A Five-microRNA Signature as Risk Stratification System in Uterine Corpus Endometrial Carcinoma

Abstract

Background: MicroRNAs (miRs) have been shown to play important roles in various cancers and may be a reliable prognostic marker. However, its prognostic value in endometrial carcinoma (UCEC) needs to be further explored.

Objectives: The aim of this study was to create a miR-based signature to effectively predict the prognosis for patients with uterine corpus endometrial carcinoma (UCEC).

Methods: Using UCEC data set in TCGA, we identified differentially expressed miRs between UCEC and healthy endometrial tissues. The LASSO method was used to construct a miR-based signature prognosis index for predicting prognosis in the training cohort. The miR-based signature prognosis index was validated in an independent test cohort. MiRNet tool was applied to perform functional enrichment analysis of this miR-based signature.

Results: A total of 208 miRs were differentially expressed between UCEC and healthy endometrial tissues. Five miRs (miR-652, miR-3170, miR-195, miR-34a, and miR-934) were identified to generate a prognosis index (PI). The five-miR signature is a promising biomarker for predicting the 5-year-survival rate of UCEC with AUC = 0.730. The PI remained an independent prognostic factor adjusted by routine clinicopathologic factors. Using the PI, we could successfully identify the high-risk individuals, furthermore, it still worked in an independent test cohort. The five miRs involved in various pathways associated with cancer.

Conclusion: We proposed and validated a five-miR signature that could serve as an independent prognostic predictor of UCECs.

Keywords: TCGA; Uterine corpus endometrial carcinoma; gynaecology tumor; microRNA signature; prognosis; risk stratification system.