Systematic assessment and optimizing algorithm of tumor mutational burden calculation and their implications in clinical decision-making

Tumor mutation burden (TMB) has been validated as a biomarker to predict the response of immune checkpoint inhibitors (ICIs) treatment in various cancers. However, the effects of different sequencing platforms, cancer types, and calculation algorithms on TMB as well as its cut-off value for predicting immunotherapy efficacy in the East Asian population still need to be further investigated. In this study, the data of 4126 samples generated by targeted panel sequencing or whole-exome sequencing (WES) in different platforms and public sequencing data from 3680 samples that contained targeted panel sequencing, WES and whole-genome sequencing (WGS) were obtained. The impact of different sequencing platforms and methods on TMB calculation was assessed. No significant bias was found in TMB calculated by different platforms. However, TMB calculated from WGS was significantly lower than those calculated from targeted panel sequencing and WES. The distribution of TMB at different sequencing depths and tumor purity were analyzed. There was no significant difference in the distribution of TMB when the sequencing depth was greater than 500, the tumor purity estimated by hematoxylin-eosin (HE) staining was between 0.1-1.0 or estimated by next-generation sequencing (NGS) was greater than 0.4. In addition, the somatic-germline-zygosity (SGZ) algorithm was optimized to calculate TMB from tumor-only sequencing samples in the East Asian population. The correlation coefficient of TMB calculated with the optimized SGZ algorithm and paired normal-tumor sequencing is 0.951. Furthermore, the optimal cut-off value of TMB in East Asian lung cancer patients treated with ICIs was determined to be 7 mut/Mb instead of 10 mut/Mb through the ROC curve and Log-rank analysis in the training cohort and validated in the test cohort. Patients with TMB ≥ 7 mut/Mb had better outcomes than patients with TMB<7 mut/Mb. In conclusion, this study systematically analyzed the factors that influenced the TMB calculation and optimized the SGZ algorithm to calculate TMB from tumor-only sequencing samples in the East Asian population. More importantly, the cut-off value of TMB for predicting immunotherapy efficacy was determined to be 7 mut/Mb instead of 10 mut/Mb in East Asian lung cancer patients, which can help in clinical decision-making.