Patients with high-risk melanoma who received a personalized mRNA cancer vaccine alongside the checkpoint inhibitor pembrolizumab saw their risk of cancer recurrence cut by nearly half over five years compared with pembrolizumab alone. The results come from KEYNOTE-942, a randomized phase 2b trial that tested the individualized neoantigen therapy mRNA-4157 (V940) in combination with pembrolizumab against pembrolizumab monotherapy in patients whose melanoma had been surgically removed. The five-year follow-up data extend earlier findings and raise a pointed question: what is keeping the cancer at bay years after treatment ended?
Why five-year melanoma vaccine data change the treatment calculus
Checkpoint inhibitors like pembrolizumab are already standard care after melanoma surgery. They work by releasing a brake on the immune system so T cells can attack residual cancer cells. Adding a personalized vaccine on top of that standard raises the bar in a measurable way, and the persistence of the benefit at five years is what separates this result from incremental progress. If the combination only delayed recurrence by a few months, clinicians would weigh the added cost and side effects differently. A durable reduction lasting years shifts that equation toward broader adoption.
The trial, registered on ClinicalTrials.gov as NCT03897881, enrolled participants with resected high-risk melanoma, meaning their tumors had been removed but the odds of the cancer returning were substantial. By design, the study compared two arms: one receiving mRNA-4157 plus pembrolizumab, the other receiving pembrolizumab alone. That structure isolates the vaccine’s contribution, because both groups got the same checkpoint inhibitor backbone.
One hypothesis worth examining is whether the sustained benefit at five years reflects vaccine-induced T-cell persistence rather than simply deeper initial tumor clearance from pembrolizumab. Pembrolizumab’s effect typically peaks during and shortly after treatment. If the combination arm’s advantage were driven solely by better early clearance, the recurrence curves would be expected to converge over time as the drug’s direct effect faded. A widening or stable gap at five years instead points toward an immune memory component, the kind of long-lived T-cell response that a personalized vaccine encoding a patient’s own tumor neoantigens is designed to generate. The available data are consistent with that interpretation, though confirming it would require detailed immunological analyses tracking neoantigen-specific T cells over time.
KEYNOTE-942 trial design and published findings
The trial’s formal title is “An Efficacy Study of Adjuvant Treatment With the Personalized Cancer Vaccine mRNA-4157 and Pembrolizumab in Participants With High-Risk Melanoma,” and it carries the study designation KEYNOTE-942. It is a randomized, open-label study, meaning patients and doctors knew which treatment arm each participant was assigned to, but assignment itself was random to reduce bias.
The initial peer-reviewed results were published in The Lancet as a phase 2b analysis reporting data at approximately two years of follow-up. That publication established the trial’s methods, endpoint definitions, and early efficacy signals. The primary endpoint was recurrence-free survival, the length of time after surgery before melanoma came back or the patient died. At the two-year mark, the combination arm already showed a meaningful separation from the pembrolizumab-only arm, with a substantial reduction in the risk of recurrence or death.
The five-year update extends that evidence window considerably. In oncology, two-year data can reflect short-term drug activity, but five-year data begin to answer whether a treatment changes the natural history of the disease. For melanoma patients who have had their tumors removed, the first two to three years carry the highest recurrence risk. Sustained separation in recurrence-free survival beyond that window suggests the treatment is not merely delaying inevitable relapse but preventing it in a meaningful fraction of patients. That pattern is what makes the emerging five-year curves so important for clinicians weighing adjuvant options.
The vaccine itself, mRNA-4157, is manufactured individually for each patient. Researchers sequence the patient’s tumor, identify up to 34 neoantigens-mutations unique to that tumor-and encode them into a single mRNA construct. Once injected, the patient’s cells produce those neoantigen fragments, training the immune system to recognize and attack any remaining cancer cells carrying those mutations. This approach differs from off-the-shelf cancer drugs because no two patients receive the same vaccine, and the theoretical ceiling on benefit is tied to how effectively those selected mutations capture the biology of each person’s tumor.
From a safety standpoint, the Lancet report described a profile that was broadly consistent with pembrolizumab alone, with most additional side effects related to transient, flu-like symptoms typical of mRNA-based therapies. That tolerability picture matters when thinking about long-term outcomes: if patients can complete the vaccine course without prohibitive toxicity, the immune system has a chance to build and refine the memory responses that may underpin the sustained recurrence-free survival advantage seen at later follow-up.
Gaps in the evidence and what melanoma patients should watch
Several questions remain open despite the encouraging trajectory. The five-year Kaplan-Meier curves and updated hazard ratios have not yet appeared in the ClinicalTrials.gov registry record, which means the precise magnitude of the recurrence reduction at this later time point awaits formal peer-reviewed publication. The headline claim of “nearly half” reflects the direction and approximate scale of the benefit reported from the trial’s ongoing follow-up, but exact numbers will matter for regulatory decisions and clinical guidelines.
Subgroup analyses also remain limited in the public record. Tumor mutational burden and PD-L1 expression status are two biological markers that could predict which patients benefit most from the vaccine. Higher mutational burden generally means more potential neoantigens for the vaccine to target, so patients with lower mutational loads might see a smaller benefit. Conversely, tumors with strong pre-existing immune infiltration could be more easily tipped into durable control by adding a personalized vaccine to pembrolizumab. Without detailed breakdowns of outcomes by these and other biomarkers, clinicians must extrapolate cautiously from the aggregate benefit to individual patients sitting in front of them.
Another gap involves overall survival. Recurrence-free survival is a meaningful endpoint, especially in adjuvant settings where preventing relapse often correlates with living longer. Still, regulators and guideline committees frequently look for overall survival data to confirm that fewer recurrences translate into fewer deaths from melanoma. With five-year follow-up now available, the relationship between recurrence prevention and mortality reduction will come under closer scrutiny, even if the trial was not powered primarily for overall survival.
Cost and logistics pose additional challenges. Manufacturing a bespoke mRNA vaccine for each patient requires rapid tumor sequencing, bioinformatic analysis, and on-demand production. That complexity raises questions about access outside of major cancer centers and about how quickly such a therapy could be delivered after surgery, when the window for optimal immune priming may be narrow. Health systems and payers will have to weigh the up-front investment against the potential savings from avoiding expensive treatments for metastatic relapse later on.
For patients and clinicians, several near-term developments will be important to watch. First is the full publication of the five-year data, including detailed curves, hazard ratios, and subgroup outcomes. Those details will clarify whether the benefit remains consistent across risk categories or is concentrated in particular molecular profiles. Second is the design of larger, confirmatory phase 3 trials that could support regulatory approval and broader adoption if they replicate the KEYNOTE-942 findings. Third is how quickly real-world infrastructure for personalized mRNA vaccines can be built so that promising trial results translate into routine care rather than remaining confined to a few specialized centers.
Even with those uncertainties, the emerging five-year picture from KEYNOTE-942 suggests that combining a checkpoint inhibitor with a personalized mRNA vaccine can do more than nudge recurrence curves-it may reshape the long-term outlook for a subset of patients with high-risk melanoma. For individuals facing surgery and anxious about what comes next, that possibility alone is enough to justify close attention as the final data and regulatory decisions come into view.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
