The Cancer Test Everyone's Watching Just Got a Split Verdict

On September 23, ten voting members of FDA's Molecular and Clinical Genetics Devices Panel did something that had never been done before: they evaluated a premarket approval application for a multi-cancer early detection test. The product was GRAIL's Galleri, a blood test designed to screen adults 50 and older for signals of more than 50 cancer types from a single draw, and the panel's verdict was less a single answer than three separate, increasingly divided ones. Unanimous, 10-0, that the test is safe. A tighter 7-2, with one abstention, that its benefits outweigh its risks. And a genuinely split 6-4 that the clinical evidence supports its effectiveness.

That's not the clean green light a company two decades and hundreds of millions of dollars into building a category-defining product would hope for, and it's not a rejection either. It's something more interesting: a panel of independent experts looking at the same data and reaching real, documented disagreement on the single question of whether a screening test designed for population-wide use is actually ready for it.

Why the safety vote and the efficacy vote diverged

The gap between a unanimous safety vote and a divided effectiveness vote isn't a contradiction — it reflects two genuinely different kinds of evidence. Safety, for a blood draw and a diagnostic algorithm, is a comparatively bounded question: does the test itself cause direct physical harm. A simple blood draw with no therapeutic intervention attached has a fairly constrained harm profile, which likely explains the unanimous vote. Effectiveness for a multi-cancer screening test is a fundamentally harder evidentiary question, because it isn't really asking "does this test detect cancer" — GRAIL has never disputed that Galleri detects cancer signals. It's asking whether detecting those signals, at scale, across a general screening population, produces a net clinical benefit once every downstream consequence is accounted for: the true positives that lead to earlier, more treatable diagnoses, but also the false positives that trigger unnecessary invasive follow-up procedures, the false negatives that create false reassurance, and the genuinely unresolved question of lead-time bias — whether earlier detection actually extends life or just extends the period during which a person knows they have a disease that would have been detected anyway, on a similar timeline, through existing screening and symptomatic presentation.

GRAIL's application leaned on real scale to answer that question: the U.S. PATHFINDER 2 study, with roughly 25,500 participants, and the U.K. NHS-Galleri trial, with more than 70,000 participants in its intervention arm — described as the largest randomized controlled study of a multi-cancer early detection test in its intended population. That's a genuinely substantial evidence base, and it's worth being precise about what the 6-4 vote does and doesn't say about it: a panel split roughly six-to-four isn't a panel that found the evidence inadequate. It's a panel that, even looking at a large, randomized dataset, couldn't reach the kind of confident consensus that makes an eventual FDA approval decision straightforward to predict.

What a split advisory vote actually means for the path forward

FDA is not bound by advisory committee recommendations, and the agency has diverged from panel votes before in both directions — approving products a panel voted against, and declining to approve products a panel endorsed. But panel votes function as a genuine signal of where the agency's own internal debate is likely to land, precisely because FDA reviewers and panel members are often working through the same underlying tensions in the data. A 6-4 vote on effectiveness suggests FDA's final decision, expected in the coming months, is genuinely live rather than a formality following a strong advisory endorsement — and it suggests that whatever the agency ultimately decides, a meaningful share of the same expert community that reviewed the data would have reasonable grounds to view the opposite outcome as also defensible.

For a product category this new — the first PMA ever reviewed for a multi-cancer early detection test — that ambiguity matters beyond GRAIL specifically. Galleri's advisory committee outcome is likely to function as a template for how FDA evaluates the next several MCED tests working their way through development, several of which are watching this specific PMA closely as a preview of the evidentiary bar they'll eventually face. A split vote here doesn't just shape GRAIL's own approval odds; it shapes what "sufficient evidence" looks like for an entire emerging category of screening technology, at a moment when several other companies are actively designing their own pivotal trials around whatever standard this review ultimately sets.

The population-screening stakes underneath the regulatory question

It's worth being explicit about why this particular evidentiary debate carries more weight than a typical PMA review. Most PMA decisions govern a device used in a defined clinical context, on patients already identified as needing it. A multi-cancer screening test aimed at the general population 50 and older is a different kind of product: if approved and adopted at scale, it doesn't just affect the relatively small number of patients enrolled in pivotal trials — it potentially reshapes screening guidelines, insurance coverage decisions, and clinical workflows for a very large share of the adult population, most of whom will never actually have cancer, and for whom the test's false-positive rate, not its true-positive rate, will be the more common personal experience. A false-positive rate that looks acceptable in a controlled trial population can translate into a very large absolute number of unnecessary follow-up procedures once applied across tens of millions of routine screening visits, which is precisely the kind of population-scale consequence an advisory panel's efficacy vote is trying to weigh, and precisely why a split vote on this specific question is more consequential than it would be for almost any other device category.

What this means for companies watching from the sidelines

For the broader MCED and liquid biopsy field, the practical takeaway isn't to wait passively for FDA's final Galleri decision before acting. Companies with their own multi-cancer or single-cancer liquid biopsy programs in active development should study the specific evidentiary gaps the split vote implies — likely centered on lead-time bias, overall survival data maturity, and the false-positive follow-up burden — rather than assuming their own eventual submission will be judged purely on detection sensitivity. Health systems and payers watching this space for coverage-policy planning purposes should treat the vote as a signal that FDA's eventual decision, whenever it lands, is likely to come with real nuance in its indication language and postmarket study commitments, rather than a simple, unqualified approval — meaning coverage and clinical-guideline decisions built on an assumption of a clean approval may need revisiting once the actual decision, and its specific conditions, are public. And for any organization already citing Galleri's advisory outcome in strategic planning, it's worth being precise internally about which of the three votes is actually driving that plan: unanimous safety and a split efficacy verdict are very different signals to build a five-year strategy around.

Sources: GRAIL, Inc., FDA Advisory Committee Votes In Favor of Approval of GRAIL's Galleri Multi-Cancer Early Detection Test, September 23, 2026; Oncology Nursing News; U.S. Food and Drug Administration, Molecular and Clinical Genetics Panel of the Medical Devices Advisory Committee Meeting Announcement; Dealroom; AllSci.