Abstract P6-06-13: Optimized prediction of clinical outcome by the PREDICT plus tool and the 70-gene signature in early stage node-negative breast cancer

Abstract

Abstract Background Established breast cancer guidelines and online tools use clinico-pathological factors including age, tumor size and grade to evaluate the risk of recurrence and select patients who are eligible to receive adjuvant chemotherapy (ACT). One of the online tools, PREDICT plus, was recently updated and is the first tool to include HER2 status and method of detection in risk assessment. Another tool to better guide AST decisions is the 70-gene signature. The 70-gene signature is a gene-expression classifier that was developed and extensively validated to predict the risk of distant recurrence in breast cancer. In clinical practice, an ad-hoc combination of clinico-pathological guidelines and gene-expression classifiers are used. The aim of this study is to evaluate the combination of the PREDICT plus tool and the 70-gene signature to optimize adjuvant systemic treatment decisions. Methods For 427 patients participating in the RASTER study (cT1-3N0M0) a 70-gene signature result was available. PREDICT risk estimates at 5 (P5) and 10 (P10) years after diagnosis were calculated using the following variables: age, method of detection, tumor size, tumor grade, number of positive nodes, estrogen receptor, and HER2 status. Patients were considered high risk if their survival probability was less than 95% at 5 years and/or 90% at 10 years. Five-year distant-recurrence-free-interval (DRFI) and distant-disease-free-survival (DDFS) probabilities were evaluated between subgroups based on the 70-gene signature and PREDICT plus. Results Median follow-up was 61.6 months. Patients with a low risk 70-gene signature (n = 219) had a 5-year DRFI of 97.0% (CI: 94.7-99.4) compared to 91.7% (CI: 87.9-95.7) for the 70-gene signature high risk patients (n = 208). The 5-year DRFI for patients with a P5 low risk (n = 228) is 96.8% (CI: 94.2-99.4) compared to 91.7% (CI: 87.9-95.7) for P5 high risk (n = 199). The 5-year DRFI for patients with a P10 low risk (n = 168) is 98.0% (CI: 95.7-100) compared to 92.1% (CI: 88.7-95.6) for P10 high risk (n = 259). ACT data, DRFI and DDFS probabilities in all subgroups are shown in table 1. Among the patients who had a low risk according to PREDICT at 5 years, but a high risk at 10 years (n = 60), the 5-year DRFI was 100% when their tumor was tested as low risk based on the 70-gene signature (13% received CT) compared to 84% in case of a high risk result (55% received CT)(p = 0.03). Table 170-gene signatureP5P10ACT (%)5-year DRFI (%) (95% CI)5-years DDFS (%) (95% CI)LowLowLow8/132 (6)97 (94-100)96 (92-100)HighLowLow21/36 (59)10097 (92-100)LowHighLowxxxxxxxxxLowLowHigh5/38 (13)100100HighHighLowxxxxxxxxxHighLowHigh12/22 (55)84 (69-100)84 (69-100)LowHighHigh20/49 (41)94 (87-100)94 (87-100)HighHighHigh136/150 (91)91 (87-96)89 (84-94) Conclusion Combining PREDICT plus with the 70-gene signature may help to identify early stage node-negative breast cancer patients for whom limited adjuvant systemic treatment might be appropriate and for whom overtreatment can be avoided. Especially in case of a low risk assessment by PREDICT at 5 years but a high risk at 10 years, the 70-gene signature may aid to select those patients at a high risk of recurrence who will benefit most from ACT. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P6-06-13.