The biological logic behind neoantigen vaccines is compelling. Neoantigens arise from somatic mutations that exist only in tumor cells, which means the immune system has not been trained to tolerate them, and the risk of off-target autoimmunity is theoretically low. Advances in next-generation sequencing, HLA typing, immunopeptidomics and synthetic vaccine platforms have made it feasible to sequence a patient’s tumor, identify candidate mutation-derived antigens and manufacture an individualized vaccine in a matter of weeks. Early clinical studies have now established feasibility, safety and immunogenicity across a striking range of solid tumors, including melanoma, pancreatic ductal adenocarcinoma, hepatocellular carcinoma, non-small cell lung cancer, bladder cancer and renal cell carcinoma, using platforms that span mRNA, peptide, DNA, viral vector and dendritic cell technologies.

The clinical momentum is real. In the phase 2b KEYNOTE-942 trial, the individualized mRNA vaccine mRNA-4157, known as V940, combined with pembrolizumab achieved 18-month recurrence-free survival of 79 percent versus 62 percent with pembrolizumab alone in resected melanoma, a hazard ratio of 0.561, with no grade 4 or 5 adverse events attributed to the vaccine. In pancreatic cancer, one of the hardest immunologic targets in oncology, responders to the personalized RNA vaccine autogene cevumeran showed markedly prolonged recurrence-free survival compared with non-responders, with a median not yet reached versus 13.4 months. A phase 3 trial in resected high-risk melanoma, INTerpath-001, is now enrolling roughly 1,089 patients, a sign that the field is moving from proof of concept toward pivotal testing.

Yet beneath these headline numbers lies an uncomfortable attrition problem that the review dissects stage by stage. Only a small fraction of the mutations identified in a tumor ever become genuinely immunogenic vaccine targets. In the autogene cevumeran pancreatic cancer trial, just 11 percent of the administered vaccine neoantigens, 25 out of 230, elicited measurable T-cell responses. In the NeoVax melanoma study, roughly 60 percent of vaccine neoantigens induced CD4-positive T-cell responses but only about 16 percent triggered CD8-positive cytotoxic T cells. A recent systematic evaluation found that only around 22 percent of predicted neoepitopes are actually presented on HLA molecules. Each false positive dilutes the effective immunogenic content of a vaccine that has limited room for passengers.

The root of the problem is that antigen presentation is a multi-step biological process, involving proteasomal cleavage, TAP-mediated transport and peptide-HLA loading, that most computational pipelines compress into a single binding-affinity prediction. Even when a peptide is presented, whether it provokes a functional T-cell response depends on the three-dimensional docking geometry between the peptide-HLA complex and a cognate T-cell receptor, governed by molecular flexibility and peptide conformation that sequence-based models capture poorly. Compounding matters, the scarcity of paired TCR-peptide-MHC datasets and the staggering diversity of the TCR repertoire make generalization across patients extraordinarily difficult. Manufacturing timelines, typically seven to eight weeks for peptide and viral vector platforms, add a further constraint, since tumors can evolve while the vaccine is being built.

This is where the review sees artificial intelligence earning its place, not as a replacement for experiments but as a way to narrow enormous biological and chemical search spaces. At the discovery stage, machine learning frameworks now integrate genomic, transcriptomic and mass spectrometry-based proteomic data to prioritize candidates with multi-layer evidence of expression and presentation. Tools such as ProGeo-neo combine proteomics with patient-specific genomic profiles to confirm protein-level expression, while ImmuneMirror merges whole-exome and RNA sequencing data, and modular workflows like TIminer and pVACtools automate reproducible end-to-end immunogenomic profiling. The authors also highlight the growing importance of noncanonical antigens, derived from alternative open reading frames, aberrant splicing, intron retention, fusion transcripts, circular RNA products and endogenous retroviral elements, some of which may be shared across patients and could inform semi-shared vaccine strategies.

Presentation prediction is where measurable gains have been most convincing. Pan-allelic tools such as NetMHCpan integrate eluted ligand data with binding affinities, and MHCflurry 2.0 folds antigen-processing features into neural network models, achieving an area under the curve of roughly 0.911 for binding classification in a benchmark of 18 predictors across 32 HLA alleles. Newer architectures such as CapHLA jointly model class I and class II presentation, and protein language model-based approaches like MUNIS improve prioritization of presented epitopes. The review also makes a pointed case for CD4-positive T cells, which clinical studies have repeatedly activated, sometimes more frequently than CD8 responses, arguing that HLA class II prediction, despite its computational challenges from open binding grooves and variable peptide lengths, deserves far more attention in vaccine design.

At the T-cell recognition frontier, deep learning is pushing into structure. AlphaFold-based modeling of peptide-HLA and TCR-peptide-HLA complexes can discriminate true target epitopes from decoys, while tools such as NetTCR-2.0, DeepTCR, TRAP and SageTCR integrate paired TCR sequences and three-dimensional structural features to predict receptor binding. Benchmarking studies, however, consistently show that these models remain strongly data-dependent and often fail to generalize to unseen epitopes, so the authors position them as prioritization and hypothesis-generating tools rather than definitive immunogenicity predictors. Downstream, AI is also reshaping formulation science: the AGILE platform uses deep learning-guided screening to identify ionizable lipids for mRNA delivery, and the LiON framework applies message-passing neural networks to large lipid nanoparticle datasets, narrowing chemical design spaces before any wet-lab testing begins.

To tie these threads together, the authors propose an evidence-weighted, closed-loop decision-gate framework spanning six stages: candidate generation, tumor relevance filtering, HLA presentation evidence, immunogenicity refinement, portfolio-level selection, and manufacturability and clinical delivery, with post-vaccination immune monitoring data fed back to refine future models. They stress that repertoire size should reflect biologically meaningful coverage rather than numerical abundance, since no clinically validated optimal neoantigen count exists. On the regulatory front, the UK’s Medicines and Healthcare products Regulatory Agency has released draft guidance for individualized mRNA cancer immunotherapies, in which AI selection algorithms may qualify as Software as a Medical Device, requiring Good Machine Learning Practice, transparent pipelines, version tracking and barcode-based traceability linking genomic data to the administered dose. The review’s bottom line is measured but optimistic: AI will not conjure effective cancer vaccines on its own, but embedded within rigorous experimental validation, scalable manufacturing and evolving regulatory oversight, it may finally make personalized neoantigen vaccination a reproducible pillar of precision oncology.

Subject of Research: The application of artificial intelligence across the translational development pipeline of personalized neoantigen cancer vaccines, including neoantigen discovery, HLA presentation prediction, T-cell recognition modeling and formulation optimization.

Article Title: Artificial intelligence for translational personalized neoantigen cancer vaccine development

Article References: Artificial intelligence for translational personalized neoantigen cancer vaccine development. (n.d.). https://doi.org/10.1186/s12929-026-01286-3

Image Credits: AI Generated

DOI: 10.1186/s12929-026-01286-3

Keywords: neoantigen cancer vaccine, artificial intelligence, personalized immunotherapy, HLA presentation, T-cell receptor, lipid nanoparticle, mRNA vaccine, precision oncology, immunopeptidomics, machine learning, cancer immunotherapy, translational oncology

Personalized cancer vaccines are designed for an individual patient rather than a broad population. Most target neoantigens, abnormal protein fragments created by mutations in tumor cells but absent from healthy tissues. Because these altered fragments can be recognized as foreign by T cells, they offer a way to direct the immune system toward malignant cells with greater precision than conventional cancer treatments. A typical development process begins with tumor and normal-tissue sequencing, followed by computational analysis to identify mutations that could generate recognizable peptides. The selected targets are then encoded in a vaccine platform, such as messenger RNA, synthetic peptides, DNA, or viral vectors. Although the concept is straightforward in principle, each stage contains variables that can influence the final treatment.

The first challenge is the quality and interpretation of tumor genomic data. Tumors are genetically diverse and often contain a mixture of malignant and nonmalignant cells, meaning that a mutation detected in a biopsy may not be present in every cancer cell. Samples can also differ in their purity, sequencing depth and storage conditions. Computational pipelines must distinguish genuine tumor mutations from technical errors and inherited variants, then determine which mutations are likely to produce peptides presented by the patient’s human leukocyte antigen molecules. These HLA proteins display intracellular protein fragments on the cell surface for inspection by T cells. Because HLA genes vary substantially between individuals, an antigen predicted to be visible in one patient may be poorly presented in another. Harmonized sequencing standards and prediction benchmarks are therefore essential for determining whether candidate neoantigens are truly comparable across studies.

A second concern is how researchers define a high-value neoantigen. Prediction algorithms commonly evaluate factors such as mutation type, gene expression, peptide binding to HLA molecules and the likelihood that T-cell receptors can recognize the displayed fragment. Yet a strong computational score does not guarantee an immune response. Some predicted peptides are produced inefficiently, degraded before reaching the cell surface or hidden by the tumor’s mechanisms of immune evasion. Others may be recognized only by a small population of T cells. The authors’ emphasis on harmonization highlights the need to combine computational predictions with experimental validation, including mass-spectrometry analysis of naturally presented peptides and functional tests using patient immune cells. Establishing shared criteria for evidence could reduce the number of weak targets entering clinical development.

The vaccine platform itself introduces another layer of variation. Messenger RNA vaccines can be manufactured rapidly and translated directly into antigenic proteins inside cells, while peptide vaccines require delivery systems and adjuvants to stimulate sufficient immune activation. Viral-vector vaccines use engineered viruses to carry tumor-antigen genes into cells, taking advantage of the strong innate and adaptive immune responses that viral infections naturally provoke. However, pre-existing immunity against a vector can limit its effectiveness, and repeated dosing may be affected by antibodies or T cells directed against the delivery virus rather than the tumor antigen. Each platform also differs in stability, manufacturing requirements, dose, timing and safety profile. Comparing these technologies requires common reporting standards that separate the effect of the antigen from the effect of the delivery system.

The third major issue is manufacturing speed and reliability. A personalized vaccine is produced for a specific patient, often after surgery or biopsy has provided sufficient tumor material. The treatment team must complete sequencing, antigen selection, design, production and quality control within a clinically useful window. Delays can be particularly consequential for patients with rapidly progressing disease. Manufacturing must also confirm the identity, purity, concentration and structural integrity of the vaccine product. For RNA-based approaches, for example, important variables include RNA sequence accuracy, chemical modification, encapsulation efficiency and resistance to degradation. For viral vectors, investigators must monitor infectivity, genetic stability, replication competence and the absence of unwanted contaminants. Harmonized release criteria could help ensure that a product made at one facility is equivalent in quality to a product made elsewhere.

The fourth challenge involves measuring immune responses in a consistent way. A vaccine may expand neoantigen-specific CD8-positive cytotoxic T cells, CD4-positive helper T cells, or both, but the presence of these cells in blood does not necessarily demonstrate that they can enter a tumor and destroy malignant cells. Researchers use tools including peptide–HLA multimer staining, interferon-gamma release assays, intracellular cytokine analysis, T-cell receptor sequencing and single-cell profiling. These methods provide different types of information and can produce different estimates of response magnitude. A patient may show a detectable immune response under one assay but not another, depending on the peptide concentration, cell culture conditions and definition of positivity. Shared reference materials, controls and reporting rules would make it easier to determine whether an immune response is robust, durable and clinically relevant.

Immune monitoring must also be connected to the biology of the tumor. Cancer cells can lose the targeted mutation, reduce antigen production or disrupt antigen presentation through defects in HLA molecules and associated processing machinery. The tumor microenvironment may further suppress immunity through regulatory T cells, myeloid-derived suppressor cells, inhibitory cytokines and checkpoint molecules such as PD-L1. For this reason, personalized vaccines are increasingly considered as components of combination treatment rather than stand-alone products. Checkpoint inhibitors may release brakes on activated T cells, while radiation or chemotherapy can alter antigen release and tumor visibility. Viral-vector vaccines may provide additional inflammatory signals that help recruit immune cells, but their effects must be distinguished from those of the accompanying therapies. Harmonized clinical designs are needed to identify which combinations truly improve outcomes.

The fifth area concerns clinical trials and regulation. Personalized vaccine studies often enroll relatively small numbers of patients because every treatment is individually designed, making conventional trial structures difficult to apply. Differences in cancer type, disease stage, prior therapy, tumor mutation burden and vaccine composition can complicate comparisons between studies. Investigators therefore need agreed definitions for endpoints, including feasibility, manufacturing success, immune response, recurrence-free survival and overall survival. Regulatory agencies must evaluate not only the final vaccine but also the computational pipeline used to select its targets and the manufacturing process used to produce it. A transparent framework could allow a platform to be validated once while individual vaccine sequences are assessed under controlled procedures, reducing duplication without compromising safety.

The authors’ message arrives as personalized oncology expands into a field where speed, precision and reproducibility must advance together. A vaccine that is biologically sophisticated but produced too slowly may not benefit a patient; a vaccine that generates immune cells but targets an irrelevant or poorly presented antigen may fail for biological reasons; and a promising clinical result that cannot be compared with other studies may delay progress across the field. Harmonization does not mean forcing every research group to use one technology. Instead, it means defining common standards for data quality, antigen selection, manufacturing, immune monitoring and clinical evaluation while preserving room for innovation. By organizing the challenges around five interconnected priorities, the review frames personalized cancer vaccines as an emerging medical system that requires coordination across genomics, immunology, bioinformatics, engineering and regulation. The prospect is not simply a faster way to make individualized vaccines, but a more reliable path toward determining which patients are most likely to benefit and why.

Subject of Research: Harmonization and standardization of personalized cancer vaccines for cancer immunotherapy, including neoantigen identification, vaccine platforms, manufacturing, immune monitoring and clinical evaluation.

Article Title: Take Five: harmonization in personalized cancer vaccines for cancer immunotherapy

Article References:

Cho, S., Lee, J., Lee, YM. et al. Take Five: harmonization in personalized cancer vaccines for cancer immunotherapy. Exp Mol Med (2026). https://doi.org/10.1038/s12276-026-01807-y

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s12276-026-01807-y

Keywords: personalized cancer vaccines, cancer immunotherapy, neoantigens, tumor sequencing, HLA presentation, viral vectors, messenger RNA vaccines, immune monitoring, vaccine manufacturing, clinical trial harmonization