Fig. 4

Prediction of outcomes of anti-PD-L1/PD-1 immunotherapy using EC.SENESCENCE.SIG. A Correlation of EC.SENESCENCE.SIG score with TMB for each cancer entity in the pan‑cancer TCGA cohort. B, C Differences in EC.SENESCENCE.SIG scores of TEC derived from patients with different responses to anti-PD-L1/PD-1 immunotherapy, p values are shown. D Flow chart describing the construction of machine learning algorithms-based predictive models for immunotherapy response. E Multiple receiver operating characteristic (ROC) plot showing the performance of different machine learning algorithms in the validation set. F ROC plot presenting the performance of the final EC.SENESCENCE.SIG model in the validation set. G Heatmap and H circus plots show the comparison between the performance of the EC.SENESCENCE.SIG model and previously published pan-cancer models for response to anti-PD-L1/PD-1 immunotherapy on different testing sets