For decades, cancer vaccines were the field's most stubborn disappointment. Tumours mutate too fast, hide too well, and look too much like the body's own cells for the immune system to reliably attack them. Trial after trial through the 1990s and 2000s ended in modest, statistically meaningless results. That equation is now changing.
The treatment, known clinically as mRNA-4157 (and being co-developed by Moderna and Merck), begins not with a drug but with a biopsy. Researchers sequence the patient's tumour DNA, identify up to 34 of its unique mutationsGlossaryA mutation is a change in DNA. Cancer cells carry mutations the body's healthy cells don't — which makes them targetable., and within weeks manufacture a custom mRNA vaccine — a personalised set of genetic instructions that train the patient's own immune system to recognise their cancer, and only their cancer.
It is the most literal example of personalised medicine ever attempted at this scale. Two patients with the same disease — say, melanoma — would receive completely different vaccines, designed against the specific mutations their tumours happen to carry.
What the trials actually showed
In a Phase 2b melanoma trial completed in late 2023 and updated through 2025, patients who received mRNA-4157 alongside Merck's checkpoint inhibitorGlossaryA drug that lifts the brakes the immune system normally puts on itself — so it can attack cancer cells. Keytruda is one example. Keytruda saw their risk of recurrence cut by roughly 44% compared to Keytruda alone. The benefit grew larger over time, not smaller — an unusual signal in oncology.
Vaccine-induced immune responses persisting for nearly four years after treatment in some patients.
Phase 3 trials are now underway in melanoma, non-small-cell lung cancer, and pancreatic cancer — three of the diseases where new tools are most urgently needed. Investigators are also tracking patients from earlier trials whose vaccine-induced immune responses are still detectable nearly four years after treatment, suggesting the protection may be durable in a way checkpoint drugs alone are not.
How a personalised vaccine actually works
The mechanism is simpler to describe than to engineer. The tumour biopsy is sequenced. Software compares the tumour's DNA to the patient's healthy DNA and flags neoantigensGlossaryProtein fragments that exist on cancer cells but not on healthy ones. They're what the vaccine teaches the immune system to recognise. — protein fragments that exist on the cancer but not on healthy cells. The vaccine encodes mRNA instructions for those specific neoantigens. Once injected, the patient's cells temporarily produce them, the immune system learns to recognise them, and the resulting T-cells go hunting.
Combined with a checkpoint inhibitor like Keytruda — which lifts the brakes the tumour places on the immune system — the result is, in some patients, an immune response substantially more focused than anything previous immunotherapy alone has produced.
What this means for patients
If you or a family member has melanoma, lung cancer, or pancreatic cancer, this is the most important trial to watch right now. These are Phase 3 trials with strong results, on track for regulatory submissions in 12–18 months. Ask your oncologist about mRNA-4157 eligibility.
What we are watching next
Phase 3 readouts in melanoma are expected in late 2026 and early 2027. If the recurrence benefit holds, the first approval — likely for high-risk resected melanoma — could come within 18 months. Lung and pancreatic readouts will follow.
The bigger question is whether this approach scales. Personalised manufacturing is expensive, slow, and bottlenecked by sequencing capacity. But if the clinical results continue to look like this, the field will solve those problems. There is too much pressure not to.