Tumor Mutational Burden Estimation with VarSome Clinical

*Research use only feature

What is Tumor Mutational Burden (TMB)?

Tumor Mutational Burden (TMB) refers to the number of somatic variants detected in a tumor sample per megabase (Mb) of the genomic sequence analyzed. It is reported in mutations per megabase (mut/Mb). TMB serves as a predictive biomarker for immune checkpoint inhibitors—immunotherapy drugs targeting pathways such as PD-1 and PD-L1 (Chan et al., 2019). In some cases, TMB is used as a companion diagnostic to determine treatment eligibility. For example, pembrolizumab is approved for tumors with TMB ≥10 mut/Mb.

Variability in TMB Calculation

TMB can be calculated using whole-exome sequencing (WES) or targeted gene panels. While WES provides a more comprehensive view, panel-based TMB estimation is more common in clinical practice. However, panel-based TMB is more susceptible to variability due to differences in panel size, gene content, and bioinformatics pipelines. To enable cross-platform comparison, statistical calibration methods are employed.

VarSome Clinical follows the Uniform TMB Calculation Method [1]. This standard uses a defined coding region and applies consistent variant selection criteria. VarSome Clinical calibrates panel-based TMB, including for the Agilent SureSelect Cancer CGP assay, against WES-derived values to ensure reliability.

How TMB is Calculated in VarSome Clinical

VarSome Clinical supports TMB estimation for somatic tumor-only analyses, launched in targeted mode, for both WES (determined as assays targeting >= 31M bp that are not WGS) and Agilent SureSelect Cancer CGP assays, with a required minimum average coverage of 50X. Two TMB scores are calculated:

  • TMB (all): Includes all coding variants, including synonymous mutations.
  • TMB (non-synonymous): Excludes synonymous mutations.

A crucial step for the accurate estimation of TMB is the removal of germline variants from somatic analyses run in tumor-only mode. Varsome Clinical's somatic pipeline is based on Sentieon TNHaplotyper2 and leverages a panel of normals and germline resources to mark germline variants. Additional filters are applied to exclude germline variants.

Where to find the TMB information in your analyzed sample

Any somatic sample analyzed either with a WES sequencing assay or an Agilent SureSelect Cancer CGP panel will get a TMB classification. In the results page of an analysis, the TMB score is visible above the Variant Table:

In addition, the TMB score can be found in the Sample/Analysis Information action:

Key Steps in TMB Calculation

  1. Variant Filtering:
    • Variants must be located within the intersection of the assay’s target regions and the CCDS-defined exome. In other words, in those assay target regions that are included in the CCDS. 
    • Only variants that pass QC control are considered.
    • Coverage at the variant site must be ≥25X, with at least 3 reads supporting the tumor allele.
    • Variant Allele Frequency (VAF) must be ≥5%.
    • Mitochondrial variants are excluded.
    • Acceptable coding impacts on the canonical transcript include: missense, nonsense, frameshift, in-frame indels, stoploss, stopgain, startloss, and startgain. Synonymous variants are only included in the TMB (all) score.
  2. TMB Calculation:
    • The final number of qualifying variants is divided by the length (in megabases) of the eligible genomic region.
    • For panel assays (e.g., Agilent SureSelect CGP), a calibration step is performed based on The Cancer Genome Atlas (TCGA) data to align panel-based TMB with WES-based results. The calibration includes a prediction curve with 95% confidence intervals (CI).

Thresholds

  • A threshold of 10 mut/Mb is applied to determine TMB status.
    • TMB-HIGH: TMB ≥ 10 mut/Mb
    • TMB-LOW: TMB < 10 mut/Mb

TMB scores are capped at 40 mut/Mb for reporting purposes, even if the calculated value exceeds this limit.

Example Cases

  1. Panel-Based Case (Agilent SureSelect CGP Assay)

    • Both TMB (all) and TMB (non-synonymous) scores are below 10 mut/Mb.
    • The sample is classified as TMB-LOW.
    • Calibrated scores are provided with a 95% CI range of 0.9–7.5 mut/Mb.
  2. WES-Based Case

    • Both TMB (all) and TMB (non-synonymous) scores exceed 10 mut/Mb.
    • The sample is classified as TMB-HIGH.
    • No calibration is needed; CI is marked as N/A.
  3. Panel-Based Case (Agilent SureSelect CGP Assay) 

    • Both TMB scores exceed 10 mut/Mb, indicating TMB-HIGH.
    • However, the CI for TMB (non-synonymous) overlaps with the 10 mut/Mb threshold, introducing ambiguity in classification.
    • This is visually flagged with a red exclamation mark in VarSome Clinical.

References

  1. Chan TA, Yarchoan M, Jaffee E et al. (2019) Development of tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic. Ann Oncol . 2019 Jan 1;30(1):44-56. Link 
  2. Merino DM, McShane LM, Fabrizio Det al. (2020) Establishing guidelines to harmonize tumor mutational burden (TMB): in silico assessment of variation in TMB quantification across diagnostic platforms: phase I of the Friends of Cancer Research TMB Harmonization Project. J Immunother Cancer. 2020 Mar;8(1). Link
  3. Vega DM, Yee LM, McShane LM, et al. Aligning Tumor Mutational Burden (TMB) quantification across diagnostic platforms: Phase 2 of the Friends of Cancer Research TMB Harmonization Project. Annals of Oncology 2021. Link