Identification of a Recurrence Signature and Validation of Cell Infiltration Level of Thyroid Cancer Microenvironment

Though many patients with thyroid cancer may remain indolent, lymph node metastases and the recurrence rates are still about 50% and 20%, respectively. There is still no ideal method to predict its relapse. Here, by analyzing the gene transcriptome profiles of eight Gene Expression Omnibus (GEO), we screened 77 commonly differential expressed genes. Next, Least Absolute Shrinkage and Selection Operator (LASSO) regression model was used to identify seven genes (i.e., FN1, PKIA, TMEM47, FXYD6, SDC2, CD44, and GGCT) highly associated with recurrence data from The Cancer Genome Atlas (TCGA) database. These patients were divided into low and high-risk groups with specific risk-score formula. Univariate and multivariate Cox regression further revealed that the 7-mRNA signature plays a functional causative role independent of multiple clinicopathological parameters. The 7-mRNA-signature integrated nomogram showed better discrimination, and decision curve analysis demonstrated that it is clinically useful. Besides, patient with lower risk score shows a relatively lower level of activated dendritic cells (DCs), resting DCs, regulatory T cells and γδT cells, and process of DCs apoptotic. In conclusion, our present immune-related classifier could produce a potential tool for predicting early-relapse.