The Data Leaked at Lunchtime. By the Close, Moderna Had Lost $24 Billion — on Results Most Oncologists Call a Win

Ninety minutes of chaos#

The data was supposed to arrive on Saturday, on a stage in Madrid, in front of the European Society for Medical Oncology. Instead it arrived on a Thursday lunchtime in New York, as screenshots.

At about 12:25 p.m. Eastern, the embargoed late-breaking abstract for INTerpath-001 — the Phase 3 trial of Moderna and Merck's personalised cancer vaccine, intismeran autogene, combined with Merck's Keytruda — appeared briefly on the Oncology Central website, days ahead of its scheduled ESMO presentation. It was taken down. By then it was circulating on X. ESMO declared the embargo breached and lifted it at 2:11 p.m.; at 2:59 p.m. Merck and Moderna issued the full results themselves. Trading in Moderna was briefly halted for volatility along the way.

The stock's verdict came faster than any oncologist's. Moderna had closed Wednesday at $245.81. Intraday on Thursday it touched roughly $183.88 — more than 25 percent down, some $24 billion of market value gone at the low. It recovered part of the fall to close at $202.63, down 17.57 percent. Merck, the senior partner, closed down about 1.8 percent. On Friday the shares dipped again in the morning before recovering to trade modestly higher by midday — $205.68 at 12:30 p.m., up 1.5 percent on the day. A leaked abstract had produced one of the largest single-day value destructions in biotech this year — and the abstract, when it was finally read in full, described a trial that succeeded.

What the trial actually showed#

INTerpath-001 randomised 1,137 patients whose stage IIB to IV melanoma had been completely surgically removed — patients at high risk of the cancer returning. Half-plus received intismeran, a vaccine manufactured from each patient's own tumour, encoding up to 34 neoantigens selected from that tumour's mutations, given with Keytruda; the rest received a placebo with Keytruda. The vaccine is the experimental addition; Keytruda, Merck's anti-PD-1 antibody, is the established backbone both arms received.

At a median follow-up of 24.6 months, the combination cut the risk of recurrence or death by 42 percent versus Keytruda alone — a hazard ratio of 0.58, statistically decisive (p below 0.001). At two years, 82.2 percent of the combination group remained recurrence-free, against 70.0 percent on Keytruda alone: an absolute difference of 12.2 percentage points, roughly one recurrence prevented for every eight patients treated. The risk of distant metastasis or death fell 38 percent (HR 0.62), with two-year distant metastasis-free survival of 88.5 percent against 80.2 percent. Overall survival data are not yet mature; the companies describe a favourable trend, nothing more. Safety was unremarkable: severe treatment-related side effects ran 14.2 percent on the combination against 13.4 percent on Keytruda alone, with no new safety signal. The manuscript has been accepted by the New England Journal of Medicine, and the full dataset is presented in Madrid on Saturday.

The good: a first for personalised cancer vaccines#

Strip the market noise away and the clinical fact is large. This is the first positive Phase 3 result for a personalised neoantigen vaccine — a therapy designed, manufactured and dosed for one patient's tumour at a time. The effect size held up, at meaningful scale, in a double-blind trial seven times larger than the Phase 2b study that first made the field pay attention: KEYNOTE-942 reported a 44 percent recurrence-risk reduction in 157 patients, later updated to 49 percent at five years. Phase 3 readouts routinely shrink Phase 2 effects; a 42 percent reduction surviving contact with 1,137 patients is the result the field needed.

What the oncologists say#

The specialists who treat melanoma — several of them investigators on the trial itself — were close to unanimous that the result matters. Professor Georgina Long of Melanoma Institute Australia, the study's principal investigator, said in the companies' release that the results "have the potential to make a foundational shift in how we treat patients with earlier-stage melanoma in the adjuvant setting," describing an approach that "tailors treatment to the unique mutational signature of a patient's tumor." Dr. Hussein Tawbi of MD Anderson, another trial investigator, was blunter with Reuters: "This is highly statistically significant. It is not by chance." Tawbi told CNBC the result is "a huge development for our patients," and emphasised that the vaccine's added toxicity is mostly mild — injection-site pain and fever rather than the severe immune side effects oncologists brace for.

Dr. Janice Mehnert, who directs melanoma oncology at NYU Langone's Perlmutter Cancer Center, framed why clinicians are excited: "This is the first time we're seeing that when you add something to immunotherapy, you actually boost the efficacy of the immunotherapy" (CNBC). Dr. Michael Postow, chief of the melanoma service at Memorial Sloan Kettering, pointed at the endgame: "If we can essentially cure more patients with earlier-stage melanoma and never have to deal with unresectable metastatic disease — that's a wonderful advance" (CNBC). CNBC's roundup of six specialists concluded the results are a clinically meaningful advance that could become standard care for patients who have surgery first. Barron's summarised the split in one line: the results "received a warm reception from the medical community" — and a cold one on Wall Street.

The named sceptic among the oncologists is Dr. Adil Daud, co-director of the UCSF Melanoma Center and a lead investigator on the trial itself. His objection is not the data but the product: "I think the downside is that it's super labor-intensive, and you're gonna have to make vaccines for each person, and so you can't just stockpile a million vaccines and then just ship it out when somebody needs it." On cost, he added, "there is the time and expense, and it's probably going to cost like a lot of money" (CNBC). Daud also expects patients whose tumours have not yet been removed to stay with the simpler approach of Keytruda before surgery, because it "doesn't put the patient through the delay" of a roughly six-week manufacturing run — which would confine the vaccine mainly to the post-surgery setting. Dr. George Ansstas of Washington University in St. Louis, who worked on the Phase 2 trial, offered the counter-reading on the softer Phase 3 hazard ratios: "It's great data actually. It's consistent with what we've proven in the phase 2 trial" — noting the Phase 3 enrolled a broader, partly lower-risk population, and that the manufacturing timeline is a plausible explanation for why the benefit only separates after six months.

The bad: three honest caveats#

The oncologists' enthusiasm comes with footnotes, and the footnotes are where Thursday's sellers were standing.

Survival is unproven. Recurrence-free survival is the trial's proven endpoint; overall survival — whether patients live longer — is still immature. Adjuvant melanoma history is littered with drugs that delayed recurrence without extending life, and regulators and payers will ask the question the data cannot yet answer.

One measure softened. Distant metastasis-free survival, the endpoint that tracks the spread that actually kills melanoma patients, came in at HR 0.62 — clearly positive, but weaker than the 0.41 the smaller Phase 2b trial posted. The Phase 3 population included earlier-stage patients, which complicates the comparison, but the decay gave sceptics their chart.

The first six months show nothing. In a prespecified cut of the data, the recurrence benefit in the first six months was nil (HR 0.99); after six months it was a 58 percent reduction (HR 0.42). The companies' reading is biological — a vaccine must teach the immune system before it can protect. The sceptical reading is that the therapy does nothing during the window when the highest-risk patients relapse, and that a cheaper sequencing alternative exists: the SWOG S1801 trial achieved a similar hazard ratio by simply giving Keytruda before surgery as well as after.

And underneath all three sits the manufacturing question no hazard ratio answers: every dose of intismeran is a bespoke product, sequenced and built from one patient's resected tumour. Whether that can be produced, shipped and reimbursed at melanoma scale — let alone across the lung, kidney, bladder and pancreatic trials now running — is an industrial problem, not a clinical one.

Why the stock fell on good news#

The leak did not reveal bad data. It revealed data at the low end of what the market had already paid for. Analysts went into Thursday hoping for a hazard ratio nearer 0.50; 0.58 landed inside the expected range but at its soft edge, attached to a softening metastasis number, in a stock that had priced in a triumph. Positioning did the rest: a crowded, momentum-held biotech name, a halted tape, and an information vacuum of roughly ninety minutes in which the only available version of the story was a screenshot. The analyst corps split along exactly that line — UBS's Michael Yee called the recurrence data "solid" and the plunge a "sell the news" event (Reuters), and Bernstein used the same phrase, while BMO's Evan Seigerman judged the results "not enough to drive confidence amidst elevated expectations" and Citi's Geoff Meacham noted that some holders had been "assigning value to an effective-cure narrative" the data do not yet support. Merck's near-indifference is the tell — Keytruda, which generated $31.6 billion in 2025, close to half of Merck's revenue, is Merck's asset either way, and intismeran's profits, if they come, are split 50/50. For Moderna, whose COVID vaccine revenue has fallen from $36 billion across 2021–22 to under $2 billion last year, the melanoma programme carries the entire platform argument. The stock was not trading on one trial. It was trading on whether personalised mRNA medicine is a business.

What happens next#

Saturday's ESMO presentation, with Moderna's investor event alongside it, puts the full dataset — including the survival trend and subgroup detail — in front of the specialists who will decide how fast this enters practice. The companies have pointed to a U.S. filing and a possible 2027 launch. Moderna reports third-quarter results on November 6, with an analyst day on November 12. And ESMO is investigating how its embargo broke. "We fully recognize the seriousness of the situation and the potential consequences," an ESMO spokeswoman told the Wall Street Journal, adding that "appropriate measures are being taken." Neither Moderna nor Merck has commented on the leak itself. The question matters beyond one stock: the leak moved $24 billion before the science could be presented in the room built for it.

Trial data are as reported in Merck and Moderna's October 22, 2026 release; the leak timeline and market figures are as reported by Reuters and market data on the dates shown. Friday's price is an intraday observation, not a close. This article is reporting and analysis, not financial advice.