Evaluation of PAM50 intrinsic subtypes and risk of recurrence (ROR) scores in premenopausal women with early-stage HR+ breast cancer: A secondary analysis of the SOFT trial.

Abstract

504 Background: The landmark Suppression of Ovarian Function Trial (SOFT) in premenopausal breast cancer patients revealed that the addition of ovarian function suppression (OFS) to adjuvant endocrine therapy with either tamoxifen (T+OFS) or exemestane (E+OFS) reduces the risk of recurrence compared with adjuvant tamoxifen alone. There are no biomarkers to aid decision-making about intensification of endocrine therapy with OFS. The purpose of this study is to assess the prognostic and predictive ability of PAM50 intrinsic subtypes and ROR scores in premenopausal women with HR+/HER2- breast cancer in the SOFT trial. Methods: Gene expression analyses were performed via the NanoString Breast Cancer 360 assay on RNA isolated from FFPE tumor samples from the SOFT trial that were HR+/HER2- (n=1245/3047). PAM50 subtype and ROR score were determined centrally by NanoString Technologies (Seattle, WA, USA), blinded to clinical characteristics, treatment and outcome. Median follow-up was 12 years. Primary endpoint in this study was distant recurrence-free interval (DRFI). Secondary endpoints were breast cancer-free interval (BCFI) and disease-free survival (DFS). Kaplan-Meier analysis and Cox proportional hazards regression models, stratified by prior chemotherapy and lymph node status, were used to evaluate the predictive performance of PAM50 intrinsic subtypes and ROR categories (low vs intermediate vs high), and secondarily ROR score as a continuous variable. We also investigated differences between very young (<40yrs) vs young (>40yrs) women. Results: Tumor samples from 1084/1245 (87%) patients successfully completed PAM50 testing and are included in this analysis. Patient characteristics in this cohort were similar to the entire SOFT trial population. Intrinsic subtype and ROR category distribution is described below. Subtype distribution significantly differed between very young vs young premenopausal women, with fewer luminal A and more luminal B and non-luminal tumors seen in the very young (p<0.001). The ROR score distribution also differed significantly: in node-negative patients there were significantly more ROR-high scores amongst very young compared to young women (36% vs 14%, p<0.001). Conclusion: Breast cancer diagnosed in very young women has aggressive disease biology. Prognostic and predictive analyses of both intrinsic subtypes and the ROR score are ongoing and will be presented at the meeting. Clinical trial information: NCT00066690 . [Table: see text]