Abstract P5-08-06: Comparison of risk prediction with the 21-gene recurrence score (oncotype DX) and the 70-gene signature (MammaPrint) in patients with estrogen receptor-positive early stage breast cancer

Abstract

Abstract Background: Gene expression profiling assays estimate prognosis & predict benefit from adjuvant chemotherapy in patients (pts) with estrogen receptor positive (ER+) early stage breast cancer (ESBC). The 21-gene recurrence score (RS) and 70-gene signature (MP) are widely used assays. Pts may have both tests performed to adjudicate discordant results. We compared RS and MP in pts with both tests and evaluated their impact on adjuvant treatment decisions. Methods: This was a retrospective clinical cohort study from 5 centers from 2007-2014. Eligibility included ER+ ESBC with RS and MP results from the same tissue sample; exclusions included carcinoma in situ, primary metastatic disease, & bilateral breast cancer. Two definitions of RS risk were evaluated (Genomic Health [GH] or TAILORx [TRx]): low risk (GH:0-17) or (TRx:0-10); intermediate risk (GH:18-30) or (TRx:11-25); and high risk (GH:>31) or (TRx:>26). Demographic, clinicopathologic (CP) data, and treatment decisions were extracted from medical records. Agreement was assessed between RS and MP using weighted kappa scores. CP factors were assessed with the t-test and chi-square test. Results: 123 pts were identified from UCSF or 4 US Oncology Network practices. 60 pts had low risk RS (GH); 27 were low risk by MP (45% agreement). 15 cases had a low risk RS (TRx); 9 were low risk by MP (60%). 10 pts had high risk RS (GH); 8 were high risk by MP (80.0%). Similarly, 18 cases had high risk (TRx) RS;16 were high risk by MP (88%). Using the RS GH risk definition, there was poor agreement for low risk RS/MP and high risk RS/MP (kappa=0.169). However, using the TRx RS risk definition, there was good agreement for low risk RS/ MP and high risk RS/MP (kappa=0.509). 53 pts had intermediate risk RS by (GH); by MP 16 were low risk (30%), and 37 were high risk (70%). 90 pts had an intermediate risk RS by TRx; by MP 34 were low risk (37.8%), and 56 were high risk by MP (62.2%). Using TRx RS increased the number of pts classified as intermediate risk compared to GH RS (73.2 vs 43.1%). Complete adjuvant treatment data were available for 90 patients. Of 40 pts with low risk RS (GH), 25% received chemotherapy associated with node status but not tumor size, grade, or MP. Of 10 pts with high risk RS (GH), 70% received chemotherapy without association with CP features or MP. For the 45 pts with intermediate risk RS (GH), 53% received chemotherapy, without correlation with CP features. Receipt of chemotherapy was positively associated with MP high risk, but this was not significant (odds ratio 2.08, 95% CI 0.54-8.01); p-0.29. Conclusion: Concordance in risk prediction between RS and MP was greater using RS defined by TRx compared to GH, and RS using TRx increases those categorized as intermediate risk. There appeared to be no correlation between CP features and decisions to use chemotherapy across risk groups. Interestingly, in those with intermediate risk RS (GH or TRx), there was a non-significant trend toward use of chemotherapy in those with high risk by MP. Although gene expression tests are used frequently to aid in treatment decisions in ESBC, considerable variation exists in their application in clinical practice. Citation Format: Jiang H, Denduluri N, Majure M, Favret A, Rugo HS. Comparison of risk prediction with the 21-gene recurrence score (oncotype DX) and the 70-gene signature (MammaPrint) in patients with estrogen receptor-positive early stage breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-08-06.