An efficient epithelial-mesenchymal transition-related gene signature for predicting the survival of patients with lung adenocarcinoma.

BACKGROUND

Epithelial-mesenchymal transition (EMT) plays an important role in the pathogenesis of lung adenocarcinoma (LUAD). In this study, we aimed to construct a prognostic signature based on EMT that could predict the prognosis of patients with LUAD.

METHODS

The messenger RNA (mRNA) expression profiles and the clinical data were downloaded from The Cancer Genome Atlas (TCGA) as the training set while data from the Gene Expression Omnibus (GEO) served as the validation set. Differentially expressed EMT-related genes (EMTGs) were identified from the training dataset. Univariate and multivariate Cox regression analyses were employed to develop a gene signature from the EMTGs to predict overall survival (OS) time. The prediction performance of the signature was tested using the time-dependent receiver operating characteristic (ROC) curve. The signature was verified in the TCGA dataset and the external dataset, GSE30219. A corresponding nomogram was also constructed to predict the prognosis of patients with LUAD. The expression of the prognostic genes at the protein level was investigated in the Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset. Gene set enrichment analysis was conducted to reveal the biological pathways associated with the high-risk group and the low-risk group.

RESULTS

A set of 79 differentially expressed EMTGs were identified. An EMT-related signature was constructed which classified patients with LUAD into two subgroups based on the median risk score. In the ROC curve analysis, the prognostic signature had a moderate discrimination accuracy for the 1-, 3-, and 5-year survival rate with areas under the curve (AUCs) of 0.732, 0.675, 0.702 in TCGA training set, respectively and 0.813, 0.672, 0.706 in the GSE30219 validation set, respectively. The established nomogram effectively predicted the OS of patients with LUAD.

CONCLUSIONS

The novel EMT-related signature established in this study is a robust and independent prognostic indicator for patients with LUAD. This signature is expected to improve the personalized management of patients with LUAD.