The immune system’s new teachers: Personalised cancer vaccines and the logistics of made-to-order medicine

Cancer immunotherapy has taught us a humbling lesson: the immune system is powerful, but it needs instruction. Checkpoint inhibitors showed that releasing the brakes on immune cells can lead to dramatic remissions, yet only for a minority of patients. For many others, there is nothing to release because the immune system does not recognise the cancer as a threat. Personalised cancer vaccines aim to solve that problem by acting as tutors, showing the immune system exactly what to attack.

These are not vaccines in the childhood sense. They do not prevent disease. Instead, they are therapeutic, bespoke treatments designed after a cancer has already taken hold. Each one is made to order, based on the unique mutations in an individual tumour. The science behind them is elegant; the logistics resemble a high-stakes relay race.

What neoantigens are – and why they matter

Every cancer accumulates mutations as it grows. Some of these mutations alter proteins in ways that produce fragments never seen in healthy cells. These fragments, known as neoantigens, can be displayed on the surface of tumour cells. In theory, they are perfect targets: unique to the cancer and absent from normal tissue.

The immune system is exquisitely sensitive to such differences. T cells patrol the body, checking protein fragments presented by cells and deciding whether they belong. When they encounter something foreign, they can mount an attack. The problem is that many cancers either hide their neoantigens poorly or never provoke a strong immune response in the first place.

Personalised vaccines try to fix this by presenting selected neoantigens to the immune system in a way that demands attention. Instead of hoping the immune system stumbles upon the right target, the vaccine provides a curated syllabus.

Why personalisation is the point

Off-the-shelf cancer vaccines have been tried before, with limited success. Tumours of the same type may share a name, but genetically they can be wildly different. A lung cancer in one patient may have little in common with a lung cancer in another. Using a standard set of antigens risks teaching the immune system the wrong lesson.

Personalisation matters because it aligns the vaccine precisely with the tumour’s vulnerabilities. By choosing neoantigens that are truly tumour-specific and likely to be visible to immune cells, researchers hope to generate a stronger, more durable response. The trade-off is complexity. Each patient becomes a manufacturing project of one.

From biopsy to vaccine: the workflow

The process begins with a biopsy or surgical sample of the tumour, paired with a blood sample to represent the patient’s normal DNA. Both are sequenced, producing vast amounts of genetic data. Bioinformatic algorithms then compare the two, identifying mutations unique to the cancer.

From hundreds or thousands of mutations, a shortlist must be made. Not every mutation produces a usable neoantigen. Some will never be presented to immune cells; others may look foreign but fail to provoke a response. Predictive models estimate which neoantigens are most likely to bind to the patient’s immune machinery and be recognised by T cells.

Once the targets are chosen, the vaccine is designed. Historically this involved synthesising peptides or loading dendritic cells in the lab, approaches that were slow and labour-intensive. The real shift came with messenger RNA.

What mRNA changed

mRNA platforms transformed the field by turning vaccine manufacture into an information problem. Instead of producing proteins directly, manufacturers produce strands of genetic instructions that tell the patient’s own cells which neoantigens to make. The same underlying process can be reused; only the sequence changes.

This flexibility shortens timelines and lowers barriers to customisation. In principle, dozens of neoantigens can be encoded in a single mRNA vaccine, broadening the immune response and reducing the chance that the tumour escapes by shedding one target. The success of mRNA vaccines in infectious disease also accelerated regulatory familiarity and manufacturing capacity.

Yet mRNA does not erase the bottlenecks. Sequencing, analysis, quality control and distribution all take time. For a patient with aggressive disease, weeks matter.

Measuring success: what counts as evidence

The evidence landscape for personalised cancer vaccines is still emerging. Early trials have focused on safety and immune responses: can the vaccine provoke T cells that recognise the chosen neoantigens? Increasingly, studies are looking at clinical endpoints such as recurrence-free survival, tumour shrinkage and overall survival.

One challenge is attribution. Many trials combine vaccines with checkpoint inhibitors, which can amplify immune responses. If a patient does well, how much credit belongs to the vaccine? Randomised studies comparing checkpoint inhibitors with and without personalised vaccines are beginning to address this, but they are complex and costly.

Another issue is time horizon. Vaccines may not cause immediate tumour regression; their value may lie in preventing relapse after surgery or keeping residual disease in check. That requires long follow-up and patience from investors and patients alike.

Combination therapy, not replacement

Few researchers see personalised vaccines as stand-alone cures. Instead, they are viewed as partners to existing immunotherapies. Checkpoint inhibitors remove the brakes; vaccines press the accelerator. Other combinations are being explored, including with radiation or targeted drugs that increase antigen presentation.

The immune system is adaptable, but tumours are too. They can mutate further, downregulate immune signals or create hostile microenvironments. A vaccine may need reinforcement, or redesign, as the disease evolves.

The hard parts

• Manufacturing a unique product for every patient, at clinical-grade quality
• Turnaround times that must fit within treatment windows
• Coordinating sequencing, analysis and production across institutions
• Integrating bespoke therapies into hospital workflows built for standard drugs
• Managing costs that challenge reimbursement models
• Protecting genomic data privacy across borders
• Scaling supply chains without losing customisation
• Proving benefit convincingly enough for regulators and payers

Cost, time and the reality of hospitals

Personalised cancer vaccines sit uneasily within health systems designed for mass-produced medicines. Each dose carries sequencing costs, computational analysis, bespoke manufacturing and cold-chain logistics. While prices may fall with scale, they are unlikely to match those of conventional drugs.

Turnaround time is equally unforgiving. From biopsy to injection can take six to eight weeks, sometimes longer. For patients with rapidly progressing disease, that may be too slow. Streamlining each step without compromising accuracy is an active area of work, but biology sets limits.

Hospitals must also adapt. Clinicians need systems to track samples, data and manufacturing slots. Errors can be catastrophic when there is no spare dose. This is medicine as choreography.

Data, trust and consent

The genomic data required to build these vaccines is intensely personal. Tumour sequences can reveal inherited variants and other sensitive information. Safeguarding that data, while allowing it to move quickly between labs and manufacturers, is a non-trivial challenge.

Patients must also understand what they are consenting to: a therapy tailored to their tumour, but also a process that may generate insights beyond their own care. Trust will be as important as technology.

Who might benefit first

In the near term, personalised cancer vaccines are most likely to benefit patients with cancers that have high mutation burdens and are treated in specialised centres. Melanoma and certain lung cancers have been early candidates, particularly in settings where surgery removes most of the tumour and the vaccine aims to prevent recurrence.

Over time, improved prediction algorithms, faster sequencing and more automated manufacturing could broaden access. The question is not whether the science can work, but whether the systems around it can keep up.

What would make it scalable

True scalability would mean reducing turnaround times to days, not weeks; lowering costs through standardised platforms; and integrating design and manufacturing into routine care. Advances in artificial intelligence for neoantigen prediction, decentralised manufacturing units and harmonised regulation could all help.

There is also a strategic choice to be made. Personalised vaccines could remain premium interventions for selected patients, or they could become part of a more general shift towards made-to-order medicine. The immune system is adaptable, but health systems must be too.

Personalised cancer vaccines ask us to rethink what a medicine is. Not a pill or a vial on a shelf, but a process that starts with a patient’s biology and ends, weeks later, with a bespoke lesson for their immune system. Whether that lesson can be delivered reliably, affordably and in time is the test that will determine whether these new teachers reshape cancer care or remain an intriguing, unfinished syllabus.