Origin associated with heterogeneity amongst Hispanic patients with breast cancer in the United States.

Abstract

e23077 Background: There are over 50 million self-identifying Hispanics in the United States. This, however, represents a heterogenous population variably affected by an array of determinants of health, including race and insurance status. We used a national registry to investigate for significant differences in outcomes of Hispanic patients with breast cancer in the United States by country of origin. Methods: We identified a cohort of Hispanic patients in the United States with breast cancer for which origin was documented using the 2004-2019 National Cancer Database (NCDB) dataset. The NCDB delineates the following specific origins: Puerto Rico, Mexico, Cuba, South or Central America (excluding Brazil), or the Dominican Republic. We performed multivariate logistic regression modeling to identify whether origin was a significant predictor of advanced cancer staging at diagnosis (AJCC stage III or IV), adjusting for age, race, receptor subtype, histology, grade, insurance status, and facility type. Consequently, we used Cox Regression survival modeling to investigate whether overall survival (OS) differed by Hispanic origin after adjusting for stage at diagnosis and other confounders of survival: age, race, stage at diagnosis, receptor subtype, and systemic therapy received. Results: Of the identified n = 60,438 Hispanic patients with breast cancer in the United States: n = 26,800 (44.3%) had Mexican, n = 9,792 (16.2%) had Puerto Rican, n = 5,283 (8.8%) had Cuban, n = 14,381 (23.8%) had South/Central American, and n = 4,172 (6.9%) had Dominican Republic origins. After adjusting for the aforementioned covariates, we found that origin was a significant predictor of staging at diagnosis (p < 0.001). Patients originally from Mexico were more likely to be diagnosed at advanced stages compared to patients from Puerto Rico (OR 1.31; 95% CI, 1.18-1.46; p < 0.001). We also identified that origin was a significant predictor of overall survival in this cohort even when controlling for various confounders, including stage at diagnosis. Compared to patients from Puerto Rico, patients from Central/South America (HR 0.78, 95% CI 0.70-0.88, p < 0.001) and patients from the Dominican Republic (HR 0.77, 95% CI 0.65-0.92, p = 0.003) exhibited significantly improved overall survival. There was no statistically-significant survival difference between patients from Mexico and patients from Puerto Rico. Conclusions: This real-world analysis showed that, for Hispanic individuals living in the US with breast cancer, origin is significantly associated with outcomes even after accounting for other known determinants of health. We suggest that region of origin should be studied further as a potential determinant of outcomes in patients with cancer.