The FDA Approved a Breast Cancer Drug Four Months After Its Advisers Voted No

On April 30, a panel of the FDA’s own oncology advisers sat through a day of slides on an AstraZeneca drug called camizestrant, listened to the agency’s own reviewers describe what troubled them, and voted 6 to 3 that the pivotal trial had not established a clinically meaningful benefit. The question was narrow. The vote was not close. 6 of the 9 advisers on the Oncologic Drugs Advisory Committee concluded the evidence wasn’t there.

On September 4, the FDA approved the drug anyway. It did so through the accelerated pathway, and in the same approval it conceded the thing the advisers had just voted on: it is, in the agency’s words, “not yet confirmed whether intervening at this point translates into clinically meaningful benefit.” That is the regulator’s own language, in the approval itself.

Camizestrant, sold as Etcamah, is an oral drug for a specific slice of advanced breast cancer: hormone-receptor-positive, HER2-negative disease that has picked up an ESR1 mutation, the genetic escape hatch tumors evolve to shrug off standard aromatase-inhibitor therapy. The FDA’s press office called it “the first FDA approval of a cancer therapy guided by detection of a resistance mutation” in a blood test. That is a genuine technological first. Whether it helps a single patient live longer is precisely the question the agency admits it cannot yet answer.

What the trial actually did

SERENA-6 was a clean, well-run, AstraZeneca-funded phase III study, and its design is the source of both its headline and its problem. Patients already doing well on a first-line aromatase inhibitor plus a CDK4/6 inhibitor had their blood screened every two to three months with a ctDNA assay. The moment an ESR1 mutation surfaced in the blood, before any scan showed the cancer growing and before the patient felt anything, 315 of them were randomized: keep the current regimen, or swap the aromatase inhibitor for camizestrant while staying on the CDK4/6 drug.

The switchers went longer without their disease progressing. Median progression-free survival was 16.0 months against 9.2, a hazard ratio of 0.44, a halving of the risk of progression on paper. It is a favorable curve, and to be fair to the drug the patient-reported outcomes moved the same way: time to deterioration in global health status ran roughly 23 months for the switchers against 6.4 for those who stayed put. Those numbers are not nothing, and a reader deciding whether to care should hold onto them.

Then comes the part the advisory committee could not get past. That progression-free-survival clock did not start at a scan or a symptom. It started the day a patient happened to test positive in the blood, a moment that landed on a different calendar date for every woman depending on when her mutation appeared and when she was next drawn. It is a timepoint the trial defined into existence rather than a standard oncology benchmark, and the FDA’s own reviewers pressed exactly this before the committee cast its vote. The study never tested switching early, on a blood test, against switching later, when the cancer visibly progresses. The control arm was kept on a regimen it was already destined to leave. SERENA-6 proved you can move the intervention forward. It did not prove that moving it forward is worth doing.

Why the pathway matters

None of this is exotic. It is the exact pattern accelerated approval was built to wave through and, repeatedly, has lived to regret. A surrogate endpoint that is “reasonably likely” to predict benefit gets a drug to market fast, with a confirmatory trial promised for later. Camizestrant’s “later” is now a mandated post-marketing study meant to show what April’s vote said hadn’t been shown.

The graveyard is well marked. Bevacizumab took the same route into metastatic breast cancer in 2008 on a progression-free-survival signal, and the FDA revoked the breast-cancer indication in 2011 when the survival benefit never materialized. Years later, a separate randomized trial adding it to HER2-positive disease found no improvement in outcomes either. Aducanumab, the Alzheimer’s drug the FDA cleared over its advisory committee’s objections in 2021, became the case study in what happens when the agency overrules its own experts. Oncologists have argued for years, in the same journals that publish these approvals, that accelerated approval is not conditional approval, and that a promised confirmatory trial is not a confirmed benefit. The committee that reviewed camizestrant said as much on the record. It had seen this film before.

Who the approval is built to serve

Desperate patients deserve proof that a drug extends their lives, not optimism about a surrogate, and the women this drug targets are exactly the ones a regulator owes that proof to. Camizestrant is not snake oil. It carries a boxed warning for dangerous heart-rhythm changes when combined with certain other medications, plus warnings for slowed heart rate and fetal harm, which tells you it is a serious pharmacologic agent doing serious things. The question was never whether it acts on the tumor. It is who the approval is engineered to serve.

It commits the health system to something new: ctDNA draws every couple of months for a large population of women who feel fine on first-line therapy, so a switch can be pulled the instant a molecular flag appears on a blood test rather than when a scan shows the cancer moving. It converts a “watch, and switch at progression” standard, which costs nothing extra, into a “screen continuously, and switch on a molecular flag” standard with a new branded drug and a recurring test attached, a cadence the diagnostics industry is glad to bill for. And it does all of that on the strength of a number that starts counting from a date the trial designed into being, with the overall survival data still immature, around twelve percent of events, and nowhere near telling anyone whether the strategy extends a life or merely a curve.

Richard Pazdur’s oncology center has made this bet before. The record of how those bets aged sits in the agency’s own withdrawal notices, filed under the drugs that reached the market first and never came back with the survival data to justify the trip.

Sources

1. FDA – Accelerated approval of camizestrant (Etcamah) for ESR1-mutated HR+/HER2- advanced breast cancer (Sept. 4, 2026)
2. CancerNetwork – FDA ODAC votes 6–3 against camizestrant (April 30, 2026)
3. Friends of Cancer Research – Stakeholder insights from the April 30 ODAC meeting
4. The ASCO Post – SERENA-6: early ESR1 detection and therapy switch, PFS 16.0 vs 9.2 months
5. Journal of Clinical Oncology – SERENA-6 primary results (LBA4), overall survival immature
6. Annals of Oncology – SERENA-6 patient-reported outcomes (time to deterioration in global health status)
7. JAMA Oncology – “Accelerated Approval Is Not Conditional Approval”
8. PharmacoEconomics & Outcomes News – FDA grants accelerated approval to aducanumab over advisers’ objections
9. Breast Cancer Research and Treatment – E1105: adding bevacizumab did not improve outcomes in HER2-positive metastatic breast cancer