Abstract LB-251: Development of a circulating microRNA risk signature predictive of breast cancer diagnosis among high-risk women

Abstract

Abstract Introduction: Current models predicting short and long term risk of breast cancer based on clinical factors, such as family history, have significant limitations. Circulating microRNAs (c-miRNAs) have emerged as measurable ‘liquid biopsy’ biomarkers for cancer detection. However, the value of c-miRNAs, an epigenetic regulator of protein levels, as a novel prediction tool of breast cancer development years before detection has not been explored. Methods: Twenty-four breast cancer cases and matched were selected from a source population cohort of 584 unaffected women enrolled in the High Risk Breast Program at the UVM Cancer Center. Women with a mutation in BRCA1 or BRCA2 were excluded. Cases were women diagnosed with stage I-III breast cancer. One control was matched to each case on age (± 3.5 yrs) and reason for high-risk status (family history or benign breast disease). Clinical data and serum samples were collected at study enrollment and at subsequent cancer-free 4-year intervals. RNA was isolated from serum and over 2500 mature human miRNAs were profiled (Affymetrix microRNA v4.0 microarray). Two sets of candidate miRNAs were identified that distinguished cases from controls and a multivariate risk score for each set calculated using Cox regression models. Set 1: 20 miRNAs with greatest area under the receiver operating characteristic curve (AUC) after univariate regression. Set 2: 19 miRNAs with significantly different expression (ANOVA p < 0.05) between cases and controls. Candidate sets 1 and 2 included 25 c-miRNAs: 14 shared, 6 unique to set 1, and 5 unique to set 2. Results: We identified 2 panels of c-miRNAs that distinguish high-risk unaffected women who later develop breast cancer from those remaining cancer-free; the AUC values of 0.896 and 0.870, greatly exceed modeled risk based on clinical parameters (Gail model AUC = 0.497; Claus model AUC = 0.507; IBIS model AUC = 0.503). Panel 1: 6 model-selected c-miRNAs from set 1. Panel 2: 5 model-selected c-miRNAs from set 2. Panels 1 and 2 included 9 c-miRNAs; 2 shared, 4 unique to panel 1, and 3 unique to panel 2. Of these 9 c-miRNAs, 6 are detected at lower levels in serum from women going onto develop breast cancer while 3 are elevated in cases as compared to cancer-free women. Pathway analysis for modeled miRNAs revealed roles in many cancer-related pathways and biological functions including Regulation of EMT, Molecular Mechanisms of Cancer, and malignant solid tumor. Conclusions: In contrast to clinical breast cancer risk models, the identified miRNA panels perform well at distinguishing cases from controls, with AUCs approaching 0.9. While these data require validation, this non-invasive ‘risk signature’ of c-miRNAs will help identify high-risk women who will most benefit from enhanced screening and prevention strategies. Pathways targeted by ‘risk signature’ c-miRNAs may provide insight into mechanisms promoting breast cancer development. Citation Format: Nicholas H. Farina, Jon E. Ramsey, Melissa E. Sands, Tiffany J. Rounds, David J. Shirley, Thomas P. Ahern, Janet L. Stein, Gary S. Stein, Jane B. Lian, Marie E. Wood. Development of a circulating microRNA risk signature predictive of breast cancer diagnosis among high-risk women [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr LB-251. doi:10.1158/1538-7445.AM2017-LB-251