Transgene's '100%' Is a 16-Patient Phase 1. The Data That Matters Doesn't Land Until 2028

"100% disease-free survival at three years" is a phrase engineered to be screenshotted, and Transgene (Euronext: TNG) — a roughly €220 million French microcap — now owns it in a peer-reviewed journal. TG4050, an individualized cancer vaccine, kept all 16 patients who received it immediately after head-and-neck surgery disease-free at three years, while three of the 16 who were instead watched relapsed. The finding is real, and Nature Communications publication lifts it well above the typical corporate press release. What it is not is proof the drug works — and that gap between a genuine early signal and an investable one carries most of the investment case.

What the "100%" is and is not

TG4050 is built from each patient's own tumor. Nec's AI predicts which mutations are most likely to be recognized by the immune system, and Transgene packages up to 30 of those predicted neoantigens into a virus-based (MVA) vector that trains T cells to hunt residual cancer. The setting is deliberately high-risk: roughly a third of patients with resected HPV-negative head-and-neck cancer relapse, so a vaccine that could keep them disease-free has a real addressable problem.

In a randomized phase 1 (NCT04183166), 32 evaluable patients were split: 16 got the vaccine right after surgery, 16 were observed and given the vaccine only if they relapsed. All 16 vaccinated patients stayed disease-free through a median follow-up of about 30 months; 3 of 16 in the observation arm relapsed. The treated arm also generated durable neoantigen-specific T-cell responses, which is the mechanistic reason to believe the signal, not just the number.

Peer review is the reason to take this seriously rather than wave it off. But read the small print as a statistical test: 16 patients, open-label, and a "control" arm that crossed over to receive the vaccine at recurrence. That design cannot cleanly prove a survival advantage even when the number is 100%. What it establishes is a supported hypothesis in one tumor type — the falsifiable prior that a real efficacy signal must be confirmed at scale in a bigger, properly randomized study.

The unit economics nobody quotes

For a personalized vaccine, product and process are the same thing. Because every dose is assembled from a specific patient's mutations, there is no inventory — each patient is a bespoke batch manufactured to order after surgery. The question that decides the product is not the efficacy headline but cost per patient and manufacturing lead time, the unit economics of a one-off production run.

Transgene's own disclosures are the tell that deployment economics remain unresolved: the company says it is still optimizing the myvac platform to reduce turnaround time, enable scalability, and increase capacity. Translation from corporate-speak: the product is not yet cheap or fast enough to presume commercial viability, and the schedule to get there is open.

The same week's context makes that gap starker, not smaller. In August 2026, Moderna and Merck announced their customized mRNA vaccine (intismeran autogene) plus Keytruda met its phase 3 endpoints in adjuvant melanoma — the first late-stage validation of the whole personalized-vaccine class, on a platform with big-pharma manufacturing muscle behind it. That is a tailwind for everyone in the field and a raised bar for a small viral-vector entrant on the economics. A competitor's success is not the challenger's proof; it is proof the market will demand the challenger match that delivery economics.

What moves the stock, and when

The decisive datum is not the 100%. It is the phase 2: Transgene completed randomization of 38 patients in early 2026 in the same head-and-neck setting, with 2-year disease-free survival as the primary endpoint. Top-line results are expected by the end of Q1 2028 — two years of waiting for the number that would actually validate the phase 1.

The partnership is built to carry the project that long: in early April 2026, Transgene and Nec signed a license agreement securing Transgene's access to Nec's AI neoantigen prediction platform and the rights to develop and commercialize the program, with Nec paying a technology access fee plus staged cash commitments running into early 2028.

Look at the balance sheet and the wait together. Transgene held about €104 million in cash at the end of Q1 2026, against quarterly net burn of roughly €8–15 million, and says funding covers it to early 2028. That puts the phase 2 readout almost exactly on top of the cash cliff — a company that must clear a registration-relevant bar at the same moment its runway runs out. Given the market cap of about €220 million against that cash, the enterprise value of the entire single-asset pipeline is little more than half the stock price. Investors are already paying for the phase 1 signal to pay off; the 3-year Nature paper is confirmation of durability, not a new catalyst that resets the valuation.

The market has also seen most of this before: the 100% figure at two years has been out since mid-2025, and the shares had already run up roughly a third across that year. The incremental new information in today's publication is that the effect held to three years — real, but the kind of confirmation that moves a thought, not the price.

So the honest reading: this headline reports a genuinely encouraging, peer-reviewed early signal from a small biotech whose deeper problem was never efficacy alone, but manufacturing economics and a two-year wait to test the signal properly, timed against a funding window that ends where the readout begins. The 100% tells you the hypothesis is alive. Whether it is investable rests on two things this paper does not answer — a randomized phase 2 that lands when the cash runs out, and per-patient economics that have yet to prove they scale. Until those resolve, the eye-catching number is a reason to track the company, not a reason to extrapolate 100% into an approval.