Cancer does not remain the same over time. As tumour cells accumulate somatic mutations, they can acquire new molecular features that distinguish them from normal cells.
What happens when a tumour develops a new molecular identity as it evolves? Some of these newly acquired mutations can generate abnormal peptides known as neoantigens. Unlike normal proteins, tumour-specific neoantigens may be recognised by T cells as new targets.
This is one of the ideas that has made personalised cancer immunotherapy particularly interesting to me: instead of asking the immune system to recognise cancer in a general way, could we identify the new molecular features created by each tumour and use them to guide a more precise immune response?
This is where personalised cancer vaccines come into the picture.
A personalised cancer vaccine is designed around the molecular features of an individual patient's tumour. Instead of giving every patient the same antigen, researchers identify tumour-specific targets, often neoantigens generated by somatic mutations, and use them to stimulate an immune response directed against that particular cancer.
But finding a mutation is only the beginning.
Sequencing can reveal hundreds or even thousands of tumour mutations, but not every mutation produces a peptide that is presented by HLA molecules, and not every presented peptide will trigger a T-cell response. This is where the integration of different technologies becomes particularly important.
Genomics can help identify tumour mutations. Computational approaches can predict which altered peptides might be recognised by the immune system. Immunopeptidomics can provide a more direct view of peptides actually presented by HLA molecules. Immunology then asks the crucial question: can T cells recognise these targets and respond to them?
The goal is to turn this molecular information into an immune response.
Dendritic cells are particularly interesting in this context because they act as professional antigen-presenting cells. In personalised dendritic-cell vaccines, these cells can be prepared to present tumour-specific information to T cells, helping initiate a response against the patient's own tumour.
From an idea to a patient
A first-in-human clinical trial of a personalised dendritic-cell vaccine, known as Neo-mDC, provides a useful example of how this concept has moved beyond theory.
The trial explored a personalised dendritic-cell vaccine in patients with resected non-small cell lung cancer. The process itself was challenging: tumour material had to be collected, potential neoantigens identified, and a patient-specific vaccine produced. Out of ten enrolled patients, six ultimately received the vaccine and completed the protocol. Importantly, the study demonstrated that this complex personalised workflow was feasible.
What happened after vaccination was even more interesting.
Five of the six vaccinated patients developed detectable T-cell responses. CD8⁺ T cells were the dominant responders, and these cells produced IFN-γ when stimulated with patient-specific neoantigens. Mutant peptides activated the T cells whereas the corresponding wild-type peptides did not, suggesting that the response was directed toward tumour-specific changes rather than simply the normal protein.
For me, this is where the concept becomes much more tangible. The mutation is no longer just a line in a sequencing report. It can become a biological signal that the immune system can recognise.
The response also showed signs of persistence. In some patients, antigen-specific T cells remained detectable for months, including long-lived memory populations. T-cell receptor analysis further revealed polyclonal responses, meaning that multiple T-cell clonotypes were involved rather than a single immune population carrying the entire response.
This matters because cancer is not a single, uniform target.
But can personalised vaccines keep up with cancer?
This is where the excitement meets some difficult practical questions.
Tumours are heterogeneous. Different tumour cells within the same patient may carry different mutations, and a tumour can continue to evolve under immune pressure. If treatment targets only a limited number of neoantigens, cancer cells that lose or stop presenting those targets may have an opportunity to escape.
Choosing the right neoantigens is therefore critical. Researchers need to consider not only whether a mutation exists, but whether it produces a relevant peptide, whether that peptide is presented by HLA molecules, and whether the patient's T cells can actually recognise it.
There is also the question of time.
The Neo-mDC experience illustrates how technically demanding personalised manufacturing can be. In that programme, the median time from surgery to the first vaccine dose was around 198 days.
For a patient with rapidly progressing disease, this is not simply a manufacturing problem. It can become a clinical problem.
This challenge has encouraged researchers to explore newer approaches that may shorten the path from tumour sampling to vaccination. One example described in the research is the Galsome-NEO platform, which combines tumour neoantigen mRNA with a lipid nanoparticle formulation containing an α-GalCer adjuvant. The proposed workflow aims to move from tumour sampling and neoantigen identification to mRNA synthesis and formulation within approximately nine weeks, substantially shorter than the earlier timeline.
A moving target
What I find particularly interesting about personalised cancer vaccines is that they address one of the defining characteristics of cancer: evolution.
The immune system is not responding to a completely static disease. Tumour cells can acquire new mutations, alter their antigenic landscape, and develop mechanisms that allow them to evade immune recognition. In this context, neoantigens represent more than just potential vaccine targets. They are molecular footprints of the changes occurring within the tumour.
This creates a fundamental challenge:
Can the immune system keep pace with a cancer that keeps changing?
Personalised vaccines offer one possible way of addressing this challenge. By identifying the molecular features that distinguish a patient's tumour and using them to guide an immune response, researchers are attempting to turn cancer evolution into a source of information rather than allowing it to remain purely an advantage for the tumour.
There are still major questions to answer. Which neoantigens are the best targets? How many should a vaccine contain? How can we account for tumour heterogeneity and immune escape? And how can personalised therapies be manufactured quickly and affordably enough to become practical for more patients?
These challenges make the field more complicated, but also more interesting.
Some of the most promising developments in cancer immunotherapy are emerging at the boundaries between disciplines. Genomics tells us what changed. Immunopeptidomics helps us understand what may actually be presented. Immunology tells us whether the immune system can respond. Clinical research tells us whether that response can matter for a patient.
Perhaps the future of cancer vaccination will not be about finding one universal cancer antigen.
Perhaps it will be about learning how to read the changes that each tumour acquires, and teaching the immune system to keep up with them.