Identification of novel immune-related molecular subtypes and a prognosis model to predict thyroid cancer prognosis and drug resistance

Background: Thyroid cancer is a common malignant tumor of the endocrine system that has shown increased incidence in recent decades. We explored the relationship between tumor-infiltrating immune cell classification and the prognosis of thyroid carcinoma. Methods: RNA-seq, SNV, copy number variance (CNV), and methylation data for thyroid cancer were downloaded from the TCGA dataset. ssGSEA was used to calculate pathway scores. Clustering was conducted using ConsensusClusterPlus. Immune infiltration was assessed using ESTIMATE and CIBERSORT. CNV and methylation were determined using GISTIC2 and the KNN algorithm. Immunotherapy was predicted based on TIDE analysis. Results: Three molecular subtypes (Immune-enrich(E), Stromal-enrich(E), and Immune-deprived(D)) were identified based on 15 pathways and the corresponding genes. Samples in Immune-E showed higher immune infiltration, while those in Immune-D showed increased tumor mutation burden (TMB) and mutations in tumor driver genes. Finally, Immune-E showed higher CDH1 methylation, higher progression-free survival (PFS), higher suitability for immunotherapy, and higher sensitivity to small-molecule chemotherapeutic drugs. Additionally, an immune score (IMScore) based on four genes was constructed, in which the low group showed better survival outcome, which was validated in 30 cancers. Compared to the TIDE score, the IMScore showed better predictive ability. Conclusion: This study constructed a prognostic evaluation model and molecular subtype system of immune-related genes to predict the thyroid cancer prognosis of patients. Moreover, the interaction network between immune genes may play a role by affecting the biological function of immune cells in the tumor microenvironment.