TMB Consortium Develops a TMB Calibration Tool

October 13, 2021

179

Estimation of tumour mutational burden (TMB) varies across different panels, with panel size, gene content, and bioinformatics pipelines contributing to empirical variability. Statistical calibration can achieve more consistent results across panels and allows for comparison of TMB values across various panel assays. The TMB Harmonization Project leveraged the expertise and insight of 16 different diagnostic laboratories to objectively evaluate the empirical variability across panel TMB values and to propose best practices for panel TMB alignment. The aim is to provide data and guidance that will help improve the consistency and reliability of panel tissue TMB estimation across platforms and facilitate the use of this complex biomarker in clinical decision making. To promote reproducibility and comparability across assays, the TMB Consortium developed a software tool and made it publicly available. The findings are described in an article published on 30 September 2021 in the Annals of Oncology.

The Friends of Cancer Research TMB Harmonization Consortium consists of several diagnostic manufacturers, academics, pharmaceutical companies, the US National Cancer Institute, Frederick National Laboratory for Cancer Research, and the US Food and Drug Administration. They previously reported results from the first phase of the TMB harmonisation project where the variability across 11 commercial and academic panel assays was described and Consortium-endorsed recommendations were proposed for the analytical validation of TMB assays. Furthermore, the TMB Consortium partnered with Quality in Pathology in Germany, to complement its approach and enrich its perspective on the variability in TMB estimates across laboratories through a technical comparability study.

TMB measurements aid in identifying patients who are likely to benefit from immunotherapy. However, there is empirical variability across panel assays and factors contributing to this variability have not been comprehensively investigated. In this study, the TMB Consortium set out to characterise the empirical variability in TMB measurements across platforms using a common set of cell lines and clinical samples tested across 16 panel assays from 16 participating laboratories.

They aimed to elucidate how certain factors such as panel size, gene content and bioinformatics pipelines impact TMB estimates, and to investigate the use of a calibration tool based on TCGA data and human tumour derived cell lines that will facilitate comparability across different panel assays.

The study team processed 29 tumour samples and 10 human-derived cell lines and distributed to 16 laboratories and each of them used their own bioinformatics pipelines to calculate TMB and compare to whole exome results. Additionally, theoretical positive percent agreement and negative percent agreement of TMB were estimated. The impact of filtering pathogenic and germline variants on TMB estimates was assessed. Calibration curves specific to each panel assay were developed to facilitate translation of panel TMB values to whole exome sequencing (WES) TMB values.

Panel sizes greater than 667Kb are necessary to maintain adequate positive percent agreement and negative percent for calling TMB high versus TMB low across the range of cut-offs used in practice. Failure to filter pathogenic variants when estimating panel TMB resulted in overestimating TMB relative to WES for all assays. Filtering out potential germline variants at >0% population minor allele frequency resulted in the strongest correlation to WES TMB.

Application of a calibration approach derived from TCGA data, tailored to each panel assay, reduced the spread of panel TMB values around the WES TMB as reflected in lower root mean squared error (RMSE) for 26 of 29 of the clinical samples (90%), although RMSE across samples at the laboratory level was less often reduced.

In this work the TMB Consortium demonstrated that the utilisation of a calibration tool based on a universal reference standard derived from TCGA data can enhance comparability of TMB across different panel assays.

The Friends of Cancer Research TMB Harmonization initiative uses a distributed research model and costs incurred are supported by each participating organisation. Additional sources of funding were provided by AstraZeneca, Bristol-Myers Squibb, EMD Serono, Genentech and Merck & Company Inc.