Neighborhood deprivation and race as predictors for cancer-related mortality in breast cancer.

Abstract

174 Background: The socioeconomic status (SES) of individuals and their neighborhoods significantly influences their access to healthcare and the outcomes of their diseases. Inequities in opportunities, education, income, and infrastructure can contribute to worse prognosis and unfavorable outcomes for certain malignancies in areas with lower SES. We studied the association of race and neighborhood deprivation with the survival of breast cancer (BC). Methods: We linked the neighborhood deprivation index (NDI) data by the National Cancer Institute with the BC cohort from the SEER database diagnosed from 2010 to 2016. NDI score of each county is created using factor analysis to identify variables from 13 SES attributes from the dimensions: wealth/income, education, occupation, housing conditions. NDI was classified into quintiles (qn) (Q1 least deprived - Q5 most deprived). The association between NDI, race/ethnicity and overall survival (OS) and disease-specific survival (DSS) was analyzed using Cox multivariate regression model. All analyses were adjusted for age, race, grade, insurance, and treatments. Results: Most of the early-stage (ES) (N= 88, 572) and advanced-stage (AS) BC (N= 12336) patients (pts) were in the Q5 (ES- 27.4%, AS- 30%) qn. There was a predominance of racial minorities in Q5 and Q4 compared to Q1 qn (ES: Blacks 13-15% vs 8%, Hispanics 15% vs 6%; AS: Blacks 19-23% vs 14%, Hispanics 13% vs 7%, p<0.001). Uninsured ES pts were higher in Q5 compared to Q1 qn (2.2% vs 1.7%, p<0.001). Triple-negative BC was higher in Q5 compared to Q1 qn (ES: 14.5% vs 11.7%; AS: 16.5% vs 8.8%, p<0.001). ES BC pts who live in Q5 qn have inferior OS and DSS compared to Q1 qn, but this trend was not observed in advanced BC (Table 1). The interaction between NDI and race for OS and DSS was significant (p-interaction 0.0030 and 0.02 respectively) for the ES BC, but not for advanced BC (p-interaction 0.67 and 0.72 respectively). In ES BC, Whites residing in Q5 qn had higher overall (OS: HR= 1.23, 95% CI= 1.14-1.34, p<0.001) and disease-specific mortality (DSS: HR= 1.29, 95% CI= 1.15-1.45, p<0.001) compared to Whites from Q1 qn. However, this disparity in mortality based on the NDI was not observed for Blacks with ES BC (OS: HR= 1.2, 95% CI= 0.98-1.46, p=0.06, DSS: HR= 1.2, 95% CI= 0.94-1.55, p= 0.12). Conclusions: Early-stage BC pts from areas with worse NDI have poor OS and DSS. Blacks with early-stage BC have poor OS and DSS regardless of the SES of the neighborhoods. To improve their clinical outcomes, personalized treatments are needed for Black pts with early-stage BC. Targeted investments and policies should prioritize improving the SES of neighborhoods with high deprivation to reduce healthcare disparities and enhance BC outcomes.[Table: see text]