Abstract P4-10-03: Racial variation in effects of gene expression profiling on chemotherapy use

Gene expression profile (GEP) testing in early hormone receptor-positive (HR+) breast cancer can inform adjuvant treatment decisions by providing refined estimates of prognosis and chemotherapy benefit. GEPs may also downshift chemotherapy use, potentially avoiding excess morbidity. Multiple studies have reported racial disparities in access to GEP testing. In this study, we examine the effect of race and GEP testing on chemotherapy utilization in early stage HR+ breast cancer. From The University of North Carolina Cancer Information and Population Health Resource (CIPHR), a statewide linkage of cancer registry and administrative claims data, we studied all women eligible for GEP testing, including those with HR+, T1-2, N0 or N1 unilateral primary breast cancers diagnosed from 2005 to 2012. Separate analyses were performed for N0 and N1 patients. Patients were required to have breast surgery within 6 months of diagnosis and no chemotherapy prior to surgery. Use of GEP testing and chemotherapy treatment were defined from insurance claims, while race was defined using cancer registry data. Propensity score adjustment by standardized mortality ratio weighting (SMRW) was used to control for differences in measured patient characteristics between treatment groups in a non-randomized cohort. Propensity score models included age, race, diagnosis year, co-morbidity, and tumor features. Propensity weighted Poisson models were then used to calculate the adjusted risk of receiving chemotherapy. To investigate differential effects of covariates on chemotherapy across racial groups, we used race-stratified models in the N0 cohort. Due to small sample sizes, race-stratified models were not used for N1 patients. The cohort included 11,958 women of whom 84.5% were non-Hispanic white (NHW), 12.9% black, and 2.6% other race/ethnicity. For stratified analyses, race was dichotomized to NHW versus non-white. In the N0 cohort (n=9,671), 11.5% of untested NHW women, 20.8% of tested NHW, 16.8% of untested non-whites, and 21.9% of tested non-whites received chemotherapy. In SMRW-adjusted models, GEP testing was associated with downshifts in chemotherapy for N0 patients (RR 0.79, 95% CI 0.72-0.86) and N1 patients (RR 0.46, 95% CI 0.40 - 0.54). In race-stratified analyses of N0 patients, use of GEP testing was associated with a decrease in chemotherapy of 19% among NHW women (RR 0.81, 95% CI 0.72 - 0.91) and 25% among non-white women (RR 0.75, 95% CI 0.57 - 0.99). Taxane-based regimens were most common, but a substantial portion of chemotherapy-treated patients (43%) received anthracycline-based regimens with or without a taxane. We conclude that GEP testing was associated with decreases in chemotherapy use after propensity adjustment for patient and tumor factors, with greater reductions among non-white patients and node-positive patients. Possible explanations include over-treatment of black women in the absence of testing due to higher perceived risk, racial differences in provider recommendation or chemotherapy choice among tested women where benefit is uncertain, such as for intermediate recurrence scores. Further studies are underway to evaluate racial differences by recurrence score group and the effect of GEPs on racial disparities in outcome. Citation Format: Reeder-Hayes KE, Meyer A-M, Baggett C, Roberts MC, Zhou X, Meng K, Wheeler SB. Racial variation in effects of gene expression profiling on chemotherapy use [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P4-10-03.