Fig. 9

The ability of NIRS to predict the responsiveness of immunotherapy and its external validation. A-D Analyses of the curative effect of four immune checkpoint inhibitors (ICIs) in high- and low- NIRS groups. E The external validation of the efficacy of NIRS in predicting the responsiveness of PD-1 and CTLA-4 inhibitors, R indicated sensitivity to ICIs of the corresponding checkpoint, and noR indicated insensitivity. Low NIRS group was significantly associated with PD-1 inhibitor responsiveness after Bonferroni correction. F–H K-M survival curve of patients with high- and low- NIRS (F), the distribution of the four-level curative effect evaluation indicators in the high- and low- NIRS groups after immunotherapy (G) and the box plot of the difference between four curative effect outcomes and NIRS (H) in the IMvigor210 cohort. I–K K-M survival curve of patients with high- and low- NIRS (I), the distribution of the four-level curative effect evaluation indicators in the high- and low- NIRS groups after immunotherapy (J) and the box plot of the difference between four curative effect outcomes and NIRS (K) in the PRJEB23709 cohort. L Differential analysis of immune exclusion between high- and low- NIRS groups, ***P < 0.001