A panel of four miRNAs (miR-190b, miR-584-5p, miR-452-5p, and miR-1306-5p) is capable of classifying luminal and non-luminal breast cancers

BACKGROUND Identifying the molecular subtypes of breast cancer (BC) plays a crucial role in enhancing the efficacy of therapy. MiRNAs (miRs) with differential expressions in different subtypes of breast tumors can be considered as non-invasive biomarkers for diagnosing BC subtypes. OBJECTIVE We aimed to investigate the efficacy of miR-190b, miR-584-5p, miR-452-5p, and miR-1306-5p as novel potent diagnostic biomarkers in discriminating patients with luminal (ER+) and non-luminal (ER–) BCs. METHODS A group of miRs significantly associated with estrogen cell receptors (ER) in breast tumors were identified using feature selection methods analysis on miR-Seq datasets retrieved from TCGA and GSE68085. Four abovementioned miRs were selected as novel potential biomarkers, and their relative expression levels were assessed within adjacent non-tumor, ER+ and ER– tumor tissues by quantitative RT-PCR. Their impact on diagnosis was also evaluated by ROC curve analysis. RESULTS In ER+ BCs compared to ER– BCs, the expression of miR-190b was remarkably increased, while the expression of miR-584-5p, miR-452-5p, and miR-1306-5p were significantly decreased. This group could discriminate ER+ and ER– BCs at an AUC of 0.973. CONCLUSIONS According to our findings, these four miRs are promising biomarkers in discriminating BC subtypes. The candidate miRs in parallel with histologic diagnosis methods can be applied for identifying patients who are most likely responding to specific therapies based on ER status.