Pancreatic cancer vaccine camupepimut misses its main goal in a phase 2 trial
The KRAS peptide vaccine from Elicio Therapeutics did not significantly extend disease free survival in AMPLIFY-7P, Nature Medicine reports. Strong T cell responders did better, a signal the company plans to use in designing a phase 3 trial.
What's new
The KRAS peptide vaccine camupepimut did not significantly extend disease free survival in the randomised phase 2 AMPLIFY-7P trial in resected pancreatic cancer, Nature Medicine reports.
Why it matters
It is the first randomised test of this lymph node targeted peptide vaccine, and strong immune responses alone did not translate into a clear clinical benefit for the whole group.
What's next
The authors say the immune and subgroup signals will shape a refined confirmatory phase 3 trial; the vaccine remains investigational.
A pancreatic cancer vaccine built from mutated KRAS peptides did not meet its primary goal in a randomised phase 2 trial, according to results published in Nature Medicine on October 23. Median disease free survival was 23.8 months with the vaccine, camupepimut, against 21.2 months with standard care or observation, a difference that was not statistically significant. The trial, AMPLIFY-7P, was sponsored and funded by Elicio Therapeutics. The authors point to secondary and exploratory signals, including strong immune responses, that they say will shape a phase 3 trial.
What the AMPLIFY 7P trial tested
Camupepimut, also known as ELI-002 7P, combines seven mutated KRAS peptide antigens with an immune stimulating adjuvant. Both parts carry an amphiphile tail that binds albumin and is designed to carry them to lymph nodes, where immune responses are trained. KRAS mutations drive most pancreatic cancers, so one off the shelf product could in principle cover many patients. The trial enrolled people with stage I to III pancreatic adenocarcinoma carrying a KRAS mutation who had finished standard treatment and had no recurrence within six months. Patients were randomly assigned two to one, 96 to the vaccine and 48 to standard of care or observation, with the groups balanced on lymph node status.
Treatment came as a priming series of injections under the skin, a break with no dosing, and then a booster series, the paper describes. The primary endpoint was disease free survival, meaning time without the cancer coming back or death. The trial is registered as NCT05726864 in the US trial registry. It followed a phase 1 study, AMPLIFY-201, in which immune responses specific to mutated KRAS were linked with longer disease free and overall survival. That earlier result was the reason for a randomised test. A randomised design can show whether any benefit comes from the vaccine rather than from differences between patients who respond well and those who do not.
The pancreatic cancer vaccine results
The main comparison failed. The hazard ratio for disease free survival was 0.85, with a P value of 0.53, so the trial could not show that the pancreatic cancer vaccine reduced recurrence. A prespecified look at six months showed 75.7% of vaccinated patients free of disease against 61.7% in the control group, a gap the authors linked to the period of active dosing, but this too fell short of significance. In a post hoc analysis adjusted for stage, the tumour marker CA19-9 and circulating tumour DNA, the vaccine was associated with a lower risk of recurrence, with an adjusted hazard ratio of 0.47. Post hoc results can suggest ideas but cannot confirm benefit.
The hazard ratio for disease free survival was 0.85, with a P value of 0.53, so the trial could not show that the pancreatic cancer vaccine reduced recurrence.
Prespecified subgroups pointed in the same direction. Patients with no detectable circulating tumour DNA at the start had a hazard ratio of 0.51, and those without prior radiation had 0.59. Subgroups in a trial of 144 people are small, and several were tested, so chance findings are likely among them. The immune data were stronger. Of 90 vaccinated patients tested, 89 mounted a T cell response against mutated KRAS. Those whose response rose more than 9.17 fold over baseline, 74 of 90, had a hazard ratio of 0.22 for recurrence against weaker responders, and their median disease free survival had not been reached, against 6.08 months for the rest, according to the Nature Medicine paper.
| Measure | Camupepimut | Control |
|---|---|---|
| Patients randomised | 96 | 48 |
| Median disease free survival | 23.8 months | 21.2 months |
| Disease free at 6 months | 75.7% | 61.7% |
| Grade 3 or higher adverse events | 10.4% | 14.6% |
| Serious adverse events | 6.3% | 8.3% |
Safety and what the responder signal means
Safety looked similar to observation. Grade 3 or higher adverse events occurred in 10.4% of vaccinated patients and 14.6% of controls, and serious adverse events in 6.3% and 8.3%, which the authors attribute mainly to cancer progression or surgical complications. For a treatment given after surgery to people who may already be cured, a light side effect profile matters. The responder finding needs care, though. Patients who mount strong immune responses may differ from weaker responders in ways that also affect survival, such as overall health or tumour biology. A comparison inside the vaccine arm cannot separate the effect of the vaccine from the effect of being a strong responder.
The paper also reports antigen spreading, in which immune responses widened to tumour targets not included in the vaccine, in 13 of 15 patients tested. That suggests the immune system went beyond the seven peptides it was shown, but the sample is very small and the finding is exploratory. In cancer vaccine research, immune readouts count as supporting evidence only, which is why randomised outcome data carry more weight than T cell counts. Readers who follow obesity drug news will know a similar rule from other fields: a headline subgroup is not a trial result, as our guide on reading trial headlines carefully explains.
What comes next for camupepimut
The authors wrote that the efficacy and immune signals will inform a refined strategy for a confirmatory phase 3 trial. They did not set out its design in the abstract, but the findings point to options such as enriching for patients without circulating tumour DNA, or for those who mount strong responses. Each choice carries trade offs. Selecting likely responders could make a benefit easier to detect but would narrow the group of patients who could use the vaccine. Elicio sponsored the study, designed it with academic authors and employed several authors, the paper discloses. No regulator has reviewed the vaccine for approval, and it remains investigational.
The result is a reminder that a promising mechanism does not guarantee a clinical win, as the two large trials in which semaglutide failed to slow Alzheimer's disease showed in another field. For pancreatic cancer, where recurrence after surgery is common and options are limited, even a modest delay in relapse would matter to patients. AMPLIFY-7P did not show that delay with statistical confidence. A larger phase 3 trial, with clearly prespecified patient groups and enough people to detect a modest effect, would be needed before camupepimut could be considered for routine use after surgery. Until then, patients with resected pancreatic cancer should discuss trial options with their oncology team.
Conclusion
The AMPLIFY-7P trial did not show that the camupepimut pancreatic cancer vaccine significantly extends disease free survival after standard treatment, with a hazard ratio of 0.85. The vaccine was well tolerated, nearly all patients mounted T cell responses, and strong responders relapsed less often. Those signals come from secondary, subgroup and post hoc analyses that cannot prove benefit. A confirmatory phase 3 trial, now being planned, will decide whether the approach has a future.
