AI Cancer Vaccine: How Moderna and Merck Target Melanoma

The idea of using Artificial Intelligence to help treat cancer is moving from futuristic speculation toward increasingly sophisticated clinical applications. Moderna and Merck are developing intismeran autogene, an investigational personalised mRNA-based cancer therapy designed around the unique mutations found in an individual patient’s tumour.

The treatment is being studied in combination with Merck’s immunotherapy Keytruda (pembrolizumab), particularly in patients with high-risk melanoma after surgery. Earlier clinical results have shown a meaningful reduction in the risk of melanoma recurrence or death compared with Keytruda alone, while longer-term follow-up has continued to support the potential of the approach.

That progress has also renewed a much bigger discussion about the role of artificial intelligence in medicine. Elon Musk described the potential of mRNA Technology in unusually broad terms, writing that artificial RNA could make curing diseases “a software problem.” The statement is provocative, but the technology behind the cancer programme illustrates why AI-assisted medicine is attracting so much attention.

What is the Moderna and Merck personalised cancer vaccine?

Intismeran autogene, formerly known as mRNA-4157 or V940, is not a conventional preventive vaccine like those used against infectious diseases. It is an individualised cancer immunotherapy designed after a patient’s tumour has been analysed.

The basic objective is to help the immune system recognise cancer cells carrying mutations that distinguish them from healthy tissue. Instead of giving every patient the same vaccine, researchers seek to construct a treatment tailored to the molecular characteristics of that person’s tumour.

This distinction is important. A conventional vaccine generally prepares the immune system to recognise a pathogen or a particular antigen. A personalised cancer vaccine, by contrast, attempts to exploit the mutations created as a tumour develops.

Moderna and Merck are studying the treatment particularly in melanoma, an aggressive form of skin cancer in which recurrence after apparently successful surgery remains a significant concern for high-risk patients.

How AI can help build a personalised cancer vaccine

The process begins with the patient’s tumour. After surgery, a sample can be genetically analysed to identify mutations present in the cancer.

A tumour may contain a large number of genetic changes, but not every mutation is equally useful for triggering an immune response. This creates a major computational problem: researchers need to identify which tumour-specific mutations are most likely to produce targets that the patient’s immune system can recognise.

That is where computational algorithms and machine-learning techniques can become valuable.

The supplied tumour information can be analysed to identify and rank potential neoantigens, which are abnormal proteins or protein fragments associated with cancer-specific mutations. The objective is to select the targets that appear most likely to produce a useful immune response.

Once the targets have been selected, an mRNA-based treatment can be designed to provide the biological instructions needed to produce them. The patient’s cells temporarily use those instructions to make the selected proteins, allowing the immune system to learn what to look for.

In simplified terms, the process resembles a chain:

• Tumour tissue is obtained and genetically analysed.
• Mutations that distinguish cancer cells from healthy cells are identified.
• Computational methods help prioritise promising neoantigen targets.
• An individualised mRNA sequence is designed around those targets.
• The treatment is administered to the patient alongside immunotherapy in the clinical programme.
• The immune system is trained to recognise cells displaying the selected cancer-related targets.

The important point is that AI does not independently “cure” cancer. It can help researchers process complex biological information and select potential targets more efficiently. The resulting treatment still has to demonstrate safety and effectiveness through clinical trials.

Why mRNA is important in cancer treatment

mRNA technology became widely recognised during the COVID-19 pandemic, but its potential applications extend well beyond infectious diseases.

Messenger RNA acts as a temporary set of biological instructions. In a therapeutic setting, researchers can use that mechanism to tell cells to produce specific proteins without permanently altering the patient’s DNA.

For cancer research, that flexibility is particularly attractive because tumours are genetically diverse. Two patients diagnosed with the same cancer can have substantially different molecular characteristics.

A personalised approach therefore attempts to move away from a purely one-size-fits-all model.

Traditional cancer treatments such as chemotherapy can attack rapidly dividing cells, which may also affect healthy tissue. Immunotherapies work differently, attempting to harness or restore the body’s own immune response against cancer.

A personalised neoantigen therapy combines these ideas in another way: it attempts to give the immune system a more precise set of targets based on the patient’s own tumour.

What the clinical evidence shows so far

The enthusiasm surrounding intismeran autogene is based on clinical evidence, but it is important to distinguish between encouraging trial results and an approved cancer cure.

In the Phase 2b KEYNOTE-942 study, patients with high-risk stage III or IV melanoma WHO had undergone complete tumour removal were treated with either Keytruda alone or the combination of Keytruda and the investigational personalised therapy.

Merck and Moderna reported that the combination produced a sustained improvement in recurrence-free survival. Five-year follow-up data announced in 2026 showed a 49% reduction in the risk of recurrence or death compared with Keytruda alone. The companies also reported a reduction in the risk of distant metastasis or death in earlier follow-up. :contentReference[oaicite:0]{index=0}

Those results are significant because the purpose of adjuvant treatment is to reduce the possibility that cancer will return after surgery. However, Phase 2b evidence is not the same as regulatory approval, and the companies are continuing to test the treatment in larger Phase 3 studies.

Phase 3 testing is the critical next step

The Phase 3 INTerpath-001 trial is designed to provide a more definitive assessment of intismeran autogene in high-risk melanoma. The trial is randomised and compares the investigational therapy plus pembrolizumab with pembrolizumab-based treatment without the personalised vaccine. ClinicalTrials.gov lists the study as a Phase 3 trial involving high-risk stage II-IV melanoma following complete resection. :contentReference[oaicite:1]{index=1}

Merck said in its 2026 reporting that the Phase 3 melanoma programme was ongoing and fully enrolled. Moderna has also identified Phase 3 intismeran readouts as an important part of its oncology development programme. :contentReference[oaicite:2]{index=2}

This stage matters because promising early results can sometimes fail to translate into successful Phase 3 outcomes. Larger trials provide a stronger test of whether a treatment’s benefits are reproducible across a broader patient population.

Why melanoma is an important testing ground

Melanoma has become an important area for cancer immunotherapy research because tumours can contain substantial numbers of mutations, creating potential neoantigens that the immune system might recognise.

For patients with high-risk melanoma, removing the visible tumour does not necessarily eliminate the possibility of microscopic disease remaining elsewhere in the body. That is why treatment after surgery can be important.

The goal of a personalised cancer vaccine is therefore not simply to shrink an existing tumour. In this setting, the broader objective is to help the immune system identify and eliminate residual cancer cells before they establish new disease.

If that strategy continues to demonstrate durable benefits, the implications could extend beyond melanoma. Moderna and Merck have been studying intismeran autogene across multiple oncology settings, including non-small-cell lung cancer. Merck’s 2026 pipeline lists the programme in Phase 3 development for melanoma and non-small-cell lung cancer. :contentReference[oaicite:3]{index=3}

Why Elon Musk calls disease treatment a “software problem”

Elon Musk’s description of medicine as a software problem is best understood as a comment on the increasing role of computation in biological research rather than a literal claim that diseases can simply be solved by writing software.

The analogy comes from the way personalised therapies increasingly depend on analysing huge amounts of biological data, identifying patterns and generating a treatment designed for a specific patient.

In the cancer-vaccine model, the biological challenge remains extremely complicated. But software can help researchers handle tasks that would be difficult to perform manually, including comparing genetic sequences, ranking candidate mutations and managing the design of individualised treatments.

That distinction matters. AI can accelerate the search for useful biological signals, but it cannot replace clinical validation, Manufacturing, regulatory review or medical care.

AI does not replace doctors or clinical trials

The excitement surrounding AI in Healthcare can sometimes lead to exaggerated expectations. A system may correctly identify a promising cancer target, but that does not prove that targeting it will improve survival.

Several additional questions must be answered before a personalised therapy can become routine:

• Does the selected target consistently generate a strong immune response?
• Does that immune response actually reduce recurrence?
• How durable is the benefit?
• Which patients are most likely to respond?
• What side effects are associated with the combination?
• Can personalised treatments be manufactured quickly and reliably?
• Can the process be scaled at a cost that makes it accessible to patients?

These questions explain why the Phase 3 programme is so important. A successful personalised therapy must work not only as a scientific concept but also as a practical medical product.

The manufacturing challenge behind personalised medicine

There is another difference between a personalised cancer vaccine and a conventional vaccine: every patient can require a different product.

That creates a manufacturing and logistics challenge. A conventional vaccine can be produced in large batches and distributed widely. An individualised cancer treatment requires patient-specific biological information to be processed before the treatment can be designed and manufactured.

For personalised medicine to become broadly useful, the entire process needs to be reliable and fast enough to fit within the patient’s treatment timeline.

Moderna has been expanding manufacturing capabilities for intismeran autogene. The company said its Marlborough facility began supplying clinical batches of the therapy in September 2025. :contentReference[oaicite:4]{index=4}

That highlights an often-overlooked part of the AI-and-medicine story: better algorithms are only one piece of the puzzle. Sequencing, software, manufacturing, quality control, distribution and clinical delivery all have to work together.

Could personalised cancer vaccines change cancer treatment?

The potential is considerable, but the word “potential” remains important.

If Phase 3 testing confirms the benefits observed in earlier studies, personalised neoantigen therapies could add a new layer to cancer treatment. Instead of relying solely on broadly targeted drugs, physicians could increasingly combine established immunotherapies with treatments designed around the molecular fingerprint of an individual’s tumour.

That would represent a significant shift in oncology. Cancer would be treated not only according to where it appears in the body, but also according to the genetic characteristics of the tumour.

Yet personalised cancer vaccines are unlikely to make surgery, chemotherapy, radiation or existing immunotherapies obsolete across the board. Cancer consists of many different diseases, and the biology varies substantially between patients and tumour types.

What happens next for Moderna and Merck?

The immediate milestone is further evidence from the Phase 3 programme. Moderna and Merck will need to establish whether the promising recurrence-related results can be reproduced in the larger trial population and whether the overall benefit-risk profile supports regulatory submissions.

The companies have already described intismeran autogene as a major part of their oncology strategy. Moderna has said that, if Phase 3 results are positive, it could prepare for a potential launch with Merck, subject to regulatory approval. :contentReference[oaicite:5]{index=5}

The broader significance goes beyond one drug. Intismeran autogene is effectively a test of whether the combination of mRNA technology, tumour sequencing, computational analysis and immunotherapy can produce a scalable new model for Cancer Care.

The bigger lesson from the AI cancer vaccine race

The most important development is not that AI has suddenly solved cancer. It has not.

What is changing is the way scientists can approach diseases that were previously too complex to personalise at scale. Modern sequencing can reveal the genetic characteristics of a tumour, computational tools can analyse those characteristics, and mRNA technology provides a flexible way of turning selected biological information into a treatment.

That combination explains why the phrase “software problem” has entered the medical conversation.

But the evidence also suggests a more measured conclusion. AI can make parts of drug discovery and treatment design faster and more data-driven. It can help researchers identify patterns that might otherwise remain hidden. Whether those discoveries ultimately become cures still depends on biology, clinical trials, manufacturing and patients’ real-world outcomes.

For Moderna and Merck, the next phase is therefore not about proving that AI can cure cancer. It is about proving that an individualised mRNA therapy can deliver a durable clinical benefit to people with melanoma and, potentially, other cancers.

If that evidence holds up, personalised cancer vaccines could become one of the clearest examples yet of computation and biotechnology converging in modern medicine.