Prognostic Value and Therapeutic Response Prediction of BRAF, RAS, and TERT Gene Mutations in Thyroid Cancer: A Large-Scale Analysis Based on the Cosmic Database

Objective : To systematically evaluate the distribution characteristics of key gene mutations (BRAF, RAS [HRAS, KRAS, NRAS], TERT) in thyroid cancer, explore their correlations with clinicopathological features, prognosis, and therapeutic response, and construct a polygenic prognostic prediction model. Methods : Thyroid cancer-related clinical sample data were downloaded from the Cosmic Database (v99, accessed on March 15, 2024), and qualified samples were included using strict screening criteria. Mutation frequency statistics, chi-square test, Kaplan-Meier survival analysis, Cox proportional hazards regression model, and pathway enrichment analysis were performed using R 4.2.1 software. Results : A total of 12,867 thyroid cancer samples were included, consisting of papillary carcinoma (72.3%), follicular carcinoma (13.5%), medullary carcinoma (8.7%), and anaplastic carcinoma (5.5%). The BRAF V600E mutation had the highest frequency (43.8%), predominantly enriched in papillary carcinoma. The RAS family mutation rate was 21.5%, with NRAS Q61R as the major variant, most prevalent in follicular carcinoma (38.2%). The TERT promoter mutation rate was 15.7%, associated with advanced tumor stage and high invasiveness. Patients with BRAF V600E mutation had significantly shortened recurrence-free survival (RFS) (HR=2.31, 95% CI:1.98-2.69, P<0.001) and poor response to radioactive iodine therapy (response rate: 31.2% vs. 68.5% in wild-type, P<0.001). Patients with concurrent BRAF V600E and TERT mutations had the worst prognosis (median overall survival [OS]: 5.2 years vs. 8.7 years in single mutation groups vs. 12.3 years in wild-type, P<0.001). The polygenic prognostic model constructed based on BRAF, RAS, and TERT achieved an AUC of 0.78 (95% CI:0.75-0.81), outperforming the traditional clinical staging model (AUC=0.65). Conclusion : BRAF, RAS, and TERT gene mutations are key predictive factors for prognosis and therapeutic response in thyroid cancer. The polygenic combined model provides a reference for clinical individualized treatment decisions.​