Galleri: what the FDA panel vote means for cancer screening

September 23 update: a favorable advisory vote

GRAIL reported votes of 10–0 on safety, 6–4 on effectiveness, and 7–2, with one abstention, that benefits outweigh risks. These concerned the proposed use in adults aged 50 and older. [11] [12]

The FDA's summary identifies the greatest value in cancers without established screening. Safety support depended on adequate instructions and education. Panelists called for clear labeling and further follow-up; some support depended on removing “early” from the detection claim. Long-term patient outcomes remain unestablished. [12]

This is a substantive regulatory milestone. Advisory recommendations do not bind the FDA, which makes the approval decision. GRAIL's September 23 announcement described that decision as still ahead. [3] [11]

The NHS-Galleri findings below deserve to be read together. The negative primary result limits claims that the screening strategy has established a benefit. The stage IV finding is a reason to continue research, because fewer diagnoses of cancer that has spread could matter to patients. A personal decision also depends on what the test misses and what happens after a positive result. [1] [2] [6]

What is established, and what remains open?

• Shown: a large randomized trial reported more cancers found through screening, with favorable early-stage and secondary stage IV findings. Its planned main comparison did not show a reduction in stage III and IV diagnoses combined. [1] [2] [6]

• Not established: these results do not show that the test reduces cancer deaths or provides a general “all clear.” [1] [4]

• What would change the assessment: longer follow-up showing how screening affects deaths, serious illness, treatment and harms, together with evidence applicable to the exact test being offered. [5] [7] [9]

What the blood test looks for

Multi-cancer early detection, or MCED, means looking for signals associated with several cancers in one sample. Galleri analyzes methylation patterns—chemical markings on DNA—in fragments circulating in blood. If it detects a cancer-associated signal, it also predicts a likely source in the body to help direct investigation. Different MCED products use different measurements and algorithms; a result for one does not validate all the others. [3] [4]

The appeal is real: many cancers have no established population screening test, and drawing blood is relatively straightforward. A positive signal, however, starts a diagnostic process. Imaging, other tests and sometimes a biopsy to examine a tissue sample are needed to determine whether cancer is present. A negative result cannot rule it out. The benefit depends on that whole pathway, not just the blood draw. [3] [4]

What NHS-Galleri compared

The trial randomly assigned 142,250 adults in England to usual care or usual care plus annual MCED screening. Recruitment targeted people aged 50–77 in NHS records. Participants had three blood collections over two years, followed by approximately another year after the last collection for the reported analysis. The study tested adding MCED to existing screening. [1] [2] [9]

Both groups gave blood. Only positive cancer-signal results were returned to participants and their general practitioners; negative results were not returned. A positive result led into an NHS diagnostic pathway. This helped keep people unaware of their assigned group until a positive result required action. [5] [9]

The primary endpoint—the main result chosen in advance—was the incidence, or rate of new diagnoses, of stage III and IV cancers combined. The analysis prioritized 12 prespecified cancer types, including lung, colorectal, pancreatic and ovarian cancers. This did not mean that the test looked for only 12 cancers. Cancer deaths were a longer-term question, not the initial primary endpoint. [2] [5]

Three results that should not share one headline

GRAIL, the trial sponsor, gave these estimates in its May 2026 conference report for the 12 prespecified cancer types. Each incidence rate ratio compares the screening group with the usual-care group. A ratio of 1 means no difference; below 1 means a lower rate in the screening group. The 95% confidence interval shows uncertainty around the estimate. [2]

“26% lower” therefore refers to third-round stage IV incidence in the specified cancer group. It is not a 26% reduction in all advanced cancers, all cancers, cancer deaths or your lifetime cancer risk. Nor is it an absolute reduction of 26 percentage points. [2]

The exact estimates above come from the sponsor's conference-results report. The investigators' institution and trial website report the same main finding and direction of the secondary results. They describe the same trial; they are not independent replications. [1] [2] [6]

Why the favorable findings still matter

Stage IV generally means that cancer has spread to distant sites. Fewer diagnoses at that stage could be clinically important if they translate into better outcomes. Investigators also reported more stage I and II cancers and fewer cancers diagnosed during an emergency presentation. These are meaningful questions for follow-up, even though the main comparison was negative. [1] [6]

The trial website reported more stage III diagnoses, especially in the first round. That helps explain how stage IV diagnoses could decrease without the combined stage III–IV result decreasing. The first round can uncover cancers already present when screening begins; later rounds face a different situation. Why the pattern occurred, and how it develops with time, needs further analysis. [6]

A secondary endpoint chosen before the results is more informative than one selected afterward because it looks favorable. But testing several outcomes increases the chance of finding an apparently positive result. With the primary endpoint unmet, the sponsor's description “nominally significant” matters: the stage IV result does not carry the same confirmatory weight as success on the planned main test. [2]

Earlier diagnosis and longer life are separate questions

Moving a diagnosis earlier can lengthen survival measured from diagnosis without changing the date of death. This is lead-time bias. Screening can also find a real cancer that would never have caused harm during that person's lifetime. That is overdiagnosis, and it can lead to unnecessary treatment. Neither concept means that every early detection lacks value. [7]

The stronger comparison asks what happens to similar groups offered different screening strategies: how many people die of cancer, become seriously ill, need treatment or experience harm. NHS-Galleri's randomized design is a major strength for studying those outcomes. Investigators said mortality follow-up would continue. The reported stage results do not establish a mortality benefit, and a model estimating deaths prevented is not an observed reduction in deaths. [1] [5] [7]

What “accurate” means in this study

The May reports describe performance across repeated screening rounds and defined follow-up windows. The measures below use different denominators, so they should not be substituted for one another. [2] [8]

High specificity and a roughly one-in-two chance of a positive result being associated with a diagnosis can coexist. Cancer is relatively uncommon in people being screened without symptoms. Specificity starts with the cancer-negative observations; positive predictive value starts with positive tests. A positive result may therefore lead to investigation that does not confirm cancer. [2] [8]

The 30.7% all-cancer sensitivity also matters: many cancers diagnosed in the following year were not signaled. A negative result should not postpone assessment of symptoms or replace recommended screening. Performance varies by cancer type and stage. These estimates are tied to the study's test, population and analysis, not every person's risk. [2] [4]

Repeated testing adds another distinction: a per-test estimate is not your cumulative probability across several years. And because this trial did not tell participants about negative results, it did not directly test whether receiving a reassuring commercial report changes later behavior. That is a limit on applying the trial to routine use, not a claim that every negative report causes false reassurance. [9]

The test version matters

The FDA's September 2026 briefing distinguishes the MCED-V2 device used during the trial from the Galleri device submitted for marketing approval. The submitted device differs in how the sample is processed, how its algorithm classifies a signal and how the data are analyzed. Its performance was studied by testing stored plasma, the liquid part of blood, from selected trial participants afterward, with statistical weighting to account for their selection. [9]

That analysis helps connect evidence between versions. It is not the same as offering the submitted version to people and observing its effects on diagnosis or health. Its first-round performance objectives should not be confused with the original trial's three-round, combined stage III–IV endpoint. [9]

The sponsor's September submission listed an NHS-Galleri outcome paper in the New England Journal of Medicine as “in press.” The May 2026 results were reported at a conference. The September regulatory documents describe trial design and device differences. [2] [9] [10]

Availability and the next regulatory decision

An earlier version had been available by prescription in the United States. Availability and laboratory oversight do not themselves demonstrate fewer cancer deaths. A US regulatory development also does not establish authorization, access or payment arrangements elsewhere. The American Cancer Society information reviewed on September 22 said more evidence was needed before widespread use in people without symptoms could be recommended. [4] [9]

How Healthy Longevity Clinic experts evaluate the evidence

For someone hoping a blood draw will help prevent a late cancer diagnosis, the key question is what the test adds to established screening and what follows either result. In the NHS-Galleri assessment described above, the 52.0% positive predictive value shows why a positive signal needs investigation and does not always lead to a cancer diagnosis. The 30.7% all-cancer sensitivity means a negative result cannot be used as permission to ignore symptoms or skip other screening. Those numbers answer practical questions that “high accuracy” leaves out. [2] [4] [8]

For Healthy Longevity Clinic, the useful question after this vote is how adding Galleri could change a person’s care. A consultation should connect the exact test version and your risk context to a concrete follow-up plan: who investigates a signal, how uncertainty is handled and what care continues after a negative result. The potential to find cancers otherwise missed is a reason to consider the test carefully; the choice should include the investigations that a result may set in motion. [2] [4] [9]

Evidence that would materially strengthen the case includes fewer cancer deaths or serious illness in comparable screened groups, with acceptable investigation and treatment burdens. It would also need to apply to the test being offered. Until then, the informed choice includes the possible benefit of earlier detection and the possibility of missed cancer, unnecessary investigation or treatment. [4] [7] [9]

Plan the next step before the blood draw

First review the screening appropriate for your age, history and local guidance. If you have symptoms, seek a diagnostic assessment; do not use a screening blood test to decide whether the symptoms deserve attention. MCED should not displace established screening. [3] [4]

Before choosing the test, identify who will coordinate investigation after a positive result, what happens if the initial workup finds nothing and who follows an unresolved result. The likely cancer source may guide testing without settling the diagnosis. Discuss possible procedures, time, cost and your willingness to undergo them. The blood-test fee may not include the cost of resolving the result. [4]

A negative result should leave recommended screening and symptom assessment in place. Someone seeking certainty should understand that testing can create uncertainty as well as relieve it. [4]

Three questions for a consultation

1. Which exact test version is proposed, and what evidence connects it to the NHS-Galleri results and my age and risk profile?

2. If a signal is detected but initial investigations find no cancer, who will coordinate the next steps, and what procedures and costs might follow?

3. What established screening will I continue after a negative result, and what demonstrated health benefit would justify adding this test for me?

Common questions

Did NHS-Galleri show 26% fewer cancer deaths?

No. That figure described lower stage IV incidence in the third screening round for 12 prespecified cancer types. The primary combined stage III–IV endpoint was not met, and the reported results did not establish fewer cancer deaths. [2]

Does a positive result mean I have cancer?

No. It means a signal needs diagnostic evaluation. In the reported assessment window, 52.0% of positive results were associated with a cancer diagnosis. That study average cannot settle your individual result. [2] [4] [8]

Can a negative result replace my usual screening?

No. The test missed cancers, and performance differed across cancer types and stages. Continue recommended screening and seek assessment for new symptoms. [2] [4]

Does the favorable vote mean Galleri is FDA-approved?

No. The advisory panel recommends; the FDA decides. GRAIL's September 23 announcement described the final decision as pending. [3] [11]

Who produced the evidence?

The academic trial unit received GRAIL funding through its university. Authors of the trial-design paper disclosed GRAIL-funded university consultancies and paid scientific advisory roles. The design also included independent oversight committees. These interests are relevant when weighing observed findings, interpretations and projections. [5]