A novel exosome‐related prognostic risk model for thyroid cancer

Abstract Aim The aim was to build an exosome‐related gene (ERG) risk model for thyroid cancer (TC) patients. Methods Note that, 510 TC samples from The Cancer Genome Atlas database and 121 ERGs from the ExoBCD database were obtained. Differential gene expression analysis was performed to get ERGs in TC (TERGs). Functional enrichment analyses including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were conducted on the TERGs. Then we constructed a model based on LASSO Cox regression analysis. Kaplan‐Meier survival analysis was applied and a Nomogram model was also built. The immune landscape was evaluated by CIBERSORT. Results Thirty‐eight TERGs were identified and their functions were enriched on 591 GO terms and 30 KEGG pathways. We built a Risk Score model based on FGFR3, ADRA1B, and POSTN. Risk Scores were significantly higher in T4 than in other stages, meanwhile, it didn't significantly differ in genders and TNM N or M classifications. The nomogram model could reliably predict the overall survival of TC patients. The mutation rate of BRAF and expression of cytotoxic T‐lymphocyte‐associated protein 4 were significantly higher in the high‐risk group than in the low‐risk group. The risk score was significantly correlated to the immune landscape. Conclusion We built a Risk Score model using FGFR3, ADRA1B, and POSTN which could reliably predict the prognosis of TC patients.