The mRNA Pancreatic Cancer Vaccine Will Work. The Question Is: For Whom?

Let me tell you a story about two timelines. The first is the scientific timeline of mRNA cancer vaccines, which is genuinely thrilling. In a phase 1 trial published in Nature¹, BioNTech's autogene cevumeran induced CD8+ T cell responses in half of 16 pancreatic cancer patients that persisted for over three years, with vaccine-induced clones showing an estimated average lifespan of 7.7 years. Responders had prolonged recurrence-free survival compared with non-responders. A phase 2 trial enrolling 260 patients² is now underway. Meanwhile, Moderna and Merck's mRNA-4157 (V940), which targets up to 34 patient-specific neoantigens, has sustained a 49% reduction in recurrence risk at five-year follow-up³ in melanoma and is expanding into multiple Phase 3 trials⁴ across lung cancer, renal cell carcinoma, and other solid tumors. Regulatory submissions are anticipated by 2026-2027⁵. These are real results in real patients with real cancers. The science is working.

The second timeline is the access timeline. And this is where things get uncomfortable, because we have a near-perfect precedent for what happens when a complex, individualized, expensive cancer therapy reaches the U.S. market without any pre-existing access infrastructure: CAR-T therapy. CAR-T was first approved in 2017. It is now 2026. And the disparities have not narrowed⁶. A 2022 analysis in Transplantation and Cellular Therapy found that "African Americans were less likely than other racial/ethnic groups to receive CAR T cell therapy," with only 1% of myeloma CAR-T patients being African American⁷ and only 7.3% of CAR-T admissions coming from neighborhoods with a mean income under $40,000. A 2025 SEER-Medicare analysis in Blood Advances⁸ confirmed that distance to the nearest CAR-T center and state income level continue to predict who gets treated. And ADVI's 2024 analysis showed⁹ "continued racial disparities in CAR-T therapy access for Black patients despite living closer to qualified treatment centers."

This is the single most important data point in the entire debate about mRNA cancer vaccine access. Nine years after approval, after Medicaid coverage decisions have been made in most states, after multiple products have entered the market, CAR-T access disparities are persistent, not narrowing. The promise that post-approval market mechanisms will "eventually" correct access inequities for individualized therapies is not borne out by the only real-world test case we have.

Now layer on what we know about pancreatic cancer specifically. This is not a disease where access gaps are minor inconveniences. The five-year survival rate for pancreatic adenocarcinoma is approximately 12.5%¹². Black Americans have a higher incidence rate¹⁰ (15.4 per 100,000 vs. 13.2 for non-Hispanic whites), are more likely to be diagnosed with late-stage disease¹¹, and are 25% less likely to undergo surgical resection¹¹. Strikingly, a study in the VA system — an equal-access integrated healthcare system — still found that Black patients were 19% more likely to have late-stage disease and 25% less likely to undergo surgery¹¹, suggesting these disparities are not purely insurance-driven.

So here is the situation we are walking into with open eyes: a therapy class that costs over $100,000 per patient¹³, requires tumor sequencing and custom mRNA synthesis currently available only at academic medical centers, will be combined with pembrolizumab (itself roughly $200,000/year), targets a disease that disproportionately kills Black Americans, and enters a healthcare system where the closest precedent therapy — CAR-T — has shown persistent access disparities nearly a decade after approval. The conventional wisdom says: approve first, fix access later. I think the conventional wisdom is wrong, and the CAR-T data tells us exactly why.

The strongest argument against my position is the investment deterrence concern. Price mandates or access conditions tied to approval, the argument goes, would scare away the R&D capital that made these vaccines possible in the first place. And this concern is not hypothetical. A study published in Health Affairs Scholar¹⁴ found that post-IRA, industry-funded post-approval oncology trials dropped 45.3% for small molecules and 32.5% for biologics. The Inflation Reduction Act's mere existence appears to have changed investment behavior. The worry is that imposing conditions before a therapy is even approved would be worse.

I take this seriously. But the argument has a critical flaw: it treats all pre-approval access frameworks as equivalent to blunt price ceilings. They are not. The UK's NHS has done something genuinely different. Through its Cancer Vaccine Launch Pad¹⁷, NHS England has partnered with BioNTech to provide up to 10,000 patients with personalized cancer immunotherapies by 2030, with 30 hospitals signed up¹⁸ to fast-track patients into clinical trials. This is not a price control. It is a distribution infrastructure investment, built in parallel with clinical development, that ensures the sequencing infrastructure and patient identification pipelines exist before approval, not after. It is exactly the kind of pre-approval access framework that does not chill investment — BioNTech is a willing partner — while ensuring the therapy reaches patients who would otherwise fall through the cracks.

The United States has nothing comparable. No Cancer Vaccine Launch Pad. No coordinated infrastructure investment linking community hospitals and safety-net systems to the sequencing and manufacturing pipelines that individualized mRNA vaccines require. And critically, the Inflation Reduction Act's drug negotiation provisions¹⁵ won't touch a newly approved biologic for at least 11 years — 13 years before maximum fair prices take effect¹⁶. A personalized mRNA cancer vaccine approved in 2027 would not be eligible for Medicare price negotiation until approximately 2040. That is not a policy framework for a disease with a median survival measured in months.

I also want to address the HIV antiretroviral precedent, because it is the go-to example for optimists about post-approval market correction. It is true that ARV prices eventually fell from $10,000+ per patient per year to under $100 through generic licensing, PEPFAR, and manufacturing scale. But that correction took roughly 15-20 years and, by UNAIDS estimates, millions of preventable deaths occurred during the window. The ARV success story is actually an indictment of the "approve then fix" approach: it documents precisely how many people die while market correction plays out. For HIV, patients had years to decades of survival to wait for prices to fall. For pancreatic cancer, the median survival after diagnosis is six to twelve months. The post-approval correction model is calibrated to the wrong disease.

The manufacturing cost trajectory is the wild card. Current per-patient costs exceed $100,000, but manufacturing improvements have already reduced production time from nine weeks to under four weeks¹³. Self-amplifying RNA constructs²⁰ could reduce the doses required. AI-driven neoantigen prediction is accelerating. But here is the thing: manufacturing cost reductions help after the infrastructure to deliver the therapy exists. If only 17 U.S. sites can administer a pancreatic cancer mRNA vaccine — as is currently the case for the BioNTech Phase 2 trial¹⁹ — then a cheaper vaccine still does not reach a patient at a rural community hospital in Mississippi.

My position is specific: the U.S. needs to build a Cancer Vaccine Launch Pad equivalent, investing in the sequencing, tumor profiling, and distribution infrastructure at safety-net hospitals and community oncology centers now, while trials are ongoing, not after approval when the rush to market will leave underserved patients behind. This does not require new FDA statutory authority or price controls. It requires CMS investment in tumor sequencing reimbursement at community hospitals, NIH-funded infrastructure grants to build vaccine delivery capacity outside academic centers, and voluntary manufacturer partnerships modeled on the NHS-BioNTech agreement. These are tractable, politically feasible interventions that do not threaten R&D returns.

Without them, I predict with high confidence that within three years of the first mRNA cancer vaccine approval in the U.S., peer-reviewed analyses will document racial and socioeconomic access disparities of comparable magnitude to what we see in CAR-T today — because the structural conditions producing those disparities are identical. The indicator to watch is straightforward: when Phase 3 data read out (expected 2026-2027 for melanoma, with pancreatic cancer on a slightly later timeline), track whether any U.S. federal agency announces a distribution infrastructure program before the FDA approval decision. If they do not, the access gap is already locked in.