Identification of an eleven-miRNA signature to predict the prognosis of endometrial cancer

Abstract

Endometrial cancer (EC) is the most common gynecological malignancy. Recent studies have uncovered miRNA acted a striking role in predicting the prognosis of multiple tumors. Over 500 EC samples were selected from the Cancer Genome Atlas (TCGA) database. Univariate, LASSO and multivariate Cox regression analysis were employed to screen out the prognosis-involved miRNAs. Kaplan-Meier (K-M) and time-dependent receiver operation characteristic (ROC) curves were conducted to reveal survival analysis and assess the accuracy of the signature. The independence of the model was verified via univariate and multivariate Cox regression analysis. Besides, qRT-PCR was conducted to testified the expression of 11 miRNAs in 16 paired tissues. A total of 514 specimens were randomly divided into the training set and the testing set, then an 11 miRNAs-based signature were determined which divided the patients into high-risk group and low-risk group. The survival was markedly different and the ROC curve exhibited a precise prediction. Meanwhile, the univariate and multivariate Cox regression analysis verified the miRNAs-based model was an independent indicator of EC. Moreove, the prediction ability of this model with clinicopathological features was more efficient. Finally, functional enrichment analysis demonstrated these miRNAs were associated with the occurrence and progression of cancer. Additionally, hsa-mir-216b, hsa-mir-363, hsa-mir-940 and hsa-mir-1301 were highly expressed in EC tissues in contrast to normal tissues through qRT-PCR. Importantly, the eleven-miRNA signature was full of robust ability to predict the prognosis of EC.