A 19‑miRNA Support Vector Machine classifier and a 6‑miRNA risk score system designed for ovarian cancer patients

Ovarian cancer (OC) is the most common gynecologic malignancy with high incidence and mortality. The present study aimed to develop approaches for determining the recurrence type and identify potential miRNA markers for OC prognosis. The miRNA expression profile of OC (the training set, including 390 samples with recurrence information) was downloaded from The Cancer Genome Atlas database. The validation sets GSE25204 and GSE27290 were obtained from the Gene Expression Omnibus database. Prescreening of clinical factors was conducted using the survival package, and the differentially expressed miRNAs (DE‑miRNAs) were identified using the limma package. Using the Caret package, the optimal miRNA set was selected to build a Support Vector Machine (SVM) classifier. The miRNAs and clinical factors independently related to prognosis were analyzed using the survival package, and the risk score system was constructed. Finally, the miRNA‑target regulatory network was built by Cytoscape software, and enrichment analysis was performed. There were 46 DE‑miRNAs between the recurrent and non‑recurrent samples. After the optimal 19‑miRNA set was selected for constructing the SVM classifier, 6 DE‑miRNAs (miR‑193b, miR‑211, miR‑218, miR‑505, miR‑508 and miR‑514) independently related to prognosis were further extracted to build the risk score system. The neoplasm cancer status was independently correlated with the prognosis and conducted with stratified analysis. Additionally, the target genes in the regulatory network were enriched in the regulation of actin cytoskeleton and the TGF‑β signaling pathway. The 6‑miRNA signature may serve as a potential biomarker for OC prognosis, particularlyfor recurrence.