Discussion
Recent studies indicated that TMB may be a new biomarker to predict the response to ICIs in different tumor entities.10-13Several clinical trials have provided different cut-off values for TMB which leads to difficulties in clinical application. The data providing clinical guidance for TMB cut-off values are scarce. Moreover, the cut-off values of TMB for pan-cancer may not be ideal for clinical use. Therefore, we calculated and compared the difference in TMB cut-offs between primary and metastatic lesions on the basis of cBioPortal mutational data.
In this study, the co-occurring mutations and TMB scores were significantly higher in metastatic lesions than in primary lesions. Several theories might explain why metastatic tissues have more gene mutations than primary tissues. First, genetic analysis of matched sets of patients’ primary and metastasis lesions revealed that metastatic lesions had mutations common to both and mutations that were distinct from primary lesions, which is the biggest cause of cancer death. Mutations in metastatic lesions were located at the chromosome ends and affected chromosome stability, gene expression regulation, and DNA repair. Previous research indicated that genomic instability was observed in metastatic subclones and/or metastatic lesions of primary lesions.14-17 Additionally, studies revealed that new genetic alterations, not seen in primary lesions, were observed in metastasis lesions.
Hong, M. K. H. et al. 18 revealed that mutation or amplification of both DNA double-strand break repair and mismatch repair gene was indicative of metastasis in prostate cancer progression. Similarly, Brastianos, P. et al. 19 demonstrated that 53% of brain metastasis harbored mutations not seen in primary lesions. These data revealed that metastasis lesions would obtain new additional mutations compared to primary lesions, which was consistent with our findings.
Additionally, the variable cut-offs of TMB between primary and metastatic lesions can be attributed to distinct tumor microenvironments and other factors that independently predict response to ICIs including immune infiltration, human leukocyte antigen genotype, clonality, and expression levels of checkpoint molecules.20-23 Our result suggested high TMB was associated with the increased OS based on the optimal cut-offs in pan-cancer types. Therefore, TMB as a new biomarker in pan-cancer may reflect fundamental mechanisms that determine the response to ICIs treatment.
Our data addresses a fundamentally important question in ICIs treatment for clinicians. Although TMB could predict survival and response to ICIs across pan-cancer types, the cut-off values of TMB differed between primary and metastatic lesions. Therefore, the TMB score for primary and metastatic lesions should be determined separately. This finding will greatly influence clinical guidance.
This study has a few limitations. This was a retrospective study using data from the cBioPortal database. Therefore, the results should be verified in a prospective study. Furthermore, clarification of appropriate cut-offs of TMB and the integration of relevant clinical variables for each cancer type will be necessary to allow for the application of TMB as a prognostic biomarker to reflect ICIs treatment in future clinical studies.