14 September 2026
Heard In AI

A cancer vaccine built from each patient's own tumor clears a phase 3 trial

Merck and Moderna said on August 19 that intismeran autogene, an mRNA therapy manufactured separately for every patient, met its main endpoints alongside Keytruda in a 1,137-patient melanoma trial. On the Moonshots podcast, the panel walked through the sequencing-and-machine-learning workflow behind it and argued about whether drug regulation can cope with a treatment that is different for each person.

A briefing reports one development when it happens. We correct or clarify it later; a new development gets a new briefing. How our formats work

Peter Diamandis slowed down to unpack the steps. A surgeon removes the melanoma. The tissue goes to a sequencing lab, which reads the tumor's whole exome and its RNA. That output is fed into a machine learning model hunting for neoantigens — surface markers that exist on the cancer and nowhere else in the patient's body. The model ranks them. Up to 34 of the top-ranked targets are encoded into a strand of mRNA, and the finished product is "manufactured and shipped in only eight weeks."

The result is a drug that exists in a single copy. "A vaccine for you is not the same as a vaccine for me," the host said during the Moonshots panel's health segment. He noted that more than 8,500 people in the United States are expected to die of melanoma this year.

What the trial actually reported

The occasion was Merck and Moderna's August 19 announcement that their individualized mRNA therapy, intismeran autogene, had met its endpoints in a phase 3 trial when given with Keytruda (pembrolizumab), an immunotherapy already used after melanoma surgery.

The trial, INTerpath-001, was randomized and double-blind. It enrolled 1,137 patients whose stage IIB to IV melanoma had been completely removed by surgery, and assigned them in a 2:1 ratio to either the combination or the pembrolizumab control regimen, for roughly a year of treatment. The prespecified interim analysis found significant improvement on two measures: recurrence-free survival, the time from randomization until the cancer came back or the patient died, and distant-metastasis-free survival, a key secondary endpoint that tracks whether the disease spreads to other organs. The companies reported no new safety signals and said evaluation of overall survival is continuing.

What the announcement did not include was numbers. No phase 3 effect sizes were published with it. The figures quoted on the podcast — a 49% reduction in recurrence or death and 59% in distant metastasis or death — come from five-year follow-up of the earlier phase 2b study, not from this trial. The host also said on air that the treatment is expected to cost as little as $5,000; that is a forecast, not a published price.

"They're just soft and they're made of fat"

Alex took the result back to a promise made to the US Congress in the early 2000s, when Eric Drexler and others helped sell the National Nanotechnology Initiative on a vision from science fiction: tiny machines patrolling the bloodstream, zapping cancer cells one by one.

"It's 2026, and we caught up with the future," he said. The machines arrived — they are simply not made of diamond. An mRNA vaccine is delivered in lipid nanoparticles, microscopic droplets of fat carrying their genetic payload into cells. "They're like soft robots," Alex said, primitive ones rather than the hard machine-phase devices Drexler described. "This is arguably what we wanted 20 years ago out of nanorobots. It's just that they're soft and they're made of fat."

He also pointed at a piece of the pipeline he thinks is going underpublicized. Sequencing the tumor is only half the job. To know which genes a tumor is expressing abnormally, you need something to compare it against — a second sequencing profile drawn from the patient's blood. Alex credited that to a company called Personalis, whose technology, as he understands it, was originally developed to detect traces of cancer in blood. His extrapolation: within a few years, personalized cancer therapy may not require sequencing the tumor at all, with the whole profile read continuously from the bloodstream instead.

He expects company, too. "This is the first, but not the last," he said of the phase 3 success, pointing listeners to Moderna's public pipeline page, which lists the clinical stage of each program across infectious disease, rare disease and cancer. Diamandis added that the company is also building toward treatments for widespread infections such as CMV and Epstein-Barr virus.

A trial design for a patient population of one

For Celine, the striking thing was not the biology but the paperwork: regulators allowed a personalized product through a trial like this at all. "We've never been able to do that before," she said, describing it as opening the floodgates for personalized medicine — the underlying problem being how a system built around standardized drugs handles what she called a sample size of n of one. She recalled Raymond McCauley describing mRNA vaccines years ago as "the first battle in the last war against all disease."

Emad Mostaque took the same observation somewhere less celebratory. Nobody on the panel was surprised by the result, he said, and nobody will be surprised by the next one. But under the current regime each new case will have to go through the same process "over and over and over again." His proposal: "Screw cancer. Like, let's actually think about this from first principles" — and upgrade the regulations so treatments this targeted reach patients faster. It remains a proposal; the approval path for the next indication is unchanged.

Dave heard the news at closer range. His daughter works at Moderna and has been sending him research reports for months. "This is not a secret," he said. He described a dramatic one-day move in the stock — the host put the surge at 110%, Dave said it almost tripled — and used it to complain about the state of Wall Street research, arguing that post-Enron rules barring analysts from trading the stocks they cover drove the people who understood this kind of science out of the job, leaving indexes and thin coverage behind. "Any good analyst studying this would have seen this coming."

Alex's one complaint was about the name. In researching it, he said, he found that the drug's name resembles a Turkish word meaning exploitation or abuse, and he suggested Merck and Moderna rename it "to something that works well in every time zone" before the rest of the world notices.

The name is likely to stay. The companies said they plan to present the INTerpath-001 data at a medical meeting and to take it to regulators, while the trial keeps following patients for the measure that settles the question in the end: whether they live longer.

Share this article

Go to the original

Sources & further reading

  1. 01

From the conversation

Podcast episodes

Article history

Updates to this article

Tags

A virtual cell that remembers what you did to it

GenBio AI's AIDO Cell simulates a human cell that holds its state across a sequence of interventions, and the Moonshots panel watched a demo and began sketching the end of medicine. The article explains what the simulator does today — prioritizing experiments in two prototype cell lines, with laboratory validation of novel predictions still underway — and separates that from the panel's proposals: an AlphaGo-style search from diseased to healthy cells, open public biology data, and frontier labs paying for all of it.

7 min read

Sanders wants an AI pause; Moonshots argues for stronger defenses

Bernie Sanders invoked the AI labs’ own safety commitments in demanding a development pause. On Moonshots, Emad Mostaque and the panel argued instead for controls on biological synthesis and faster detection of threats—but a proposal to monitor every AI prompt raised a different question: how much privacy would those defenses cost?

6 min read

A robot said to outrun Usain Bolt splits a panel over what bodies are for

On Moonshots, Peter Diamandis described a new Unitree humanoid as hitting 12.66 metres per second after three months of development. Salim Ismail said the lesson is to stop copying humans; Alex argued general-purpose robots will swallow the specialists; Emad Mostaque predicted extreme machines will be kept off the street, and Alex proposed consumer, industrial and military classes.

5 min read