A large retrospective study published in Nature in October 2025 found that cancer patients receiving immune checkpoint inhibitors lived substantially longer if they had received a COVID-19 mRNA vaccine within 100 days of starting treatment. The finding has attracted both significant interest and methodological scrutiny, and a randomized trial is now being designed to test it prospectively.
What the study found
The paper, published in Nature on October 22, 2025, was led by Adam J. Grippin and Steven H. Lin at the University of Texas MD Anderson Cancer Center, with contributions from the University of Florida and other institutions. The researchers analyzed retrospective data from multiple cancer patient cohorts at MD Anderson, focusing on patients who received immune checkpoint inhibitors (ICIs), a widely used class of cancer immunotherapy drugs.
The primary analysis examined 884 patients with unresectable stage III or stage IV non-small cell lung cancer (NSCLC). Among these, 180 patients had received at least one mRNA COVID-19 vaccine within 100 days of starting ICI therapy, while 704 had not. The vaccinated group had a median overall survival of 37.33 months, compared with 20.6 months in the unvaccinated group, an adjusted hazard ratio of 0.51, meaning the risk of death was approximately half as high in those who had received the vaccine. Three-year overall survival was 55.7% in the vaccinated group and 30.8% in the unvaccinated group. These results were consistent across patients with stage III unresectable disease (adjusted HR 0.37) and stage IV disease (adjusted HR 0.52).
A second cohort of 210 patients with metastatic melanoma showed a similar pattern: 43 patients who received a COVID vaccine within 100 days of starting immunotherapy had improved survival compared with the 167 who did not. The benefit was most pronounced in patients with immunologically cold tumors, defined by very low PD-L1 expression on tumor cells. In this subgroup, three-year overall survival improved by nearly five-fold among vaccinated patients.
The survival benefit was consistent regardless of which vaccine manufacturer was used, whether patients received one or two doses, and whether the dose was a primary or booster dose. It was not seen with influenza or pneumococcal vaccines in NSCLC cohorts, a comparison the researchers used to argue that the effect is specific to mRNA vaccines rather than to vaccination or immune activation generally.
The proposed mechanism
While developing personalized mRNA-based cancer vaccines for brain tumors, Grippin and colleagues noticed that mRNA vaccines trained immune systems to attack cancer cells even when the mRNA did not encode a tumor antigen. This observation led them to hypothesize that other mRNA vaccines, including COVID-19 vaccines, might produce the same effect through a non-antigen-specific route.
Their proposed mechanism centers on type I interferons, specifically interferon-alpha. When mRNA is delivered via lipid nanoparticles, it triggers an innate immune response that produces large quantities of IFN-alpha. This interferon activates antigen-presenting cells (APCs), which in turn prime CD8+ T cells to recognize and attack tumor-associated antigens. The primed T cells then infiltrate tumors. In response to this infiltration, tumor cells upregulate PD-L1, an immune checkpoint protein that normally suppresses T cell activity. Immune checkpoint inhibitors work by blocking exactly this suppression mechanism, specifically by blocking PD-1 or PD-L1 interactions. The combination of mRNA-driven T cell priming and ICI-driven checkpoint blockade is required for maximal tumor regression.
In preclinical mouse models, SARS-CoV-2 mRNA vaccines produced significant increases in IFN-alpha, increased myeloid-lymphoid activation, and measurably increased PD-L1 expression on tumors. The same correlates were found in healthy human volunteers who received COVID mRNA vaccines: increased type I interferon levels and myeloid-lymphoid activation signatures consistent with innate immune priming. Concomitant ICI treatment was required for tumor regression in immunologically cold tumor models; the vaccine alone was insufficient.
As Grippin has noted in public commentary, the effect appears to be driven by the innate immune response to the mRNA molecule itself, not by immunity to the spike protein. The implication is that any mRNA vaccine, including future mRNA vaccines designed for other purposes, might produce the same tumor-sensitizing effect.
What subsequent studies found
Two independent real-world replication attempts have been published since the Nature paper appeared.
A UK cohort study, posted to medRxiv in January 2026, analyzed patients treated with ICI therapy at a UK cancer center using the same 100-day exposure window. This study did not replicate the survival benefit seen in the MD Anderson cohort. The authors noted several methodological differences between the two cohorts, including differences in vaccine uptake patterns, timing of ICI initiation relative to vaccination, and population-level confounding factors specific to the UK vaccination rollout, which was tightly correlated with age, health status, and priority group.
A French real-world cohort study published in Annals of Oncology in May 2026 found results more consistent with the original MD Anderson finding, reporting improved overall survival in patients who received mRNA COVID vaccines concurrently with PD-(L)1 inhibitor therapy. This study used French national health data and included multiple cancer types.
Why this is harder to interpret than it appears
Retrospective cohort studies of this kind carry several inherent interpretive challenges, and the UK non-replication underscores them.
The most important is healthy vaccinee bias, also called healthy user bias. Patients who chose to receive COVID vaccines may systematically differ from those who did not in ways that affect cancer outcomes independent of the vaccine. Vaccinated patients may be more likely to adhere to treatment, have fewer comorbidities, or have better access to supportive care. In the US context, vaccination rates also correlated strongly with demographic and socioeconomic factors that independently predict survival. The researchers attempted to control for these confounders using multivariate adjustment, but residual confounding in retrospective data cannot be fully eliminated.
The timing criterion (vaccine within 100 days of ICI start) also creates an implicit selection effect: patients healthy enough to receive a vaccine and begin immunotherapy within the same 100-day window may already be better-performing patients than those who were vaccinated well before or not at all. The UK authors raised this as a specific concern.
The strength of the effect (adjusted HR 0.51 in the primary cohort) is large enough that healthy vaccinee bias alone is unlikely to fully explain it, but the absence of replication in the UK cohort means the effect cannot yet be taken as established.
What comes next
The MD Anderson team is designing a randomized Phase III clinical trial to test prospectively whether COVID mRNA vaccines improve responses to immune checkpoint inhibition in patients with advanced NSCLC. A randomized design would eliminate the healthy vaccinee confounding that complicates the retrospective analysis. If the trial confirms the retrospective finding, the vaccine could potentially be added to the standard of care for patients beginning ICI therapy, representing a relatively low-cost and widely accessible intervention given that mRNA COVID vaccines are already approved, manufactured at scale, and inexpensive compared with the immunotherapy drugs they would augment.
Three of the study authors, Elias Sayour, Adam Grippin, and Duane Mitchell, hold patents related to University of Florida-developed mRNA vaccines that are licensed by iOncologi Inc., a biotech company in which Mitchell holds an interest. This conflict of interest is disclosed in the paper and is relevant context for evaluating the authors’ enthusiasm about next-generation mRNA therapeutics.
References
- Grippin AJ, Marconi C, Copling S, et al. SARS-CoV-2 mRNA vaccines sensitize tumours to immune checkpoint blockade. Nature. 2025;647(8089):488-497. doi:10.1038/s41586-025-09655-y.
- Association of mRNA COVID-19 vaccination with overall survival in patients receiving immune checkpoint inhibitors: a retrospective cohort analysis. medRxiv, January 2026. medrxiv.org.
- Overall survival according to COVID-19 mRNA vaccination co-exposure in patients treated with PD-(L)1 inhibitors: a French real-world cohort study. Annals of Oncology. 2026. doi:S0923-7534(26)00181-X.
- UT MD Anderson Cancer Center. ESMO 2025: mRNA-based COVID vaccines generate improved responses to immunotherapy. October 19, 2025. mdanderson.org.