Immunotherapy fails because tumors are not static targets. Cancer cells mutate, reshape their local environment, and exploit backup immune-suppression pathways to escape the immune attack that checkpoint inhibitors initially unleashed. Roughly half of patients with certain cancers never respond at all, and among those who do, a significant fraction eventually see their disease progress. Understanding the specific reasons behind that progression matters, because the “what’s next” depends heavily on the “why.”
Two Distinct Ways Resistance Develops
Oncologists broadly sort immunotherapy resistance into two categories. Primary resistance describes patients whose tumors never respond to checkpoint inhibitors in the first place. These patients start treatment, and the disease either holds steady or grows. Acquired resistance is different: the patient initially benefits, sometimes dramatically, but the tumor eventually finds a way around the immune response and starts progressing again.1PubMed Central. Acquired Resistance to Immune Checkpoint Inhibitors The distinction matters practically because the biological mechanisms tend to differ, and so do the treatment options that follow.
Primary resistance often traces to features already present in the tumor before treatment started: a hostile local environment, low mutation burden, or signaling pathways that keep immune cells out entirely. Acquired resistance, by contrast, usually involves the tumor evolving under the selective pressure of an active immune response. The immune system kills the vulnerable cancer cells, and the survivors are the ones that happened to carry traits letting them dodge detection. Over weeks or months, those survivors repopulate the tumor.
Tumors That Learn to Hide
For your immune system to kill a cancer cell, it first has to recognize it as dangerous. That recognition depends on molecules on the tumor cell surface that display fragments of abnormal proteins. One critical component of this display system is a protein called beta-2-microglobulin (B2M). When the gene for B2M is mutated or lost, the entire surface display machinery breaks down. The result is that immune cells, particularly the CD8+ T cells that checkpoint inhibitors rely on, can no longer “see” the tumor. The cancer cell becomes effectively invisible to immune surveillance.2PubMed Central. The role of B2M in cancer immunotherapy resistance: function, resistance mechanism, and reversal strategies
This is not a rare curiosity. Studies have documented B2M alterations across melanoma, lung cancer, colorectal cancer, and other tumor types, and the pattern consistently links to poor immunotherapy outcomes.3PubMed. Beta2-microglobulin(B2M) in cancer immunotherapies: Biological function, resistance and remedy What makes this mechanism particularly frustrating is its finality: once a tumor population loses its ability to present antigens, simply continuing the same checkpoint inhibitor accomplishes nothing. The drug can release the brakes on the immune system all it wants, but if the target is invisible, there is nothing to attack.
When T Cells Run Out of Steam
Even when tumors remain visible, the immune cells sent to destroy them can become functionally worn out. Chronic exposure to tumor signals drives CD8+ T cells into a state called exhaustion, where they lose their ability to kill effectively and stop multiplying. This is not just fatigue in a metaphorical sense. Exhausted T cells have a fundamentally different gene-expression profile and a locked-in pattern of chemical modifications to their DNA that distinguishes them from healthy, functioning immune cells.4PubMed Central. TOX transcriptionally and epigenetically programs CD8(+) T cell exhaustion
A protein called TOX sits at the center of this process. When T cells are stimulated continuously, as happens inside a growing tumor, TOX locks them into the exhaustion program. Recent research has shown that TOX acts as a durable barrier: it keeps the exhausted state in place at the level of how DNA is packaged and read, not just which genes happen to be turned on at a given moment. Intriguingly, when researchers experimentally removed TOX from already-exhausted T cells, those cells regained some flexibility to become more functional again.5PubMed. Continuous expression of TOX safeguards exhausted CD8 T cell epigenetic fate That finding is still in the lab, but it points toward a possible future where exhaustion might be partially reversed rather than accepted as permanent.
Backup Brakes the Tumor Can Pull
Most approved checkpoint inhibitors work by blocking PD-1, PD-L1, or CTLA-4. But these are not the only “off switches” the immune system has. T cells carry multiple other inhibitory receptors on their surface, including LAG-3, TIM-3, and TIGIT. When a drug blocks PD-1, the tumor can compensate by leaning harder on these alternative checkpoints. Multiple studies have found that these receptors are frequently co-expressed on tumor-infiltrating T cells in cancers ranging from melanoma to lung cancer to liver cancer, and higher expression of LAG-3 in particular has been linked to worse outcomes.6PubMed Central. LAG-3, TIM-3, and TIGIT: distinct functions in immune regulation
This compensatory checkpoint upregulation is one of the clearest explanations for why a single-agent immunotherapy can work for a while and then fail. The drug successfully lifts one brake, but the tumor reroutes suppression through a different receptor. It is also why combination therapies targeting more than one checkpoint are a major area of development, as discussed below.
A Hostile Neighborhood
The area immediately surrounding a tumor, called the tumor microenvironment, can become deeply hostile to immune function. Tumors recruit suppressive immune cells, including regulatory T cells and myeloid-derived suppressor cells, that actively shut down the very killer T cells and natural killer cells that immunotherapy depends on.7PubMed Central. The Functional Crosstalk between Myeloid-Derived Suppressor Cells and Regulatory T Cells within the Immunosuppressive Tumor Microenvironment Think of it as the tumor hiring its own security force from within the immune system itself.
On top of that, tumor cells rewire their metabolism in ways that starve immune cells of fuel. Cancer cells consume glucose at high rates, leaving nearby T cells without the energy they need to function. They also break down amino acids like tryptophan to produce byproducts that further suppress immune activity and promote the expansion of regulatory cells that dampen the immune response.8PubMed Central. Role of metabolic transformation in cancer immunotherapy resistance: molecular mechanisms and therapeutic implications The accumulation of lactate and the depletion of other key amino acids contribute to an environment where T cells simply cannot do their job, regardless of how many checkpoint inhibitors you throw at the problem.9PubMed Central. Tumor-intrinsic metabolic reprogramming and how it drives resistance to anti-PD-1/PD-L1 treatment
Oncogenic Pathways That Lock Immune Cells Out
Some tumors are called “cold” because they have very few immune cells inside them to begin with. This immune exclusion can be driven by specific signaling pathways within the cancer cells themselves. Activation of the WNT/beta-catenin pathway, for example, has been shown to correlate with immune exclusion across multiple human cancer types.10PubMed Central. WNT/β-catenin pathway activation correlates with immune exclusion across human cancers Loss of a tumor suppressor gene called PTEN has been linked to similar immune-cold environments, particularly in melanoma, where PTEN loss was associated with resistance emerging in a patient who had initially responded to anti-PD-1 therapy for two years before the disease progressed.
These findings suggest that for some patients, the problem is not that the immune system got tired or was suppressed but that it was never really allowed into the tumor in the first place. Treatments aimed purely at releasing immune brakes will not help if the tumor’s internal wiring prevents immune cells from arriving at the scene.
What Your Gut Has to Do With It
One of the more surprising developments in immunotherapy research has been the discovery that the bacteria living in your gut influence whether checkpoint inhibitors work. The composition of the gut microbiome has been linked to anti-PD-1 effectiveness in both lab models and human patients. In a clinical trial for melanoma patients whose disease had progressed on anti-PD-1 therapy, researchers tried transplanting gut bacteria from patients who had responded well. The combination of fecal microbiota transplant plus continued anti-PD-1 therapy produced clinical benefit in 6 of 15 patients with previously refractory disease. Responders showed increased T cell activation in their tumors and shifts toward bacterial species previously associated with immunotherapy success.11PubMed Central. Fecal microbiota transplant overcomes resistance to anti-PD-1 therapy in melanoma patients
This line of research has since expanded beyond melanoma. In lung cancer patients with acquired resistance to PD-1 therapy, fecal transplants similarly led to increased gut microbial diversity in those who went on to respond.12PubMed Central. Fecal microbiota transplantation for advanced non-small cell lung cancer with secondary PD-1 resistance efficacy prognostic factors and microbiome diversity analysis The field is still in early stages, and nobody is yet prescribing specific bacteria as a routine part of cancer care. But the idea that the gut ecosystem shapes the tumor immune environment is now well-established enough to drive multiple ongoing clinical trials.
Catching Resistance Before It Shows on Scans
Waiting for a tumor to grow large enough to show up as progression on a CT scan means resistance has already been underway for some time. Liquid biopsies, which analyze fragments of tumor DNA circulating in the blood, can detect signs of resistance earlier. Longitudinal tracking of circulating tumor DNA can reveal emerging resistance mechanisms before conventional imaging picks up changes, giving oncologists a potential head start on switching therapies.13Journal for ImmunoTherapy of Cancer. Liquid biopsy approaches to capture tumor evolution and clinical outcomes during cancer immunotherapy
Imaging itself presents unique challenges during immunotherapy. A phenomenon called pseudoprogression can occur, where the tumor appears to grow on a scan because immune cells are flooding into it, even though the treatment is actually working. On the other end of the spectrum is hyperprogression, a genuine and alarming acceleration of tumor growth. One large retrospective analysis found that hyperprogression occurred in about 14% of lung cancer patients on immunotherapy, compared with about 5% on chemotherapy. Reported rates of hyperprogression across studies vary widely, from 6% to 43%, with a pooled estimate around 13%.14PubMed Central. iRECIST and atypical patterns of response to immuno-oncology drugs The lack of a standard definition contributes to that wide range, but the clinical reality is clear: not every scan that looks worse actually is worse, and some cases of apparent progression are genuinely dangerous acceleration.
Combining Checkpoint Inhibitors
The logic of compensatory checkpoints leads directly to combination therapy. If a tumor can dodge one checkpoint blockade by leaning on another receptor, blocking both receptors at once should cut off that escape route. The first combination to reach approval targeting a newer checkpoint was relatlimab, a LAG-3-blocking antibody, combined with nivolumab (an anti-PD-1 drug). In a major trial of untreated advanced melanoma, the combination roughly doubled progression-free survival compared with nivolumab alone: about 10 months versus under 5 months. At the one-year mark, about 48% of patients on the combination had not progressed, compared with 36% on nivolumab alone.15PubMed Central. Relatlimab and Nivolumab versus Nivolumab in Untreated Advanced Melanoma
These are not miracle numbers, and the improvement is modest for some patients. But the principle that blocking a second checkpoint can meaningfully extend disease control is now proven. Trials targeting TIM-3, TIGIT, and other receptors in combination with PD-1 blockade are ongoing, with mixed early results depending on the tumor type and the specific combination.
Adding Radiation and Anti-Angiogenesis Drugs
When immunotherapy alone or even dual-checkpoint blockade is not enough, oncologists increasingly look at combining it with treatments from different categories. One promising approach pairs checkpoint inhibitors with drugs that block blood vessel growth in tumors, known as VEGF inhibitors. The rationale goes beyond simply starving the tumor of blood supply. VEGF inhibitors appear to normalize the chaotic tumor blood vessels, making it easier for immune cells to physically get into the tumor, while also reducing the myeloid-derived suppressive cells that dampen immune function.16Targeted Oncology. Overcoming Resistance to Immune Checkpoint Inhibitors in Lung Cancer: Is There a Role for VEGF Inhibitor Treatment?
A phase 2 trial in liver cancer patients whose disease had progressed on prior immunotherapy tested the combination of targeted radiation with an anti-PD-1 drug and a VEGF inhibitor. Among the 21 patients treated, about a third of non-irradiated lesions responded, the disease control rate was about 67%, and estimated median overall survival reached over two years.17PubMed Central. Stereotactic body radiotherapy with sintilimab and bevacizumab biosimilar in anti-PD-1 refractory hepatocellular carcinoma: the ReUNION-1 phase 2 trial These are small numbers, but for patients who had already failed immunotherapy, any meaningful response signals that resistance can sometimes be overcome by attacking the problem from multiple angles simultaneously. Preclinical work on triple combinations adding a TGF-beta inhibitor to PD-1 and VEGF blockade has shown even stronger effects in mouse models of colorectal cancer.18Cancer Research. Abstract 1562: Rewiring the immune-excluded tumor microenvironment: Vactosertib/anti-PD-1/VEGF inhibitor triplet therapy reinstates antitumor immunity in CRC
Cell Therapies as a Next Step
When checkpoint inhibitors fail entirely, some patients may be candidates for adoptive cell therapy, which involves either engineering a patient’s own T cells or harvesting immune cells directly from the tumor, expanding them in the lab, and infusing them back. The approval of lifileucel in 2024 marked the first time a cellular therapy was authorized for a solid tumor, a significant milestone given that previous cell therapy approvals had been limited to blood cancers.19PubMed Central. CAR-T and TIL therapies in solid tumors: barriers, clinical lessons, and convergent solutions
That said, the obstacles are real. Therapeutic T cells in solid tumors face three escalating barriers: getting into the tumor through its physical and vascular defenses, recognizing malignant cells despite antigen variability and immune evasion, and persisting long enough in a suppressive microenvironment to achieve durable control. Each of these maps to a resistance mechanism described above, which is why cell therapies are being designed to address them in combination rather than one at a time.
Trying Immunotherapy Again After a Break
One counterintuitive option is retreatment with the same class of drugs after a gap. This approach has been studied most extensively in melanoma. In a cohort of advanced melanoma patients who had stopped immunotherapy (whether after a response, stable disease, or for other reasons), reintroduction of checkpoint inhibitors upon disease progression yielded an objective response rate of about 59%, with durable ongoing responses.20PubMed Central. Discontinuation of Immune Checkpoint Inhibition in Patients with Advanced Unresectable Melanoma Achieving CR, PR, or SD
In lung cancer, immunotherapy rechallenge after first-line treatment failure showed a median overall survival of about 19.5 months and a median progression-free survival of about 7 months. Adding anti-angiogenesis drugs to the rechallenge regimen appeared to extend progression-free survival further compared with immunotherapy alone.21PubMed Central. The efficacy analysis of immunotherapy rechallenge after progression from first-line chemo-immunotherapy in advanced non-small cell lung cancer Retreatment strategies are being classified into distinct clinical scenarios: restarting after stopping while still responding, restarting after a break forced by side effects, and restarting after progression on prior immunotherapy. Each carries different expected outcomes and considerations.22PubMed. Immune checkpoint inhibitors rechallenge in urological tumors: An extensive review of the literature
The fact that rechallenge works at all tells us something about resistance. In some cases, the “resistance” may be less about the tumor permanently escaping immune control and more about the immune system needing a reset. A treatment holiday can allow T cell populations to recover from exhaustion, particularly if the tumor has not acquired hard-wired genetic escape mechanisms like B2M loss.
The Corticosteroid Dilemma
Immunotherapy can trigger serious side effects where the unleashed immune system attacks healthy organs. Corticosteroids are the standard treatment for these immune-related adverse events, but they are fundamentally immunosuppressive, which creates an obvious tension: you are dampening the very immune response you were trying to boost. An analysis across six clinical trials found that higher peak corticosteroid doses were associated with worse outcomes. Patients who received the equivalent of about 2 mg/kg prednisolone had roughly 66% higher risk of death compared with those given 0.5 mg/kg. The total cumulative dose over time, however, was not linked to survival differences.23PubMed. Corticosteroids for Immune-Related Adverse Events and Checkpoint Inhibitor Efficacy: Analysis of Six Clinical Trials
The practical implication is that oncologists try to manage side effects with the lowest effective steroid dose and taper quickly. There is growing interest in steroid-sparing alternatives for managing immune-related toxicity precisely because of this trade-off. For patients, the key takeaway is that managing side effects and maintaining treatment effectiveness are not separate conversations; they are deeply intertwined.
Predicting Resistance Before It Happens
If you could predict which patients will resist immunotherapy and which will respond, you could spare non-responders the side effects and direct them to alternatives sooner. Traditional biomarkers like PD-L1 expression and tumor mutation burden help but are imperfect. Machine learning approaches that integrate multiple data types are showing promise. One model that combined mutation profiles using both random-forest and neural-network methods identified 55 tumor mutations predictive of durable benefit in lung cancer, outperforming both PD-L1 and tumor mutation burden as predictors. It even identified responders among patients with low mutation burden, a group that conventional biomarkers would have written off.24ScienceDirect (Fundamental Research). Harnessing multi-omics and machine learning for predicting immune checkpoint blockade responses: Advances, challenges, and future directions
Meanwhile, the role of the patient’s own immune genetics remains an open question. One might expect that people with greater diversity in their HLA genes (the genes that help immune cells recognize foreign proteins) would respond better to immunotherapy. But a large analysis of patients treated with pembrolizumab found no meaningful association between HLA diversity or specific HLA genotypes and treatment response across multiple solid tumor types. The researchers concluded that HLA genotype alone should not be used to make treatment decisions for patients on pembrolizumab.
Access to Testing Is Part of the Problem
Even when good biomarkers exist, not every patient gets tested. In metastatic colorectal cancer, for example, testing for microsatellite instability, which identifies patients most likely to benefit from immunotherapy, is not performed uniformly. Disparities in testing rates mean that some patients who could benefit from immunotherapy never get the chance because the biomarker that would have flagged them as candidates was never checked.25PubMed Central. Opportunities for Improving System-Level Barriers to Biomarker Testing for Metastatic Colorectal Cancer This is a systemic problem, not a scientific one. The biology of resistance is complicated enough without adding missed opportunities at the point of care.