Using the 21-gene recurrence score and the recently developed Recurrence Score-Clinical-Pathologic to assess recurrence risk in patients with node-negative, ER-positive early-stage breast cancer receiving aromatase inhibitor treatment alone.

Abstract

592 Background: The 21-gene Oncotype DX recurrence score (RS) is widely used for assessment of distant recurrence risk and prediction of chemotherapy benefit in patients with early stage ER-positive breast cancer. The Recurrence Score-Clinical-Pathologic (RSPC) risk assessment tool assesses distant recurrence risk when tamoxifen (TAM) therapy is used without chemotherapy, integrating RS with tumor grade, tumor size, and patient age (Tang et al., J Clin Oncol 28:15_suppl, 2010 [suppl; abstr 509]). Neither RS nor RSPC risk assessment have significant treatment interaction with TAM versus aromatase inhibitor (AI) use. Individualized risk assessment when AI therapy is planned would be desirable. Methods: A recent meta-analysis (Dowsett et al, JCO 2010) compares the relative efficacy of AIs and TAM in reducing the risk of distant recurrence. The AI:TAM hazard ratio from the meta-analysis was combined with 10-year risk of distant recurrence as assessed by RS or RSPC for individual N0, ER+ patients to assess the patient’s distant recurrence risk with planned AI therapy. Results: From log rank statistics reported in the AI meta-analysis, we derived a treatment hazard ratio estimate of 0.82 (AI:TAM). Assessments of recurrence risk with AI depend, as expected, on the distant recurrence risk with TAM alone as assessed by RS or RSPC. Confidence interval widths for the AI-alone and TAM-alone risk assessments did not differ greatly. The estimated absolute benefit of AI relative to TAM for 10-year distant recurrence risk in individual patients ranges from about 1% to about 5%. Conclusions: The AI:TAM meta-analysis hazard ratio estimate can be combined with individual patient’s RS and RSPC distant recurrence risk assessment to assess the patient’s distant recurrence risk with planned treatment with AI alone.