The randomized phase 2 AMPLIFY-7P trial evaluated a lymph node–targeted amphiphile vaccine, camupepimut (ELI-002 7P), in patients with resected, stage I–III KRAS-mutant pancreatic ductal adenocarcinoma who had no recurrence within 6 months of standard therapy. Investigators randomized patients 2:1 to receive camupepimut (n = 96) or standard-of-care/observation (n = 48); nodal status was a stratification factor. The trial tested disease-free survival (DFS) as the primary endpoint.
The primary DFS endpoint was not reached. Median DFS in the vaccine arm was 23.8 months versus 21.2 months in observation (HR 0.85, P = 0.5329). Those numbers indicate no statistically significant separation by the prespecified primary analysis. However, short-term outcomes showed a numerical advantage at 6 months in favor of camupepimut: DFS 75.7% versus 61.7%, a 14.0% absolute difference that corresponded with the dosing period (HR 0.55, P = 0.0557). The study team reported this early divergence alongside additional prespecified and exploratory analyses that were considered informative for interpretation.
Investigators conducted prespecified subgroup analyses and a post hoc multivariate Cox regression to explore prognostic and predictive factors. The multivariate model adjusted for baseline clinical variables and indicated an improved DFS associated with camupepimut (adjusted HR 0.47, P = 0.0086). The same model identified several adverse prognostic markers: stage III disease (adjusted HR 2.29, P = 0.0127), elevated baseline CA19-9 (adjusted HR 3.16, P < 0.0001) and baseline circulating tumor DNA (ctDNA) positivity (adjusted HR 5.99, P < 0.0001).
In prespecified subgroup testing, patients who were ctDNA negative derived a DFS signal with camupepimut (HR 0.51, P = 0.0280). A similar signal was observed among patients without prior radiation therapy (HR 0.59, P = 0.0221). The study report emphasized these subgroup results as potentially hypothesis-generating for patient selection and stratification in later trials.
The trial produced clear immunogenicity data. Ex vivo mKRAS-specific T cell responses were detected in 98.9% (89/90) of assessed camupepimut-treated patients. Investigators also observed antigen spreading in a high proportion of those tested (86.7%, 13/15), indicating expansion of immunity beyond the targeted epitopes in available samples.
A predefined immunologic correlate further linked magnitude of response with clinical outcome. Patients who mounted an mKRAS-specific T cell response exceeding 9.17-fold over baseline (observed in 82.2%, 74/90) had markedly improved DFS: the median DFS was not reached for the high-responder group versus 6.08 months for lower responders, and the associated HR was 0.22 (P < 0.0001). The study team reported this correlation as a strong translational signal connecting vaccine-induced immunity to disease control.
Camupepimut had a safety profile similar to observation. Grade ≥3 adverse events occurred in 10.4% (10/96) of vaccine recipients versus 14.6% (7/48) in the observation group. Serious adverse events were infrequent (6.3% in the camupepimut arm versus 8.3% in observation) and reflected disease progression or surgical complications rather than a clear vaccine-specific toxicity signature. Investigators noted tolerability consistent with prior phase 1 data; the source did not report additional unexpected immune-related toxicities.
The combination of early DFS separation at 6 months, multivariate-adjusted DFS benefit for the vaccine, and robust mKRAS-specific immune responses informed the authors' interpretation that the data warrant a refined confirmatory approach. Investigators reported that these efficacy and immunogenicity signals will inform the design of a confirmatory phase 3 trial. The source did not provide specific phase 3 design parameters, dosing changes, or enrollment criteria for that confirmatory study.
The primary endpoint was not met by the prespecified analysis, and the reported DFS signals were derived from prespecified subgroup tests, post hoc adjustment and exploratory immunologic correlates; these analytic layers introduce risk of bias and require prospective confirmation. The study report highlighted strong associations between baseline ctDNA and outcomes, yet did not include prospective ctDNA-directed randomization. The sample sizes for some biomarker subgroups and antigen-spreading assessments were small, and the source did not detail long-term overall survival outcomes beyond the reported analyses. Finally, the authors indicated that the observed signals will guide a phase 3 strategy but did not report the confirmatory trial's planned endpoints, patient selection criteria or timeline—key details that remain unanswered and must be resolved in future protocols.