A personalized ovarian cancer vaccine faces tumor escape

A personalized ovarian cancer vaccine is designed around the mutations found in an individual patient's tumor. In a phase 1 clinical trial, researchers found signs of biological activity in six of ten vaccinated women. But the study also exposed a challenge that could determine whether such treatments ultimately succeed: cancer cells can retain their target mutations while escaping immune recognition.

Published in Nature Medicine on October 14, 2026, the trial evaluated NeoVax, an experimental therapeutic vaccine developed specifically for patients with high-grade epithelial ovarian cancer.

Unlike preventive vaccines, therapeutic cancer vaccines are given after a cancer diagnosis. Their purpose is to stimulate an immune response against malignant cells, potentially including those that remain after conventional treatment.

A vaccine designed by the tumor's own mutations

Two patients with the same type of cancer may have very different collections of mutations in their tumors.

Some mutations generate altered protein fragments known as neoantigens. Because these fragments differ from normal human proteins, they may provide the immune system with recognizable targets.

However, not every mutation produces a useful immune target. Whether a neoantigen can be recognized depends on several factors, including how it is processed and presented to T cells.

The NeoVax team began by analyzing tumor samples to identify patient-specific mutations. Computational methods helped select candidate neoantigens likely to be recognized by each patient's immune system.

A central consideration was the patient's HLA profile.

HLA molecules display fragments of proteins on cell surfaces, allowing T cells to inspect what is happening inside those cells. Differences in HLA molecules affect which antigen fragments can be presented effectively.

Researchers used this information to manufacture an individualized vaccine containing pools of synthetic long peptides corresponding to selected tumor neoantigens.

The formulation also included poly-ICLC, an immune-stimulating adjuvant intended to enhance the response to vaccination.

Patients subsequently received nivolumab, an antibody that blocks the PD-1 immune checkpoint.

PD-1 is part of the system that normally limits T-cell activity. Tumors can exploit inhibitory pathways to suppress antitumor immune responses. Blocking PD-1 can help restore T-cell function in appropriate circumstances.

The experimental strategy therefore combined two interventions: directing immunity toward patient-specific tumor targets and later reducing an inhibitory signal that might limit the resulting immune response.

What happened in the ten vaccinated patients?

The trial enrolled patients with high-grade epithelial ovarian cancer treated with standard platinum-based combination chemotherapy.

Ten women ultimately received NeoVax. One had newly diagnosed disease, while nine had experienced their first recurrence of platinum-sensitive ovarian cancer.

The primary goals of this phase 1 study were feasibility and safety.

Feasibility is particularly important for individualized vaccines. Each treatment must be designed and manufactured using information from the patient's tumor, and the entire process must fit within a clinically meaningful treatment window.

Among evaluable patients with recurrent disease, nine of twelve received vaccination, representing 75%. In the newly diagnosed cohort, only one of nine evaluable patients received the vaccine.

This difference illustrates an important practical limitation. Producing and delivering an individualized vaccine depends not only on technical capabilities but also on the patient's treatment course and the timing of tissue collection, chemotherapy, and vaccination.

The investigators reported no unexpected safety signals, and no patient discontinued treatment because of toxicity. Adverse events did occur, however, and a study of this size cannot rule out uncommon risks.

The most closely watched findings involved immune activity.

Biomarker analyses suggested biological activity of the vaccine in six of the ten vaccinated patients. These findings were associated with broader vaccine-specific and tumor-specific immune responses in blood and tumor samples.

The distinction between immune activity and clinical benefit is essential.

Six of ten does not mean that six patients were cured, nor does it represent a confirmed objective tumor response rate.

The biomarkers provided evidence consistent with biological effects, but the trial was not designed to demonstrate that NeoVax prolonged survival or prevented recurrence.

There was no randomized control group, and patients received both standard chemotherapy and subsequent PD-1 blockade. Consequently, the clinical contribution of vaccination cannot be isolated from the other treatments.

The immunological observations are therefore informative, but they are not proof of therapeutic efficacy.

When the target remains but the immune system loses sight of it

One particularly revealing question was whether the mutations chosen for vaccine design would still be relevant by the time patients received their vaccines.

The median interval between the diagnostic tumor biopsy and vaccination was 6.5 months. During that period, patients received chemotherapy, which could alter the populations of cancer cells within their tumors.

Despite this delay, more than 70% of selected mutation targets remained detectable through cell-free DNA analysis.

That result suggests many of the mutations identified during vaccine design persisted over a clinically relevant interval.

Yet persistence of a mutation is not sufficient for successful immune recognition.

A cancer cell carrying a target mutation must also process the corresponding protein and display its antigenic fragments through HLA molecules. Without effective antigen presentation, even properly activated T cells may fail to recognize the cell.

The researchers identified several mechanisms by which resistant tumor clones could disrupt this process.

Some tumors had already lost parts of their HLA system before vaccination, potentially preventing the presentation of several vaccine-targeted antigens.

In another case, a tumor sampled after vaccination carried a structural alteration affecting JAK1, a gene involved in interferon-gamma signaling.

Interferon-gamma helps regulate antigen presentation and other components of antitumor immunity. Disrupting JAK1 signaling can therefore interfere with a cancer cell's ability to respond to immune signals and display antigens.

Laboratory experiments strengthened this interpretation. Restoring JAK1 function in the affected tumor model restored relevant signaling responses and improved antigen presentation and T-cell recognition.

The findings are consistent with immune selection: as immune pressure increases, tumor populations that are less visible to T cells may gain a survival advantage.

They do not establish that vaccination directly caused every observed resistance mechanism. Some defects were present before treatment, while others were detected afterward.

This distinction changes how personalized cancer vaccines should be evaluated.

Selecting mutations that distinguish malignant cells from healthy tissue is only the first step. Vaccine developers must also consider whether the tumor retains the molecular machinery necessary to present those mutations to the immune system.

The NeoVax study demonstrates that individualized peptide vaccines can be manufactured and administered in selected ovarian cancer patients and can generate measurable immune responses. Whether the approach improves clinical outcomes remains unresolved.

Its most instructive finding may be the difference between carrying an immune target and making that target visible. A tumor can retain the mutation used to design its vaccine while losing the ability to display the corresponding antigen to T cells.