Establishing a Prognostic Signature Based on Epithelial-Mesenchymal Transition-Related Genes for Endometrial Cancer Patients.

Backgrounds: Epithelial-mesenchymal transition (EMT) is a sequential process where tumor cells develop from the epithelial state to the mesenchymal state. EMT contributes to various tumor functions including initiation, propagating potential, and resistance to therapy, thus affecting the survival time of patients. The aim of this research is to set up an EMT-related prognostic signature for endometrial cancer (EC). Methods: EMT-related gene (ERG) expression and clinical data were acquired from The Cancer Genome Atlas (TCGA). The entire set was randomly divided into two sets, one for contributing the risk model (risk score) and the other for validating. Univariate and multivariate Cox proportional hazards regression analyses were applied to the training set to select the prognostic ERGs. The expression of 10 ERGs was confirmed by qRT-PCR in clinical samples. Then, we developed a nomogram predicting 1-/3-/5-year survival possibility combining the risk score and clinical factors. The entire set was stratified into the high- and low-risk groups, which was used to analyze the immune infiltrating, tumorigenesis pathways, and response to drugs. Results: A total of 220 genes were screened out from 1,316 ERGs for their differential expression in tumor versus normal. Next, 10 genes were found to be associated with overall survival (OS) in EC, and the expression was validated by qRT-PCR using clinical samples, so we constructed a 10-ERG-based risk score to distinguish high-/low-risk patients and a nomogram to predict survival rate. The calibration plots proved the predictive value of our model. Gene Set Enrichment Analysis (GSEA) discovered that in the low-risk group, immune-related pathways were enriched; in the high-risk group, tumorigenesis pathways were enriched. The low-risk group showed more immune activities, higher tumor mutational burden (TMB), and higher CTAL4/PD1 expression, which was in line with a better response to immune checkpoint inhibitors. Nevertheless, response to chemotherapeutic drugs turned out better in the high-risk group. The high-risk group had higher Conclusion: We constructed the ERG-related signature model to predict the prognosis of EC patients. What is more, it might offer a reference for predicting individualized response to immune checkpoint inhibitors and chemotherapeutic drugs.