IPM Take
This may become one of the most consequential screening decisions in years. But the argument cannot be reduced to “one blood draw finds cancer.” Galleri’s promise is enormous precisely because the current evidence leaves difficult questions: which cancers are found early enough to matter, how many diagnostic work-ups follow positive signals, whether mortality or late-stage disease actually falls, and who pays if the test enters routine screening.
Executive Summary
The FDA’s Molecular and Clinical Genetics Devices Advisory Committee voted in favour of GRAIL’s Galleri multi-cancer early detection test for prescription screening in adults aged 50 and older. The panel voted 10–0 that there was reasonable assurance of safety, 6–4 on effectiveness, and 7–2 with one abstention that the benefits outweighed the risks. The vote is advisory and does not constitute FDA approval. Galleri uses cell-free DNA methylation patterns to detect a cancer signal and predict where that signal may have originated. The divided effectiveness vote reflected continuing uncertainty over clinical utility, particularly because sensitivity varies substantially by cancer and stage and the NHS-Galleri trial did not demonstrate a significant reduction in late-stage cancer diagnoses after three annual rounds.
Why it matters
- Patients / advocates: A single blood test could expand screening to cancers with no established population programme, but false reassurance and unnecessary work-ups matter.
- Policymakers: Approval would force decisions on how MCED testing fits beside mammography, colonoscopy, cervical screening and lung screening.
- Diagnostics / pathology: A positive signal is the beginning of a diagnostic pathway, not a diagnosis.
- Payers: Screening millions of asymptomatic adults creates a major affordability and downstream-resource question.
The FDA panel said yes. It did not say the argument was over.
Galleri is the test cancer screening has been waiting for and worrying about at the same time: one blood draw, dozens of cancers, a molecular signal before symptoms, and an algorithm that attempts to identify where the cancer is coming from.
That is an extraordinary proposition.
It is also why the evidentiary bar should be extraordinary.
The advisory committee’s voting pattern tells the story better than any marketing line. Safety received unanimous support. Effectiveness passed only 6–4. Overall benefit-risk passed 7–2 with one abstention. The panel was not deciding whether the technology is interesting. It was deciding whether the evidence is sufficient to put it into screening.
The attraction is obvious. Most cancers still have no organised population screening programme. Pancreatic, ovarian, liver and esophageal cancers are often detected too late. A blood test capable of moving some of those diagnoses earlier could transform cancer control.
But “detects cancer” and “improves cancer outcomes” are not interchangeable claims.
Galleri’s reported specificity is extremely high, limiting false-positive signals, but overall sensitivity remains much lower and varies by cancer type and stage. More advanced cancers are generally easier to detect than stage I disease. And the large NHS-Galleri trial created an uncomfortable result: the test did not significantly reduce stage III and IV cancer incidence after three rounds of screening.
That does not make the technology useless. It makes the implementation question harder.
If FDA eventually approves Galleri, health systems will need protocols for positive results, diagnostic imaging, specialist referral, incidental findings, informed consent and what to tell a person who receives a negative result despite the test not detecting all cancers.
Then comes the political question: reimbursement.
A test intended for millions of healthy people can create enormous population benefit. It can also create enormous spending and a two-tier screening system if access begins mainly through self-pay.
Galleri may be approaching the regulatory front door. Cancer policy now needs to decide what happens on the other side.
