Abstract B59: Patterns of breast cancer mortality over 10 years by tumor phenotype and Hispanic versus non-Hispanic white ethnicity

Introduction: Breast cancer tumor phenotypes, defined by expression of hormone receptor status, have been found to be biologically distinct and to influence treatment and survival. Tumor phenotypes are relevant to therapeutic response and length of survival and could explain ethnic-related variation in breast cancer survival. Some studies have reported that Hispanic (H) women are more likely to have ER- and triple negative (ER-, PR-, HER2-) tumors than non-Hispanic white (NHW), and have poorer survival. The purpose of this study was to determine if tumor phenotype influences the difference in survival between H and NHW women with breast cancer.

Study Population and Methods: Data from the New Mexico site of a collaborative, population-based cohort, the ‘Health, Eating, Activity, Lifestyle’ (HEAL) Study, were analyzed to compare Hispanic to non-Hispanic white women for patterns of early and late mortality over approximately 10 years of follow-up, and to assess the influence of tumor phenotype as defined by ER, PR, and HER2 hormone receptor status (+/-). The analysis included 533 women (126 H, 407 NHW) diagnosed with their first primary of breast cancer between 1996 and 1999. Hormone receptor status was determined by immunohistochemistry. Other prognostic tumor characteristics were abstracted from medical records by the SEER New Mexico Tumor Registry.

Results: The risk of breast cancer mortality was higher in H than NHW women, adjusting for significant covariates, including tumor phenotype, age at baseline interview and SEER summary stage during the first four to six years after baseline interview (HR_4yrs=2.33 (95%CI 1.27–4.26); HR_6yrs=2.07 (95%CI 1.24–3.45), but was attenuated at 10 years; HR_10yrs=1.51 (1.00–2.28). This time-dependent pattern of mortality was also found for ER-compared to ER+ tumors (HR_4yrs=2.14 (95%CI 1.12–4.10); HR_6yrs=1.89 (95%CI 1.08–3.32); HR_10yrs=1.39 (95%CI 0.87–2.23), adjusting for age at baseline interview, stage and ethnicity. The pattern was similar for the triple negative phenotype, but the HR estimates were more unstable, due to the smaller sample sizes across phenotype categories. Adjustment for tumor phenotype had little impact on the estimates for ethnicity (<10% change), after adjustment for age and stage. Similarly, adjustment for ethnicity had little impact on the estimates for tumor phenotypes.

Conclusion: Ethnicity and tumor phenotype appear to have independent effects on breast cancer mortality in this population-based cohort of women with breast cancer. The effects of both ethnicity and tumor phenotype are greater for early mortality and diminish with time, and are independent of age and stage of disease. This time-dependent effect has not been previously reported for H compared to NHW women. These results suggest the need to explore factors other than tumor phenotype that may explain the difference between H and NHW women for breast cancer survival.

Citation Information: Cancer Epidemiol Biomarkers Prev 2011;20(10 Suppl):B59.