Tumor Mutational Burden (TMB) Test: Cancer Immunotherapy Marker and Results

A tumor mutational burden test estimates how many acquired DNA mutations are present per megabase of analyzed tumor sequence. The result is usually reported as mutations per megabase, or mut/Mb, and may be categorized as TMB-high, intermediate, or low according to the laboratory’s validated method. TMB is used as one biomarker when considering immune checkpoint inhibitors because a tumor with more mutations may produce more abnormal proteins that the immune system can recognize.

TMB is not a direct count of immune-visible targets and is not a guarantee of response. The value changes with panel size, genes covered, mutation types counted, germline filtering, tumor purity, sequencing quality, and bioinformatics. Tissue TMB and blood TMB are also not interchangeable. A threshold such as 10 mut/Mb is linked to specific assays and treatment indications rather than being a universal boundary for every cancer. The result should be integrated with tumor type, stage, prior therapy, microsatellite instability, mismatch-repair status, PD-L1, driver alterations, clinical evidence, and the current drug label in the relevant country.

• TMB estimates somatic mutations per megabase; it does not measure immune response directly.
• A high value can enrich for immunotherapy benefit in selected settings but cannot predict an individual outcome with certainty.
• Different laboratories can produce different TMB values from the same tumor because methods and counting rules vary.
• Low tumor content, poor DNA quality, or a small panel can make a result unreliable or indeterminate.
• Tissue TMB and blood TMB require separate validation, thresholds, and interpretation.

Table of Contents

• What TMB Measures
• Why TMB May Relate to Immunotherapy
• How the Test Is Performed
• Understanding High, Low, and Indeterminate Results
• The 10 Mutations per Megabase Threshold
• TMB, MSI, PD-L1, and Other Biomarkers
• Tissue TMB, Blood TMB, and Repeat Testing
• Limitations and Questions to Ask

What TMB Measures

Tumor mutational burden is an estimate of the number of somatic sequence variants in a defined amount of tumor DNA. Somatic means acquired in the cancer rather than inherited in every cell. The denominator is the number of megabases that the assay considers evaluable, so the result is normalized as mutations per megabase.

The original research standard was often whole-exome sequencing with a matched normal sample. Whole-exome sequencing analyzes most protein-coding regions, but it is resource intensive. Clinical laboratories usually estimate TMB from a validated targeted next-generation sequencing panel covering a smaller genomic footprint.

The calculation is not simply “all mutations divided by panel size.” Laboratories apply rules about which variants count. These may address:

• Nonsynonymous substitutions that change a protein.
• Synonymous substitutions, which some validated algorithms include to improve statistical precision.
• Small insertions and deletions.
• Known or suspected driver mutations.
• Variants with low allele fraction.
• Germline polymorphisms.
• Formalin-related artifacts and sequencing noise.
• Regions with inadequate coverage.

Because these rules differ, two assays can return different values from the same specimen. The value should therefore be interpreted within the assay that produced it. A result copied into a medical note without the laboratory name, panel, and threshold loses important context.

TMB is a proxy, not a direct measurement of neoantigens. A DNA mutation can create an altered protein fragment, but many mutations are never expressed, processed, presented by human leukocyte antigen molecules, or recognized by T cells. Conversely, a tumor with modest TMB can contain a highly immunogenic mutation.

Mutation clonality also matters. A clonal mutation present in most cancer cells may be more relevant to an immune response than a subclonal mutation found in a small fraction. Standard clinical TMB often counts both without fully weighting their biological importance.

The cause of the mutation burden can vary. Ultraviolet light, tobacco exposure, mismatch-repair deficiency, polymerase proofreading defects, prior therapies, and other processes can produce different mutational patterns. Tumors with the same numeric TMB are not necessarily biologically equivalent.

TMB is distinct from the number of genes altered on the report. One gene can contain several counted mutations, while a clinically important amplification or fusion may not contribute to the TMB calculation at all. A broad solid tumor NGS panel may report both individual targets and TMB, but the two outputs answer different questions.

Why TMB May Relate to Immunotherapy

Immune checkpoint inhibitors release inhibitory signals that restrain T cells. Drugs targeting PD-1, PD-L1, CTLA-4, and other checkpoints can allow an existing antitumor immune response to become more active. They do not directly attack a mutation.

The biological rationale for TMB is probabilistic. More somatic mutations create more opportunities for altered peptides, or neoantigens. Some neoantigens may be presented on the tumor-cell surface and recognized as foreign. A tumor with many potential targets may be more likely to respond once checkpoint inhibition restores T-cell activity.

This association has been observed across several cancers and clinical datasets, but it is incomplete. High TMB tumors can fail to respond because:

• The mutations do not create immunogenic peptides.
• The tumor does not express or present the relevant proteins.
• Human leukocyte antigen molecules are lost or dysfunctional.
• T cells are excluded from the tumor microenvironment.
• Other suppressive cells or pathways dominate.
• The tumor carries resistance alterations affecting interferon signaling or antigen presentation.
• The high TMB is concentrated in minor subclones.

Low TMB tumors can still respond. They may have viral antigens, one strong clonal neoantigen, an inflamed microenvironment, high PD-L1 expression, or another source of immune recognition. TMB should therefore modify probability rather than function as an absolute gate.

The predictive value also depends on cancer type. Some tumors naturally have higher mutation distributions than others. A value that is unusually high for one histology may be typical in another. The relationship between TMB and benefit can differ according to the checkpoint inhibitor, combination regimen, treatment line, and study population.

TMB may be prognostic in some datasets, but prognostic and predictive are different. A prognostic biomarker is associated with outcome regardless of treatment. A predictive biomarker identifies differential likelihood of benefit from a particular therapy. Clinical use requires evidence that the measurement helps select treatment, not only that it correlates with survival.

Real-world and trial studies continue to refine where TMB adds value. Large pan-cancer analyses can show broad associations, but they do not erase tumor-specific differences or replace a validated indication. The treating oncologist should use the evidence that matches the patient’s cancer and therapy.

TMB is most informative when it is one part of an integrated immune profile. MSI, mismatch-repair protein status, PD-L1, tumor-infiltrating lymphocytes, driver mutations, and clinical features may strengthen or weaken the rationale for immunotherapy.

How the Test Is Performed

Tissue TMB is usually calculated from formalin-fixed tumor tissue analyzed by a sufficiently broad DNA sequencing panel. The pathologist selects a tumor-rich region and estimates tumor percentage. The laboratory extracts DNA, prepares a sequencing library, sequences the targeted regions, calls variants, filters artifacts and likely germline changes, and divides the qualifying mutation count by the validated genomic footprint.

Preanalytic quality affects every step. Delayed or prolonged fixation can damage DNA. Acid-decalcified bone specimens may perform poorly. Necrosis and abundant normal cells lower tumor purity. A small biopsy can be consumed by diagnostic stains before molecular testing is ordered.

The laboratory’s analytic validation should establish:

• Minimum DNA input and tumor content.
• Genomic footprint used for TMB estimation.
• Depth and uniformity of coverage.
• Variant allele-fraction threshold.
• Mutation types included or excluded.
• Germline filtering strategy.
• Handling of known drivers and artifacts.
• Precision near the clinical cutoff.
• Reportable range and indeterminate criteria.
• Concordance with a reference method or calibrated standard.

Panel size is important because TMB is a sampling estimate. A very small panel may be disproportionately influenced by a few mutations and produce unstable values. Larger validated footprints generally improve precision, although design, coverage, and bioinformatics remain critical.

Matched-normal sequencing uses blood or another normal sample to remove inherited variants directly. Tumor-only assays usually filter likely germline variants with population databases and computational rules. Tumor-only filtering can leave rare germline variants in the count, particularly for ancestries underrepresented in reference databases, leading to overestimation.

The specimen’s tumor fraction can also alter the value. Low purity may cause true mutations to fall below the detection threshold and underestimate TMB. Copy-number changes and subclonality complicate variant detection. A laboratory should not force a numeric result when the quality is insufficient.

TMB may be bundled with comprehensive genomic profiling. The report should state the exact score, unit, category, assay-specific cutoff, and any limitations. A category without a numeric value makes cross-review difficult. Some reports also state whether the assay is an approved companion diagnostic for a particular therapy.

There is no special diet, fasting, or medication preparation for the molecular calculation itself. Preparation relates to the biopsy or blood draw. A biopsy plan should consider bleeding risk, anesthesia, lesion accessibility, and whether enough tissue will remain for pathology and additional biomarkers.

Turnaround time depends on specimen retrieval and laboratory workflow, often ranging from one to several weeks. When treatment is urgent, other validated biomarkers may be tested in parallel rather than waiting for a broad panel.

Understanding High, Low, and Indeterminate Results

A TMB report should be read in layers: the numeric score, the laboratory category, specimen quality, assay name, and clinical interpretation.

TMB-high

TMB-high means the value meets or exceeds the assay’s validated threshold. It may support consideration of an immune checkpoint inhibitor in a setting where clinical evidence, guidelines, and regulatory authorization apply.

It does not mean immunotherapy will work, that it is the best first treatment, or that every checkpoint inhibitor is equivalent. The oncologist still considers cancer type, approved indication, prior treatment, contraindications, autoimmune disease, transplant history, organ function, pace of disease, and alternative therapies.

A very high value may prompt review for a biological cause such as mismatch-repair deficiency or a pathogenic polymerase proofreading alteration. Those mechanisms can have separate treatment and hereditary implications.

TMB-low

TMB-low means the score is below the assay threshold. It lowers confidence in TMB as a reason for immunotherapy but does not prove that immunotherapy cannot help. Many approved indications use PD-L1, tumor type, clinical setting, or combination therapy rather than TMB.

Treatment should not be withheld solely because of low TMB when another established indication is present. Conversely, a low result can be important when the proposed therapy depends specifically on a TMB-high indication.

Intermediate or borderline

Some laboratories report an intermediate category or a value close to the cutoff. Measurement imprecision is most consequential near a threshold. A one- or two-mutation difference may reflect sampling or analytic variation rather than a meaningful biological divide.

Borderline results should be interpreted with assay precision, tumor quality, and other biomarkers. Repeating the same specimen on another panel may produce a different number without resolving the clinical question.

Indeterminate or not evaluable

An indeterminate result means the laboratory could not produce a reliable estimate. Causes include low tumor purity, insufficient DNA, degraded material, inadequate coverage, assay failure, or a specimen type outside validation.

Indeterminate is not low. A new tissue specimen, alternative block, or validated plasma assay may be considered if TMB would change treatment. The potential benefit must justify another procedure.

Result	Reasonable interpretation	What it does not prove
High	May increase the probability of checkpoint-inhibitor benefit in a validated setting	Guaranteed response or universal eligibility
Low	TMB alone does not support a TMB-high treatment indication	Immunotherapy can never work
Borderline	Small analytic differences may change the category	A sharp biological boundary
Indeterminate	The specimen or assay could not provide a reliable value	A low mutation burden

The 10 Mutations per Megabase Threshold

The value of 10 mut/Mb is widely recognized because a United States tumor-agnostic indication linked pembrolizumab to TMB-high unresectable or metastatic solid tumors meeting that threshold on an approved test, after prior treatment and when satisfactory alternatives were lacking. The exact current label, age range, formulation, and regulatory status should be verified when treatment is considered.

The threshold should not be generalized beyond its evidence base. Important qualifications include:

• The original clinical evidence involved particular assays, cancer cohorts, prior-treatment requirements, and trial designs.
• Different sequencing panels can classify the same tumor differently near 10 mut/Mb.
• Tumor types were not equally represented, and response rates varied.
• Some cancers have separate biomarker cutoffs or stronger tumor-specific evidence.
• A drug’s approved use can change over time and differs among countries.
• Blood TMB should not automatically use a tissue threshold.

A threshold is needed for clinical decisions, but mutation burden is continuous. A tumor at 9 mut/Mb is not biologically opposite to one at 10 mut/Mb. Laboratory precision and clinical context are especially important near the boundary.

Some studies use percentile-based or tumor-specific cutoffs, and some trials select higher values. These approaches may better account for differences in baseline mutation burden, but they are not interchangeable with an approved companion-diagnostic threshold.

The laboratory should report how its assay was calibrated. A panel may be analytically harmonized to a reference but still require clinical validation for the intended use. Correlation between two assays across many samples does not guarantee agreement for every individual near the cutoff.

Patients should ask whether the test named in the report is approved or validated for the proposed treatment. A laboratory-developed result can be clinically useful, but the treating team may need confirmatory testing when the drug label requires a specific companion diagnostic.

Cost and access also matter. Meeting a numeric cutoff does not guarantee insurance coverage or drug availability. The oncology team should document the treatment indication, assay, prior therapy, and absence or presence of alternatives.

TMB, MSI, PD-L1, and Other Biomarkers

TMB, microsatellite instability, mismatch-repair deficiency, and PD-L1 describe different aspects of tumor immunity. They overlap but are not substitutes.

MSI-high or mismatch-repair deficient tumors often accumulate many mutations and therefore may have high TMB. However, a tumor can be TMB-high without MSI-high status, and rare technical or biological discordance occurs. MSI or mismatch-repair testing can carry hereditary implications for Lynch syndrome.

PD-L1 is a protein-expression biomarker measured by immunohistochemistry. It can change by tumor site, treatment, time, and assay. Some treatment indications require a tumor proportion score, combined positive score, or immune-cell score. TMB cannot supply that score.

Driver alterations can affect the choice and sequence of therapy. In some oncogene-driven cancers, an effective targeted therapy may be preferred even when TMB is elevated. Certain drivers are associated with lower immunotherapy benefit despite PD-L1 or mutation findings.

Tumor-infiltrating lymphocytes and immune gene-expression signatures describe the existing immune environment. They are used routinely in some settings and investigationally in others. A tumor may have many mutations but few infiltrating T cells.

Human leukocyte antigen genotype and antigen-presentation defects influence whether neoantigens are visible to T cells. These factors are not usually included in standard TMB reports.

An integrated interpretation might read: TMB-high with MSI-high and mismatch-repair loss, which creates a coherent immune-sensitive profile; or TMB-high with a dominant targetable driver and low PD-L1, which requires a more nuanced sequencing decision. No single combination guarantees response.

The clinical team should prioritize biomarkers according to the cancer-specific guideline. In one disease, PD-L1 may determine first-line therapy; in another, MSI or a targetable driver may be more decisive. TMB is one layer of the profile, not a universal ranking system.

Tissue TMB, Blood TMB, and Repeat Testing

Tissue TMB is calculated from a tumor specimen. Blood TMB is estimated from circulating tumor DNA in plasma. Blood testing can be faster and avoids an invasive biopsy, but it adds challenges involving tumor fraction, shedding, and interference from blood-cell clones.

A plasma sample with very little circulating tumor DNA may underestimate mutation burden or be non-evaluable. Some cancers shed less DNA into blood, and isolated central nervous system disease may be poorly represented. A negative or low blood result should not automatically overrule an adequate tissue result.

Blood assays use their own panel size, variant filters, and cutoffs. A blood TMB value should be interpreted only with clinical evidence for that specific assay and treatment setting. Numerically identical blood and tissue scores do not necessarily represent equivalent biology.

Clonal hematopoiesis can introduce mutations from white-blood-cell clones into plasma. Matched white-blood-cell sequencing or computational filtering can reduce this problem. Failure to remove these variants can falsely elevate the apparent tumor mutation count.

TMB can evolve. Therapy may select resistant subclones and add treatment-related mutations. However, routine serial TMB testing is not established for every cancer. Repeating the test is most reasonable when a new specimen is already being obtained, the original result was inadequate, or a later treatment decision specifically depends on current profiling.

Spatial heterogeneity also matters. Different regions of one tumor or separate metastases can have different subclonal mutations. Core clonal mutations are more stable, but TMB estimates may vary by site and sample purity.

A repeat result that crosses a cutoff should be reviewed for assay and specimen differences before being interpreted as true biological change. Switching from tissue to blood, changing laboratories, or using a larger panel can alter the number independently of tumor evolution.

Limitations and Questions to Ask

TMB has limitations at every level. Analytically, the score depends on panel footprint, depth, counting rules, germline subtraction, tumor purity, and artifact filtering. Clinically, associations differ by cancer and treatment. Biologically, mutation quantity does not capture neoantigen quality, clonality, antigen presentation, immune infiltration, or suppressive pathways.

A high result can create false confidence. Immunotherapy can cause serious immune-related toxicities involving the lungs, bowel, liver, endocrine glands, skin, nervous system, heart, and other organs. The expected benefit must justify these risks. Transplant recipients and people with autoimmune disease require particularly individualized discussion.

A low result can also be overinterpreted. It should not cancel an established non-TMB indication or a combination regimen proven in that cancer. The report must be integrated into the full treatment pathway.

Before using the result, ask:

• Was TMB measured in tissue or blood?
• Which panel and genomic footprint were used?
• What mutations were counted, and how were germline variants removed?
• Did the specimen meet tumor-content and quality requirements?
• What is the numeric score, not only the category?
• Which cutoff was validated for this assay?
• Is the proposed therapy approved or guideline supported for this cancer and treatment setting?
• How do MSI, mismatch repair, PD-L1, driver alterations, and prior therapies affect the interpretation?
• Is the score close enough to the cutoff that analytic variation matters?
• Would confirmatory or alternative testing change management?

The report should be retained with the pathology and sequencing documents. New evidence can change interpretation, but reanalysis cannot correct an inadequate specimen. A molecular pathologist or molecular tumor board can help when the result conflicts with other biomarkers or when an off-label treatment is being considered.

TMB is most useful as a carefully measured, assay-specific biomarker rather than a stand-alone number. Its role is to improve the probability estimate for treatment benefit, not to replace clinical judgment.

References

• Recommendations for Tumor Mutational Burden Assay Validation and Reporting: A Joint Consensus Recommendation of the Association for Molecular Pathology, College of American Pathologists, and Society for Immunotherapy of Cancer 2024 (Consensus Recommendation)
• Harmonization of tumor mutation burden testing with comprehensive genomic profiling assays: an IQN Path initiative 2024 (Study)
• Tumor mutational burden and survival on immune checkpoint inhibition in >8000 patients across 24 cancer types 2025 (Cohort Study)
• The role of neoantigens and tumor mutational burden in cancer immunotherapy: advances, mechanisms, and perspectives 2025 (Review)
• Tumor Mutational Burden (TMB) Harmonization Project 2025 (Official Project)
• FDA approves pembrolizumab for adults and children with TMB-H solid tumors 2020 (Official Page)

Disclaimer

This article provides general educational information and is not a substitute for personalized medical advice or cancer treatment planning. TMB interpretation depends on the assay, specimen, tumor type, treatment setting, jurisdiction, current drug labeling, and other biomarkers. An oncologist and qualified molecular pathology team should determine whether a result is valid and clinically actionable.