A novel prognostic model for papillary thyroid cancer based on epithelial–mesenchymal transition‐related genes

Abstract

Abstract Background The frequent incidence of postsurgical recurrence issues in papillary thyroid cancer (PTC) patients is a primary concern considering the low cancer‐related mortality. Previous studies have demonstrated that epithelial–mesenchymal transition (EMT) activation is closely related to PTC progression and invasion. In this study, we aimed to develop a novel EMT signature and ancillary nomogram to improve personalized prediction of progression‐free interval (PFI). Methods First, we carried out a differential analysis of PTC samples and pairwise normal thyroid samples to explore the differentially expressed genes (DEGs). The intersection of the DEGs with EMT‐related genes (ERGs) were identified as differentially expressed EMT‐related genes (DE‐ERGs). We determined PFI‐related DE‐ERGs by Cox regression analysis and then established a novel gene classifier by LASSO regression analysis. We validated the signature in external datasets and in multiple cell lines. Further, we used uni‐ and multivariate analyses to identify independent prognostic characters. Results We identified 244 prognosis‐related DE‐ERGs. The 244 DE‐ERGs were associated with several pivotal oncogenic processes. We also constructed a novel 10‐gene signature and relevant prognostic model for recurrence prediction of PTC. The 10‐gene signature had a C‐index of 0.723 and the relevant nomogram had a C‐index of 0.776. The efficacy of the signature and nomogram was satisfying and closely correlated with relevant clinical parameters. Furthermore, the signature also had a unique potential in differentiating anaplastic thyroid cancer (ATC) samples. Conclusions The novel EMT signature and nomogram are useful and convenient for personalized management for thyroid cancer.

Keywords

• bioinformatics
• epithelial–mesenchymal transition
• nomogram
• papillary thyroid cancer
• predictive model
• recurrence