Eliza Ward: Hey. So this week kind of broke my brain a little.
Brian Reed: Yeah, mine too — three in 48 hours?
Eliza Ward: Three FDA approvals, 48 hours. September 3rd, the FDA approves Zanvastro — that's zilganersen — for Alexander disease. First disease-modifying treatment ever for AxD. And then September 4th, Etcamah gets accelerated approval for a specific breast cancer indication. Plus Simtriyo for ADHD somewhere in that same window.
Brian Reed: Hold on — Alexander disease. How many people actually have it?
Eliza Ward: Fewer than one in a million. So this trial — NCT04849741 — enrolled 54 patients. That was basically the whole accessible population. And they approved it.
Brian Reed: 54 patients. That's — okay, so what does the FDA even do with a dataset that small? Like, what's the threshold here?
Eliza Ward: That's exactly the question. And it's not unique to Zanvastro — each of these three approvals is resting on a different kind of bet. But wait, actually — let's just name what happened first before we get into what it means.
Brian Reed: Right — and the Etcamah one, the camizestrant approval, is the one that really stopped me. First cancer drug approved based on liquid biopsy results alone. That's — that's a genuinely different kind of regulatory move.
Eliza Ward: Yeah, and the camizestrant piece is genuinely wild — but zilganersen first, because that 54-patient number is doing a lot of work. Ionis Pharmaceuticals built this whole drug around one idea: GFAP mRNA is the instruction manual for making a faulty protein. Zilganersen intercepts that manual before the protein ever gets assembled.
Brian Reed: Wait, so the drug isn't cleaning up the mess — it's stopping the mess from being made in the first place?
Eliza Ward: That's exactly it. Before Rosenthal fibers ever form in the astrocytes. That's the root-cause move — not symptom management.
Brian Reed: Okay, so here's what I keep getting stuck on. The trial — NCT04849741 — measured walking speed and a gross motor composite at 61 weeks. And those improved. But like... I mean, does that mean the damage already done to the nervous system is actually reversing? Or just that it stopped getting worse? Because those are very different things for a family watching a nine-year-old count steps across a kitchen.
Eliza Ward: We don't know. That's confirmed — we don't have that answer from 61 weeks of data.
Brian Reed: Right — and actually, I want to push back gently on the idea that 54 is a shortcut. Alexander disease hits fewer than one in a million people. You can't hold this to a three-thousand-patient standard. But that's also why the aseptic meningitis signal from post-market surveillance is — it's not optional. With a trial this small, that's where you find what the trial couldn't.
Eliza Ward: And Zanvastro is a 50mg intrathecal injection. Quarterly. Into the spine. In kids. So any meningitis signal, even from social media flags, carries real weight.
Brian Reed: Four designations — Orphan Drug, Fast Track, Breakthrough Therapy, Rare Pediatric — plus a Priority Review Voucher for Ionis when it approved. That's the full incentive stack. Which is fine, the disease earns it. But none of those designations answer whether functional benefit at 61 weeks holds at year five.
Eliza Ward: That's the honest version of this approval — defensible, not proven. Post-market surveillance is load-bearing here, not a formality.
Brian Reed: Camizestrant, though — that's the one where the framing in the coverage is actually wrong. Not incomplete. Wrong.
Brian Reed: The headline is 'first cancer drug approved on liquid biopsy alone.' That part's accurate. But the coverage is treating accelerated approval like it's full approval. AstraZeneca got the green light September 4th — confirmatory data still pending. So we don't actually know yet if Etcamah works better than what's already on the market.
Eliza Ward: Wait — and there's a second layer. It's not just 'is the drug good.' It's 'is the Guardant360 CDx test good enough to define who gets the drug.' Those are separate bets.
Brian Reed: Right, but — hang on, I want to be fair here. The ESR1 mutation is treatment-emergent. It shows up during aromatase inhibitor and CDK4/6 inhibitor therapy. So tissue biopsy structurally can't catch it the way liquid biopsy can. You'd have to re-biopsy every few months. That's not practical.
Eliza Ward: I'll concede that completely — that's actually the genuinely innovative part. The Guardant360 CDx authorization makes sense because the mutation only exists in circulating tumor DNA during treatment. Tissue biopsy misses the window.
Brian Reed: So then what's the problem?
Eliza Ward: The problem is — okay, actually — if the confirmatory trial comes back and ctDNA missed clinically meaningful progression that tissue would've caught, we've defined the eligible patient population on an incomplete signal. We built eligibility criteria on a test we haven't fully validated against outcomes yet. That's not a proven first. That's a conditional one.
Brian Reed: And that question — what happens when confirmatory data arrives and who's actually watching the clock — that's the thread that runs through all three of these approvals, which we'll get to in a minute.
Eliza Ward: That clock — and who actually owns it — is the thing nobody's been specific about. Because you've got three different evidentiary logics, approved in the same 48 hours, and the FDA hasn't published a unified framework explaining how they relate. There's no document that says 'here's how a 54-patient rare disease trial and a liquid-biopsy-only oncology approval and a new mechanism without a head-to-head — here's how those standards talk to each other.'
Brian Reed: Wait, like formally? There's just... nothing?
Eliza Ward: Nothing public. Three silos. Each defensible in its own lane. But the post-market risk is — actually, this is the part that keeps me up — it's asymmetric across all three. Zanvastro's safety signals are underpowered because 54 patients can't catch a one-in-five-hundred adverse event. Etcamah's confirmatory data is still pending while the drug is already being prescribed. And for Simtriyo — Otsuka's triple reuptake inhibitor, blocking dopamine, norepinephrine, and serotonin simultaneously — real-world prescribing data is the only thing that will ever tell us whether someone stable on a stimulant actually needs to switch.
Brian Reed: That Simtriyo piece — I mean, 'first new mechanism in years' sounds meaningful. But the trial beat placebo. It didn't beat Adderall. So picture a pediatrician deciding whether to move a seven-year-old off something that's working onto a once-daily extended-release capsule — for kids six and up weighing at least 20 kilograms — based on mechanism novelty. That decision lands entirely on post-approval prescribing patterns. The FDA bet that real-world data fills that gap.
Eliza Ward: No head-to-head. That's confirmed.
Brian Reed: Right — and the part I don't get resolved by anything in the approval documents is: who's aggregating across all three simultaneously? Like, is there one office watching the Zanvastro meningitis signal, the camizestrant confirmatory clock, and Simtriyo prescribing drift — together — and asking whether the combined load on post-market surveillance is actually manageable?
Eliza Ward: That's — wait, that's the structural question. Because if the answer is no, then each approval looks fine individually and the system is still overextended.
Brian Reed: And it's not hypothetical. The FDA has pulled accelerated approvals before when confirmatory trials disappointed. So the concrete thing to watch — for camizestrant specifically — is when AstraZeneca's confirmatory data actually drops, and whether the Guardant360 CDx liquid biopsy population matches the patients who actually responded.
Eliza Ward: That's the signal. Not the approval date. The confirmatory readout — and whether the ESR1 mutation detected in blood actually predicted the outcome, or just predicted eligibility for a drug we haven't fully validated yet.
Brian Reed: And that's where I land — three genuine firsts, 48 hours apart, September 3rd Zanvastro, September 4th Etcamah. Each one defensible on its own logic. But the part I can't resolve is whether the post-market infrastructure was built for what they're now asking it to carry.
Eliza Ward: Yeah — and I don't think that question has an answer yet. The FDA made the bets. What I can't tell you is who's actually watching the clock on all three simultaneously, or how fast the system can move if one of them doesn't land.
Brian Reed: That's — honestly, that's the number I'll be watching. Not the approval date.
Eliza Ward: The confirmatory readout on camizestrant. When it comes. If the Guardant360 CDx population matches who actually responded — that's when we know whether the liquid-biopsy-only bet was earned or just... I mean, provisional. We genuinely don't know yet.
Brian Reed: And if it doesn't match — that's when we find out whether the system actually learned anything.
