Development and validation of a prognostic model based on RNA binding proteins in patients with esophageal cancer.

Background: RNA-binding proteins (RBPs) play a crucial role in regulating RNA turnover and are associated with cancer development. However, little is known about the role of RBPs in esophageal cancer (ESCA). The present study focuses on the association between RBP gene expression and survival in ESCA, addressing the clinical relevance of an RBPs-based prediction model for prognosis. Methods: RNA-sequencing data and clinical information of patients with ESCA were obtained from The Cancer Genome Atlas (TCGA) database. We identified differentially expressed genes in ESCA and intersected them with RBP-encoding genes. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed with the identified differentially expressed RBPs. Then, a protein-protein interaction (PPI) network was constructed through the STRING database to determine the hub RBPs. Univariate Cox regression analysis and multivariate Cox regression analysis were applied to construct a novel prognostic model based on RBPs. Based on the R package "Caret", we divided patients into the training set and validation set. The efficacy of the prognostic model was evaluated by the area under the receiver operating characteristic (ROC) curve. A nomogram was developed for the prediction of patient survival outcomes. Results: A total of 158 ESCA patients from the TCGA database were included in our analysis. We screened out five prognostic RBPs (CLK1, CIRBP, MRPL13, TNRC6A, and TYW3) through univariate and multivariate Cox regression analysis. CLK1, CIRBP, TNRC6A and TYW3 were downregulated in tumor samples, while MRPL13 was upregulated. A prognostic model constructed with these five RBPs in the training data set accurately stratified ESCA patients into high- and low-risk groups. When the same prognostic model was applied to the test data set and entire cohort, the 5-RBP signature remained an independent prognostic factor in multivariate analysis. The areas under the time-dependent ROC curve of the prognostic model for predicting one-year survival in the training data set, test data set, and entire cohort were 0.789, 0.753, and 0.764, respectively, confirming that this model is a good prognostic model. The nomogram based on the five RBPs and clinical variables could improve individualized outcome predictions and highlight the importance of RBPs in the outcomes of patients with ESCA. Conclusions: Our study provides a potential prognostic model for predicting the prognosis of ESCA patients. The prognostic nomogram could improve individualized outcome predictions for patients with ESCA, therefore providing novel insights into future diagnosis and treatment.