Genomic instability‐associated two‐miRNA signature as a novel prognostic biomarker in breast cancer

Abstract

AbstractBackgroundBreast cancer (BC) is the most common cancer among women worldwide and a leading cause of cancer‐associated deaths among women. However, there is a lack of accurate prognostic biomarkers for BC. In the present study, we aimed to identify a genomic instability (GI)‐associated microRNA signature as a novel potential prognostic biomarker in BC.MethodsWe performed an integrative analysis to investigate the relationship between GI and BC and identify GI‐associated microRNAs (miRNAs). Subsequently, we conducted a discovery and validation study using multicenter cohorts. The GI‐associated miRNA signature was developed in the discovery cohort and independently validated in internal and external cohorts.ResultsGI‐associated miRNAs expression in BC showed heterogeneity and was significantly correlated with BC prognosis. We identified a GI‐associated two‐miRNA signature (miR‐105‐5p and miR‐767‐5p), termed GI2miR, that stratified BC patients into high‐risk and low‐risk groups with significantly different clinical outcomes (log‐rank p = 0.027) in The Cancer Genome Atlas (TCGA) discovery cohort (n = 763). The prognostic value of GI2miR was further validated in internal TCGA validation cohort (n = 253) (log‐rank p = 0.035) and independent GSE22216 cohort (n = 210) (log‐rank p = 0.036). The GI2miR demonstrated independent prognostic value in multivariate Cox proportional hazard regression analyses and stratification analysis.ConclusionsWe have developed a novel prognostic signature based on GI‐associated two miRNAs for BC, which may lay the foundation for BC to improve prognosis prediction.