Results
We analyzed the genomic and survival data of 1661 patients treated with ICIs including 731 and 930 primary and metastatic lesions, respectively. We found that the tissues obtained from the metastatic lesions had more co-occurring gene mutations than those obtained from the primary lesions (P <0.05, Q <0.05) (Figure 1A). Importantly, tissues from patients with metastasis lesions had a higher TMB score than those from patients with primary lesions (P =0.024) (Figure 1B).
The Kaplan-Meier method was applied to analyze the OS. According to the median cut-off values (5.58) of TMB from the primary lesions, we compared the survival in the lesions with low TMB score and the high TMB score. The result revealed that there was no significant difference in OS between the low and high TMB groups in primary lesions (P = 0.416) (Figure 2A). However, using the same cut-off values for TMB from the primary lesions, there was a significant difference in survival between the high TMB and low TMB group in metastatic lesions, and patients with a high TMB had improved survival (P <0.001) (Figure 2B).
To further verify the difference in the effect of TMB scores in guiding the prognosis of ICIs between primary and metastatic lesions, we selected 350 non-small cell lung cancer (NSCLC) patients who were treated with ICIs including 171 and 179 primary and metastatic lesions, respectively. The results of CheckMate227 and CheckMate568 trials9 suggested that in NSCLC patients with TMB ≥10 mutations/MB, progression-free survival was significantly prolonged after immunotherapy. Therefore, according to the cut-off value (cut-off=10), we analyzed the survival in NSCLC patients with different TMB scores for primary and metastatic lesions. In patients whose tissue samples came from primary lesions, there was no significant difference in OS (P = 0.474) between the low TMB and the high TMB groups (Figure 2C). However, the TMB was significantly associated with OS in patients whose tissue samples came from metastasis lesions (P =0.012) (Figure 2D). This suggested that the role of TMB as a prognostic biomarker for the response to ICIs was different in patients with tissues originated from primary lesions and metastasis lesions.
We predicted the optimal cut-off values of TMB in primary and metastatic lesions in 1661 pan-cancer cases using the X-tile model. The optimal cut-off values of TMB that predicted survival was 20.19 in primary lesions (P < 0.001) (Figure 2E) and 10.18 in metastasis lesions (P < 0.001) (Figure 2F). The data revealed that the optimal score of TMB in metastatic lesions was significantly lower than that in primary lesions. Therefore, the optimal cut-off value varied to vary for primary and metastasis lesions in all pan-cancer tissues.