Abstract 4004: Prognostic value of miRNAs in breast cancer varies with patient race and molecular subtype

Abstract

Abstract Background: MicroRNAs (miRNAs) may serve as biomarkers for breast cancer (BC) diagnosis and prognosis. Since their clinical value based on race/ethnicity and molecular subtypes has not been studied, we evaluated the prognostic significance of a panel of miRNAs in luminal and triple-negative BCs (TNBCs) of African Americans (AAs) and non-Hispanic Caucasians (CAs). Methods: TaqMan® assays were used to quantify levels of miR-21, miR-106a, miR-206, miR-155, miR-210, and miR-335 in 151 BCs and corresponding normal tissues of 80 AAs (luminal = 24 and TNBCs = 56) and 71 CAs (luminal = 27 and TNBC = 44). Fold changes in levels between tumor-normal pairs were used to categorize the tumors into high- and low-expression groups. The levels of miRNAs were correlated with molecular subtype, patient race/ethnicity, and overall survival by Kaplan-Meier univariate and Cox regression multivariate analyses. Further, the added predictive value of these miRNAs to a baseline prediction model consisting of standard clinical risk factors was evaluated at 2-, 5-, and 8-years post-surgery based on positive predictive curves. Results: In BC tissues, miR-21, miR-106a, miR-155, miR-206, and miR-210 were up-regulated, and miR-335 was down-regulated. For the combined population, univariate and multivariate analyses showed that high levels of miR-21 (adjusted HR = 2.91, 95% CI: 1.27-6.62) and miR-106a (adjusted HR = 2.49, 95% CI: 1.04-5.97) and down-regulation of miR-335 (adjusted HR = 2.97, 95% CI: 1.35-5.54) were associated with poor survival. In multivariate analyses by race, high miR-21 (HR = 7.35, 95% CI: 2.46-21.98) and low miR-335 (HR = 14.13, 95% CI: 2.93-68.13) were indicators of poor survival in only CA patients. Additionally, expression levels of miR-21 aided in identifying risk of death in CA patients with either molecular subtype. In CAs, differences in the area under the positive predictive curve (ΔAUC) between the models with and without miR-21 were 0.077 (95% CI = 0.007-0.154) and 0.053 (95% CI = 0.002-0.128) at 5 and 8 years after surgery, respectively. In TNBCs of both racial groups, addition of miR-106a increased predictive accuracy at 5 (ΔAUC = 0.026, 95% CI = 0.001-0.068) and 8 years (ΔAUC = 0.022, 95% CI = 0-0.061) post-surgery. No miRNA was helpful in identifying patients at high risk of death due to luminal BCs or to AA ethnicity. Conclusions: High levels of miR-106a and miR-21 and low levels of miR-335 are independent prognostic markers of BCs. Furthermore, miR-21 has added predictive value above the standard clinical risk factors for CAs and miR-106a has added predictive value for TNBCs. In evaluation of the clinical utility of miRNAs, patient race and molecular types of BCs should be considered. This work was supported by a pilot grant from UAB Breast SPORE grant of the National Institutes of Health/National Cancer Institute (5P50CA089019). Citation Format: Balananda-Dhurjati Kumar Putcha, Trafina Jadhav, Michael P. Behring, Sejong Bae, Andra R. Frost, Isam-Eldin Eltoum, Li Chen, Heidi L. Weiss, William E. Grizzle, Upender Manne. Prognostic value of miRNAs in breast cancer varies with patient race and molecular subtype. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4004. doi:10.1158/1538-7445.AM2015-4004