Personalized mRNA cancer vaccines moved from proof of concept to phase 3 trials — and the early data is striking

When BioNTech and Moderna demonstrated that mRNA could be manufactured, stabilized, and injected into humans in under a year to produce a COVID vaccine, it validated a platform that researchers had been developing for decades. The more ambitious application was always cancer: using the same mRNA technology to train the immune system to recognize and attack a patient's specific tumor. That application has now cleared the proof-of-concept stage.

In 2023, Moderna and Merck published phase 2b results for mRNA-4157/V940 — an individualized neoantigen therapy (INT) for high-risk melanoma. When combined with pembrolizumab (Keytruda, Merck's PD-1 inhibitor), the vaccine reduced the risk of recurrence or death by 49% compared to pembrolizumab alone. That result, published in The Lancet, was striking enough that the FDA granted the combination Breakthrough Therapy designation. Phase 3 enrollment is now underway.

How a personalized cancer vaccine is built

The process starts when a surgeon removes a tumor sample. That sample is sequenced — both the tumor's genome and the patient's normal cells — to identify somatic mutations: the DNA changes that exist only in the cancer cells, not in healthy tissue. A subset of these mutations produce proteins that stick out on the cancer cell's surface, called neoantigens. These are targets the immune system could recognize as foreign, but hasn't, because cancer cells have evolved mechanisms to evade immune detection.

An AI system analyzes the sequencing data and selects the 34 most immunogenic neoantigens — the ones most likely to trigger a strong T-cell response. The mRNA vaccine encodes synthetic versions of these neoantigens. When injected, the mRNA instructs the patient's own cells to produce the neoantigen proteins, which the immune system then learns to recognize. Once trained, T-cells can seek out and destroy any cell displaying those markers — which means the cancer cells.

The manufacturing process currently takes six to eight weeks from tumor biopsy to finished vaccine, and each vaccine is unique. A manufacturing facility cannot batch-produce inventory; it produces one patient's vaccine at a time using automated synthesis and quality control. Both Moderna and BioNTech are investing heavily in automation and distributed manufacturing to reduce this timeline and cost.

What the data shows so far

The melanoma phase 2b results are the most mature dataset. The 49% improvement over pembrolizumab monotherapy — itself already a highly effective immunotherapy — is a meaningful signal. Pembrolizumab alone had already transformed melanoma treatment; combining it with a neoantigen vaccine appears to produce additive immune activation.

Moderna has expanded its INT program to other cancer types. Non-small cell lung cancer (NSCLC) is in phase 2 trials. Colorectal cancer, bladder cancer, and renal cell carcinoma programs are at earlier stages. BioNTech's equivalent program, called individualized neoantigen-specific immunotherapy (iNeST), is running parallel trials in several solid tumor types. The trial designs vary, but the core hypothesis is consistent: patients whose tumors have strong neoantigen profiles should respond to a vaccine that primes T-cells against those neoantigens.

Not all cancers are equally good candidates. Some tumors — particularly certain subtypes of breast cancer, prostate cancer, and pancreatic cancer — have low tumor mutation burdens, meaning they accumulate fewer somatic mutations and present fewer neoantigens. These "cold" tumors also tend to have immunosuppressive microenvironments that resist T-cell infiltration even when the immune response is primed. The vaccine approach is most promising for "hot" tumors with high mutation burdens: melanoma, NSCLC, colorectal cancer with microsatellite instability, and bladder cancer.

The manufacturing and cost challenge

Personalized medicine is expensive by definition. The current per-patient cost of manufacturing mRNA-4157/V940 is not publicly disclosed, but analogous individualized immunotherapies have historically cost $100,000–$400,000 per treatment course. CAR-T therapies, which also require patient-specific manufacturing, are priced at $300,000–$500,000 per infusion in the US.

The mRNA manufacturing process is more automatable than CAR-T (which requires harvesting and engineering a patient's T-cells), which gives personalized mRNA vaccines a clearer path to cost reduction. Moderna's manufacturing roadmap projects that scaled automation could reduce synthesis time below four weeks and bring costs significantly below current CAR-T pricing — though "significantly below $300,000" still puts this out of reach for most of the world without insurance coverage or health system purchasing power.

Payer coverage will be the critical gating factor. Insurers and national health systems price cancer treatments against survival benefit. If phase 3 trials confirm the phase 2b results, a 49% reduction in recurrence in high-risk melanoma is almost certainly cost-effective by standard health economic thresholds. The harder cases will be earlier-stage cancers where the absolute benefit is smaller, or cancer types where the patient population is smaller and the evidence base thinner.

What comes after solid tumors

The longer-term aspiration — which researchers are careful to describe as still theoretical — is applying personalized mRNA vaccines to prevent cancer recurrence after surgery, before any evidence of metastasis. This adjuvant setting, treating patients with no detectable disease but known high recurrence risk, is where the impact could be largest: if you can prevent a stage 2 melanoma from becoming stage 4, you've potentially cured the patient rather than just extending life.

BioNTech has also begun exploring whether the neoantigen vaccine approach could extend beyond cancer to other diseases where the immune system needs to be retrained. Autoimmune applications — teaching the immune system to tolerate its own tissues — and chronic infectious disease programs are both in early research stages.

What the phase 2b melanoma data established is that personalized mRNA cancer vaccines are not a speculative technology. They produce measurable, clinically meaningful immune responses that translate to real reductions in cancer recurrence. Whether they achieve the regulatory approvals, manufacturing scale, and payer coverage to become a standard treatment is a question the next three to five years of phase 3 data will answer.