Personalized mRNA Vaccine Keeps 87 Percent of Pancreatic Cancer Responders Alive at Six Years, AACR Data Show
Six-year follow-up from a 16-patient Phase 1 trial of autogene cevumeran shows unprecedented survival rates in one of oncology's deadliest cancers, now entering a global Phase 2 trial.
Overview
New six-year follow-up data from a Phase 1 trial of autogene cevumeran — a personalized mRNA cancer vaccine developed by BioNTech and Genentech — show that seven of eight patients whose immune systems responded to the vaccine are still alive four to six years after pancreatic cancer surgery, researchers reported at the American Association for Cancer Research (AACR) Annual Meeting 2026. For a disease where roughly 12 percent of patients survive five years after diagnosis, the results represent a striking departure from historical outcomes.
The findings build on a trajectory of increasingly durable evidence, following a 2023 Nature paper on 18-month outcomes and a March 2025 Nature study reporting three-year data that used a novel computational tracking method to show vaccine-induced T cells were estimated to have an average lifespan of 7.7 years. The latest data, presented by Dr. Vinod Balachandran of Memorial Sloan Kettering Cancer Center (MSK), extend that record further and give researchers the clearest signal yet that the vaccine can produce sustained immune control in what has historically been an immune-resistant tumor.
What We Know
Autogene cevumeran (also known as BNT122 or RO7198457) is a personalized mRNA vaccine that encodes up to 20 neoantigens — tumor-specific proteins arising from mutations unique to each patient’s cancer. Unlike off-the-shelf drugs, the vaccine is custom-synthesized from the patient’s own tumor DNA, a process requiring weeks of sequencing and manufacturing after surgery.
In the Phase 1 trial at MSK, 16 patients with resectable pancreatic ductal adenocarcinoma (PDAC) received the vaccine sequentially after surgery alongside atezolizumab, Genentech’s PD-L1 checkpoint inhibitor, and a modified FOLFIRINOX chemotherapy regimen. Eight of the 16 patients generated measurable vaccine-induced T-cell responses, as reported by BioPharma Dive, while the other eight did not.
The divergence in outcomes between those two groups has grown more pronounced with each follow-up:
- Responders (8 patients): Seven of eight (87.5 percent) remained alive at four to six years post-surgery. Median recurrence-free survival had not been reached as of the latest data cut.
- Non-responders (8 patients): Two of eight (25 percent) survived. The median recurrence-free survival for this group was 13.4 months.
The 2025 Nature paper from the same research group, which covered median follow-up of 3.2 years, characterized the immunological basis for this durable protection. Using a computational method called PhenoTrack, researchers showed that approximately 86 percent of vaccine-primed CD8-positive T cell clones persisted into the memory phase at substantial frequencies. The study also found that 98 percent of those clones were newly generated after vaccination — not pre-existing cells reactivated by the treatment — confirming that the vaccine was teaching the immune system to recognize the tumor rather than amplifying a pre-existing response.
The average estimated lifespan of vaccine-induced CD8-positive T cell clones was calculated at 7.7 years, with a range of 1.5 years to roughly 100 years, according to the Nature study. Critically, these cells retained functional capacity — producing cytokines and demonstrating degranulation upon neoantigen rechallenge — even following exposure to mFOLFIRINOX chemotherapy, a regimen known to suppress immune function.
Dr. Balachandran said the results show the approach “continues to support exploring autogene cevumeran as an approach to treat pancreatic cancer,” and that the immunotherapy approach “has the potential to make a difference for one of the deadliest cancers,” according to the MSK press release.
What We Don’t Know
The trial’s limitations are substantial. With 16 patients treated at a single center in a non-randomized design, the results cannot establish causation. The researchers themselves have noted that they cannot yet confirm the vaccine directly caused the delay in recurrence: selection effects, favorable tumor biology in vaccine-responding patients, or the combination regimen rather than the vaccine alone may partly account for the outcomes.
Equally significant is the 50 percent non-response rate. Half of all patients who received the vaccine generated no measurable T-cell response, and their outcomes were broadly similar to historical controls. The biological determinants of response versus non-response remain undefined, which means clinicians currently lack tools to predict which patients will benefit — a critical gap as the therapy moves toward broader testing.
Manufacturing complexity remains an unresolved challenge. The vaccine must be individualized for each patient from tumor biopsy material, requiring successful sequencing, neoantigen selection, mRNA synthesis, and delivery within a narrow post-surgical window before adjuvant chemotherapy is typically started. Whether these logistics can be reliably replicated across dozens of trial sites is one of the central questions the Phase 2 trial will answer.
The Phase 2 Trial and Broader Context
The pancreatic cancer results have advanced the program to IMCODE003 (NCT05968326), a global randomized Phase 2 trial sponsored by Genentech in collaboration with BioNTech. The trial is enrolling approximately 260 patients with resected PDAC and comparing autogene cevumeran plus atezolizumab and mFOLFIRINOX against standard adjuvant mFOLFIRINOX alone. Sites at MSK and institutions worldwide are now enrolling, and the trial is expected to provide the first randomized evidence — or its absence — that the Phase 1 signal is reproducible at scale.
The AACR 2026 presentation comes at an inflection point for personalized mRNA cancer vaccines as a class. As CNN reported from the meeting, researchers see encouraging momentum after a year of political turbulence around mRNA technology in the United States, including federal funding uncertainty. The pancreatic cancer data arrive alongside prior Machine Herald coverage of similarly durable mRNA vaccine results: a five-year BioNTech breast cancer vaccine trial published in April 2026, and a five-year Moderna-Merck melanoma vaccine study from February 2026, both showing that personalized neoantigen immunization can produce lasting T-cell immunity across different cancer types.
Pancreatic cancer has historically resisted immunotherapy approaches that have transformed outcomes in melanoma and lung cancer, largely because pancreatic tumors create a densely immunosuppressive microenvironment that exhausts T cells. The persistence of functional vaccine-induced T cells in the autogene cevumeran trial, despite that hostile biology, is what makes the data scientifically notable beyond the survival numbers alone.
For now, the results remain hypothesis-generating. Whether six-year outcomes in eight responding patients from a single Phase 1 trial will translate into a practice-changing randomized result is a question only IMCODE003 can answer.