A growing number of studies is providing evidence that a tumor’s mutational burden (TMB) is a good marker for treatment of tumor patients. For instance, a current phase I trial in patients with NSCLC (Non Small Cell Lung cancer) has shown that immune therapy markedly benefits patients with a high TMB value (fig 1).
Fig.1, Results from the CheckMate 227 study: Patients were stratified by TMB into low (<10) and high (>=10 Mut/Mb).
TMB denotes the number of mutations found in the patient’s tumor but absent from the patient’s healthy tissue. The tumor-specific mutations evaluated for TMB are called coding somatic sequence mutations (fig 2). They are determined by sequencing a part of the tumor’s genome. The number of such mutations is normalized to the total size of all known genes, i.e., to the number of coding bases in the genome. The resulting value is represented as mutations per millions of base bairs (Mut/Mb).
Fig. 2, The tumor mutation burden is computed based on somatic sequence mutations. Larger genomic changes such as chromosome aberrations, copy number changes and structural variants are not considered.
The clinical importance of this marker is due to the fact that a higher number of tumor-specific mutations leads to a higher number of abnormal proteins in the tumor cells and, finally, to a stronger “foreignness” of the tumor tissue in relation to healthy tissue. Human cells, including tumor cells, present part of their protein content on their surfaces. Immune system cells regulary check these “presented” protein parts. The higher a tumor’s TMB, the more likely it is that the immune system will recognize the tumor as foreign and attack it. Some tumors block detection by immune cells. In these cases, immune therapy (checkpoint inhibition) can lead to drastic improvements.
Fig.3, Presentation of tumor cell-derived somatic peptides. Somatic mutations arise frequently in cancer and alter permanently the genomic information. These genetic changes can result in expression of proteins with altered amino acid sequence. These proteins are processed within the tumor cell by the proteasom into short peptides which are subsequently loaded onto MHC class I molecules. In the following, the peptide-MHC class I complex arrives at the tumor cell surface for presentation to immune effector cells (CD8+ T cell). In this process, peptides which carry a somatic change, and thus display a particularly strong immunostimulatory potential, can be presented on the tumor cell surface and cause an effective anti-tumor immune response.
To determine the TMB, a tissue sample of the patient’s tumor is taken (e.g., as a biopsy), the genomic material is extracted and sequenced using NGS (next generation sequencing). By bioinformatic analysis of the data, the number of mutations in the tumor tissue is determined. It is important to also analyse healthy tissue from the same patient, as the resulting datasets can be compared to distinguish new mutations in the tumor from naturally occuring variants that are part of the patient’s genomic makeup (fig 4a).
Fig. 4a, Methods for determining TMB. (A) Using tumor and healthy tissue allows for a safe detection of somatic variants.
There are methods that try to determine TMB without analyzing a healthy tissue sample. These are based on the fact that a tumor tissue sample always contains small amounts of healthy tissue. As a result, sequencing a tumor tissue sample in fact analzyes a mixture of two tissues. Using statistical models, such methods try to separate the mixture data into two datasets (fig 4b). The performance of such methods improves with higher admixture of healthy tissue, but always suffers from a margin of error in the area of 30% (Jones 2015, Yelinski 2018, value for 50% contamination).
Fig. 4b, Algorithmic determination of somatic variants based on only one tissue sample is possible but carries a risk of false-positive results.
To improve sensitivity, sequencing is not performed on the whole genome but limited to a subset of a few hundred genes, called a gene panel. Mutations are detected based on this dataset and TMB is then extrapolated to all genes. However, the gene panel must not be too small, as this would lead to very imprecise results. Panels should not fall below a minimum size of 1.5 Mb (Buchhalter 2018). CeGaT’s Tumor Immuno-Oncology Analysis, at 2.2 Mb, is safely above this boundary and thus ensures a robust estimation of TMB (fig 5).
Fig. 5, Extrapolation of TMB from CeGaT’s Tumor Immuno-Oncology Analysis (2.2 Mb) to the complete coding region (38 Mb). Classification into low (<10 Mut/Mb) and high TMB (>=10 Mut/Mb) is shown by color.
When extrapolating from a gene panel to all genes, detected mutations are first split into so-called driver and passenger mutations. While driver mutations are necessary for the development and maintenance of the tumor and affect known tumor genes, passenger mutations are a consequence of the tumor’s genetic instability and occur uniformly distributed over all genes. A tumor gene panel examines known and therapeutically relevant tumor genes and thus will be significantly enriched for driver mutations with respect to the rest of the genome. Driver mutations must not be used in extrapolation as this would lead to overestimated TMB. Passenger mutations, on the other hand, occur at the same rate in tumor genes as in all other genes and are the basis for extrapolation. The extrapolated TMB is thus the sum of the unchanged number of detected driver mutations and extrapolated the number of passenger mutations in all genes, normalized by the total size of all known genes.
To draw clinical conclusions from the TMB result, the value has to be reliable. This means that the whole process from sample preparation, to data generation and finally bioinformatic analysis needs to be quality assured. Quality control and standardization are of paramount importance to ensure reproducible results and to make sure adequate measures are in place to prevent errors (e.g., sample swapping).
CeGaT is accredited by German DAkkS ISO 15189 and American CAP/CLIA accreditation and offers quality assured TMB evaluation for clinical use. We recommend TMB analysis based on tumor and healthy tissue on our in-house developed gene panel which assures sensitive and reliable determination of TMB.
High-quality determination of TMB is the basis for therapeutic decisions on Checkpoint Inhibition and should be regarded as part of routine diagnostics in tumor therapy.