The idea sounds like science fiction stated plainly: sequence a patient’s tumour, identify the mutations unique to it, manufacture a vaccine encoding exactly those targets, and teach the immune system to hunt the cancer down. Personalised mRNA cancer vaccines have now been doing this in human trials for several years — and the results have held up well enough that the field has moved from asking whether it works to asking how widely, and how fast it can be made.
The strongest data come from melanoma. In a randomised trial of patients whose high-risk melanoma had been surgically removed, adding a personalised vaccine (Moderna and Merck’s mRNA-4157/V940) to standard immunotherapy cut the risk of recurrence or death by roughly half compared with immunotherapy alone, with the benefit holding at three-year follow-up. Pancreatic cancer — a far harder target — produced the field’s most-discussed early signal: in a small study, half of the patients mounted strong immune responses to a personalised vaccine after surgery, and responders stayed cancer-free far longer than non-responders.
How a Vaccine Gets Personal
The process is a relay race against the cancer’s clock. A tumour sample and a blood sample are sequenced; algorithms compare them to find neoantigens — the abnormal protein fragments that mark the tumour as foreign. The top candidates, sometimes thirty or more, are encoded into a single mRNA strand, manufactured, quality-checked and shipped back for injection, typically within six to nine weeks of surgery. Research groups reporting through journals indexed by the National Institutes of Health have driven that turnaround down steadily, because weeks matter: the vaccine’s job is to eliminate the microscopic disease that surgery left behind, before it can regrow.
The Obstacles Are Industrial, Not Just Scientific
Each vaccine is a batch of one. That inverts pharmaceutical economics: instead of one factory making millions of identical doses, the model needs many small, fast, heavily automated productions — closer to a network of specialised kitchens than a brewery. Regulators have had to build frameworks for approving a process rather than a fixed product, and costs remain high enough that health systems are already debating which patients benefit most.
There are scientific limits too. Tumours with few mutations offer fewer targets; some cancers hide from the immune system in ways a vaccine alone cannot fix, which is why nearly all trials combine vaccines with checkpoint inhibitors that release the immune system’s brakes.
Equity questions follow close behind. A treatment manufactured for one patient will not be cheap, and the early candidates are cancers of wealthier health systems. Trialists counter that costs fall with automation — the first CAR-T cell therapies were similarly bespoke and similarly expensive — and that the platform, once built, generalises across cancers. Whether health systems fund personalised vaccines broadly will likely depend on the phase-three effect sizes: halving recurrence buys a great deal of manufacturing.
What Comes Next
Larger phase-three trials in melanoma and lung cancer will deliver verdicts over the next two years, and off-the-shelf vaccines targeting shared mutations — which would ease the manufacturing problem — are in early testing. The NHS in England has moved notably fast, launching a vaccine trials platform that matches patients to studies at diagnosis. As our health coverage has tracked, the direction is consistent: cancer treatment becoming less like a standard protocol and more like a bespoke commission.
The Bottom Line
Personalised vaccines will not replace surgery, chemotherapy or immunotherapy; they are being woven into them, aimed at the residual disease those tools miss. For patients, the shift is profound in a quieter way — treatment designed from their own tumour’s fingerprint. It is the same pattern as the heating transition covered in our heat pump report: a technology matures in specialist hands, then reorganises an entire field around itself.
