Biotech & Longevity · Wednesday, 14 October 2026

01 Briefing what happened

BioNTech's mRNA vaccine for one kind of head and neck cancer misses its first main goal in a final-stage trial

Biotech & Longevity 23 sources

Adding BioNTech's BNT113 to the immune drug Keytruda did not slow tumour growth, and the trial now waits on survival; seven deaths are linked to Neurocrine's Prader-Willi drug, and Roche pays $100 million for a Chinese obesity drug.

1 of 2

main goals missed: tumours did not take longer to grow

the other goal, how long patients live, is still being counted [2]

fall in BioNTech's share price early on Tuesday

from $95.84 at Monday's close, Fierce Biotech reported a fall to $91.32 before trading opened [1][2]

20%

rise in BioNTech's shares in August, on a rival's results

Merck and Moderna's personalised cancer vaccine had succeeded in its own final-stage trial [2]

The lead story — what happened

BioNTech, the German company that made the first approved mRNA vaccine, against COVID, with Pfizer, said on Tuesday its cancer vaccine BNT113 missed one of two main goals in a final-stage trial. [1][2]
The trial gave BNT113 plus Keytruda, a widely used immune drug made by Merck, or Keytruda alone, as the first treatment for advanced head and neck cancer. [1][2]
Every patient had a cancer driven by HPV16, a high-risk strain of the human papillomavirus. [1][2]
BNT113 is made of mRNA, a set of instructions that cells read to make proteins. It makes the body produce two HPV16 proteins, E6 and E7, so the immune system learns to attack cancer cells that carry them. [2]
At a planned early look at the results, adding BNT113 did not lengthen the time before tumours grew again. That time is called progression-free survival. [1][2]
An independent committee that watches the trial's data told BioNTech to carry on, to find out whether patients on BNT113 live longer. [1][2]
The trial's other main goal is how long patients live. Under the plan written before it began, a win on that measure alone would make the trial a success. [2]
Ozlem Tureci, BioNTech's co-founder and chief medical officer, said immune effects against a tumour can build and deepen over time. [2]
There is a case like it: Keytruda itself once missed a tumour-growth goal in head and neck cancer, then showed that patients lived longer. [2]
Analysts at BMO Capital Markets, a bank, said the miss made them significantly less confident that BNT113 will lengthen survival. [2]
Daina Graybosch, an analyst at the bank Leerink Partners, had called the evidence for BNT113 sparse when the trial began. Her team expected it to fail, partly because it enrolled people with lower levels of PD-L1, a protein linked to how well Keytruda works. [1]
BioNTech's shares fell about 6% early on Tuesday, and shares of Moderna, which is testing its own mRNA cancer vaccine with Merck, dipped 3%. [1][2]

Who is involved

BioNTech

a German drug company that made the first approved mRNA vaccine, against COVID, with Pfizer, and is now turning back to cancer; it runs the BNT113 trial
Merck

a large US drugmaker whose immune drug Keytruda was given to every patient in the trial
Ozlem Tureci

BioNTech's co-founder and chief medical officer; she said the drug is working the way it was designed to
The trial's monitoring committee

independent experts who see the results as they come in; they said the trial should continue

How it unfolded

August Merck and Moderna's cancer vaccine succeeds in a final-stage trial, and BioNTech's shares rise 20% [2]
August Leerink Partners calls the evidence behind BNT113 sparse [1]
13 Oct BioNTech says BNT113 did not delay tumour growth; shares fall about 6% [1]
Next patients keep being followed to see whether they live longer [1][2]

Where this points

The next test is the survival count: BioNTech gave no date for it, and the trial goes on as planned until the committee says the numbers are ready. [1][2]

What is pushing on the whole day

The bar and the word are our reading of how hard each one is pushing today. The arrow is where it is heading. The evidence is in the stories below.

Drugs falling short in tests High↑

BioNTech's cancer vaccine did not slow tumour growth in its final-stage trial. [1] Nurix, a California company, stopped work on its lymphoma drug zelebrudomide after reviewing early trial data. [14] Neurocrine's Vykat XR, approved 18 months ago, is now linked to seven deaths. [3]

Big drugmakers buying ideas Building↑

Roche, a Swiss drugmaker, agreed to pay up to $100 million soon for an obesity drug from Rona, a Chinese company. [5] Pfizer signed a deal with Enodia, a small drug company, that could be worth more than $500 million. [13] Investors put about $350 million behind Velaris, a new heart and lung drug company. [11]

Drug-company money near medical decisions Building→

Nearly 40% of meetings of the US drug regulator's advice panels included a patient or consumer member tied to a drug-funded group. [7] A review of dozens of studies found that gifts and meals from drugmakers lead to more prescriptions that should not have been written. [8] US officials say Martin Shkreli, a former drug executive, kept trading drug-company shares despite a lifetime ban. [9]

The rest of the day

17 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

02
Seven deaths linked to a Prader-Willi drug

Vykat XR, a drug sold by the US drugmaker Neurocrine Biosciences, has been linked to seven deaths and nearly 100 hospital stays, STAT reported on Tuesday. [3] Vykat XR was approved 18 months ago for Prader-Willi syndrome, a rare genetic condition that can cause a hunger that never stops. [3] STAT obtained the reports sent to regulators through a freedom-of-information request. [3] Two months ago an open letter raised concerns about the drug's safety. [3][4]

7Deaths reported
nearly 100Hospital stays
Reports sent to regulators in the 18 months since Vykat XR was approved. They show what happened, not what caused it.

Why it matters — The reports do not prove the drug caused the deaths, since many patients were very ill and took several medicines. [3] Some experts say the evidence is already enough to conclude the drug is causing unacceptable harm. [3]
03
Roche buys a Chinese obesity drug

Roche, the Swiss drugmaker, will pay Rona Therapeutics of China up to $100 million soon for the rights to an experimental obesity drug, and up to $1.06 billion more if it passes later goals. [5] The drug silences ALK7, a protein found mostly in fat tissue. [5][6] People born with a weaker ALK7 tend to carry less fat around the waist. [5] Neither company said how far the drug has got in testing. [5]

$100mPaid soon
$1.06bnPaid only if goals are met
Most of the money in Roche's deal is promised for later, and is paid only if the drug keeps passing tests.

Why it matters — It is Roche's fifth deal with a Chinese drug company since the end of August and its tenth since early 2025, BioPharma Dive counts. [6] Alnylam, Arrowhead, GSK and Novo Nordisk are chasing the same protein, hoping to cut harmful belly fat alongside weight-loss drugs like Wegovy. [5][6]
04
Drug money near the US regulator's panels

Nearly 40% of meetings held by advice panels of the FDA, the US drug regulator, over a recent 20-year period included a patient or consumer member tied to a group funded by drugmakers, an analysis found. [7] More than one meeting in ten had a member from a group funded by the company whose drug was being judged. [7] Most of these links were not disclosed at the meetings. [7] The study was published in the journal Health Affairs Scholar. [7]

40 of 100
of every 100 panel meetings had a member tied to a drug-funded group
Nearly 40 in every 100 FDA advice meetings, over about 20 years.

Why it matters — These panels advise the FDA when it reviews a company's medicine. [7] The FDA has firm conflict-of-interest rules for the scientists on its panels, but looser ones for patient and consumer members. [7]
05
US officials accuse Shkreli of breaking his ban

The US Federal Trade Commission, which polices unfair business, and seven US states asked a court on Tuesday to hold Martin Shkreli in contempt. [9][10] In 2022 a judge banned him from the drug industry for life, after finding he had illegally blocked cheaper copies of Daraprim, a drug for parasite infections whose price he raised by 5,000%. [9][10] Officials say he has since traded shares in dozens of drug companies and commented on the industry online. [9]

Why it matters — They want a judge to make him sell those shares and hand over any profits. [9] A working phone number for Shkreli could not be found, and his former lawyers did not reply. [9]
06
Gifts from drugmakers change prescriptions

A review by Cochrane, a network that pools medical research, looked at dozens of earlier studies of gifts and payments from drugmakers to doctors. [4][8] It found they lead to more inappropriate prescribing and more prescriptions, which likely raises health costs. [4][8] The more meals a doctor got from a company, the more of that company's drug they prescribed. [8] Drug companies spent more than $250 million on meals for doctors in 2022. [8]

Why it matters — Patients and health systems pay for prescriptions that were not needed. [8] The review also found that rules limiting contact between companies and doctors probably improved prescribing. [4]
07
US science agency may hide grant scores

The NIH, the US government's main funder of medical research, is considering no longer telling scientists the score their grant application received. [16] Today at least three expert reviewers score each application from 1 to 9, and the average is shared. [16] Under the plan, applicants would only learn whether they landed in the top 25%, the next 25% or the bottom half. [16] Public comments on the plan closed on Tuesday. [16]

Why it matters — Critics, including the biologist Joshua Weitz, say hidden scores raise the risk that political priorities override expert review. [16] An earlier plan to let political appointees pick grant winners was put on hold after the US Congress rejected it. [16]
08
Nurix drops one of its two lymphoma pills

Nurix Therapeutics, a California company, said on Monday it will stop developing zelebrudomide, a pill for aggressive forms of lymphoma, a blood cancer. [14] The drug removes BTK, a protein cancer cells need, instead of just blocking it. [14] Nurix had been reviewing data from an early trial since July. [14] Patients already taking it may carry on. [14]

Why it matters — Nurix is putting its effort into a second BTK pill, bexobrutideg, for which Roche paid $700 million up front in June. [14] That drug is now in a final-stage trial against Jaypirca, a leukaemia pill from Eli Lilly. [14]
09
A drug company factory takes another firm public

Velaris Therapeutics, a new company working on heart and lung drugs, joined the stock market on Tuesday by being bought by Alaunos Therapeutics, a listed company that had dropped its own drugs in 2023. [11] Investors led by Fairmount Funds put in about $350 million alongside the deal. [11] Velaris is the ninth company spun out of Paragon, a drug-idea factory Fairmount founded in 2021. [11][12] Paragon also hired the leaders of Apogee, one of its earlier spinouts, which AbbVie bought for $10.9 billion. [12]

Why it matters — Joining the market by slipping into an existing listed company is called a reverse merger, and investors long saw it as a deal for weak companies. [11] More than two dozen have been announced in 2026, against 10 in all of 2025, about as many as new share sales this year. [11]
10
Pfizer signs a deal to block a cell doorway

Enodia Therapeutics, a small company, signed a research deal with Pfizer that could be worth more than $500 million. [13] Enodia makes small-molecule drugs that block Sec61, a doorway in the cell that many proteins pass through as they are made. [13] Blocking it can stop a disease-causing protein before it does harm. [13] Enodia will find the drug candidates, and Pfizer will run the trials and sell any medicine. [13]

Why it matters — Earlier this year Enodia bought a similar research programme from Kezar Life Sciences for $1 million up front. [13] The deal starts with finding drug candidates, before any testing in people. [13]
11
Routine blood tests spot cancer weight loss

Researchers studied the records of 52,236 people with cancer to track cachexia, a severe weight and muscle loss that some cancer patients suffer. [21] Bouts of it came when the cancer was getting worse and were linked to lower survival. [21] Ordinary lab tests showed a steady pattern when it happened. [21] Tumours with changes in a gene called TP53 were more likely to cause it. [21]

Why it matters — Models trained on the data could predict a patient's coming risk of cachexia. [21] The study, in Nature Medicine, used routine hospital data rather than a new trial. [21]
12
A gene on the X may tip autism toward boys

Autism is diagnosed about four times more often in boys than in girls. [17] Researchers at Baylor College of Medicine in Texas found small changes in autistic males in the stretches of DNA that set how much of MECP2, a gene on the X chromosome, is made. [17] One change, inherited from an unaffected mother, cut the gene's output by about 30% in human nerve cells grown in a lab. [17] Boys have one X chromosome, while girls have two, so a girl's second copy may cover for a weak one. [17]

Why it matters — The work, in the American Journal of Human Genetics, explains only part of the gap between boys and girls. [17] The authors say it gives a way to look for similar changes in other X chromosome genes. [17]
13
Blocking knee blood vessels eases pain for a year

Doctors at Charite hospital in Berlin treated 194 people with knee arthritis by blocking small, abnormal blood vessels around the joint with particles that dissolve within hours. [19] Median pain fell from 7 out of 10 before treatment to 3 at six months, and stayed at 3 after a year. [19] At one year, 80% of those checked had pain relief big enough to notice in daily life. [19] No serious side effects were recorded. [19]

Before7
6 weeks4
6 months3
1 year3
Median knee pain on a 0 to 10 scale, before treatment and at each check after it. Most of the drop came in the first six weeks.

Why it matters — It could help people for whom injections no longer work and who cannot or will not have a knee replaced. [19] Every patient in the study got the treatment, so there was no group to compare against, and it did not show the treatment slows damage to the joint. [19]
14
Cancer-killing cells made inside mice

Researchers made tiny fat bubbles that carry mRNA instructions into natural killer cells, a type of immune cell. [20] Injected into mice, the bubbles turned those cells into CAR cells that hunt a chosen target. [20] One version wiped out the B cells in the blood of healthy mice, and others shrank two kinds of human blood cancer grown in mice. [20] The work was published in Nature Biotechnology. [20]

Why it matters — Today's CAR treatments are made by taking a patient's cells out, changing them in a lab and putting them back, which takes weeks. [20] This was only in mice, and most things that work in mice never work in people.
15
Female mice cloned from a male

Researchers used a gene-editing tool to remove the Y chromosome from cloned male mouse embryos, which turned them female. [18] The clones were born healthy and fertile, and they had young with the male clones. [18] The idea is to rescue endangered animals when only one male is left. [18] The study has not yet been checked by other scientists. [18]

Why it matters — It would not work for every animal: females with one X chromosome are fertile in mice, but not in horses or humans. [18]
16
An AI helper for people with epilepsy

About a third of the world's 70 million people with epilepsy have seizures that medicines do not control. [22] Researchers linked brain-wave recordings to an AI chat program on a phone, which tells patients about likely seizures, missed medicine and sleep. [22] It was tested on 13 patients in hospital. [22] Patients' replies helped the seizure detector make fewer false alarms. [22]

Why it matters — Implanted devices can already sense coming seizures but cannot tell the patient in real time. [22] Thirteen hospital patients is a very small first test, published in Nature Medicine. [22]
17
UCB marks its $2 billion US drug plant

UCB, a Belgian drugmaker, dedicated the site in Gwinnett County, Georgia, where it will build its first US factory, at a cost of $2 billion. [15] It will make biologics, medicines grown in living cells. [15] The plant should open in 2032 or 2033 with about 330 permanent jobs. [15] The county has promised $174 million in incentives and building work. [15]

Why it matters — It is UCB's first factory in the US, where it already employs more than 2,000 people. [15] Gwinnett County carries part of the cost through the $174 million it has promised. [15]
18
An eye chip lets a Stargardt patient read a word

Science Corporation, a US company founded by a co-founder of the brain-implant firm Neuralink, has put its PRIMA eye chip into its first patient with Stargardt disease. [23] Stargardt is an inherited disease that slowly kills the eye's light-sensing cells; about 150,000 people worldwide have lost sight to it. [23] The 2mm chip sits at the back of the eye in place of those cells, fed by a camera on glasses. [23] The patient, treated in May at Sydney Eye Hospital, read a word in his second training session. [23]

Why it matters — PRIMA is already sold in Europe for an advanced form of age-related macular degeneration, a common cause of blindness in older people. [23] The sight it gives is a small black-and-white patch, like looking through a straw, and this is one patient's early result, reported by the company's own chief executive, Max Hodak. [23]

02 Lesson why it matters

Why an immune treatment can lose on tumour growth and still win on survival

A cancer vaccine teaches the immune system slowly, so an early scan can show tumours growing before the body fights back, while survival is counted for years.

The twist

A scan in the first months can miss a defence that is still building. Counting how long people live gives that defence time to show, so the makers of BNT113 wrote both measures into the plan before the trial began.

The picture

Tumour growth checkmissed

Survival countstill running

BioNTech's trial named two main measures before it started. The early one missed. The slower one is still being counted.

How it works

A cancer vaccine shows the immune system what the tumour's cells look like
Immune cells take weeks or months to multiply and reach the tumour
Meanwhile the tumour can keep growing on scans
So an early check of tumour growth can show no gain
Survival is counted over years, long enough for a slow defence to matter
Naming both measures before the trial starts lets a survival win still count

The same force, elsewhere today

Where this chain is also running, in today's other stories.

Blocking knee blood vessels

Pain kept falling after treatment, from 7 to 4 at six weeks and to 3 by six months, so a check at six weeks would have shown less than the one at six months.
The Prader-Willi drug Vykat XR

Its trials ended before approval, and the reports of deaths and hospital stays piled up over the 18 months of use that followed.

Where you've seen this

Flu jabs

the body needs about two weeks after the jab to make enough antibodies to protect

Strength training

muscle builds over weeks, so a test on day three shows almost nothing

Planting an orchard

a tree's first harvest says little about what it will give in ten years

The catch

A late survival gain is not promised. Many trials that miss on tumour growth go on to miss on survival too, and BMO's analysts said this miss made them less hopeful.

And the whole of it

The patients still in the BNT113 trial, their doctors and the people who sold BioNTech shares on Tuesday are all waiting for the same count. None of them can see it yet. An independent committee will say when the survival numbers are ready.

03 Truth what's really going on

What is really going on

Roche agreed to pay up to $100 million soon for an obesity drug from Rona without either company saying how far it has been tested, and Pfizer signed a deal with Enodia for drugs that have not been found yet. [5][13] The same day, two drugs already given to patients ran into trouble: BioNTech's cancer vaccine did not slow tumour growth, and seven deaths are now linked to Neurocrine's Vykat XR. [1][3]

Why it works on us — A deal is announced at its largest possible value, $1.2 billion for Roche and Rona, while the part paid soon is about a twelfth of that. [5]

Who gains

Merck — Keytruda alone stays the treatment to beat in this cancer, for now, because adding BNT113 did not slow tumours. [1][2]
Rona Therapeutics and its early backers, including Lilly Asia Ventures — Roche pays up to $100 million soon for one drug, after Rona raised about $68 million in its first two funding rounds. [5][6]
Fairmount Funds — It led the $350 million behind Velaris and will pick four of the five directors of the combined company. [11]
Pfizer — Enodia does the early search for drug candidates, and Pfizer's later payments depend on goals being reached. [13]
Construction and plant workers in Gwinnett County, Georgia — UCB's $2 billion plant is expected to bring more than 1,000 building jobs and about 330 permanent ones. [15]

Who pays

BioNTech's shareholders — The shares fell about 6% on Tuesday, after rising 20% in August on a rival's results. [1][2]
People with Prader-Willi syndrome and their families — They rely on a drug now linked to seven deaths, with no settled answer on whether it caused them. [3]
Patients and health systems — Gifts and meals from drugmakers lead to prescriptions that should not have been written, and someone pays for each one. [8]
Scientists applying for NIH grants — Under the plan they would learn only a broad band, not the score experts gave their work. [16]
Gwinnett County — It has promised $174 million in incentives and building work to win the plant. [15]

What nobody knows yet

Open questions from across today’s stories — ours included.

01

Whether patients given BNT113 will live longer.

The survival data are not mature, BioNTech gave no date for them, and many trials that miss on tumour growth also miss on survival. [2]
02

Why BNT113 did not slow tumour growth.

BioNTech released no numbers, and Leerink Partners had expected a miss partly because the trial enrolled people with lower levels of PD-L1. [1]
03

Whether Vykat XR caused the seven deaths.

The case reports do not prove cause: many patients were very ill and took several medicines, and some experts read the same reports as proof of harm. [3]
04

How far Rona's obesity drug has got in testing.

Neither company gave its code name or stage, and none of Rona's public drugs with named targets act on ALK7. [5]
05

Whether members tied to drug-funded groups changed how FDA panels voted.

The analysis found the ties, most of them undisclosed, and said they raise questions about voting; it did not show votes changed. [7]
06

What the NIH will decide on hiding grant scores.

Public comments closed on Tuesday, and the US health department did not answer a request for comment. [16]
07

Whether blocking knee blood vessels works better than doing nothing.

All 194 patients got the treatment, and about 1 in 5 of them was not checked at one year. [19]
08

How much of autism's tilt toward boys the MECP2 changes explain.

The researchers call it the tip of the iceberg; so far they report changes in two stretches of DNA that control the gene. [17]
09

Whether the cloning trick would work in an animal that needs saving.

The study has not been checked by other scientists, and females with one X chromosome are not fertile in horses or humans. [18]

04 Hope carry this

Doctors in Berlin blocked small blood vessels around the knee in 194 people with arthritis. A year later, 80% of those checked still had pain relief they could feel in daily life, with no serious side effects.

Also true today

Researchers turned cloned male mouse embryos into females by removing the Y chromosome, and the female clones were born healthy and fertile.
A team at Baylor College of Medicine found changes in the MECP2 gene on the X chromosome that may explain part of why autism is diagnosed about four times more often in boys.