Fig. 4

Constructed, evaluated, and validated a valid risk model. a Results of univariate Cox regression analysis. b Effective coefficient feature of LASSO model (For the accuracy of the model, lambda.min (0.008711847) was chosen as λ for model construction). c λ selection graph in LASSO model (The two dotted lines indicate two special lambda values, lambda.min on the left and lambda.1se on the right. lambda.min is more accurate and uses a larger number of genes). d The expression of five prognostic genes in tumor and normal tissues. e Risk curve for high and low risk groups (The risk graph is composed of two parts, with the horizontal coordinate being the same, which the patient’s risk value increasing from left to right). f Heat map of two risk groups (The expression of biomarkers in high and low risk groups was mapped using pheatmap package of R software). g, h K-M survival curve of the two risk groups. i, j ROC curves evaluate the effectiveness of risk models (The higher the AUC value, the higher the prediction accuracy)