Abstract LB-139: Prognostic value of a BCSC-associated microRNA signature in hormone receptor-positive her2-negative breast cancer

Abstract

Abstract Background: Breast cancer patients with high proportion of cancer stem cells (BCSCs) are more chemotherapy-resistant and have unfavorable clinical outcomes. MicroRNAs (miRNAs) can function as oncogenes or tumor suppressor genes, which regulate key features of BCSCs. Since BCSCs have unique miRNA expression profile, we hypothesized that a biology-driven model based on BCSC-associated miRNAs could provide prognostic information for low recurrent-risk HR+HER2- breast cancer patients. Methods: A systematic literature reviewing suggested 50 BCSC-associated miRNAs that regulate the BCSC features. We excluded 15 of these miRNAs as they were not constantly detectable in the 40 paraffin embedded tumor samples by qRT-PCR (A pilot assay). A total of 35 BCSC-associated miRNAs were identified. We randomly assigned 303 patients into a training group (n = 202) and an internal validation (n = 101) group. We built a 10-miRNA-based classifier using a LASSO Cox model in the training group and validated its prognostic in the internal validation group, and two external independent groups of 308 patients. Results: In this multicenter study, we developed build a 10-miRNA classifier to predict disease-relapse free survival (DRFS). With this classifier, HR+HER2- breast cancer patients were classified into high-risk and low-risk group of diseases recurrence. The DFS was significantly different between these two groups in all of the patient groups. In the training group, the 5-year DRFS was 90.83% and 35.0% for the low-risk and the high-risk group, respectively (hazard ratio [HR] 15.85, 95% CI 8.15-30.82; p<0•0001). In the internal validation group, the 5-year DRFS was 91.95% and 7.14% for the low-risk and the high-risk group, respectively (HR 20.62, 95% CI 8.31-51.20; p<0•0001). Furthermore, in the 1st external validation group, the 5-year DRFS for low-risk group and high-risk group were 87.50% and 37.50%, respectively (HR 6.623, 95% CI 3.54-12.39; p<0•0001). In the 2nd external validation group, it was 91.74% for the low-risk group and 57.14% for the high-risk group (HR 6.99, 3.12-15.64; p<0•0001). Moreover, the 10-miRNA-based classifier outperformed the traditional clinicopathological risk factors and IHC4 scoring in predicting the 5-year DRFS. In addition, the patients in the high-risk group were associated with a more favorable response to chemotherapy than the low-risk group (HR 1.806, 95% CI 1.10-2.98; p = 0.0278). Interpretation: Our 10-miRNA-based classifier provides a reliable prognostic model for disease recurrence in HR+HER2- breast cancer patients. This classifier may facilitate patient counseling and serve in personalized treatment for breast cancer patients. Citation Format: Chang Gong, Weige Tan, Kai Chen, Na You, Shan Zhu, Wei Luo, Xinhua Xie, Yunjie Zeng, Nengtai Ouyang, Zhihua Li, Mushen Zeng, ShiMei Zhuang, Wan-Yee Lau, Qiang Liu, Fengxi Su, Xueqin Wang, Erwei Song. Prognostic value of a BCSC-associated microRNA signature in hormone receptor-positive her2-negative breast cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-139.