Galleri Found More Cancers, but Not Fewer Late Ones

In 142,250 people in England, about 1 in 100 per group got a late-stage cancer of 12 types in 3 years. Half of positives were cancer; it missed 7 in 10 cancers.

---

Galleri is a blood test, sold to healthy adults, that looks for signs of many cancers at once. Its maker, GRAIL, paid for the largest randomized trial of such a test (15). It ran inside England's NHS, with a Cancer Research UK trials unit (1). The trial was built to answer one question: with the test added, are fewer people diagnosed with cancer at a late stage?

Cancer stage runs from I to IV. Stage I is small and local. Stages III and IV have grown or spread beyond where they started, and stage IV has reached distant parts of the body.

The answer was no (1). Over about three years, 706 people offered the test were diagnosed with a stage III or IV cancer of 12 types that cause most cancer deaths (1). So were 688 people in the comparison group, whose blood was stored but not tested (1). That is about 1 in 100 people in each group. The trial's own website says it "did not meet its main goal" (11).

The test did find more cancers, and more of them early (1). Fewer people in the tested group had stage IV cancer, but more had stage III, and the rise outweighed the fall (1). The stage IV drop is the number GRAIL now leads with. It was a key secondary endpoint, set in advance, but the trial's fixed order of tests stopped at the primary endpoint, so it was not formally tested (1). On its own it cannot show the test works; the authors say it suggests "the possibility" of fewer stage IV diagnoses in later rounds (1).

About half of positive results were cancer (2). The authors note that is higher than for urgent NHS referrals for suspected cancer, where about 6 in 100 find one (2).

The bigger costs lie elsewhere. About 7 in 10 people diagnosed with cancer within a year of a test had tested negative (2), and a false alarm can mean months of scans (4). Whether the test means fewer cancer deaths is not yet known (1). The FDA has not approved it (17).

Archival Photograph

The Trial Was Built to Answer One Question

Most screening looks for one cancer at a time: a mammogram looks at the breast, and the bowel kit looks for hidden blood in a stool sample.

Galleri reads chemical marks, called methylation patterns, on fragments of DNA that tumours shed into the blood (2). If it finds a cancer pattern, it also predicts where in the body the cancer started (2). GRAIL says it looks for more than 50 cancer types, including pancreatic, ovarian, and liver and bile duct cancers, which have no screening test (14).

The NHS-Galleri trial invited 1,496,311 people aged 50 to 77 in eight regions of England (1). Of those, 142,250 gave a first blood sample and were randomized (1). None had been diagnosed with or treated for cancer in the previous three years (1).

Half, 71,122 people, had their blood tested, and positive results were sent back (1). The other 71,128 formed the comparison group: their blood was stored, untested (1). Both groups gave blood three times, a year apart, and around 9 in 10 came back each time (1). Negative results were not returned (1), so no one in the trial was told they were clear.

Everyone kept usual NHS care, including screening, and 71.8 percent of those offered the test said they attended every screening they were invited to (1). So the trial tested Galleri on top of normal screening.

The primary endpoint is the one result a trial is built and sized to test. Our article on why rapamycin's first longevity trial is judged on its primary endpoint explains why that choice carries so much weight.

Here, it was the rate of stage III or IV cancer across 12 types (1). They are lung, head and neck, bowel, pancreas, myeloma, liver and bile duct, stomach, oesophagus, anus, lymphoma, ovary and bladder. The paper says they cause more than 60 percent of cancer deaths (1).

The designers expected about 20 percent fewer stage III or IV cancers of those types (1). About 140,000 people were enough to detect a drop that size at least nine times in ten, if it was real (1). Only if this first result was clear would the trial go on to test broader groups of cancers (1).

The Main Result: 706 Against 688

Stage III or IV cancer of the 12 types was diagnosed in 706 people offered the test and 688 in the comparison group (1). The rate ratio was 1.03 (1).

A rate ratio compares how fast events piled up in the two groups, not a count of people. 1.0 is the no-difference point. 1.03 means these cancers turned up about 3 percent more often in the tested group, a gap small enough to be chance.

The range of values compatible with this data ran from 0.92 to 1.14 (1). The P value was 0.63 (1). That is the chance of seeing a gap at least this large if there were genuinely no difference.

The range says something stronger too. Its lower end, 0.92, means the data allow a drop of at most about 8 percent. The 20 percent drop the trial was designed around sits well outside it. The trial ruled out the large effect it was hoped to show (1).

Because this first step failed, no broader group was formally tested (1). Those groups showed no detectable difference either (1).

The authors explain that their planning models "projected a larger and quicker stage shift than observed" (1).

More Stage III, Fewer Stage IV

The flat result hides two movements. Counting each person once at their highest stage, 360 people in the tested group had a stage III cancer of the 12 types, against 287 in the comparison group (1). Stage IV went the other way: 341 against 394 (1).

Per 10,000 people over the trial, that is about 10 more with stage III and about 7 fewer with stage IV. This person-by-person breakdown was done after the fact (1).

Most of the extra stage III came in the first round (1). The first screen picks up cancers that have been growing for a while. In round one, stage III or IV cancers ran 1.19 times as often in the tested group, range 0.98 to 1.43 (1). In rounds two and three the ratios were 0.95 and 0.88, both ranges including 1.0 (1).

The first two of those round figures were done after the fact. All three are descriptive, meaning reported without a formal test (1).

The authors offer a reading that favours the test. The extra stage III cancers may be ones still silent in the comparison group, which "may later present at stage III or IV" there (1). Cancer counts in the two groups have "not yet converged", they write (1).

The stage IV figure

Stage IV alone was a prespecified key secondary endpoint: planned in advance and reported, but under the trial's plan not formally tested (1). Its rate analysis uses slightly different counts, 342 against 397, and gives a rate ratio of 0.86, about 14 percent fewer (1). The range ran from 0.74 to 1.00, so its top touches 1.00, no difference (1).

Leading with a result that was not formally tested is a pattern we found in rapamycin's PEARL trial too. The paper is clear about how to read it. Secondary results, "including stage IV cancer incidence reduction, were interpreted descriptively" (1). Their ranges "should not be used for hypothesis testing" (1).

A result is called statistically significant when a gap that size would be unlikely by chance under the usual test. GRAIL's release of 30 May 2026 called the stage IV result "nominally statistically significant" (14). Nominally is the right word for a number that passes that test alone but was not one of the trial's planned tests.

What the label leaves out is the order of the evidence: the result the trial was built to test found no detectable difference.

To be fair to the test, the stage IV gap widened each round. For the 12 types, the ratio was 0.91 in round one, 0.78 in round two and 0.74 in round three (1). A lower ratio means fewer stage IV cancers in the tested group. The round three range, 0.57 to 0.95, sits below 1.0 (1). Across all routinely staged cancers, rounds two and three gave 0.78 and 0.80 (1).

That is what you would expect if the test were starting to catch cancers before they spread. It is also descriptive, untested, and drawn from small numbers (1). The paper says "Further follow-up is planned" (1), and GRAIL said in February that it plans to extend follow-up by 6 to 12 months (13).

Cancer by cancer

Counts by cancer type were planned in advance but are small and untested, and the trial "was not powered to assess clinical utility" within single cancer types (1). Lung cancer shows the swap. Stage IV was lower in the tested group, 99 against 123, and stage III higher, 101 against 58 (1).

Stage IV oesophageal cancer fell, 19 against 38 (1). Pancreatic cancer showed no drop: 59 against 43 with stage IV (1). With numbers this small, that gap may be chance.

Does fewer stage IV mean fewer deaths?

That depends on the cancer. A correlation measures how closely two things move together, on a scale from minus 1 to 1; near 1 means they rise and fall together. A 2024 review of 41 screening trials compared drops in late-stage cancer with drops in deaths (5).

The correlation was 0.99 for ovarian and 0.92 for lung cancer, but 0.39 for bowel and minus 0.69 for prostate (5). Late stage "may be a suitable alternative end point" for some cancers, "but is not suitable for others", the authors concluded (5).

A 2025 review of 57 trials raised a worry specific to blood tests. Cancers that shed more DNA may be the more aggressive ones, so fewer late diagnoses might not mean fewer deaths (6).

The NHS-Galleri design paper gives its own warning. In a large UK ovarian screening trial, a 10.2 percent drop in late-stage cancer shrank to a 4.4 percent drop in deaths, and neither was statistically significant (3).

The UK National Screening Committee, which advises the NHS, calls a fall in late-stage cancer "the most promising surrogate outcome", a stand-in measured sooner than deaths (7). It also says such stand-ins "cannot yet replace mortality as the definitive outcome" (8). Our article on epigenetic clocks covers the same problem with lab scores that stand in for aging.

What the Test Did Find

Over the trial, 3,637 people in the tested group were diagnosed with a cancer of any type, against 3,400 in the comparison group (1). That is about 51 against 48 per 1,000.

Stage I or II cancers of the 12 types ran 647 against 559, in an analysis done after the fact (1). The authors write that the extra early cancers are "likely due to earlier detection of clinically significant disease" (1).

Screening of any kind found 1,173 cancers in the tested group, 32.3 percent of its cancers (1). In the comparison group, usual NHS screening found 290, or 8.5 percent (1). That four-to-one gap is the source of GRAIL's "four-fold" claim. It measures how cancers were found, not how they turned out.

Fewer cancers were found outside any screening, the usual way: 2,464 against 3,110 (1).

Fewer people were diagnosed after an emergency: 225 against 286, a 21.3 percent drop, or about 32 against 40 per 10,000 (1). That comparison was reported without a formal test (1). GRAIL's releases cite 25 percent (14, 17); that figure adds a review, requested by GRAIL, of 39 cases from the tested group only (1).

Finding more cancers can mean catching dangerous ones sooner. It can also mean overdiagnosis: finding cancers that would not have caused harm in a person's lifetime, then treating them. The trial cannot yet tell the two apart, and that analysis "will be published separately" (1). An analysis of 43 screening trials, cited by the UK screening committee, found that finding more cancers did not line up with fewer deaths (8).

What a Result Means

A positive result

Across three rounds, 197,146 tests gave 1,801 positive results, about 9 in every 1,000 tests (2). Of those, 937 turned out to be cancer and 864 did not (2). So about half of positives were cancer: 52 percent overall, falling from 58 percent in round one to 46 percent in round three (2).

Put the other way, 99.55 percent of tests in people without cancer came back negative, so 0.45 percent were false alarms (2). The authors note that single-cancer screening tests have false-alarm rates of 0.98 to 19.9 percent (2).

Any test run on healthy people flags some of them; our guide to your body's biomarkers shows how a lab's normal range does the same.

When the test found a cancer, its first or second guess at where it started was right 92.5 percent of the time (2). Of 303 people with a false alarm in round one, 54 were diagnosed with cancer in the next two rounds, and 44 of those cancers matched the site the test had predicted, in an analysis done after the fact (2).

The real cost of a false alarm is the search that follows, through the NHS's urgent referral route (2). The trial did not systematically record it: details of the workup "were neither prospectively specified nor systematically ascertained" (2). Results on procedures, radiation and anxiety "will be published separately" (1).

Inside the trial, side effects linked to taking part were 371 against 321, none serious; the excess was mainly anxiety after a positive result (1).

The clearest picture of the search comes from PATHFINDER, an earlier US study of 6,621 people tested with an earlier version of Galleri (4). Of 92 positive results, 57 were false alarms (4). Ruling cancer out took a median of 162 days, meaning half of those 57 waited longer than about five months (4). Confirming a cancer took a median of 57 days (4).

Of the 57 people with false alarms, 53 had imaging and 17 had a procedure; one had a testicle removed (4). The study's sponsor paid for the scans and tests (4).

A negative result

A negative result is weaker reassurance. Of 3,051 people diagnosed with cancer within 12 months of a test, 937 had tested positive, about 3 in 10 (2). The other 2,114, about 7 in 10, had tested negative (2). For the 12 types that cause most cancer deaths, it caught between 47.6 and 63.4 percent, depending on the round (2).

It varied widely by cancer: 32 of 36 liver and bile duct cancers, but 4 of 157 melanomas (2). Of the cancers it missed, 713 were prostate, 301 breast and 130 bowel (2). Adding the ones it flagged gives the totals, by our sums. It missed 301 of 357 breast cancers and 130 of 316 bowel cancers diagnosed within a year of a test (2).

The paper says a negative result "should not be interpreted as ruling out any particular cancer" (2). It adds that testing of this kind "is intended to complement, not replace, existing screening programs" (2).

How It Is Sold

19 February 2026. GRAIL's headline read: "Landmark NHS-Galleri Trial Demonstrates a Substantial Reduction in Stage IV Cancer Diagnoses, Increased Stage I and II Detection of Deadly Cancers, and Four-Fold Higher Cancer Detection Rate" (13). The missed primary endpoint was in the subheading (13).

20 February 2026. Patient Care Online reported "NHS-Galleri Data Show Fewer Stage IV Cancers, but Primary Endpoint Missed" (12).

30 May 2026. GRAIL's release for the ASCO cancer meeting opened with a subheading saying the combined primary endpoint showed no reduction (14). The body called the stage IV result "nominally statistically significant" (14). Its "About Galleri" text calls Galleri the only test of its kind "clinically proven through a randomized controlled trial" to find cancer earlier and "reduce Stage IV diagnoses" (14).

18 August 2026. The Health Suite, a private clinic in Leicester, announced it would offer the test in the UK (23). Its announcement said the trial missed its primary goal (23).

4 November 2026. GRAIL's third-quarter results cited "a substantial reduction of >20% in Stage IV diagnoses of 12 prespecified aggressive cancers" (17). That fits rounds two and three. Across all three rounds, the drop was 14 percent (1).

Clinic pages, read 7 November 2026. Human Longevity says GRAIL's US studies "approximately doubled the number of cancers detected", then "most in early, more treatable stages" (20). We did not check that against the US studies. In NHS-Galleri, 39.1 percent of test-found cancers were stage I or II, and 68.3 percent were stage I to III (2).

Superpower says it screens "over 50 types of cancer at their earliest, most treatable stages" (22). The trial did not measure how well it catches cancers by stage (2).

What it costs

Clinic prices, read on 7 November 2026: Human Longevity lists "$1,200 List Price" and "$798 Exclusive Price" (20). Mito Health, through Quest, charges $890.57 for non-members and $636.12 for members (21). Superpower lists "$699+$199 membership" (22).

GRAIL started a promotional pricing program in September 2026 "ahead of the FDA advisory committee meeting" (18). We could not open GRAIL's own price page, and found no UK price.

Where It Stands, and What Would Change the Verdict

United States. Galleri has not been approved. GRAIL says it "has not been cleared or approved by the Food and Drug Administration" (17). It is sold as a laboratory developed test, one a single lab designs and runs itself (18).

GRAIL applied for FDA approval on 29 January 2026 (15). The application rests on PATHFINDER 2, a US study, and on first-round data from NHS-Galleri's tested group (15, 18). A third study checks that a newer version of the test performs like the earlier one (15, 18). GRAIL says it also includes NHS-Galleri results, including the drop in stage IV diagnoses (13).

An FDA advisory panel met on 23 September 2026, the day after the main trial paper went online (1, 16). According to GRAIL, it voted 7 to 2, with one abstention, that the benefits outweigh the risks (16). It was unanimous on safety and split 6 to 4 on effectiveness (16). GRAIL expects a decision "by early first quarter of 2027" (18).

The version on sale since late 2024 is "an updated version of Galleri", GRAIL says, so it is not exactly the one the trial tested (18).

United Kingdom. The test is not available on the NHS. On 29 May 2024, GRAIL reported that NHS England would wait for the final results (19). NHS England's words, as GRAIL quoted them: "the data so far do not support moving at such a fast pace" (19).

GRAIL now says the results "have met certain success criteria and missed others", and a joint steering committee will discuss deployment "if at all" (18). The UK National Screening Committee has set out the evidence it would need first, and notes that no country has yet brought these tests into an organised screening programme (7).

What would change it. Results on cancer deaths. The trial was not sized to detect a difference in overall cancer deaths. An earlier planned comparison of deaths has not been reported yet (9). The main death analyses come 3 and 6 years after the final visit (1).

The final blood samples were taken in July 2024 (19), so by our estimate the main death analyses fall around 2027 and 2030. The registries list cancer deaths followed for as long as 8 years, with completion estimated for January 2031 (9, 10).

A missed primary endpoint next to a hopeful secondary result is a pattern we have met before. It turned up in a fracture screening trial in our bone health article, and in trials of NMN, urolithin A and other supplements.

The Scorecard

Each row reads tested group against comparison group, from NHS-Galleri unless noted.

What	Result	How sure
Stage III or IV, 12 types (the primary endpoint)	706 against 688 people; rate ratio 1.03, range 0.92 to 1.14 (1)	No detectable difference; a drop bigger than about 8 percent ruled out
Stage IV, 12 types	341 against 394 people (after the fact); rate ratio 0.86, range 0.74 to 1.00 (1)	Planned key secondary result; descriptive, not tested
Stage III, 12 types	360 against 287 people; about 10 more per 10,000 (1)	After the fact
Cancers diagnosed, any type	3,637 against 3,400 people (1)	More found; not formally tested; overdiagnosis not yet measured
Emergency diagnoses	225 against 286 people (1)	Descriptive, not tested
Positive results that were cancer	937 of 1,801, about half (2)	In this trial
False alarms in people without cancer	0.45 percent of tests (2)	In this trial
Cancers flagged within a year of a test	937 of 3,051, about 3 in 10 (2)	In this trial
Follow-up after a false alarm	Not systematically recorded; PATHFINDER median 162 days (4)	Not known for this trial
Cancer deaths	Not reported (1)	Main analyses around 2027 and 2030, our estimate
FDA approval	Not approved; decision expected early 2027 (17, 18)	Pending

What You Can Do Now

None of this is a reason to change your care on your own. It is a set of numbers to weigh and questions to ask.

Keep your usual screening: breast, cervical and bowel, and lung screening if you are at high risk and eligible (2). NHS-Galleri tested the blood test on top of these, not instead of them (1). It missed 301 of 357 breast cancers diagnosed within a year of a test (2).

If you are offered the test, weigh these numbers from the trial (2):

• About 9 in 1,000 tests come back positive.
• About half of those positives are cancer.
• Of every 10 people diagnosed with cancer within a year of a test, about 3 had a positive result first.

Take these questions to the clinic's doctor, and tell your own doctor if you have the test:

1. If my result is positive, who arranges the scans, who pays, and how long might it take to rule cancer out?
2. If it is negative, what does that tell me? In the trial, about 7 in 10 cancers diagnosed within a year had a negative test first (2).
3. Which version of the test is this, and has the FDA approved it? As of 4 November 2026, it had not (17).
4. Has any trial shown it lowers cancer deaths? Not yet (1).

If the clinic cannot tell you who arranges follow-up scans, how long they take and what they cost, you cannot weigh the test yet.

A negative result does not explain away a symptom. A new lump, bleeding, a cough that will not clear or weight loss you cannot explain still needs a doctor, whatever a blood test said.

If you track your health with lab tests, our guide to your body's biomarkers explains what a normal range does and does not tell you.

Frequently Asked Questions

Does the Galleri test work?

It finds cancer: 3,637 people in the tested group were diagnosed, against 3,400 in the group whose blood was not tested. It did not lower stage III or IV cancers, the trial's main goal: 706 against 688 people. Its effect on deaths is not yet known.

Is Galleri FDA approved?

No, as of GRAIL's release of 4 November 2026. An FDA advisory panel voted in favour on 23 September 2026, according to GRAIL, which expects a decision in early 2027.

How accurate is the Galleri test?

About half of positive results were cancer: 937 of 1,801 across three rounds. In people without cancer, 99.55 percent of tests came back negative. It flagged about 3 in 10 cancers diagnosed within a year of a test. For the 12 types that cause most cancer deaths, it flagged 47.6 to 63.4 percent, depending on the round.

What happens if my Galleri test is positive?

In the trial, positives went through the NHS urgent referral route, and about half were cancer. In an earlier US study, ruling cancer out after a false alarm took a median of 162 days.

Can Galleri replace a mammogram or a colonoscopy?

No. The trial tested it on top of usual NHS screening. It missed 301 of 357 breast cancers and 130 of 316 bowel cancers diagnosed within a year of a test. The trial paper says testing of this kind is meant to complement existing screening, not replace it.

Does Galleri reduce cancer deaths?

Not shown. The trial has not reported deaths and was not sized to detect a difference in overall cancer deaths. The main death analyses are planned 3 and 6 years after the last trial visit, which by our estimate means around 2027 and 2030; an earlier planned comparison has not been reported.

How much does the Galleri test cost?

On US clinic pages read on 7 November 2026, prices ran from $636.12 for members at Mito Health to a $1,200 list price at Human Longevity. We found no UK price.

Is Galleri available on the NHS?

No. NHS England chose in 2024 to wait for the final results. GRAIL says they met some agreed criteria and missed others.

Sources

1. Sasieni P, et al. Effect of Screening with Multicancer Early-Detection Test on Late-Stage Cancer Diagnosis. New England Journal of Medicine. 2026. PMID: 42776074. DOI: 10.1056/NEJMoa2505723. Author manuscript (PMC7619610) and Supplementary Appendix read in full; published abstract read on PubMed.
2. Neal RD, et al. Performance of a multi-cancer early detection test in the randomized controlled NHS-Galleri trial. Nature Medicine. 2026. PMID: 42773209. DOI: 10.1038/s41591-026-04652-8. Full text read (PMC13668664); supplementary tables not opened.
3. Neal RD, et al. Cell-Free DNA-Based Multi-Cancer Early Detection Test in an Asymptomatic Screening Population (NHS-Galleri): Design of a Pragmatic, Prospective Randomised Controlled Trial. Cancers. 2022. PMID: 36230741. DOI: 10.3390/cancers14194818. Open access.
4. Schrag D, et al. Blood-based tests for multicancer early detection (PATHFINDER): a prospective cohort study. The Lancet. 2023. PMID: 37805216. DOI: 10.1016/S0140-6736(23)01700-2. Accepted manuscript read in full (PMC11027492).
5. Feng X, et al. Cancer Stage Compared With Mortality as End Points in Randomized Clinical Trials of Cancer Screening: A Systematic Review and Meta-Analysis. JAMA. 2024. PMID: 38583868. DOI: 10.1001/jama.2024.5814. Abstract and key points read.
6. Rebolj M, et al. Late-Stage Outcomes as Surrogates for Mortality in Cancer Screening Trials: A Systematic Review and Meta-analysis. Cancer Epidemiology, Biomarkers and Prevention. 2025. PMID: 40694037. DOI: 10.1158/1055-9965.EPI-25-0201. Full text read (PMC12491949).
7. Batson S, et al. UK National Screening Committee position statement on evidence required for multicancer detection tests. The BMJ. 2026. PMID: 42185045. DOI: 10.1136/bmj-2026-089868. Full text read (PMC13224043).
8. Batson S, et al. UK National Screening Committee position statement on surrogate outcomes in cancer screening trials. The BMJ. 2026. PMID: 42185047. DOI: 10.1136/bmj-2026-629407. Full text read (PMC13224044).
9. ISRCTN registry. ISRCTN91431511, the NHS-Galleri trial. Record version 66, last updated 24 September 2026. DOI: 10.1186/ISRCTN91431511. Full record read.
10. ClinicalTrials.gov. NCT05611632, the NHS-Galleri trial. Last update posted 10 July 2026. Registry record. Current version read.
11. NHS-Galleri trial. What the trial found. nhs-galleri.org. Read on 7 November 2026.
12. Patient Care Online. NHS-Galleri Data Show Fewer Stage IV Cancers, but Primary Endpoint Missed. 20 February 2026. patientcareonline.com. Read on 7 November 2026.
13. GRAIL, Inc. Landmark NHS-Galleri Trial Demonstrates a Substantial Reduction in Stage IV Cancer Diagnoses, Increased Stage I and II Detection of Deadly Cancers, and Four-Fold Higher Cancer Detection Rate. Press release, 19 February 2026. grail.com. Read on 7 November 2026.
14. GRAIL, Inc. GRAIL Reports Full Results From NHS-Galleri Trial Demonstrating Substantial Reduction in Stage IV Cancer Diagnoses at 2026 ASCO Annual Meeting. Press release, 30 May 2026. grail.com. Read on 7 November 2026.
15. GRAIL, Inc. GRAIL Announces FDA Advisory Committee Meeting to Review Premarket Approval Application for the Galleri Multi-Cancer Early Detection Test. Press release, 7 August 2026. grail.com. Read on 7 November 2026.
16. GRAIL, Inc. FDA Advisory Committee Votes In Favor of Approval of GRAIL's Galleri Multi-Cancer Early Detection Test. Press release, 23 September 2026. grail.com. Read on 7 November 2026.
17. GRAIL, Inc. GRAIL Reports Third Quarter 2026 Financial Results. Press release, 4 November 2026. grail.com. Read on 7 November 2026.
18. GRAIL, Inc. Form 10-Q for the quarter ended 30 September 2026, filed 5 November 2026. SEC EDGAR. Key sections read.
19. Kumar H. GRAIL Update on the Accelerated Implementation of Multi-Cancer Early Detection Technology by NHS England. GRAIL, 29 May 2024. grail.com. Read on 7 November 2026. Quotes NHS England; we could not open NHS England's own post.
20. Human Longevity. Galleri Multi-Cancer Early Detection Test. humanlongevity.com. Clinic sales page, read on 7 November 2026.
21. Mito Health. Galleri Multi-Cancer Early Detection Test. mitohealth.com. Clinic sales page, read on 7 November 2026.
22. Superpower. Galleri Multi-Cancer Test. superpower.com. Clinic sales page, read on 7 November 2026.
23. The Health Suite. The Health Suite among first in UK to offer blood test that screens for more than 50 cancers. The clinic's own press release, published on News-Medical, 18 August 2026. news-medical.net. Read on 7 November 2026.

Funding Transparency

LSD is editorially independent. We receive no funding from pharmaceutical, supplement or longevity companies, or from diagnostic test makers. These are the funding ties behind each source above, taken from each paper's own funding and conflict statements where we could read them. Sources 13 to 23 are the maker's and clinics' own pages, cited for what they claim, not as evidence.

• Source 1 (Sasieni, 2026): Funded by GRAIL, which makes Galleri. In the paper's words, "The funder participated in the study design, data curation, data interpretation, and drafting and approval of the publication." Data analysis was run independently at Queen Mary University of London and verified by GRAIL. GRAIL, with the Cancer Research UK Cancer Prevention Trials Unit, was responsible for the study's design, conduct, reporting and analysis.
• Source 1, continued: GRAIL paid for medical writing, graphics and copy editing. Three authors list GRAIL Bio UK as their affiliation. We did not open the per-author NEJM disclosure forms; the ties below come from source 2, which has the same 13 authors.
• Source 2 (Neal, 2026): Same funding statement in substance; medical writing was by Citrus Health Group, funded by GRAIL. Kumar, GRAIL's Chief Scientific Officer (13), is employed by and owns stock in GRAIL; Liang and Smittenaar are GRAIL employees who own GRAIL stock. Y Lee and Warwick have salaries supported by GRAIL research funding, and Round has GRAIL research funding.
• Source 2, continued: Neal has a consulting or advisory role with GRAIL Bio UK and is an inventor on another cancer test with a revenue share. Sasieni has consulting or advisory roles with GRAIL Bio UK and MEDIAN Technologies, institutional research funding from GSK, GRAIL Bio UK and Merck, and a pending patent on HPV testing.
• Source 2, continued: Swanton is a co-chief investigator funded by GRAIL and a paid member of GRAIL's Scientific Advisory Board. From March 2026 he is a nonexecutive director of Novartis, and he lists many other industry ties. Dolly, Johnson, Jones and L Lee declared none. Johnson's affiliation is NHS England.
• Source 3 (Neal, 2022): The two co-chief investigators served through university consultancy funded by GRAIL. Three authors were GRAIL employees holding Illumina stock. GRAIL funded medical writing, and also funded extra cancer registration officers to record the trial's cancers.
• Source 4 (PATHFINDER, 2023): Funded by GRAIL, which "contributed to the study design, data analysis, interpretation, drafting, review, and approval". Four authors are GRAIL employees and two are GRAIL consultants. The lead author reports GRAIL research funding to Dana-Farber, where she was a study investigator. The sponsor paid for the imaging and lab tests in the study.
• Source 5 (Feng, 2024): The authors report no conflicts of interest. The funder is not named in the part of the paper we could read. We read the abstract and key points only.
• Source 6 (Rebolj, 2025): Funded by Cancer Research UK and the UK National Institute for Health and Care Research. Sasieni, the first author of source 1, reports personal fees from GRAIL; Janes reports grants from GRAIL; Duffy reports occasional GRAIL consultancy. R Smith leads the American Cancer Society's initiative on multi-cancer tests.
• Sources 7 and 8 (UK National Screening Committee, 2026): The committee's chair, Mike Richards, chaired a clinical advisory group for GRAIL and NHS England jointly, from 2021 to 2022. He is senior counsel at Incisive Health, which has advised GRAIL; he says he was not involved in that work. Five other authors are committee members; the rest are secretariat staff, subgroup members or researchers.
• Sources 9 to 11 (registries and the trial website): Public records of a trial funded by GRAIL (1).
• Source 12 (Patient Care Online): A news report. We found no funding statement relevant to the facts we cite.
• Sources 13 to 19 (GRAIL): GRAIL speaking for itself, with a full commercial interest in the test. GRAIL faces a securities class action, Robbins v. GRAIL, filed on 5 June 2026, alleging false and misleading statements about NHS-Galleri; these are allegations, not findings (18).
• NHS England: GRAIL has an agreement with the NHS under which the trial results were judged against success criteria for a possible rollout (18). One trial author, Johnson, lists NHS England as his affiliation (1).
• Sources 20 to 23 (clinics): Businesses that sell the test. Source 23 is The Health Suite's own announcement, published on News-Medical. We cite them only to show how the test is described for sale, not for any trial number.

Related Reading

• Rapamycin's First Longevity Trial Missed Its Primary Endpoint, why the result a trial was built to test matters most
• A Bone Drug Cut Fractures by a Third. Vitamin D Did Not., including a screening trial that missed its main goal and hit a secondary one
• Epigenetic Clocks Disagree With Themselves by 8.6 Years, another test sold to healthy people ahead of outcome data
• 1 of 8 NMN and Urolithin A Trials Improved How Bodies Work, what happened when supplement trials were judged on their own goals
• Your Body's Biomarkers: Measuring What Matters for Longevity, what a lab result can and cannot tell a healthy person

---

In 142,250 people in England, the Galleri blood test found more cancers. Stage III and IV cancers of 12 types stayed level, about 1 in 100 in each group over three years. Whether it saves lives is not yet known.

---

Written with the help of AI tools, shaped and verified by humans. Every result in this article was checked against the trial's own papers, its registry records, or GRAIL's own releases and filings. The one source read as an abstract only is marked.

This article is for information only and is not medical advice. Do not start or stop any screening or medicine on the basis of an article.