HEALTH
Somewhere in America this week, a patient with a rare tumour growing quietly in their gut is lying on a table while a lead-shielded syringe empties a radioactive isotope into their vein. The isotope will travel through their bloodstream, find the tumour, dock onto it like a key into a lock, and burn it from the inside.
That treatment has existed since 2018. Until last Monday, exactly one company in the United States was allowed to sell it.
That changed on 14 September, when the U.S. Food and Drug Administration approved a drug called BEXLUTRY — lutetium Lu 177 dotatate injection, made by Curium — for adults with somatostatin receptor-positive gastroenteropancreatic neuroendocrine tumours. The company says the product is already shipping. Prescribers can order it now.
And here is the part that has the pharmaceutical industry talking rather than the patient community cheering: BEXLUTRY is not a new drug. It is the same molecule, at the same dose, on the same schedule, as the product that has held the American market alone since January 2018. No new pivotal trial was run. None was required.
Meanwhile, five weeks earlier, a genuinely novel radioactive drug for the very same disease — one that ran its own head-to-head Phase 3 trial and won — was stopped dead at the FDA's door. Not over safety. Not over whether it worked. Over manufacturing paperwork and a third-party facility.
Three companies. One molecule family. One disease. And the order in which they reached patients had almost nothing to do with the science.
THE DISEASE NOBODY HAS HEARD OF — UNTIL IT IS THEIRS
Neuroendocrine tumours are the cancer that most people learn about the day they are diagnosed with one. They arise from hormone-producing cells scattered through the gut, the pancreas, the small intestine, the stomach and the rectum. Curium estimates that the gastroenteropancreatic variety — GEP-NETs — account for roughly 60 to 70 per cent of all neuroendocrine tumours.
For years they were filed under "rare" and largely ignored. That filing is now out of date. Analyses of United States cancer registry data show the incidence of neuroendocrine tumours nearly doubled between 2000 and 2021, climbing from 4.6 to 8.2 cases per 100,000 people. Among Americans over 65 the rate went from 18.0 per 100,000 in 2000 to 27.4 per 100,000 in 2021. Go back further and the trajectory is startling: the age-adjusted incidence rate rose more than six-fold between 1973 and 2012.
Some of that is better imaging finding tumours that were always there. Some of it is not. Either way, a cancer once described as an orphan disease now shows up often enough that a competitive market can exist around it.
What makes these tumours different from most cancers is their pace. Well-differentiated neuroendocrine tumours are slow. A patient may live with one for years, even decades, cycling through treatment after treatment. That changes the maths of drug development entirely. When a cancer kills in months, the question is survival. When it creeps, the question is control — how long can each option hold the line before the next one is needed, and how many options are there in total?
That is why a second supplier of a radioactive drug matters more here than the phrase "generic competition" usually suggests.
HOW YOU IRRADIATE A TUMOUR FROM THE INSIDE
The technology behind these drugs has a name that sounds like an engineering specification: peptide receptor radionuclide therapy, or PRRT.
The logic is elegant. Most well-differentiated neuroendocrine tumours plaster their surfaces with a receptor called SSTR2 — a docking port for the hormone somatostatin. That receptor is, in effect, a postal address the tumour has helpfully printed on its own front door.
So chemists build a delivery package. They take DOTATATE, a synthetic look-alike of somatostatin that binds to the receptor, and chemically bolt a radioactive isotope — lutetium-177 — onto it. Injected into a vein, the package circulates until it finds cells carrying the address. It sticks. And then the lutetium does what unstable isotopes do: it emits beta radiation, at close range, directly into the tumour.
There are two consequences of this design that patients feel immediately.
The first is that you must prove the address exists before you post the letter. A pathology report saying "neuroendocrine tumour" is not enough to qualify a patient. The tumour has to light up on a specialised scan — typically a gallium-68 dotatate PET — showing that the receptors are actually there in sufficient density. No receptor, no treatment.
The second is that the drug is delivering radiation through the bloodstream, which means the bone marrow and the kidneys are in the blast radius. The administration protocol builds in a countermeasure: an amino acid solution is infused before, during and after the dose, specifically to protect the kidneys. Patients are told to drink and urinate frequently. The standard course is four doses, eight weeks apart, to a cumulative 29.6 gigabecquerels.
This is not a pill you collect from a pharmacy. It is a radioactive medicine that requires a licensed facility, trained handling, renal protection and post-dose radiation safety management. The number of American centres that can do it is limited — and that limitation is about to become the entire story.
WHAT THE EVIDENCE ACTUALLY SHOWS
Because BEXLUTRY was approved through the FDA's 505(b)(2) pathway, it leans on evidence generated for the original product. It is worth looking hard at that evidence, because it is more nuanced than the headlines.
The pivotal study was NETTER-1, a Phase 3 trial that enrolled 231 patients between September 2012 and January 2016 — 117 randomised to lutetium Lu 177 dotatate plus octreotide, 114 to high-dose octreotide alone. Crucially, the comparison was not against nothing. It was against more of the existing treatment.
On stopping the cancer from advancing, the result was emphatic. The hazard ratio for progression-free survival was 0.21 — a 79 per cent reduction in the risk of progression or death. Median progression-free survival was not even reached in the treatment arm, against 8.5 months in the control group. A later post-hoc analysis put the figures at 25.0 months versus 8.5 months. The estimated 20-month progression-free survival rate was 65.2 per cent versus 10.8 per cent. Objective response: 18 per cent versus 3 per cent.
Then comes the part that rarely makes the press release.
The final overall survival analysis, after more than six years of median follow-up, showed median survival of 48.0 months in the treatment arm versus 36.3 months in the control arm — a gap of 11.7 months that did not reach statistical significance. The hazard ratio was 0.84, with a p-value of 0.30.
Before anyone concludes the drug does not extend life, note the detail that explains most of it: 36 per cent of patients in the control group went on to receive radioligand therapy after the trial ended. When a third of your comparison group eventually gets the drug you are testing, the survival gap narrows by design. "It did not work" and "the difference became impossible to measure" are very different statements, and the trial cannot cleanly distinguish them.
A second body of evidence comes from the ERASMUS study — single-arm, open-label, with safety data drawn from 1,214 patients and efficacy assessed in 360 with GEP-NETs. The investigator-assessed response rate there was 16 per cent. Modest. But among the 58 patients who did respond, the median duration of response was 35 months — close to three years of control from a four-dose course.
That asymmetry is the whole point. In a slow cancer, the prize is not shrinkage. It is time.
THE RISKS THAT ARRIVE YEARS LATE
The safety profile carries over from the reference product, and it is not trivial.
In NETTER-1, Grade 3–4 adverse reactions more common with the radioligand than with high-dose octreotide included lymphopenia in 44 per cent of patients, elevated gamma-glutamyltransferase in 20 per cent, vomiting in 7 per cent, and nausea and elevated AST in 5 per cent each.
The labelled warnings are blunter: myelosuppression, secondary myelodysplastic syndrome and leukaemia, renal toxicity, hepatotoxicity, hypersensitivity reactions, neuroendocrine hormonal crisis, embryo-foetal toxicity and infertility. In the long-term NETTER-1 safety data, myelodysplastic syndrome occurred in two patients — 2 per cent — in the treatment arm.
Secondary blood cancers do not show up in the weeks after a dose. They surface years later. And in a disease where patients survive a long time, that timing matters enormously. The honest question a patient and oncologist have to sit with is not "is this safe?" but "is the control I gain worth a delayed risk I may live long enough to meet?"
THE THREE DOORS: A LESSON IN HOW DRUGS ACTUALLY REACH PEOPLE
Here is where this story stops being about one approval and starts being about how American medicine works.
Over roughly six months, three different companies pursued essentially the same treatment for the same patients through three different regulatory doors. Watch the order in which they arrived.
Door one: the novel molecule. ITM Isotope Technologies Munich developed 177Lu-edotreotide, known as ITM-11 — a genuinely new agent with its own comparative Phase 3 programme, the COMPETE trial, which pitted it against everolimus in patients with inoperable, progressive Grade 1 or 2 GEP-NETs. It met its primary endpoint, showing a clinically and statistically significant improvement in progression-free survival. On 7 August 2026, the FDA issued a Complete Response Letter. The agency identified no concerns with the clinical data, the non-clinical data or the safety profile. It cited chemistry, manufacturing and controls issues, and items relating to a third-party commercial facility. The company says it intends to resubmit.
Door two: the generic. Lantheus pursued PNT2003 as a straightforward copy through the abbreviated pathway. It received tentative approval on 2 March 2026 — first past the post as a radioequivalent. But a tentative approval is not a licence to sell. It could not convert to final approval until the 30-month patent-litigation stay expired. That stay has now lapsed, and there is still no announced final approval or launch.
Door three: the branded copy. Curium filed BEXLUTRY as a branded 505(b)(2) — same active ingredient, supported by published evidence plus targeted bridging data showing a similar biological and chemical profile. Final approval. Immediate supply.
Three roads. The one carrying the newest science is stalled over a factory inspection. The one carrying the cheapest formal route is stuck in patent limbo. The one that arrived first is a copy of an eight-year-old drug — and it got there because the company that made it knew how to manufacture radioactive medicine at scale.
WHY A SIX-DAY HALF-LIFE CHANGES EVERYTHING
To understand why manufacturing dominates this field, consider a number: 6.6 days.
That is the half-life of lutetium-177. You cannot make this drug in March and warehouse it for October. Every dose is produced, shipped and administered against a clock that starts ticking the moment the isotope is made. Stockpiling is physically impossible.
In conventional pharmaceuticals, a supply problem means a drug is hard to get — frustrating, sometimes dangerous, usually survivable with substitution. In radiopharmaceuticals, a supply problem means a specific patient, booked for a specific Tuesday, cannot be dosed that Tuesday. The batch either exists or it has decayed.
Curium has positioned itself squarely on this point, describing itself as a vertically integrated manufacturer of lutetium-based neuroendocrine tumour therapy and foregrounding reliable supply at scale. Group Chief Executive Officer Renaud Dehareng called the approval "a defining milestone for Curium as we expand our offering into oncology therapeutics," and said the company intends to use its nuclear medicine infrastructure to widen access to radioligand therapy for eligible patients.
Strip away the corporate language and the claim is straightforward: with the molecule held constant, the competition is now about whether the dose shows up on the day.
There is a geopolitical shadow here too. Lutetium-177 feedstock depends on a limited set of reactor facilities. Irradiation capacity is finite. Transport networks for short-lived isotopes are few and fragile. Whoever controls that chain effectively controls who gets treated — a quiet fact that sits underneath every one of these approvals.
THE MONEY, AND WHAT IT MIGHT MEAN FOR YOUR BILL
The financial stakes are not small. The reference product generated approximately $816 million in full-year 2025 revenue, with its manufacturer citing growing demand in the United States, Europe and Japan. Together with a sister radioligand therapy for prostate cancer, that company's radioligand franchise reaches roughly $2.8 billion annually. The broader radioligand therapy market was valued at somewhere between $2.6 billion and $3.2 billion in 2025.
For patients, the relevant figure is the list price. The wholesale acquisition cost of the reference product has been reported at $55,896 per dose — and most patients need four. That puts a full course somewhere in the region of $204,000 to $224,000 for the drug alone, before scanning, infusion, facility fees and the amino acid protection protocol.
Copay assistance programmes mean many commercially insured patients pay far less out of pocket, sometimes nothing. But list prices are not a fiction. They shape what insurers pay, what uninsured and underinsured patients are billed, what hospitals budget for, and whether a treating centre decides the programme is viable at all.
This approval marks the first time same-molecule competition has existed in this market. Whether that translates into lower prices, faster scheduling, or simply two companies charging similarly is the question the next twelve months will answer. What is already true is that a single point of failure has become two.
AN EXPERT VIEW: NOVELTY IS NOT THE SAME AS ACCESS
The uncomfortable lesson buried in this week's news is one that rarely gets said out loud in patient advocacy circles.
We are trained to believe that medical progress is driven by discovery — that the best science wins, and the newest molecule reaches the most people. This sequence of events says otherwise. A drug with a fresh Phase 3 victory sits blocked. A drug that is a copy of an eight-year-old product is in patients' veins this week. The difference between them was not scientific merit. It was manufacturing maturity and the choice of regulatory pathway.
For anyone who has watched a family member wait for a treatment that "exists" but cannot be obtained, that will not come as a surprise. An identical indication on a label delivers precisely nothing without the supply and the sites to match it. A drug you cannot be dosed with is, from where the patient is lying, not a drug at all.
The same structure is settling in worldwide. The reference product is approved in Japan, where its maker cites the country as a growth driver. A PSMA-targeting radioligand therapy received Chinese regulatory approval in late 2025, bringing that class into the domestic market there for the first time, as China shifts from importer toward producer of radiopharmaceuticals. Across the United States, Europe, Japan and China alike, uptake of these therapies is governed less by whether a regulator has said yes than by whether the isotope arrives on the day it is needed.
WHAT THIS MEANS FOR YOU
If you or someone close to you has a neuroendocrine tumour: There are now two approved sources of lutetium Lu 177 dotatate in the United States for adults with SSTR-positive GEP-NETs. Ask your oncology team directly whether this changes scheduling at your treating centre — the practical benefit of a second supplier shows up as appointments that do not get postponed.
Note the age gap in the labels. The reference product's indication was expanded in 2024 to include patients aged 12 and over. BEXLUTRY's current indication is for adults only. For a paediatric or adolescent patient, that distinction is not academic.
Confirm the receptor first. This treatment only works if the tumour expresses somatostatin receptors. If nobody has mentioned a gallium-68 dotatate PET scan, ask why.
Ask about the long tail. Myelodysplastic syndrome and leukaemia are labelled risks that appear years after treatment. In a slow-moving cancer, you may well be around to encounter them. Ask what monitoring is planned, and for how long.
Ask about cost before, not after. Ask what your insurer will pay, what the facility fee is, whether copay assistance applies to the specific product being used, and — a new question as of this week — whether more than one manufacturer's product is available to you.
If you are simply a taxpayer or a bill-payer: watch what happens to the price. This is the first genuine test of whether competition can move the needle on a class of cancer drugs that has never had any.
THE BOTTOM LINE
Nothing about how this treatment works changed on 14 September. The molecule is the same. The dose is the same. The four infusions, the amino acid drip, the eight-week gaps, the radiation-safety briefing — all the same.
What changed is that the United States went from one supplier of a radioactive cancer medicine to two, and it happened because a company that knows how to make isotopes filed the right kind of paperwork, while the company with the newer science filed the wrong kind.
For patients, that is probably good news — a second source, a possible price ceiling, and one fewer single point of failure in a supply chain that measures its margin for error in days.
But it is worth holding onto the shape of it. The system that decides which cancer treatments reach real people is not, in the end, a competition of ideas. It is a competition of factories, licences, transport schedules and legal filings. The best molecule does not automatically win. This week, it did not even place.
