Optimized Marker System for Early Diagnosis of Breast Cancer.

Changes in the methylation levels of 21 microRNA genes in 91 breast cancer samples in comparison with paired samples of histologically unchanged tissue were studied by quantitative methylation-specific PCR. For 19 microRNA genes, a significant increase in the methylation level in tumors in comparison with normal tissues was shown (Mann-Whitney test). When considering the data for breast cancer samples only from patients with clinical stages I and II (59samples), 17 genes with a significantly increased level of methylation were identified. Increased methylation level for 11 genes (MIR124-1, MIR124-3, MIR125B-1, MIR127, MIR129-2, MIR132, MIR137, MIR193a, MIR34B/C, MIR375, and MIR9-1) compared to the paired norm was highly significant (p<0.001, FDR=0.01). The ROC analysis was used to optimize a set of markers for diagnosing breast cancer at the early stages consisting of 4 microRNA genes: MIR125B1, MIR127, MIR1258, and MIR132; the system is characterized by 100% specificity, 85% sensitivity, and AUC=0.924. Importantly, 100% specificity eliminates false positive results. Detection of methylation of at least one of the 4 genes of this set is sufficient to classify the patient's sample as breast cancer.