Research Paper Volume 16, Issue 3 pp 2494—2516
Risk model based on genes regulating the response of tumor cells to T-cell-mediated killing in esophageal squamous cell carcinoma
Figure 4. Construction and validation of the GRTTK based risk model. (A) Optimal regularization parameter (λ) selection (λ = 0.0407) using minimum partial likelihood deviance for TCGA cohort. (B) Coefficients of the risk model corresponding to the optimal λ. (C) Coefficient values for the eight genes in the risk model. (D) Kaplan-Meier analysis of overall survival (OS) based on risk groups in TCGA cohort. (E) ROC curves evaluating risk model performance in TCGA cohort. (F) Multivariable Cox analysis of the TCGA-ESCC cohort. (G) OS in the high-risk and low-risk groups in the GEO cohort (P = 0.0010). (H) ROC curves for the risk model for the GSE53622 cohort. (I) Multivariate Cox analysis of the GSE53622 cohort. (J) A nomogram displaying the predictive value of clinical stage and risk models. (K) AUC for the combined evaluation of patient prognosis using clinical stage and risk scores.