Abstract
BACKGROUND: Circulating microRNAs (miRNAs) prove to be potential non-invasive indicators of cancers. The purpose of this study is to profile serum miRNA expression in breast cancer (BC) patients to find potential biomarkers for BC diagnosis. METHODS: The miRNA expression patterns of serum samples from 216 BC patients and 214 normal control subjects were compared. A four-phase validation was conducted for biomarker identification. In the screening phase, the Exiqon miRNA qPCR panel was employed to select candidates, which were further analyzed by quantitative reverse transcriptase PCR in the following training, testing, and external validation phases. RESULTS: A 12-miRNA (let-7b-5p, miR-106a-5p, miR-19a-3p, miR-19b-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, and miR-16-5p) panel in serum was constructed. The diagnostic performance of the panel was assessed using ROC curve analyses. The area under the curves (AUCs) were 0.952, 0.956, 0.941 and 0.950 for the four separate phases, respectively. Additionally, the expression features of the 12 miRNAs were further explored in 32 pairs of BC tumor and para-tumor tissues, and 32 pairs of serum exosomes samples from patients and healthy subjects. miR-16-5p, miR-106a-5p, miR-25-3p, miR-425-5p, and miR-93-5p were highly overexpressed and let-7b-5p was conversely downregulated in tumor tissues. Excluding miR-20a-5p and miR-223-3p, the 10 other miRNAs were all significantly upregulated in BC serum-derived exosomes. CONCLUSION: A signature consisting of 12 serum miRNAs was identified and showed potential for use in non-invasive diagnosis of BC.