Personalizing treatment in early-stage breast cancer: The role of standard clinical factors and genomic information in adjuvant chemotherapy decision making

572 Background: The Oncotype DX recurrence score (RS) independently predicts the likelihood of benefit from adjuvant chemotherapy. However, the clinical factors that influence chemotherapy recommendations in addition to RS are not well characterized. We sought to determine how clinicians integrate the RS and standard clinicopathologic data when choosing adjuvant chemotherapy. Methods: We identified women with ER+, HER2-, LN- breast cancer seen at DFCI in whom RS testing was performed between November 2004 and October 2008. Clinical and pathological characteristics, RS and chemotherapy treatment were identified from electronic medical records. A multivariable model was used to examine which factors drove the decision to administer chemotherapy. Results: RS was performed on 269 women with the following case distribution: RS low (<18) 50%, RS intermediate (18–30) 41%, RS high (>30) 9%. Chemotherapy was given to 7% of women with low RS, compared to 42% and 86% of women with intermediate and high RS, respectively. Tumor grade, T stage, progesterone receptor expression and RS were associated with receipt of chemotherapy in univariate analyses but age, LVI and menopausal status were not. In a multivariable logistic regression model, tumor grade, size, and RS were independent predictors of chemotherapy administration. Conclusions: Oncotype DX RS plays a critical role in medical decision making for women with early stage breast cancer at this single academic institution. However, other tumor and clinical features independently contribute to chemotherapy decisions, suggesting that tailored treatment does, and should, integrate both traditional and molecular pathological factors. [Table: see text] No significant financial relationships to disclose.