Why can't we reliably predict which neoantigens will work?
The biggest bottleneck is that current algorithms often pick the wrong targets. Neoantigens are unique mutations on a patient's tumor, but only a fraction of them actually get presented on the cell surface and recognized by T cells. A 2022 review notes that many computational tools exist to predict neoantigens, but they frequently miss the mark—especially for CD4+ T cell targets, which are harder to predict than CD8+ ones [5][8]. In practice, this means vaccines may include 'ignored' neoantigens that the immune system never responds to, wasting time and resources [5].
The numbers from clinical trials illustrate the problem starkly. In a pancreatic cancer trial, researchers screened tumors for mutations and synthesized up to 20 neoantigens per patient, yet only half of patients (8 out of 16) mounted a detectable T cell response [1]. Even in a more immunogenic cancer like renal cell carcinoma, where all 9 patients generated T cell responses, the study was too small to know which prediction methods worked best [3]. A 2023 review bluntly states that 'proof of unequivocal efficacy has remained beyond reach for the majority of clinical trials' because of these prediction gaps [6].
Can neoantigen vaccines work in cancers with very few mutations?
Yes, but the evidence is mixed and the path forward is uncertain. Cancers like pancreatic and liver cancer have relatively few mutations, meaning fewer potential neoantigens to target. In the pancreatic cancer trial, only 8 of 16 patients responded to the vaccine, and those who did had a much longer recurrence-free survival (median not reached vs. 13.4 months) [1]. A follow-up study at 3.2 years showed that vaccine-induced T cells could persist for an average of 7.7 years, suggesting that even a small number of good targets can produce durable immunity [2].
However, not all low-mutation tumors are equal. In renal cell carcinoma, another moderately mutated cancer, all 9 patients generated T cell responses and none had recurred at 40 months [3]. But in hepatocellular carcinoma (liver cancer), a study found that only about 50% of predicted neoantigens were actually immunogenic in mice, and the number of usable targets per patient was small (median of 13 for HLA class I) [9]. A 2024 study on colorectal cancer—which often has very few mutations—proposed using circular RNAs as an alternative source of neoantigens, showing that T cells trained on these could kill tumor organoids [4]. This suggests that expanding the pool of targets beyond mutations may be necessary for low-mutation cancers.
Do neoantigen vaccines actually delay cancer recurrence?
The short answer is that we don't know yet—the evidence is promising but comes from small, early-phase trials. The most compelling data comes from the pancreatic cancer trial: among 16 patients, the 8 who generated vaccine-induced T cells had a median recurrence-free survival that was not reached at 18 months, compared to 13.4 months for non-responders [1]. At 3.2 years, that difference persisted, and the vaccine-induced T cells showed multiyear longevity [2]. In renal cell carcinoma, none of the 9 patients had recurred at 40 months [3].
But these are all phase I trials with no control group—meaning we can't be sure the vaccine caused the outcome. A 2021 review emphasizes that 'deeper evaluation of the phenotypes, functionality and long-lasting memory potential' of vaccine-induced T cells is needed, and that larger randomized trials are essential [8]. A 2023 review echoes this, noting that despite over 100 clinical trials, 'proof of unequivocal efficacy has remained beyond reach' [6]. The vaccines also face practical hurdles: they are expensive, time-consuming to manufacture, and require sophisticated delivery systems like lipid nanoparticles, which may limit access [7][10].
About These Sources
This answer is built on 10 peer-reviewed studies — published from 2021 to 2025, 5 from 2024 or later, 8 in Q1 journals, collectively cited 3,120 times — selected as the most relevant from 11 studies that passed quality screening, drawn from 38 papers retrieved from a database of over 500 million.
Sources used in this answer
Personalized RNA neoantigen vaccines stimulate T cells in pancreatic cancer
In a phase I trial of 16 pancreatic cancer patients, an mRNA neoantigen vaccine induced T cell responses in 8 patients; those responders had longer recurrence-free survival (not reached vs. 13.4 months at 18-month follow-up).
RNA neoantigen vaccines prime long-lived CD8+ T cells in pancreatic cancer
At 3.2-year follow-up of the same pancreatic cancer trial, vaccine-induced CD8+ T cells persisted with an average lifespan of 7.7 years and retained effector function; responders had prolonged recurrence-free survival.
A neoantigen vaccine generates antitumour immunity in renal cell carcinoma
In a phase I trial of 9 renal cell carcinoma patients, a neoantigen vaccine generated T cell responses in all patients, including against driver mutations, and no recurrences occurred at 40-month median follow-up.
Circular RNA as a source of neoantigens for cancer vaccines
In colorectal cancer, circular RNA-derived neoantigens (circMYH9, circRAPGEF5) elicited T cell responses that killed tumor organoids, offering an alternative neoantigen source for low-mutation tumors.
Identification of neoantigens for individualized therapeutic cancer vaccines
Review of computational neoantigen prediction tools, noting that many algorithms exist but often fail to identify immunogenic targets, especially for CD4+ T cells; proposes a classification of neoantigens as guarding, restrained, or ignored.
Challenges in developing personalized neoantigen cancer vaccines
Review highlighting that despite over 100 clinical trials, neoantigen vaccines lack unequivocal proof of efficacy due to obstacles in vaccine design and the tumor microenvironment.
Advancements and challenges in personalized neoantigen-based cancer vaccines
Review discussing challenges in neoantigen prediction, delivery (e.g., lipid nanoparticles), and the need for better bioinformatics tools and predictive models.
Advances in the development of personalized neoantigen-based therapeutic cancer vaccines
Review of personalized neoantigen vaccines, noting robust immunogenicity in early trials but emphasizing the need for larger studies, better CD4+ epitope prediction, and optimized delivery platforms.
Neoantigens as potential vaccines in hepatocellular carcinoma
In 14 hepatocellular carcinoma patients, about 50% of predicted neoantigens were immunogenic in HLA-transgenic mice; co-immunization with CD4 and CD8 epitopes enhanced CD8 responses.
Tailoring nanovectors for optimal neoantigen vaccine efficacy.
Review of nanovector delivery systems for neoantigen vaccines, noting challenges in efficiently expanding neoantigen-specific T cell populations despite advances in prediction.