: .560 to .682, p< .001) mortality hazard compared to Whites. Tumor grade was the strongest predictor of survival (p< .001). Treatment was also significant: Brachytherapy had the lowest mortality hazard for radiation treatments and surgery was associated with lower mortality hazard compared to no surgery. 37.8%, 44.1%, and 29.1% of Blacks, Whites, and Japanese-Americans underwent surgery (p< .05), 27.3%, 23.8%, and 38.2% of Blacks, Whites, and Japanese-Americans underwent EBRT, (p< .05), and 4.9%, 6.7%, and 4.7% of Blacks, Whites, and Japanese-Americans underwent brachytherapy, with Whites significantly different from the other groups( p<.05). Conclusions: SEER database analysis revealed race impacts prostate cancer survival. Treatment selection for prostate cancer differs among racial groups and race is associated with differential treatment outcomes. Comparisons between treatments must be viewed with caution as confounding variables likely exist not accounted for in the SEER database. Future research on race, treatment, and socioeconomic impact on cancer-specific mortality is warranted.