Personalized Cancer Vaccine Combo Is Feasible in Platinum-Sensitive Ovarian Cancer

A personalized neoantigen-targeting vaccine (NeoVax) followed by nivolumab (Opdivo) was feasible and had no unexpected safety signals in patients with platinum-sensitive, high-grade epithelial ovarian cancer, according to results from the phase 1 NeoVax trial (NCT04024878) published in Nature Medicine.¹

What were the feasibility and safety results?

Between November 2020 and September 2023, 22 patients consented to prescreening, 12 consented to treatment, and 10 received the vaccine. Reasons for not proceeding included inadequate tissue from core needle biopsies (n = 4), progression on standard-of-care therapy (n = 2), and patient preference (n = 2). Nine of 10 vaccinated patients received all planned doses. Nine of 12 evaluable patients in Cohort B (first platinum-sensitive recurrence) were vaccinated (75%; 90% CI, 47%-93%), compared with 1 of 9 in Cohort A (newly diagnosed disease; 11%; 90% CI, 0.6%-43%).

The most common vaccine-associated adverse effects (AEs) were injection site reactions (70%), myalgia (50%), chills (50%), and neutropenia (50%). Two grade 3 AEs occurred in 1 patient each (allergic reaction and neutropenia). One patient had grade 3 abdominal pain and diarrhea associated with nivolumab. No patients discontinued treatment because of toxicity. Investigators reported that progression-free survival (PFS) was highly variable following treatment, with a median of 3.3 months (range, 1-not reached).

“We assert that introducing [personalized neoantigen-targeting cancer vaccines] with first-line treatment, at times of truly minimal residual disease may be more beneficial,” wrote study author Panagiotis A. Konstantinopoulos, MD, PhD, director of the Mellen and Eisenson Family Center for BRCA and Related Genes and director of Translational Research in the Division of Gynecologic Oncology, as well as Velma Eisenson Endowed Chair for Clinical and Translational Research at Dana-Farber Cancer Institute, and professor of medicine at Harvard Medical School, with co-study authors.¹ “Our work thus argues for (1) early vaccine approaches to target low-volume disease; (2) improved vaccine antigen selection to obtain broad and balanced responses; and (3) targeting of antigens that tumors retain the [human leukocyte antigen] HLA machinery to present.”

What did biomarker and immune analyses show?

Dynamic changes in CA125 and cell-free DNA (cfDNA) suggested biologic activity in 6 of 10 patients. Three had transient disease control, and 3 had prolonged control, while 4 had progression. One patient with prolonged control had cfDNA levels that decreased below the limit of detection and remained disease free through data cutoff.

Despite a median of 6.5 months between the diagnostic biopsy used to design the vaccine and the start of vaccination, more than 73.1% of target single nucleotide variant mutations were detectable by cfDNA at the last boost. Interferon (IFN)-γ reactivity to at least 1 vaccine pool was detected in 8 of 10 patients, and 85 of 144 vaccine peptides (59.0%) were immunogenic. Patients with transient or prolonged control had immune responses to more vaccine target mutations than patients with progression (mean, 9.3 mutations vs 5.5; P = .029).

What explained resistance to the vaccine?

Among resistance mechanisms, loss of antigen presentation was the most robust link to therapy resistance. Two of 4 patients with progression had loss of HLA haplotype heterozygosity before vaccination, which likely rendered more than half of vaccine antigens off target. A third patient's tumor had treatment-emergent deficiency in interferon signaling.

Analysis of a The Cancer Genome Atlas (TCGA) ovarian cohort (n = 160) showed approximately 31% of tumors had allele-specific HLA loss. “Considering that [approximately] 31% of tumors analyzed in both our cohort and a TCGA ovarian cohort exhibit allele-specific HLA loss due to either genomic deletion or transcriptional repression, our work strongly supports recent calls to take into account HLA-[loss of heterozygosity] and mechanisms of antigen presentation loss in vaccine design….”^1,2

How was the NeoVax trial designed?

The open-label trial enrolled patients at Dana-Farber Cancer Institute who were 18 years or older with histologically confirmed epithelial ovarian, primary peritoneal, or fallopian tube cancer. In cohort A, patients were newly diagnosed, and in cohort B, they had platinum-sensitive first recurrence of disease, with platinum-sensitivity status defined as disease progression more than 6 months after last dose of first-line platinum-based chemotherapy.

NeoVax contained 4 personalized pools of synthetic long peptides targeting patient-specific neoantigens with the immunostimulant poly-ICLC, given as 5 priming doses over 3 weeks, with boosters at 12 and 20 weeks. Nivolumab started at week 14, and was administered at a 240 mg flat dose for up to 2 years. Feasibility and safety were the primary end points. Vaccines targeted a median of 14 mutations (range, 7-15).

CancerNetwork® previously reported on NeoVax plus pembrolizumab (Keytruda) in newly diagnosed glioblastoma.³

References

1. Sellars MC, Frank JA, Shim B, et al. Personalized neoantigen-targeting cancer vaccines in ovarian cancer: a phase 1 trial. Nat Med. Published online October 14, 2026. doi:10.1038/s41591-026-04717-8
2. Huber F, Bassani-Sternberg M. Defects in antigen processing and presentation: mechanisms, immune evasion and implications for cancer vaccine development. Nat Rev Immunol. 2026;26(1):23-34. doi:10.1038/s41577-025-01208-8
3. Reardon DA, Ghannam JY, Kovarsky D, et al. A personalized neoantigen vaccine reprograms the immune landscape of glioblastoma. J Clin Oncol. 2026;44(suppl 16):2006. doi:10.1200/JCO.2026.44.16_suppl.2006