On August 19, 2026, two drug companies announced a result that cancer researchers had waited years to see. In a large Phase 3 trial, a personalised mRNA treatment, given together with an immunotherapy drug, reduced the risk of melanoma coming back after surgery more than the immunotherapy drug did on its own. Merck and Moderna described it as the first positive Phase 3 result for an individualised neoantigen therapy and for an mRNA-based cancer therapy.
The phrase “cancer vaccine” suggests a jab that stops you getting cancer. This is something different, and the science behind it is easier to grasp than the jargon suggests. This article explains how it works, what the evidence shows, and what is still unproven.
First, what it is and what it is not
A preventive vaccine trains your immune system before an enemy arrives. A therapeutic vaccine trains it to fight a disease that is already there, or may still be lurking. The mRNA cancer vaccines in the news are therapeutic. In the melanoma trial, patients received the treatment after surgery had removed their tumour, to go after any cancer cells left behind that scans cannot see. Doctors call this “adjuvant” treatment.
Nor is it a cure or a one-size-fits-all product. Each dose is built from one patient’s own tumour. As of the companies’ August announcement, it had not been approved anywhere. Merck and Moderna said they would present the data at an upcoming medical meeting and discuss filing with regulators.
mRNA in plain language
Your cells follow instructions stored in DNA, which stays inside the cell’s nucleus. To make a protein, a cell copies the relevant instruction onto a short-lived messenger, called mRNA, which carries it out to the cell’s protein-making machinery. Think of DNA as the master cookbook locked in the library, and mRNA as a single recipe photocopied for the kitchen. Once the dish is made, the photocopy is thrown away.
An mRNA vaccine delivers a synthetic photocopy. Cells that take it up make the protein it describes, then break the mRNA down. According to MedlinePlus, mRNA from vaccines does not enter the nucleus and does not alter DNA, and the National Human Genome Research Institute notes that mRNA is short-lived and that your cells make their own mRNA all the time.
Bare mRNA is fragile, and the immune system would destroy it. So, as the National Cancer Institute explains, it is wrapped in lipid nanoparticles, tiny fat bubbles that protect it on the way in. This is the same basic technology used in the COVID-19 vaccines. Those vaccines were possible so quickly partly because mRNA is, in the words of one researcher quoted by NCI, easy, fast and scalable: the same manufacturing process works for any mRNA sequence.
The problem: cancer hides in plain sight
Your immune system is good at spotting things that are foreign, like viruses and bacteria. Cancer is harder, because cancer cells are your own cells gone wrong. Many tumours also find ways to switch off the immune attack.
But there is a weakness. Cancer arises from mutations, which are typos in the genetic instructions. Some of those typos make cells produce abnormal proteins that healthy cells never make. Fragments of these proteins are displayed on the cell’s surface, like name tags. Immune cells called T cells inspect those tags, and a tag they have never seen before can trigger an attack. These abnormal tumour-only markers are called neoantigens, and NCI notes that because they are not found on normal cells, they are promising targets.
Here is the catch. Every person’s tumour has its own set of typos. That is why this approach is personal: the vaccine has to be designed for one patient’s cancer.
How a personalised vaccine is made
- Surgery. The tumour is removed, and a sample goes to the lab.
- Sequencing. The tumour’s genes are read and compared with healthy tissue to find the mutations that are unique to the cancer.
- Picking targets. Computer algorithms predict which neoantigens are most likely to bind to T-cell receptors and trigger a response. The Merck–Moderna vaccine, called intismeran autogene, can include instructions for up to 34 of them.
- Making the mRNA. The chosen sequences are written into mRNA and packaged. NCI reports this takes one to two months from tissue collection, and that Moderna makes each of its personalised vaccines in about six weeks. In a pancreatic cancer trial, the time from surgery to the first dose was nine weeks.
- Injections. In the Phase 3 melanoma trial, the vaccine was given at 1 mg every three weeks for up to nine doses.
- Training the immune system. Cells that take up the mRNA make the neoantigen proteins. Among the likeliest to do so are dendritic cells, which NCI describes as sentinels that “act as teachers,” showing the new targets to T cells. T cells that recognise them multiply, circulate, and hunt for cells carrying those markers. Some become long-lived memory cells.
A helpful way to picture it: the vaccine hands the immune system a wanted poster of the tumour’s own faces.
Why it is given with another drug
In the trials, the vaccine is combined with Keytruda (pembrolizumab), a type of drug called a checkpoint inhibitor. Tumours can exploit a brake on T cells known as PD-1. Keytruda blocks that brake, which Merck says activates T cells. If the vaccine is the wanted poster, Keytruda removes the handbrake so the T cells can act on it. The vaccine adds aim, and the drug adds freedom to attack.
What the evidence shows
Melanoma. In a randomised Phase 2b trial called KEYNOTE-942, patients with high-risk melanoma that had been completely removed received either the vaccine plus Keytruda or Keytruda alone. Five-year follow-up, presented at the 2026 ASCO meeting, showed a 49 per cent lower risk of recurrence or death (hazard ratio 0.51, 95 per cent confidence interval 0.294 to 0.887) and a 59 per cent lower risk of distant spread or death (hazard ratio 0.411, interval 0.200 to 0.843).
A quick guide to reading those numbers: a hazard ratio of 0.51 roughly means that, at any given moment, the combination group had about half the risk of an event compared with the Keytruda-only group. It does not mean that 49 per cent of patients were cured.
Overall survival was an earlier, exploratory comparison, with a 2.5-year rate of 96.0 per cent against 90.2 per cent. The confidence interval (0.114 to 1.584) was wide and included no difference, so that figure was suggestive and not conclusive.
The Phase 3 trial, INTerpath-001, enrolled 1,137 patients with completely removed stage IIB to IV melanoma, randomly assigned 2:1 to the vaccine plus Keytruda, or Keytruda alone. At a pre-specified interim analysis, it met its main goal, recurrence-free survival, and a key secondary goal, distant metastasis-free survival. The companies said safety matched earlier studies with no new safety signals. They have not yet published how large the benefit was, and the trial continues to measure overall survival.
Pancreatic cancer. This cancer is among the hardest to treat, with a five-year survival rate of around 13 per cent, according to Memorial Sloan Kettering. A BioNTech and Genentech vaccine, autogene cevumeran, was tested with an immunotherapy drug and chemotherapy in a Phase 1 trial of patients who had surgery. As published in Nature, eight of the 16 evaluable patients developed strong vaccine-driven T-cell responses. At an extended follow-up of 3.2 years, those responders had longer cancer-free survival than the non-responders, and the vaccine-induced T cells were long-lived. MSK later reported that seven of the eight responders were alive four to six years after treatment. The authors also showed that responders and non-responders mounted equal immune responses to a COVID-19 mRNA vaccine, which argues against the difference being just a matter of stronger immune systems.
The caveats are large. It was a small, early trial without a control group, and it showed that response correlated with outcome, which is not the same as proving the vaccine caused the benefit. A randomised Phase 2 trial called IMCODE003, which aims to enrol 260 patients, is under way to test that.
Not every cancer responds. In an early head and neck cancer study, NCI reported in 2022 that two of the first ten patients had their tumours disappear and five had their tumours shrink, but a colorectal cancer group in the same study did not appear to benefit.
Safety and side effects
In the Phase 2b melanoma trial, the most common side effects linked to the vaccine were fatigue (60.6 per cent of patients), injection-site pain (56.7 per cent) and chills (49.0 per cent). Most were mild or moderate. Fatigue was the most common severe (grade 3) event, and there were no grade 4 or 5 events related to the vaccine. At about three years, severe treatment-related events (grade 3 or higher) were reported in 25 per cent of the combination group versus about 20 per cent of the Keytruda-only group. Immune-related side effects were similar (37.5 per cent versus 36 per cent), according to Merck. Keytruda itself can cause immune-related problems in many organs, which is a known and monitored risk of that drug. Detailed Phase 3 safety data have not yet been released.
Questions people ask
Will it change my DNA? No. As above, mRNA does not enter the nucleus and is broken down soon after it delivers its message.
Is it the same as the vaccines that prevent cancer? No. Preventive vaccines, such as those against viruses linked to some cancers, work by stopping the infection. A therapeutic vaccine like this one works on cancer that already exists or may have been left behind.
Can I get it? Not yet. It is available only in clinical trials. Patients who are curious should ask their oncologist about trials rather than seeking unproven treatments.
How much will it cost? No price has been announced.
The real limits
- Only one Phase 3 result so far, in one cancer, and the detailed numbers have not been released. The result came from an interim analysis, and overall survival is still being measured.
- Choosing targets is imperfect. NCI notes that finding the best neoantigens remains a challenge for the field, and researchers say it is not yet clear how best to combine the vaccines with other treatments.
- Time and logistics. Each vaccine needs a tumour sample, sequencing and weeks of manufacturing, which is a hurdle for patients whose cancer is moving fast.
- Tumours fight back. Immune responses can be switched off inside a tumour, and even the pancreatic trial produced a strong response in only half of the patients.
What comes next
Merck and Moderna’s INTerpath programme now has nine Phase 2 and Phase 3 trials across melanoma, non-small cell lung cancer, bladder cancer and kidney cancer, plus earlier studies in pancreatic, gastric and perioperative lung cancer. BioNTech, as of its 2024 update, had randomised Phase 2 trials of its own vaccine in pancreatic cancer, melanoma and colorectal cancer. The answer to whether this approach works beyond melanoma will come from those studies.
Key terms
| Term | Meaning |
|---|---|
| mRNA | A short-lived copy of a gene’s instructions that cells use to make a protein |
| Neoantigen | An abnormal protein marker made only by tumour cells because of mutations |
| T cells | Immune cells that inspect markers on cells and destroy those that look dangerous |
| Checkpoint inhibitor | A drug, such as Keytruda, that releases a brake on T cells |
| Adjuvant | Treatment given after surgery to deal with leftover cancer cells |
| Hazard ratio | The relative risk of an event over time; below 1 means lower risk |
No responses yet