Differences in 21-Gene and PAM50 Recurrence Scores in Younger and Black Women With Breast Cancer

Genomic tests, such as the Oncotype Dx 21-gene and Prosigna risk of recurrence (ROR-P) assay, are commonly used for breast cancer prognostication. Emerging data suggest variability between assays, but this has not been compared in diverse populations.

RNA sequencing was performed on 647 previously untreated stage I-III estrogen receptor-positive/human epidermal growth factor receptor 2-negative tumors in the Carolina Breast Cancer Study, which oversampled Black and younger women (age <50 years at diagnosis), using research versions of two common RNA-based prognostic assays: ROR-P_R and the 21-gene recurrence score (RS_R). Relative frequency differences and 95% CIs were estimated for associations with race and age, and hazards of 5-year local or distant recurrence were modeled with Cox regression. Proliferation and estrogen module scores from each assay, representing broad activity of genes in those pathways, were examined to guide interpretation of differences between tests.

Among both younger and older individuals, Black women had higher frequency of intermediate and high ROR-P_R scores than non-Black women. Race was not significantly associated with RS_R in either age group. High (hazard ratio [HR], 4.67 [95% CI, 1.73 to 12.70]) and intermediate (HR, 2.12 [95% CI, 0.98 to 4.62]) ROR-P_R scores were associated with greater risk of recurrence, but RS_R did not predict recurrence. RS_R emphasized estrogen over proliferation modules, whereas ROR-P_R emphasized proliferation. Higher proliferation scores were associated with younger age and Black race in both assays. Modifications to the RS_R algorithm that increased emphasis on proliferation improved prognostication in this diverse population.

ROR-P_R and the 21-gene RS_R differentially emphasize estrogen-related and proliferative biology. The emphasis of 21-gene RS on estrogen-related biology and lower endocrine therapy initiation among Black women may contribute to poorer prognostic ability in heterogeneously treated populations.