Enhancing the quality of panel-based tumor mutation burden assessment: a comprehensive study of real-world and in-silico outcomes

Yuanfeng Zhang,Duo Wang,Zihong Zhao,Rongxue Peng,Yanxi Han,Jinming Li,Rui Zhang

DOI: https://doi.org/10.1038/s41698-024-00504-1

2024-01-23

npj Precision Oncology

Abstract:Abstract Targeted panel-based tumor mutation burden (TMB) assays are widely employed to guide immunotherapy for patients with solid tumors. However, the accuracy and consistency of this method can be compromised due to the variability in technical details across different laboratories, particularly in terms of panel size, somatic mutation detection and TMB calculation rules. Currently, systematic evaluations of the impact of these technical factors on existing assays and best practice recommendations remain lacking. We assessed the performance of 50 participating panel-based TMB assays involving 38 unique methods using cell line samples. In silico experiments utilizing TCGA MC3 datasets were performed to further dissect the impact of technical factors. Here we show that the panel sizes beyond 1.04 Mb and 389 genes are necessary for the basic discrete accuracy, as determined by over 40,000 synthetic panels. The somatic mutation detection should maintain a reciprocal gap of recall and precision less than 0.179 for reliable psTMB calculation results. The inclusion of synonymous, nonsense and hotspot mutations could enhance the accuracy of panel-based TMB assay. A 5% variant allele frequency cut-off is suitable for TMB assays using tumor samples with at least 20% tumor purity. In conclusion, this multicenter study elucidates the major technical factors as sources of variability in panel-based TMB assays and proposed comprehensive recommendations for the enhancement of accuracy and consistency. These findings will assist clinical laboratories in optimizing the methodological details through bioinformatic experiments to enhance the reliability of panel-based methods.

oncology

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the accuracy and consistency issues caused by the differences in technical details of tumor mutation burden (TMB) detection methods based on targeted panels. Specifically, the paper focuses on the following aspects:

1. **Influence of technical factors on TMB assessment**: There are significant differences among different laboratories in terms of panel size, somatic mutation detection methods, and TMB calculation rules. These differences may lead to inconsistent TMB assessment results and a decline in accuracy.

2. **Best practice recommendations**: Currently, there is a lack of systematic evaluation and technical guidelines to standardize these detection methods. Therefore, through systematic experiments and analysis, the paper aims to put forward best practice suggestions for improving the accuracy and consistency of TMB detection.

3. **Specific research content**:
- **Experimental design**: The paper used 50 participating targeted panel TMB detection methods, involving 38 unique detection methods, and carried out experiments using cell line samples. In addition, computer - simulated experiments were also carried out through the TCGA MC3 data set.
- **Main findings**:
- **Panel size**: Panels exceeding 1.04 Mb and 389 genes are necessary conditions for basic discrete accuracy. For reliable continuous accuracy, the panel size should reach at least 6.02 Mb and 2,126 genes.
- **Somatic mutation detection**: The reciprocal gap between the recall rate and the precision rate of somatic mutation detection should be less than 0.179 to ensure reliable psTMB calculation results.
- **Mutation types**: Including synonymous mutations, nonsense mutations, and hot - spot mutations can improve the accuracy of panel TMB detection.
- **Variant allele frequency (VAF) threshold**: A 5% VAF threshold is suitable for samples with a tumor purity of at least 20%.

4. **Conclusion**: This multi - center study clarifies the main technical factors affecting panel - based TMB detection and proposes comprehensive improvement suggestions to improve accuracy and consistency. These suggestions will help clinical laboratories optimize method details through bioinformatics experiments, thereby enhancing the reliability of panel methods.

In summary, through detailed experiments and analysis, the paper reveals the main technical factors affecting panel - based TMB detection and provides specific improvement measures to enhance its reliability and consistency in clinical applications.