Skip to main content
Fig. 3 | Journal of Translational Medicine

Fig. 3

From: Construction of a predictive model for immunotherapy efficacy in lung squamous cell carcinoma based on the degree of tumor-infiltrating immune cells and molecular typing

Fig. 3

Construction of an immune infiltration prediction model based on 17 genes. A Venn diagram of the intersection of differentially expressed genes, genes affecting overall survival and disease-free survival of patients in the training set, as well as in the two validation sets GSE126044 and GSE135222; B Lasso coefficient distribution diagram of 20 genes with x-coordinate log (λ) for screening the best tuning parameter (λ); C screening of the tuning parameter in the lasso regression model based on tenfold cross-validation; Plotting was performed based on this value and the AUC value of the ROC curve. A vertical dashed line was drawn at the best value by using the minimum standard and 1 standard error of the minimum standard (1-SE standard); D nomogram plotted based on the IPTS. IPTS could predict whether the patient had an NMF type of either C1 or C2. The higher the IPTS, the higher the probability of the patient having type C2. When IPTS = 0.6369, the probability of the patient in the C1 and C2 types was 50%. E Calibration curve plot according to lasso regression analysis. The x-coordinate represented the probability that the model predicts type C2 for patients, and the y-ordinate represented the actual probability. F Sankey diagram plotted according to NMF or IPTS subtyping, and 17 gene signatures. The low IPTS group tended towards type C1, while the high IPTS group was more representative of type C2

Back to article page