Development of a Costimulatory Molecule Signature to Predict Prognosis, Immune Landscape, and Response to Immune Therapy for Hepatocellular Carcinoma

Figure 3

The establishment of prognostic signatures on the basis of costimulatory molecule genes. (a, b) LASSO regression was performed on prognostic costimulatory molecule genes preliminarily selected by univariate Cox regression, and eight genes were identified for constructing prognostic model. (c) Kaplan-Meier survival for OS of patients in both groups from TCGA data. (d) AUC of time-dependent ROC curves confirmed that our risk model has predictive property. (e) Distribution and correlation of survival time, survival status, and risk score of patients from TCGA data. (f) t-SNE and PCA analysis demonstrated the distribution of patients in two groups from TCGA data.

(a)

(b)

(c)

(d)

(e)

(f)