Personalized Peptide Vaccines Activated T Cells in Pancreatic Cancer. They Did Not Prove Longer Survival.

A personalized vaccine made from each patient's own pancreatic tumor mutations activated targeted T cells in two early trials. That is a real biological result. It is not evidence that the vaccines prolonged life.

Researchers at Washington University and Johns Hopkins tested two ways to show the immune system what a pancreatic tumor looks like. One vaccine used synthetic long peptides. The other used DNA encoding selected tumor mutations. Twenty people received at least one dose after surgery and chemotherapy, and all 16 who completed treatment and had evaluable immune samples responded to at least one vaccine target.

The survival numbers are less convincing. Median overall survival was 4.4 years among vaccine recipients and 3.5 years in a matched institutional comparison group. The difference was not statistically significant. Disease-free survival was also no better in the vaccine group.

This is the tension worth holding onto: the investigators proved that a custom peptide or DNA vaccine can wake up mutation-specific T cells in a cancer that usually resists immunotherapy. They did not prove that those T cells changed the course of the disease.

What the two phase 1 trials tested

Pancreatic ductal adenocarcinoma, or PDAC, is often described as an immunologically cold tumor. Its dense tissue, sparse effective immune infiltration, and suppressive local environment help explain why checkpoint inhibitors that work in melanoma and some lung cancers have had limited impact in most PDAC.

A personalized cancer vaccine tries to start earlier in the immune process. Researchers sequence a patient's tumor and normal tissue, identify mutations found only in the cancer, predict which altered protein fragments might be visible to T cells, and build a vaccine around those fragments. The targets are called neoantigens.

The new Science Advances paper combines results from two closely related phase 1 studies. Both enrolled people whose pancreatic tumors had been surgically removed and who had completed adjuvant chemotherapy. Fourteen entered the synthetic long peptide trial, NCT03956056. Eighteen entered the DNA trial, NCT03122106. Recurrence, withdrawal, and problems with tumor RNA meant that only nine peptide-vaccine participants and 11 DNA-vaccine participants received at least one dose.

The peptide group received seven injections. Each personalized product contained pools of long peptides representing selected mutations, mixed with the immune stimulant poly-ICLC. The DNA group received six monthly doses delivered with an electroporation device, which uses brief electrical pulses to help cells take up DNA.

This was not a head-to-head test designed to crown a winning platform. The primary endpoint was safety. Immune response was secondary. Disease-free and overall survival were exploratory.

The immune result was strong, with an important laboratory caveat

The team used whole-exome and RNA sequencing, then pVACtools software, to rank possible neoantigens. A typical peptide vaccine included 10 targets. A typical DNA vaccine included 15.

Among the 16 people who completed all planned doses and could be assessed, every participant developed an interferon gamma response to at least one targeted neoantigen. Two people who recurred before finishing treatment also had detectable responses after receiving most of their doses.

The breadth differed between platforms. The median share of selected targets classed as immunogenic was 41% in the synthetic long peptide group and 12% in the DNA group. That comparison is interesting, but the study was too small and was not designed to establish that peptides are superior to DNA.

There is another detail that keeps the 100% response rate in perspective. The main analysis tested blood cells after 12 days of culture with the relevant peptides. This can expand rare responsive cells until they become easier to detect. The authors say this method may partly explain why they found a response in every evaluable participant. A response detected after laboratory expansion does not tell us how many active cells were circulating in the body or whether enough reached the tumor.

The researchers went beyond a single assay. They found both CD4 and CD8 T-cell activity, measured cytokines and granzyme B, tracked expanded T-cell receptor clones, and functionally confirmed that selected receptors recognized their intended neoantigen. That makes the immunology more persuasive. It still leaves the clinical question open.

Safety looked manageable in this small cohort

No participant had a treatment-related adverse event of grade 3 or higher. Everyone who received a vaccine reported at least one grade 1 event, while one peptide recipient and two DNA recipients had grade 2 events.

For the peptide vaccine, common events included limb edema, fatigue, headache, fever, injection-site reactions, and muscle pain. Injection-site pain and muscle pain were the most common events in the DNA group. The electroporation delivery method matters here because some discomfort may reflect the procedure as well as the vaccine construct.

These findings support short-term tolerability in 20 treated people. They cannot define rare toxicity, safety in broader pancreatic-cancer populations, or the effects of combining a vaccine with checkpoint blockade. Phase 1 safety data are a starting point, not a finished safety profile.

The survival signal does not clear the bar

The investigators compared vaccine recipients with patients treated at the same institution during the same period. They used propensity matching for clinical and pathological factors and required comparison patients to remain disease-free for 24 weeks after surgery, mirroring the time vaccine recipients needed to reach treatment.

Median overall survival was 4.4 years with vaccination and 3.5 years in the matched cohort. The P value was 0.23, so the study did not show a statistically reliable overall-survival difference. Median disease-free survival was 1.7 years with vaccination and 2.2 years in controls, with a P value of 0.77.

An exploratory analysis found longer survival after recurrence, 2.0 years versus 0.7 years, with a P value of 0.01. That is the most eye-catching clinical number in the paper. It is also vulnerable to chance, treatment differences after recurrence, and residual selection bias. It was not the primary endpoint of a randomized trial.

The magnitude of the measured immune response did not correlate with survival. That disconnect matters. A vaccine can produce assay-positive T cells without producing enough tumor control to improve outcomes.

The authors are direct about the limit: neither trial was powered to detect a survival benefit. Their clinical comparisons are hypothesis-generating.

Manufacturing is part of the treatment

Personalization sounds elegant until the clock enters the room. Tumor tissue had to be sequenced, candidate targets ranked, and an individual product manufactured and tested after surgery and chemotherapy.

Only 13 of the 20 administered vaccines were produced within six months. Median manufacturing time was 119 days for the peptide platform and 184 days for DNA. COVID-era disruptions affected the studies, so those times should not be treated as fixed limits of current technology. They still expose a basic problem in aggressive cancer: some patients recur while their vaccine is being made.

That happened here. Six enrolled patients withdrew before dosing because their disease returned, one in the peptide study and five in the DNA study. A therapy that arrives too late cannot help, even if its immune design is sound.

Faster prediction, manufacturing, and release testing may improve feasibility. Shared vaccines against common driver mutations could shorten the process for some patients. A separate 2026 phase 1 trial, for example, used pooled synthetic long peptides targeting six mutant KRAS variants with nivolumab and ipilimumab. Eleven of 12 participants developed increased average T-cell responses to the KRAS antigens. That approach is less individualized, but it addresses a mutation common in pancreatic cancer.

How this fits with the mRNA vaccine story

The result joins a small but coherent body of human evidence. A 2023 Nature study of a personalized mRNA neoantigen vaccine in resected pancreatic cancer reported vaccine-induced T cells in half of treated patients. Follow-up published in 2025 found that vaccine-expanded CD8 T-cell clones could persist for years.

The new study matters because it reaches a similar biological destination with two different platforms. Peptides deliver the target fragments directly. DNA asks cells to produce them. mRNA uses another temporary set of instructions. Across platforms, researchers can repeatedly induce mutation-specific immunity after pancreatic-cancer surgery.

What remains unsettled is whether these responses prevent recurrence or extend survival. The studies use different vaccines, immune assays, combinations, and treatment schedules. Small early cohorts can identify feasibility and biological activity. They cannot supply the randomized comparison needed for efficacy.

The Oria take

The cleanest conclusion is that personalized vaccination worked as immunology. It has not yet worked as proven pancreatic-cancer therapy.

The synthetic long peptide result is especially relevant to peptide research because it shows a complete clinical workflow: sequence a tumor, select private mutations, manufacture several long peptides, administer them with an adjuvant, and recover functional T cells that recognize the chosen targets. This is a far more demanding standard than showing that a peptide binds a receptor in a dish.

The next test needs a randomized control group, a manufacturing window short enough for an aggressive disease, and clinical endpoints set before the data are examined. It should also measure whether vaccine-induced T cells enter residual tumors, not only whether they expand in cultured blood samples. The team has already registered a window study, NCT05111353, that gives a peptide vaccine after preoperative chemotherapy but before surgery, creating an opportunity to examine the tumor directly.

Until efficacy data arrive, patients should not read this study as a reason to seek peptide vaccines outside an oncology trial. These were individually manufactured investigational products given after standard surgery and chemotherapy under institutional protocols. They are unrelated to generic research peptides sold online.

Sources

1. Perkins FZ, et al. "Synthetic long peptide and DNA personalized cancer vaccines induce robust neoantigen-specific T cell responses in pancreatic cancer." Science Advances, 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13544235/

2. ClinicalTrials.gov. Synthetic long peptide personalized cancer vaccine trial, NCT03956056. https://clinicaltrials.gov/study/NCT03956056

3. ClinicalTrials.gov. DNA personalized cancer vaccine trial, NCT03122106. https://clinicaltrials.gov/study/NCT03122106

4. Rojas LA, et al. "Personalized RNA neoantigen vaccines stimulate T cells in pancreatic cancer." Nature, 2023. https://pubmed.ncbi.nlm.nih.gov/37165196/

5. Sethna Z, et al. "RNA neoantigen vaccines prime long-lived CD8+ T cells in pancreatic cancer." Nature, 2025. https://pubmed.ncbi.nlm.nih.gov/39972124/

6. Chen Z, et al. "A Neoantigen-Based Peptide Vaccine for Patients With Advanced Pancreatic Cancer Refractory to Standard Treatment." Frontiers in Immunology, 2021. https://pubmed.ncbi.nlm.nih.gov/34484187/

7. "Mutant KRAS vaccine with dual checkpoint blockade in resected pancreatic cancer: a phase I trial." Nature Communications, 2026. https://pubmed.ncbi.nlm.nih.gov/41667470/

Evidence grade: C. Two phase 1 trials provide direct human evidence that individualized synthetic long peptide and DNA vaccines can be manufactured, administered with acceptable short-term tolerability, and induce mutation-specific T-cell responses. The evidence does not establish clinical benefit because only 20 people were treated, there was no randomized vaccine control, survival was exploratory, and the overall-survival difference was not statistically significant.

Medical disclaimer: This article is for educational purposes and is not medical advice. Personalized neoantigen vaccines for pancreatic cancer remain investigational. Do not buy unapproved peptide products, change cancer treatment, or delay oncology care based on this report. Pancreatic cancer requires individualized treatment from a qualified multidisciplinary oncology team. Oria exists to help readers understand emerging peptide research, not to diagnose disease or replace medical care.