Tumor mutational burden, or TMB, is a count of how many mutations sit within a tumor’s DNA, and it has become one of oncology’s most talked-about biomarkers because tumors carrying more mutations tend to respond better to immunotherapy. The logic is straightforward: more mutations mean more abnormal proteins on the tumor’s surface, giving the immune system more targets to attack. That simple idea has already earned TMB a role in drug approvals and treatment decisions, but the reality in the clinic is messier than the concept suggests.
Why More Mutations Can Be Good News
Every time a cancer cell divides, its DNA can pick up errors. Some of those errors change the instructions for building proteins, producing altered versions that the body has never seen before. These altered proteins, called neoantigens, get displayed on the surface of cancer cells. Because they look foreign, the immune system can recognize them as threats and mount an attack through T cells, the body’s trained killers of abnormal cells.1PubMed Central. The role of neoantigens and tumor mutational burden in cancer immunotherapy: advances, mechanisms, and perspectives
A tumor with hundreds of mutations will, on average, present far more neoantigens than one with only a handful. That greater neoantigen load translates into a busier battlefield: more activated CD4 and CD8 T cells infiltrate the tumor, and the surrounding immune environment ramps up expression of checkpoint molecules like PD-1 and PD-L1, along with other signals of an ongoing immune fight.2PubMed Central. Mutations in DNA repair genes are associated with increased neo-antigen load and activated T cell infiltration in lung adenocarcinoma Immune checkpoint inhibitors work by releasing the brakes on that fight, so a tumor already swarming with immune cells has more to gain from those drugs.
Not All Mutations Are Created Equal
TMB is usually reported as the total count of mutations per megabase of DNA sequenced. But that headline number hides an important distinction: where those mutations sit within the tumor matters enormously. Some mutations are clonal, meaning they exist in virtually every cancer cell in the tumor because they arose early and were passed down as the tumor grew. Others are subclonal, present in only a fraction of the cells because they appeared later in a particular branch of the tumor’s family tree.
Research in advanced bladder cancer found that responders to immune checkpoint inhibitors had roughly twice the proportion of clonal mutations compared to non-responders, and that computing a “clonal TMB” sharpened the separation between the two groups far more than total TMB alone.3PubMed Central. The impact of mutational clonality in predicting the response to immune checkpoint inhibitors in advanced urothelial cancer This makes biological sense: a clonal neoantigen is displayed on every cancer cell, giving T cells a uniform target across the entire tumor. A subclonal neoantigen is only on some cells, so even a vigorous immune response against it leaves much of the tumor untouched. High levels of neoantigen heterogeneity have been linked to relapses after immunotherapy, while tumors with clonally expressed neoantigens tend to be more sensitive to checkpoint blockade.4Translational Oncology. The thin red line between the immune system and cancer evolution
This is one reason two tumors with identical total TMB can behave differently on the same drug. One tumor’s mutations may be mostly clonal and immunologically visible; the other’s may be scattered across subclones, offering the immune system no single clean target.
What Drives High TMB
Certain exposures and genetic defects reliably push mutation counts up. In lung cancer, smoking is the dominant factor, and high TMB is more common in squamous-cell tumors and in men.5PubMed Central. Tumor mutational burden in lung cancer: a systematic literature review In melanoma, ultraviolet radiation from sun exposure or indoor tanning drives a distinctive pattern of DNA damage that makes cutaneous melanomas among the most mutation-heavy cancers.6PubMed Central. Melanoma Tumor Mutational Burden and Indoor Tanning Exposure
Beyond environmental exposures, internal breakdowns in the cell’s own DNA-repair machinery are another major source of high TMB. Mismatch repair deficiency, often detected through microsatellite instability testing, is perhaps the best-known example. Tumors that have lost the ability to correct small copying errors during cell division accumulate mutations rapidly.7PubMed. Microsatellite instability and high tumor mutational burden detected by next generation sequencing are concordant with loss of mismatch repair proteins by immunohistochemistry Mutations in homologous recombination genes and in POLE, the gene encoding the main proofreading enzyme of DNA replication, also predict elevated mutation counts and increased immune cell infiltration.2PubMed Central. Mutations in DNA repair genes are associated with increased neo-antigen load and activated T cell infiltration in lung adenocarcinoma
POLE and POLD1 Mutations as a Special Case
Tumors carrying mutations in the proofreading domains of POLE or POLD1 deserve separate mention because they can reach mutation loads that dwarf even mismatch-repair-deficient cancers. These mutations knock out the cell’s ability to proofread freshly copied DNA, leading to an “ultra-hypermutated” state.8PubMed Central. POLE/POLD1 mutation and tumor immunotherapy The clinical implications are striking: pooled data show that patients whose tumors carry POLE or POLD1 mutations had an objective response rate to immunotherapy of about 35%, compared with roughly 20% in patients without those mutations.9PubMed Central. Distinctive genomic characteristics in POLE/POLD1-mutant cancers can potentially predict beneficial clinical outcomes in patients who receive immune checkpoint inhibitor
In colorectal cancer specifically, POLE/POLD1 proofreading-deficient tumors appear to produce durable responses to checkpoint inhibitors, sometimes exceeding what is seen even in the well-established mismatch-repair-deficient population.10PubMed Central. Immune checkpoint inhibitors in POLE/POLD1 proofreading-deficient CRC: from molecular basis to clinical practice and future directions These tumors are rare, but identifying them matters because they represent some of the strongest candidates for immunotherapy across all of oncology.
TMB in the Clinic and the Pembrolizumab Approval
The most consequential regulatory decision involving TMB came in 2020, when the U.S. FDA approved the checkpoint inhibitor pembrolizumab for any solid tumor with a TMB of 10 or more mutations per megabase, regardless of where in the body the cancer originated. In the trial data supporting that approval, patients whose tumors met the threshold had an objective response rate of about 29%, compared with roughly 6% among patients below it. Most responders maintained their response for at least two years.11PubMed Central. FDA Approval Summary: Pembrolizumab for the treatment of tumor mutational burden-high solid tumors
A large real-world study of more than 8,400 patients across 24 cancer types reinforced the general pattern. After adjusting for other factors, patients with the highest TMB (20 or more mutations per megabase) had roughly half the risk of death compared to those with the lowest counts. In 9 out of 10 cancer types that were powered for individual analysis, TMB above 10 was associated with better survival on checkpoint inhibitors. The benefit held even in microsatellite-stable tumors, except for colorectal cancer.12PubMed Central. Tumor mutational burden and survival on immune checkpoint inhibition in >8000 patients across 24 cancer types
Still, the relationship between TMB and immunotherapy benefit is not uniform. A systematic review found that high TMB consistently predicted improved outcomes in lung cancer and melanoma, but results in breast and prostate cancer were inconclusive.13PubMed Central. Systematic Review Evaluating Tumour Mutational Burden as a Key Biomarker in Personalized Cancer Immunotherapy: A Pan-Cancer Systematic Review And some cancers respond well to immunotherapy despite having relatively low mutation counts. Merkel cell carcinoma, kidney cancer, and mesothelioma all show higher-than-expected response rates to checkpoint inhibitors, likely because the quality of their antigens compensates for quantity, whether through viral antigens, structural mutations, or complex chromosomal rearrangements.14European Journal of Cancer. Tumour mutational burden as a biomarker for immunotherapy: Current data and emerging concepts
When High TMB Predicts Worse Outcomes
It is worth stressing that TMB’s meaning changes depending on the treatment context. The connection between high TMB and better outcomes applies specifically to immunotherapy. When the treatment is conventional chemotherapy, the picture can flip. An exploratory study of non-small cell lung cancer patients receiving platinum-based chemotherapy after surgery found that patients with high TMB actually had shorter survival than those with low TMB.15PubMed Central. Prognostic Significance of Tumor Mutation Burden among Patients with Non-small Cell Lung Cancer Who Received Platinum-based Adjuvant Chemotherapy: An Exploratory Study In that setting, a heavily mutated tumor may simply be a more aggressive, genetically unstable cancer with more routes to drug resistance, and there is no immune-system boost from the chemotherapy to offset that disadvantage.
In colorectal cancer treated with fluoropyrimidine and oxaliplatin, though, high TMB was tied to better recurrence-free survival even with conventional chemotherapy, and it outperformed microsatellite instability status as a prognostic marker in that cohort.16Clinical Cancer Research. Tumor Mutation Burden and Prognosis in Patients with Colorectal Cancer Treated with Adjuvant Fluoropyrimidine and Oxaliplatin The takeaway is that TMB is not a simple “higher is better” marker; its value depends on both the cancer type and the treatment being given.
How TMB Is Measured and Why That Is Harder Than It Sounds
The gold standard for TMB measurement is whole-exome sequencing, which reads essentially all the protein-coding regions of a tumor’s genome. In practice, that is expensive and time-consuming, so most clinical TMB testing uses targeted gene panels that sequence a curated list of a few hundred cancer-related genes and extrapolate the mutation rate to the whole exome.
This shortcut introduces real problems. Because cancer gene panels are built around genes that are already known to be frequently mutated in cancer, they tend to overestimate TMB. One study found panel-derived TMB was significantly inflated compared to whole-exome values, and that a large fraction of patients were misclassified above the 10-mutations-per-megabase cutoff depending on which panel was used, with overestimation rates reaching up to 27% in some cancer types for one commonly used panel.17Journal of the National Cancer Center. Tumour mutational burden is overestimated by target cancer gene panels A panel built from randomly selected genes, rather than cancer-associated ones, showed no such inflation, confirming the bias is driven by the gene selection itself.
Panel size matters as well. Panels covering at least about 1 megabase of genomic territory produce TMB estimates that correlate reasonably well with whole-exome results, but accuracy drops sharply below about 0.5 megabases.18Journal for ImmunoTherapy of Cancer. Tumor mutational burden quantification from targeted gene panels: major advancements and challenges Simulation studies confirm that both panel size and gene content meaningfully affect the result, and that two panels of nearly identical size can still produce different TMB estimates because they cover different genes.19PubMed Central. Assessments of tumor mutational burden estimation by targeted panel sequencing: A comprehensive simulation analysis
The practical consequence is that a TMB score of 10 on one panel does not mean the same thing as a score of 10 on another. One analysis found that to correctly classify patients at the 10-mutations-per-megabase threshold, the effective cutoff could vary by more than 20% between commonly used test platforms.20PubMed Central. Impact of panel design and cut-off on tumour mutational burden assessment in metastatic solid tumour samples Harmonization efforts coordinated by the Friends of Cancer Research consortium have developed calibration tools that allow better comparison across panels, but standardization remains a work in progress.21PubMed. Aligning tumor mutational burden (TMB) quantification across diagnostic platforms: phase II of the Friends of Cancer Research TMB Harmonization Project
Tumor Purity and Sequencing Depth
Even with an appropriate panel, the quality of the tumor sample affects the result. A biopsy is never purely cancer cells; it contains varying amounts of normal tissue, immune cells, and connective tissue. If the proportion of tumor cells, called tumor purity, is too low, mutations get diluted into the background and TMB can be underestimated. One study found that tumor purity significantly influenced whole-exome-based TMB estimates but had less effect on targeted-panel TMB, likely because panels sequence their targeted regions much more deeply and can therefore pick up mutations even in lower-purity samples.22Journal for ImmunoTherapy of Cancer. Clinical advantage of targeted sequencing for unbiased tumor mutational burden estimation in samples with low tumor purity Similarly, sequencing depth, or how many times each region of DNA is read, needs to be sufficient; TMB estimates stabilize once the depth exceeds about 500x.23PubMed Central. Systematic assessment and optimizing algorithm of tumor mutational burden calculation and their implications in clinical decision-making
Liquid Biopsy as an Alternative
Not every patient has a tumor that can be easily biopsied. Blood-based TMB testing, which measures mutations in fragments of tumor DNA circulating in the bloodstream, offers an alternative. A study from the SCRUM-Japan MONSTAR-SCREEN project found that when enough tumor DNA was present in the blood sample (at least 1% of total cell-free DNA), there was strong agreement between blood-based and tissue-based TMB, with a correlation coefficient of 0.74 and sensitivity above 80% for identifying TMB-high patients.24PubMed Central. Validity and utility of blood tumor mutational burden (bTMB) is dependent on circulating tumor DNA (ctDNA) shed: SCRUM-Japan MONSTAR-SCREEN The catch is that not all tumors shed enough DNA into the blood, and when the tumor fraction drops too low, the test becomes unreliable. For now, blood-based TMB is a promising complement to tissue testing rather than a wholesale replacement.
TMB and PD-L1 Are Telling Different Stories
Oncologists already had a biomarker for checkpoint inhibitors before TMB entered the picture: PD-L1 expression on tumor cells. A natural question is whether TMB just duplicates what PD-L1 testing already tells you. The short answer is no. Analysis across multiple tumor types has shown that PD-L1 expression and TMB are largely independent of each other, with non-overlapping effects on response rates to checkpoint inhibitors.25PubMed Central. PD-L1 expression and tumor mutational burden are independent biomarkers in most cancers A patient can be PD-L1-high and TMB-low, TMB-high and PD-L1-low, both, or neither, and each combination carries different odds of benefit.
In lung adenocarcinoma, combining the two markers identified a group of responders with a median overall survival of 32 months, compared with 8.5 months for non-responders defined by the same combination. That separation was not seen in squamous-cell lung cancer, underscoring that even combined biomarkers work differently across subtypes.26PubMed Central. PD-L1 expression and tumor mutational burden status for prediction of response to chemotherapy and targeted therapy in non-small cell lung cancer In rare or refractory cancers like pancreatic cancer, where immunotherapy has historically performed poorly, checking both PD-L1 and TMB can occasionally identify individual patients who benefit from checkpoint blockade, even though the cancer type overall is not considered immunotherapy-responsive.27PubMed Central. Quantification of PD-L1 expression and tumor mutational burden in biologically distinct advanced pancreatic cancers responding to pembrolizumab: case reports
Chromosomal Chaos and Its Complicating Role
TMB captures point mutations and small insertions or deletions, but tumors also acquire large-scale chromosomal abnormalities, a condition called aneuploidy, where entire chromosome arms are gained or lost. Research in metastatic non-small cell lung cancer found that highly aneuploid tumors responded less well to immunotherapy regardless of their TMB. Among patients with high TMB, those whose tumors had low aneuploidy scores achieved an objective response rate of about 76%, compared with roughly 63% for those with high aneuploidy.28Scientific Reports. Clinical and molecular correlates of tumor aneuploidy in metastatic non-small cell lung cancer Aneuploidy may help tumors evade the immune system through mechanisms that TMB does not capture, adding yet another layer to predicting who benefits from checkpoint inhibitors.
How Tumors Escape Despite High TMB
Even among patients with genuinely high mutation burdens, immunotherapy does not always work. Modeling work has identified two evolutionary routes tumors can use to resist checkpoint blockade. The first is preexisting resistance: within a high-TMB tumor, some cell populations may carry mutations that happen not to produce recognizable neoantigens, effectively hiding in plain sight from the immune system. The second is immunoediting, where the selective pressure of a successful immune response gradually eliminates the most visible cancer cells and leaves behind those that have lost their neoantigens or antigen-presenting machinery.29PubMed. Tumor Mutational Burden Shapes Success and Resistance in Cancer Immunotherapy Over time, the tumor reshapes itself to become invisible even as the total mutation count remains high.
This explains a frustrating clinical pattern: a patient responds well to immunotherapy initially, only to relapse months or years later. The mutations are still there in the cancer’s DNA, but the cells that displayed them have been replaced by descendants that found ways to hide. TMB measured at the start of treatment may not reflect the immunological landscape at the time of relapse.
The Cost Question
Genomic testing is not cheap, and the economics of TMB testing have drawn scrutiny. An analysis modeling the cost-effectiveness of biomarker testing in non-small cell lung cancer found that while immunotherapy was cost-effective across all groups studied, using PD-L1 testing to select patients offered better economic value than using TMB testing. TMB testing was actually not cost-effective compared with no testing at all in that model, though tissue-based TMB was the most cost-effective of the various TMB detection methods evaluated.30Journal of Clinical Oncology. Cost-effectiveness of PD-L1 testing and tumor mutational burden testing of immune checkpoint inhibitors for non-small cell lung cancer That finding should be read with caution, since it was based on a single clinical trial’s data in one cancer type, and the economics shift as testing costs fall and treatment options expand. But it underscores that a biomarker’s clinical value and its economic value are not always the same thing, and health systems weigh both when deciding what to fund.
The Immune Microenvironment Beyond TMB
High TMB does not just attract T cells. It reshapes the broader immune neighborhood of the tumor in ways that can help or hinder treatment. In gastric cancer, tumors with high TMB showed increased infiltration by activated memory T cells, follicular helper T cells, and both M0 and M1 macrophages. But that macrophage infiltration came with a twist: high macrophage presence was associated with worse survival outcomes.31PubMed Central. Comprehensive analysis of tumor mutation burden and immune microenvironment in gastric cancer In lung cancer, researchers found that tumors with both high TMB and high numbers of CD8-positive T cells expressing PD-L1 had an immune environment that was paradoxically immunosuppressive despite looking “hot” on paper.32PubMed Central. Massive PD-L1 and CD8 double positive TILs characterize an immunosuppressive microenvironment with high mutational burden in lung cancer These findings are a reminder that TMB tells you about the quantity of potential immune targets, not about whether the tumor’s local environment actually permits an effective immune attack. The next generation of biomarker strategies will almost certainly need to integrate TMB with measures of the immune microenvironment to get closer to predicting who truly benefits from immunotherapy.