Biotech & Longevity · Saturday, 8 August 2026
01 · Briefing · what happened
A living cancer drug: the FDA clears an engineered virus that copies itself inside tumors
After two rejections, regulators approved Replimune's Tudriqev, a virus reengineered to multiply inside melanoma and burst it from within - part of a two-approval week that also cleared the first mRNA flu shot.
1 in 4
patients responded
advanced, treatment-resistant melanoma
~14 mo
median response duration
how long responses lasted
$450k
price per course
before rebates and discounts
3rd
attempt at approval
rejected twice before
At a glance
- The FDA approved Tudriqev, an engineered virus, for advanced melanoma that no longer responds to standard immunotherapy.
- The virus is injected into the tumor, copies itself, bursts the cancer cells, and exposes them to the immune system.
- It was rejected twice before an outside advisory panel overruled FDA scientists' doubts.
- About one in four patients responded; the trial had no comparison group.
- Approval is provisional - a confirmatory Phase 3 trial reads out in late 2027; the price is $450,000 per course.
- The same week the FDA cleared the first mRNA-based flu vaccine, Moderna's mFluSiva.
Forces in play
few options once PD-1 drugs fail
single-arm, no control group
panel overruled FDA scientists
first mRNA flu shot cleared same week
How it unfolded
- Twice before the FDA rejects RP1
- Last week an advisory panel votes to back it
- Thursday accelerated approval granted
- Late 2027 confirmatory Phase 3 trial reads out
Where this points
Watch the 2027 confirmatory trial - a clean win locks the approval in, a miss could pull the drug and cool the whole self-amplifying-therapy field.
Full briefing
On Thursday the FDA granted accelerated approval to Tudriqev, an engineered oncolytic virus, for advanced melanoma that has stopped responding to standard immunotherapy
An oncolytic virus is a virus rebuilt to attack cancer. Tudriqev is injected straight into a tumor, where it copies itself, destroys the cancer cells, and spills out proteins that flag the tumor to the immune system
The caveats are large. Approval rests on a single-arm trial with no comparison group, and FDA scientists had questioned how the effect was measured
The same week brought a second landmark: the FDA cleared Moderna’s mFluSiva, the first flu vaccine built on mRNA - the code-based technology behind the COVID shots
Quietly, the flip side showed too. Aurora Therapeutics, a startup trying to scale custom CRISPR gene therapies for single patients, abruptly ended its program
02 · Lesson · why it matters
The medicine that makes more of itself
Almost every drug only dilutes from the moment it enters you. This one multiplies - and that single fact changes what the whole science is about.
How it works
- A dosed drug spreads out and clears - the amount only falls
- An oncolytic virus is injected once into the tumor
- Inside cancer cells it copies itself, so the dose rises where it lands
- It bursts the cells and flags them to the immune system
- Engineering keeps it multiplying in cancer, not healthy tissue
The twist
Most drugs are things you dole out and they only dilute; a self-copying therapy manufactures itself at the target, so the whole game becomes selectivity and control, not dosage.
Where you've seen this
Chain reactions
one split atom triggers more - power or bomb depending only on containment
Wildfire vs a controlled burn
the same self-spreading force, useful only where you can bound it
A viral idea
it grows on its own once seeded, so you steer it, you don't ration it
The catch
Self-amplifying only helps if it stays where you aimed it - the power and the danger are the same property, which is why the engineering is the whole drug.
Full lesson
A drug that grows
Think about how any pill you have ever taken works. You swallow a fixed amount. Your body absorbs it, spreads it around, then breaks it down and clears it out. From the first minute, the amount of drug in you only ever goes down. The art of medicine, for a century, has been picking that starting amount - enough to work, not enough to harm.
The virus the FDA just approved does the opposite. A doctor injects it once into a melanoma tumor. Inside the cancer cells it does what viruses do: it copies itself. Where it lands, the dose does not fall. It rises. The virus bursts the cell, spills out, and infects the cells next door. The drug manufactures more of itself, at the exact spot you wanted it.
Why that flips the problem
Once a therapy can reproduce, the old question - how much do we give? - almost stops mattering. You are not really dosing a quantity. You are seeding a process and letting it run.
So the danger moves. A normal drug is dangerous if you give too much. A self-copying drug is dangerous if it copies in the wrong place. The entire risk collapses onto one question: can you keep it multiplying inside the cancer and nowhere else? That is why almost all the engineering goes into making the virus thrive in tumor cells and stall in healthy ones. The selectivity is not a feature of the drug. It is the drug.
A pattern older than medicine
You have seen this shape before, far from any hospital. A single splitting atom knocks loose others that split in turn. Bounded inside a reactor, it is a city’s electricity. Unbounded, it is a bomb. Same reaction. The only difference is containment.
A match in a firebreak is a controlled burn that clears dead brush. The same flame in dry wind is a wildfire. An idea that spreads on its own does not need you to ration it out; once it is loose, you can only try to steer it. Money left to compound grows on a curve, not a line, for the same reason - each unit makes the next.
Anything that reproduces itself obeys different rules than anything you hand out in portions. The power and the danger are never two things you can separate. They are one property - self-amplification - seen from two sides. You cannot keep the growth and remove the risk of runaway. You can only bound where it happens.
The arrangement under the approval
None of this reached patients on the science alone. This virus was rejected twice. FDA scientists doubted how the trial measured its effect - a single group of patients, no comparison arm, about one in four responding. Then a panel of outside experts looked at the same data, weighed the fact that these patients had almost nothing left to try, and voted to approve. The regulator agreed, provisionally, with a bigger trial still years away.
So a life-or-death line - is this evidence enough? - was not a fact waiting to be read off the numbers. It was a judgment call, made by people in a room, under pressure from patients with no other options and a price tag of $450,000. Someone drew that line. It could have landed elsewhere, and twice it did.
What any single seat can see
The reader is inside this too. The next patient offered this virus, or the mRNA flu shot cleared the same week, meets it as a settled fact. It is a product on a shelf, a thing that simply exists. What they do not see is how contingent it was. Behind the shelf sit the two rejections, the split between the agency’s own scientists and its advisers, and the confirmatory trial that has not yet reported.
That is the honest shape of a frontier. A self-copying medicine is genuinely new, and no single seat - not the company’s, not the regulator’s, not the panel’s - can yet see how it plays out at scale. The people closest to it are not certain; they are betting, carefully, that a therapy that grows on its own can be kept where it was aimed. Knowing that the line was drawn by hand, and not by nature, is what lets you hold the result a little more loosely.
03 · Lab · your turn
Aim the virus
Rehearse why a self-copying cancer virus is governed by selectivity, not dose - push it too loose and the same spread that clears the tumor runs into healthy tissue.
04 · Hope · carry this
We are learning to turn even a virus into a medicine, and a drug the system twice refused still reached the patients who had nothing left to try.
More from Biotech & Longevity