Checkpoint inhibitor immunotherapy — drugs like pembrolizumab (Keytruda) and nivolumab (Opdivo) — can train the immune system to fight cancer so effectively that the protection persists long after patients stop receiving treatment. This is not a theoretical hope. In the KEYNOTE-024 trial for advanced lung cancer, 82 percent of patients who completed two years of pembrolizumab were still alive three years after their last dose, and nearly half had no disease progression at all without any further therapy. In melanoma, the story is similarly striking: among patients who achieved a complete response on pembrolizumab in the KEYNOTE-001 trial, 90 percent remained disease-free two years after discontinuation. No chemotherapy drug has ever produced numbers like these.
The mechanism behind this durability comes down to immune memory. Checkpoint inhibitors do not kill cancer cells directly. Instead, they release the brakes on T cells, allowing the immune system to recognize tumors as threats. Once that recognition happens, the body generates memory T cells — the same type of long-lived immune cells that make vaccines work for years. These memory T cells persist in and around the tumor site, continuing to patrol for cancer cells long after the drug has cleared the body. This article examines the clinical evidence behind immunotherapy’s lasting effects across multiple cancer types, what the latest 2025 research from Dana-Farber and Memorial Sloan Kettering reveals about patients who stop treatment early due to side effects, which patients are most likely to benefit, and the important caveats that come with these otherwise remarkable findings.
Table of Contents
- How Does Immunotherapy Keep Working After You Stop Taking It?
- What the Long-Term Trial Data Actually Show
- What Happens When Patients Stop Early Due to Side Effects
- How Durable Is Immunotherapy Across Different Cancer Types?
- Who Does Not Benefit — and the Limits of Immune Memory
- Fixed-Duration Therapy Is Becoming the Standard
- What This Means for the Future of Cancer Treatment
- Conclusion
- Frequently Asked Questions
How Does Immunotherapy Keep Working After You Stop Taking It?
To understand why checkpoint inhibitors have lasting effects, it helps to compare them with traditional chemotherapy. Chemotherapy drugs are cytotoxic — they poison rapidly dividing cells while they circulate in the bloodstream. The moment you stop infusions, the killing stops. There is no memory, no ongoing protection, no biological mechanism that persists. Immunotherapy works on an entirely different principle. Anti-PD-1 and anti-PD-L1 drugs block a protein interaction that tumors exploit to hide from the immune system. Once that camouflage is stripped away and T cells are activated against tumor-specific antigens, the immune system expands populations of intratumoral memory T cells.
According to the [National Cancer Institute](https://www.cancer.gov/about-cancer/treatment/types/immunotherapy/checkpoint-inhibitors), this PD-1 blockade fundamentally reshapes the immune landscape within and around tumors. Think of it this way: the drug is the key that unlocks a door, but the immune system is the one that walks through it. Once memory T cells have learned to identify cancer cells, they do not need the drug to keep doing their job. This is the same principle that allows your body to fight off measles decades after a childhood vaccination. The critical difference is that immunotherapy teaches your immune system to recognize your own mutated cells rather than an outside pathogen — a far more difficult task biologically, which is why not every patient responds, and why the patients who do respond often see benefits that last for years. It is worth noting that this durability was not something researchers confidently predicted when checkpoint inhibitors were first approved. The early trials capped treatment at two years partly for safety and cost reasons, and the long-term follow-up data that have since emerged were, by many accounts, better than expected. Eight- and ten-year survival data from the KEYNOTE-001, -010, -024, and -042 trials were presented at the [ESMO Congress 2025](https://www.pharmacytimes.com/view/keynote-trial-data-presented-at-esmo-2025-congress-reiterate-pembrolizumab-s-benefits-in-nsclc), reaffirming that pembrolizumab’s benefits are not temporary blips but sustained, measurable survival advantages.
What the Long-Term Trial Data Actually Show
The most robust evidence comes from the KEYNOTE family of trials studying pembrolizumab. In KEYNOTE-010 and KEYNOTE-024 — both focused on advanced non-small cell lung cancer — [five-year follow-up data published in the Journal of Clinical Oncology](https://ascopubs.org/doi/10.1200/JCO.21.00174) showed that 83 percent and 82 percent of patients who completed two full years of treatment remained alive three years after stopping. Roughly 48 percent and 46 percent, respectively, had not experienced any disease progression and required no further cancer therapy during that period. For a disease where the historical five-year survival rate for metastatic cases hovers around 12 percent, these figures represent a fundamental shift. Melanoma data tell a complementary story. In KEYNOTE-001, which enrolled 655 patients with advanced melanoma, 17 percent achieved a complete response — meaning no detectable cancer remained.
Among the 67 patients who then stopped pembrolizumab while in complete response, the [two-year disease-free survival rate was 90 percent](https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2022.983581/full). Older data on ipilimumab (Yervoy), which targets CTLA-4 rather than PD-1, show that [20 percent of melanoma patients](https://ascopubs.org/doi/10.1200/EDBK_240837) maintain durable disease control five to ten years after starting treatment. However, these numbers come with an important caveat. Not every patient who stops immunotherapy stays in remission. In KEYNOTE-189, which studied pembrolizumab combined with chemotherapy for NSCLC, [half of 56 patients who completed approximately two years of treatment experienced disease progression after stopping](https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2022.983581/full). In KEYNOTE-010, 32 percent of patients progressed after discontinuation. The biology of immune memory is powerful but not universal, and factors like tumor mutation burden, PD-L1 expression, and the patient’s underlying immune health all influence whether the protection holds.
What Happens When Patients Stop Early Due to Side Effects
One of the most clinically meaningful questions about immunotherapy’s durability concerns patients who do not complete the standard two-year course — not because their cancer progressed, but because the treatment’s side effects became intolerable. Immune-related adverse events (irAEs) such as colitis, hepatitis, pneumonitis, and thyroid dysfunction force roughly 10 percent of patients to discontinue checkpoint inhibitors prematurely. Until recently, the outcomes for these patients were poorly understood. A [2025 study from researchers at Dana-Farber Cancer Institute and Memorial Sloan Kettering](https://www.dana-farber.org/newsroom/news-releases/2025/effects-of-immunotherapy-can-last-after-treatment-stops-according-to-dana-farber-study) addressed this gap directly. The study focused on patients with advanced NSCLC who stopped checkpoint inhibitors because of irAEs and found that disease control often continued after discontinuation. Patients who had received immunotherapy for at least six months before stopping fared best.
Those whose tumors had PD-L1 expression of 50 percent or higher, non-squamous histology, and evidence of tumor shrinkage during treatment were most likely to maintain long-term benefit. One particularly reassuring finding: the use of immunosuppressive medications like corticosteroids to manage side effects did not negatively impact post-discontinuation cancer outcomes — a fear that had lingered among both oncologists and patients. This matters because the decision to stop immunotherapy due to side effects is one of the most stressful moments in a patient’s cancer journey. Knowing that the immune system may have already been sufficiently “trained” provides a meaningful degree of reassurance, though it is not a guarantee. Patients who stopped within the first three months of treatment had far shorter median survival — 22 months — compared with 87 months for those who made it past six months. The takeaway is that more exposure to the drug gives the immune system more time to build a robust anti-tumor memory, but even partial courses can yield lasting benefit for the right patients.
How Durable Is Immunotherapy Across Different Cancer Types?
The durability of immunotherapy responses varies by cancer type, and the differences are significant enough to affect treatment planning. In NSCLC, published data from April 2025 in *Clinical Cancer Research* showed a [median progression-free survival of 12.7 months and median overall survival of 43.7 months](https://www.medicalnewstoday.com/articles/stopping-immunotherapy-for-lung-cancer-after-2-years-doesnt-increase-mortality-risk) — more than three and a half years — after stopping checkpoint inhibitors. Forty-six percent of patients were alive at five years, compared with the roughly 12 percent five-year survival rate typical for metastatic lung cancer treated with chemotherapy alone. These numbers transform what was once a rapidly fatal diagnosis into something that, for a meaningful minority, begins to resemble a chronic condition. Colorectal cancer patients with specific molecular features (typically microsatellite instability-high tumors) have shown some of the most impressive post-discontinuation durability.
According to data presented by the [American Association for Cancer Research](https://www.aacr.org/patients-caregivers/progress-against-cancer/study-shows-continuing-benefit-after-stopping-immunotherapy/), progression-free survival after stopping immunotherapy was 98 percent at one year, 91 percent at two years, and 84 percent at three years. In kidney cancer, a disease where checkpoint inhibitors have become a mainstay of first-line treatment, [two-thirds of patients who stopped treatment due to side effects showed no cancer progression six months later](https://actionkidneycancer.org/clinical-outcomes-of-kidney-cancer-patients-who-stopped-immunotherapy-due-to-immune-related-side-effects/). The tradeoff patients and oncologists face is straightforward but difficult: longer treatment increases the chance of durable immune memory but also increases cumulative exposure to potentially serious side effects. For cancers like MSI-high colorectal cancer, where response rates are exceptionally high, shorter courses may be entirely reasonable. For tumor types with lower overall response rates, the calculus shifts toward completing the full course when tolerable.
Who Does Not Benefit — and the Limits of Immune Memory
The durability story is compelling, but it would be irresponsible to present it without frank discussion of its limits. Immunotherapy does not work for everyone, and among those who do respond, not all will maintain that response after treatment ends. The biology of immune memory is complex, and some patients’ immune systems fail to generate or maintain adequate populations of memory T cells against their specific tumor antigens. Patients with low PD-L1 expression, low tumor mutation burden, or immunosuppressive tumor microenvironments are less likely to achieve the kind of deep, durable response that persists after drug discontinuation. Impaired production of memory T cells — whether from the cancer itself, prior treatments, age, or other health factors — can weaken checkpoint inhibitor effects over time and contribute to treatment resistance.
In KEYNOTE-010, 32 percent of patients experienced disease progression after completing their course of pembrolizumab. In KEYNOTE-189, the figure was closer to 50 percent. These are not small numbers, and they underscore that stopping immunotherapy is a decision that must be made with careful monitoring, not blind optimism. There is also the practical reality that even patients who maintain disease control after stopping may need to restart treatment at some point. Some trial protocols allow for retreatment at the time of progression, and early evidence suggests that retreatment can recapture responses in a subset of patients. But the data on retreatment are far less mature than the data on initial treatment, and the assumption that the immune system can simply be “rebooted” with another course of checkpoint inhibitors does not always hold.
Fixed-Duration Therapy Is Becoming the Standard
The accumulating evidence has shifted clinical practice. Fixed-duration immunotherapy — typically up to two years or 35 cycles of pembrolizumab — is now considered a reasonable and evidence-supported strategy for patients who are responding well. This is a departure from the early days of checkpoint inhibitor use, when many oncologists were reluctant to stop a treatment that appeared to be working and no firm guidelines existed on optimal duration.
The [DANTE trial](https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(24)00539-X/fulltext), published in *The Lancet’s eClinicalMedicine*, provides supporting evidence specifically for metastatic melanoma, showing that time-limited therapy is a viable approach for patients who remain progression-free after two years on checkpoint inhibitors. This has real implications for quality of life and healthcare costs. Two years of immunotherapy already carries significant financial and physical burdens; knowing that a defined stopping point does not compromise long-term outcomes is meaningful for patients navigating both the medical and practical dimensions of cancer treatment.
What This Means for the Future of Cancer Treatment
The concept of training the immune system to provide lasting cancer surveillance represents a genuine paradigm shift in oncology. As eight- and ten-year follow-up data continue to mature — including the long-term KEYNOTE results presented at ESMO 2025 — researchers are moving beyond asking whether immunotherapy’s effects last, and toward identifying who will maintain durable responses and who will need additional or alternative strategies.
The next frontier involves biomarker research to predict, before treatment starts, which patients are most likely to achieve lasting immune memory. If clinicians can reliably identify these patients, it opens the door to personalized treatment durations — shorter courses for those with favorable immune profiles and extended or combination therapies for those at higher risk of relapse. The goal is not just survival, but survival with the fewest possible months spent on treatment and the highest quality of life in the years that follow.
Conclusion
Checkpoint inhibitor immunotherapy has established something that no previous systemic cancer treatment could credibly claim: durable disease control that persists for years — sometimes a decade or more — after the drug is stopped. The evidence spans multiple cancer types, multiple trials, and now extends to patients who were forced to stop treatment early. The mechanism is biologically coherent, rooted in the same immune memory that protects us from infections encountered decades earlier. For the subset of patients whose immune systems respond robustly, these drugs do not just delay cancer; they appear to fundamentally alter the body’s relationship with it.
That said, immunotherapy’s lasting benefits are neither universal nor guaranteed. A significant minority of patients will progress after stopping treatment, and the factors that determine durable response are still being defined. Patients considering immunotherapy, or currently on it and approaching a potential stopping point, should have detailed conversations with their oncologists about their specific tumor biology, treatment response, and risk profile. The field is moving quickly, and the data that inform these decisions are becoming richer every year.
Frequently Asked Questions
How long do immunotherapy effects last after stopping treatment?
Clinical trial data show effects lasting years to a decade or more in responding patients. In KEYNOTE-024, 82 percent of patients who completed two years of pembrolizumab for lung cancer were alive three years after stopping, and long-term KEYNOTE data presented at ESMO 2025 confirm benefits extending to eight and ten years.
Can you stop immunotherapy and restart it later if cancer returns?
Some clinical trial protocols allow retreatment at the time of progression, and a subset of patients can recapture their initial response. However, retreatment data are less mature than initial treatment data, and success is not guaranteed.
Does stopping immunotherapy early because of side effects mean the treatment failed?
Not necessarily. The 2025 Dana-Farber and Memorial Sloan Kettering study found that patients who received at least six months of checkpoint inhibitors before stopping due to side effects often maintained long-term disease control. Using immunosuppressive drugs to manage side effects did not worsen cancer outcomes.
What percentage of patients stay in remission after stopping immunotherapy?
It depends on the cancer type and trial. For NSCLC, approximately 46 to 48 percent of patients who completed two years of pembrolizumab had no progression three years later. For colorectal cancer with specific molecular features, 84 percent remained progression-free at three years. However, in some trials, up to half of patients progressed after stopping.
Is two years the standard length of immunotherapy treatment?
Two years (or 35 cycles of pembrolizumab) has become a commonly used fixed-duration approach supported by trial data and clinical consensus. The DANTE trial further supports time-limited therapy for melanoma patients who remain progression-free at two years.
Who is least likely to benefit from immunotherapy’s lasting effects?
Patients with low PD-L1 expression, low tumor mutation burden, or those who stopped treatment within the first three months tend to have shorter-lasting benefits. In the Dana-Farber study, patients who discontinued within three months had a median survival of 22 months compared with 87 months for those who received at least six months of treatment.
