Summary
In the randomized AMPLIFY-7P phase 2 trial, camupepimut did not significantly improve disease-free survival across all participants after pancreatic cancer surgery. Prespecified subgroup analyses and immune-response findings provide signals for testing in a larger trial.
A randomized phase 2 trial of camupepimut, a vaccine targeting cancer-linked KRAS mutations, did not meet its primary endpoint of improving disease-free survival across all participants with resected pancreatic cancer. In AMPLIFY-7P, median disease-free survival was 23.8 months with the vaccine and 21.2 months with observation; the difference was not statistically significant (hazard ratio 0.85, P = 0.5329).
The study, published in Nature Medicine on 23 October 2026, also reported favourable signals in prespecified patient subgroups and strong vaccine-induced T-cell responses. Those findings will inform a planned phase 3 trial, but the primary result remains the central test of the vaccine in this study.
What the vaccine and trial tested
Camupepimut, also called ELI-002 7P, is an off-the-shelf vaccine designed for pancreatic tumours carrying one or more of seven KRAS mutations: G12D, G12R, G12V, G12A, G12C, G12S and G13D. Its peptide antigens and immune-stimulating CpG adjuvant are modified with albumin-binding lipids. This amphiphile design is intended to carry vaccine components through lymphatic vessels into lymph nodes, where immune responses are organised and T cells can be activated against the targeted mutations.
AMPLIFY-7P enrolled 144 adults with stage I–III KRAS-mutated pancreatic ductal adenocarcinoma. Participants had undergone surgery and standard locoregional treatment, and had no radiographic evidence of recurrence within six months of completing that treatment. They were randomly assigned in a 2:1 ratio to camupepimut (96 people) or observation (48). The open-label trial took place at 24 centres in the United States. Its primary outcome was investigator-assessed disease-free survival—the time without cancer recurrence or death.
The vaccine course involved six priming doses over eight weeks, a two-month period without dosing, and four weekly booster doses. At the data cutoff of 20 April 2026, median follow-up was 19.4 months.
Primary result and signals for further study
The primary disease-free-survival endpoint was not met. During the six-month dosing period, the disease-free-survival rates were 75.7% with camupepimut and 61.7% with observation, a 14.0 percentage-point difference. An exploratory comparison of outcomes from randomisation to six months gave a hazard ratio of 0.55 (P = 0.0557). This early separation did not change the overall trial result.
Prespecified subgroup analyses found a lower rate of relapse or death among participants whose circulating tumour DNA (ctDNA) was undetectable at baseline: median disease-free survival was not reached with camupepimut and was 21.2 months with observation (hazard ratio 0.51, P = 0.0280). In participants who had not received prior radiation, median disease-free survival was 23.8 months with the vaccine and 10.4 months with observation (hazard ratio 0.59, P = 0.0221). These subgroup findings are signals for further testing, not a replacement for the result in the full trial population. The analyses were not adjusted for multiple comparisons, and prior radiation was more common in the observation group.
Among 90 vaccine-arm participants assessable for immune response, 89 developed a measured KRAS-specific T-cell response. In an exploratory analysis, responses above a 9.17-fold increase from baseline were associated with longer disease-free survival. Because this comparison grouped people by their response after treatment, it shows an association rather than establishing that a larger immune response caused the difference in outcomes.
Cam upepimut was generally well tolerated. Grade 3 or higher adverse events were reported in 10.4% of vaccine recipients and 14.6% of the observation group; serious adverse events occurred in 6.3% and 8.3%, respectively. Overall survival data were not mature at publication. The study authors say a phase 3 trial is planned, with design changes informed by these results.