The Galleri blood test finds deadly cancers earlier, and the effect grew each year in a 142,000-person trial

Evidence check, September 27, 2026

The claim: "A yearly Galleri blood test can catch deadly cancers early, before symptoms."

Our grade: Promising. In the first randomized trial, 142,000 people in England, the test found four times as many cancers through screening and fewer people were diagnosed at stage 4. It missed its main goal, it misses about two in three cancers, and no one knows yet if it saves lives, so use it on top of regular screening.

You're 55. You get your mammogram or PSA test, your colonoscopy is up to date. Then a friend mentions a blood test that looks for more than 50 cancers at once, including pancreatic and ovarian cancer, which nothing else screens for.

Does it actually find those cancers in time to make a difference?

That's the question that matters. Most cancers found at stage 1 or 2 can be cured with surgery or radiation. Most found at stage 4, after they've spread, can't. This week the first big randomized trial of the Galleri test was published in the New England Journal of Medicine, and an FDA advisory panel voted on it. The trial was built to answer exactly this question.

Where the claim came from

Galleri is made by GRAIL. It reads tiny bits of DNA that tumors shed into the blood. Cancer DNA carries a pattern of chemical tags (methylation) that normal DNA doesn't, and the test can often tell which organ it came from. It costs $949, or $799 ordered online, and most insurance doesn't cover it [5, 7].

The idea took off with the longevity crowd. Peter Attia has made the case for screening more widely than the guidelines ask, saying the benefit of cancer screening has been "pretty clear and consistent" [1]. The reasoning is simple. Today's recommended tests cover only a few cancers: breast, colon, cervix, lung for heavy smokers, and prostate by choice. The cancers that kill most people, like pancreas, liver, esophagus and ovary, often aren't found until they cause symptoms.

Clue one: it finds cancers nothing else looks for

PATHFINDER 2 gave the test to 35,878 adults aged 50 and up in the US and Canada, with no signs of cancer, and followed them for a year [4]. It wasn't randomized. It measured what the test finds.

Of the cancers the test caught, 73% were types with no recommended screening test. More than half (53%) were at stage 1 or 2. Adding Galleri to the usual screening found about 6.5 times as many cancers through screening as usual screening alone.

That's the promise in one number. These are cancers people usually learn about only when something hurts.

Clue two: when it says "cancer", it's usually right

A screening test that cries wolf does harm. False alarms mean scans, biopsies and fear.

Galleri rarely cries wolf. Fewer than 1 in 200 people without cancer got a positive result (specificity 99.55% to 99.64%) [3, 4]. When the test did say "cancer signal detected", a cancer was confirmed in 52% of cases in the English trial and 60% in PATHFINDER 2. When it named the organ, it was right about 9 times in 10 [3, 4].

Clue three: the big trial, and why the headline says "failed"

NHS-Galleri is the test that counts. About 142,000 people aged 50 to 77 in England gave blood once a year for three years. Half, chosen at random, had their blood tested; the other half had it stored and got usual care [2]. The main goal was fewer cancers found at stage 3 or 4, across 12 deadly types.

That goal wasn't met. Stage 3 and 4 cancers came out about even: 706 in the tested group, 688 in the usual-care group [3]. That's why the BMJ ran the headline "Galleri cancer test fails to meet main aim" [5].

But look at the years one by one. In the first year, the test found more late-stage cancers, not fewer. That's expected. A first screen catches cancers that were already growing, and some had already reached stage 3. The early catch shows up in later years, once the test is looking at new cancers.

And that's what happened. Compared with usual care, stage 4 diagnoses were 9% lower in year one, 22% lower in year two, and 26% lower in year three [3].

That's the punchline. The test did what screening is supposed to do, and the effect grew each year.

Over three years, stage 4 cancers fell 14% overall (342 against 397), just at the edge of what statisticians count as a real result [2, 3]. Stage 1 and 2 cancers rose 16% (647 against 559). Four times as many cancers were found through screening (1,173 against 290). And 21% fewer people learned they had cancer in an emergency room, which is often the worst way to find out [3]. There were no serious side effects from the test or the follow-up [2].

So back to your blood draw. Yes, it finds some of the deadliest cancers early, often before any other test would, and the benefit built over the three years.

Our grade

The case for the other side is real. The test missed about two in three of all cancers found within a year (sensitivity 30.7%). For the 12 deadliest types it was better, about 55% [3]. So a negative result doesn't mean you're clear. The trial missed its main goal. Both studies were paid for by GRAIL. And finding a cancer early only helps if treatment then changes the outcome. At the FDA panel, one expert who voted against effectiveness put it this way: for patient outcomes, "we don't have that data" [5]. The panel still voted 6 to 4 that the test works and 7 to 2 that its benefits outweigh its risks [6].

What would raise the grade: fewer cancer deaths in the NHS-Galleri follow-up, or in REACH, a GRAIL study of about 50,000 Medicare patients [6]. What would lower it: stage 4 reductions fading with longer follow-up.

What this means for you

• Keep your regular screening. Mammograms, colon screening, cervical screening, and a lung CT if you smoked. Most of the cancers Galleri missed were types those tests already cover, like early breast and prostate cancer [4]. It adds to them. It doesn't replace them.
• If you're 50 or older and can afford it, a yearly Galleri test is a reasonable add-on. The trial's benefit came from testing every year, not once.
• Treat a negative result as "no signal found", not "no cancer". New symptoms still need a doctor.
• If you get a positive result, the next step is imaging of the organ it names. About half of positives turn out to be cancer, so don't panic, but don't wait.
• Ask your doctor whether your insurance covers it. Most plans don't, but HSA and FSA money can often be used [7].

What's next

The FDA decides whether to approve Galleri in the coming months. The NHS-Galleri team is extending follow-up to see whether the stage 4 drop holds and grows, and later whether fewer people die of cancer. GRAIL's REACH study in Medicare patients is running.

In our next cancer screening check: a new analysis of all eight big mammogram trials says overdiagnosis (finding a cancer that never would have hurt you) is under 5%, not the 30% to 50% often quoted.

Sources

1. Attia P. AMA #56: Cancer screening: pros and cons, screening options, interpreting results, and more. The Peter Attia Drive, February 12, 2024. https://peterattiamd.com/ama56/
2. Sasieni P, Johnson P, Round T, et al. Effect of screening with multicancer early-detection test on late-stage cancer diagnosis. N Engl J Med, 2026. About 142,000 participants aged 50–77, randomized, three yearly screens, industry-funded (GRAIL). https://doi.org/10.1056/NEJMoa2505723
3. NHS-Galleri: primary results from a randomised controlled trial. J Clin Oncol 44 (17 suppl): LBA100, 2026 (ASCO abstract). Same trial as [2], full stage and performance numbers, industry-funded (GRAIL). https://ascopubs.org/doi/10.1200/JCO.2026.44.17_suppl.LBA100
4. Nabavizadeh N, et al. Performance and safety of a multi-cancer early detection test: the PATHFINDER 2 study. Nat Med, 2026. 35,878 participants aged 50+ (32,007 analyzed), single-arm, 12 months, industry-funded (GRAIL). https://www.nature.com/articles/s41591-026-04618-w
5. Mahase E. Galleri cancer test fails to meet main aim in NHS trial, but FDA still endorses it. BMJ, 2026;394:e100975. News. https://doi.org/10.1136/bmj-2026-100975
6. Ault A. FDA panel votes on Grail's multicancer test, highlights data gaps. Medscape, September 25, 2026. News. https://www.medscape.com/viewarticle/fda-panel-votes-grails-multicancer-test-highlights-data-gaps-2026a1000zvy
7. GRAIL. Galleri pricing and payment options. galleri.com, accessed September 27, 2026. https://www.galleri.com/patient/the-galleri-test/cost