Multi-cancer early detection blood tests

Also known as Multi-cancer early detection (MCED) tests, MCED blood tests, Galleri test

Multi-cancer early detection (MCED) blood tests look for signals from many cancer types in one sample. The main commercial example, GRAIL's Galleri, is under US regulatory review, and the trial evidence so far is mixed: the UK NHS-Galleri trial missed its primary endpoint, and whether earlier detection reduces deaths is not yet shown.

Pilot and early commercialDisputedReviewed 21 Oct 2026
AI illustration · Illustrative scene, not a photograph of the specific device or experiment.

In plain language

A multi-cancer early detection (MCED) test is a single blood test that looks for signs that cancer may be present, aiming to catch many different cancer types at once. Its developers aim to find cancers earlier, when they are usually easier to treat. In the PATHFINDER study, the test flagged a signal in 1.4% of participants, and 0.5% had a confirmed cancer.

The central question is whether finding cancers earlier saves lives. Blood tests can be wrong in both directions. A positive result is often a false alarm that leads to further tests, and many cancers are missed. The large studies so far give different pictures. The UK NHS-Galleri trial reported a nominally significant reduction in stage IV diagnoses but did not meet its primary endpoint. A US expert panel voted in September 2026 on the test's benefit and effectiveness, with split results. Reduced cancer deaths have not been shown.

Technical explanation

Galleri and similar MCED tests analyse a blood sample and return a positive or negative result. A positive result also points to a likely tissue of origin, which the developer calls cancer signal origin (CSO). Company-reported CSO accuracy was 97% for an earlier test version and 88% for a refined version in PATHFINDER (The Lancet, 2023, as reported in GRAIL's press release), and 92.5% in the NHS-Galleri full results presented at ASCO in 2026.

Published figures depend on the population, the test version and the endpoint, so they should not be compared directly. In PATHFINDER, 6,662 adults aged 50 or over were enrolled. A signal was found in 1.4% of participants, 0.5% had a confirmed cancer, and the positive predictive value (PPV) was 43.1% for the refined test and 38% for the earlier version. In NHS-Galleri, GRAIL reported episode sensitivity of 54.7% for the 12 prespecified cancers and 30.7% across all cancer types, with specificity of 99.55% and PPV of 52.0%. In the PATHFINDER 2 analysis reported by Targeted Oncology from the FDA panel, sensitivity was 35.0% (95% CI 29.8% to 40.5%), specificity 99.85% and PPV 77.0% among 303 participants diagnosed with cancer within 12 months. The article says these performance analyses used retrospective testing of stored plasma, while the studies prospectively used an earlier investigational assay version.

NHS-Galleri tested whether adding the blood test to standard screening reduces later-stage cancers across three screening rounds. It enrolled 142,250 people aged 50 to 77 in England. Its primary endpoint, a combined reduction in stage III and IV diagnoses for 12 prespecified cancers, was not met (incidence rate ratio 1.03, 0.92 to 1.14). A secondary endpoint, stage IV diagnoses, fell by 14% overall (incidence rate ratio 0.86, 95% CI 0.744 to 0.998, which GRAIL describes as nominally statistically significant), with reductions of 9%, 22% and 26% in successive rounds. Sensitivity varied sharply by cancer type and stage, which the FDA panel discussed.

What the evidence says

Each statement is typed by the strength of its evidence and linked to the source that supports it.

Demonstrated Shown in a published experiment, test or shipping product.

The NHS-Galleri trial enrolled 142,250 people aged 50 to 77 in England and compared adding the Galleri test to standard screening; its primary endpoint, a combined reduction in stage III and IV diagnoses for 12 prespecified cancers, was not met (incidence rate ratio 1.03, 0.92 to 1.14).

Sources: [5] · Primary endpoint and participant numbers

In the PATHFINDER study published in The Lancet in 2023, 6,662 adults aged 50 or over were enrolled, the refined test had a PPV of 43.1% (38% for the earlier version), and 48% of confirmed cancers were stage I or II; the study is described as a pilot non-randomised study, and its lead investigator says larger controlled trials are needed to establish clinical utility.

Sources: [1] · Study results and stated limitations

GRAIL submitted the final module of its premarket approval application on 29 January 2026, relying on PATHFINDER 2 data from 25,490 consented participants with one year of follow-up and first-year NHS-Galleri data; the FDA designated the test a Breakthrough Device in 2018, and it has not been cleared or approved.

Sources: [3], [4] · PMA submission announcement

Reported, unverified Stated by a company or team but not independently confirmed.

In a prespecified secondary endpoint, GRAIL reports that stage IV diagnoses for the 12 cancers fell by 14% overall (incidence rate ratio 0.86, 95% CI 0.744 to 0.998), described as nominally statistically significant, with reductions of 9%, 22% and 26% across three rounds, and that stage I and II diagnoses rose by 16%.

Sources: [5] · Stage IV and stage I-II results · medium confidence

In NHS-Galleri, GRAIL reports an episode sensitivity of 54.7% for the 12 prespecified cancers and 30.7% across all cancer types, a specificity of 99.55% and a PPV of 52.0%; the release does not report NPV.

Sources: [5] · Performance metrics · medium confidence

In an FDA panel analysis of PATHFINDER 2, among 303 participants diagnosed with cancer within 12 months, the 12-month episode sensitivity was 35.0% (95% CI 29.8% to 40.5%), specificity 99.85% and PPV 77.0%; these came from retrospective testing of stored plasma samples.

Sources: [6] · PATHFINDER 2 performance · medium confidence

In the PATHFINDER 2 analysis, 72.9% of participants with a positive signal had at least one invasive procedure, including 47.8% of those without a cancer diagnosis; in NHS-Galleri, the test's sensitivity was 10.7% for prostate cancer and 13.6% for stage I cancers.

Sources: [6] · Panel discussion of downstream workup and sensitivity · medium confidence

The FDA Molecular and Clinical Genetics Devices Panel voted 7 to 2 (one abstention) that benefits outweigh risks, 6 to 4 on effectiveness and unanimously on safety, for adults aged 50 or older.

Sources: [6] · Panel votes · medium confidence

Disputed Contested by other credible sources.

Experts disagree on whether NHS-Galleri shows benefit. GRAIL describes the stage IV results as a substantial and clinically meaningful reduction. The BMJ reported in 2024 that leaked documents suggested the trial was not suitable to justify a national screening test, and that, according to emails it had seen, the UK National Screening Committee chair had privately raised serious concerns with NHS England's chief executive.

Sources: [4], [7] · Benefit debate · medium confidence

Limitations and open problems

  • The NHS-Galleri trial did not meet its primary endpoint. The stage IV reduction is a prespecified secondary endpoint, and GRAIL describes it as nominally statistically significant, so it may not hold up in further analysis.
  • Sensitivity varies widely by cancer type and stage. In NHS-Galleri analyses discussed by the FDA panel, sensitivity was 10.7% for prostate cancer, 63.8% for lung cancer and 13.6% for stage I cancers.
  • Many positive results lead to invasive investigation. In PATHFINDER 2, 72.9% of participants with a positive signal had at least one invasive procedure, including 47.8% of those who did not have a cancer diagnosis.
  • Most performance and trial results so far come from company announcements and trade press. The sources reviewed here did not include a peer-reviewed NHS-Galleri paper, and the PATHFINDER 2 figures here come from a trade press report of the FDA panel rather than a full publication.
  • No reduction in cancer deaths has been shown. GRAIL's February 2026 top-line release says follow-up may be extended by 6 to 12 months; no extended results are available yet.
  • It is not clear whether NHS-Galleri results apply to the FDA-reviewed version of the test. The PMA submission includes a bridging analysis between the study versions and the updated PMA version. The test has not been cleared or approved by the FDA.

Possible future uses

Speculative

Ideas, not predictions. None of these has been demonstrated.

  • Use alongside single-cancer screening for people with a higher cancer risk, such as those with a strong family history, if benefit is shown.
  • Blood-based triage that sends people with a positive signal to targeted imaging, if signal-origin accuracy holds up in routine use.
  • Monitoring for recurrence in people already treated for cancer, which has not been tested in the sources reviewed here.

Timeline

  1. 5 Oct 2023GRAIL reports final PATHFINDER results in The Lancet: 6,662 adults screened, refined test PPV 43.1%, and 48% of confirmed cancers at stage I or II. Sources: [1]
  2. 18 Jun 2025GRAIL releases top-line PATHFINDER 2 results from 25,578 analysed participants without numerical sensitivity, specificity or PPV. Sources: [2]
  3. 29 Jan 2026GRAIL submits the final module of its premarket approval application to the FDA, relying on PATHFINDER 2 and first-year NHS-Galleri data. Sources: [3]
  4. 19 Feb 2026GRAIL reports top-line NHS-Galleri results: the primary endpoint is not met, and stage IV reductions of more than 20% are reported in the second and third screening rounds. Sources: [4]
  5. 30 May 2026Full NHS-Galleri results are presented at the ASCO Annual Meeting, with stage IV incidence rate ratio 0.86 for 12 cancers. Sources: [5]
  6. 24 Sept 2026Trade press reports that the FDA Molecular and Clinical Genetics Devices Panel voted 7 to 2 (one abstention) that benefits outweigh risks and 6 to 4 on effectiveness. The article does not state the meeting date. Sources: [6]

Futures this could help bring about

AI illustration
How far away?

Screening for many cancers with one blood test

Pilots · 6 of 9 steps done

Related breakthroughs

Glossary

Sensitivity
The share of people who have a condition that the test correctly flags as positive.
Specificity
The share of people without the condition that the test correctly flags as negative.
Positive predictive value (PPV)
The share of positive results that are confirmed as true cancers. It depends on how common the condition is in the group tested.
Stage IV cancer
A cancer that has spread to other parts of the body. Later-stage cancers are usually harder to treat.

Sources

  1. company announcement
    Final Results From PATHFINDER Study of GRAIL's Multi-Cancer Early Detection Blood Test Published in The Lancet
    GRAIL, 5 Oct 2023. Accessed 21 Oct 2026.
  2. company announcement
    GRAIL Announces Positive Top-Line Results From The Galleri PATHFINDER 2 Registrational Study
    GRAIL, 18 Jun 2025. Accessed 21 Oct 2026.
  3. company announcement
    GRAIL Submits FDA Premarket Approval Application for the Galleri Multi-Cancer Early Detection Test
    GRAIL, 29 Jan 2026. Accessed 21 Oct 2026.
  4. company announcement
    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
    GRAIL, 19 Feb 2026. Accessed 21 Oct 2026.
  5. company announcement
    GRAIL Reports Full Results from NHS-Galleri Trial Demonstrating Substantial Reduction in Stage IV Cancer Diagnoses at 2026 ASCO Annual Meeting
    GRAIL, 30 May 2026. Accessed 21 Oct 2026.
  6. news
    FDA Panel Votes in Favor of Galleri Multi-Cancer Early Detection Test
    Andrea Eleazar. Targeted Oncology, 24 Sept 2026. Accessed 21 Oct 2026.
  7. news
    New evidence casts doubt on a much-hyped blood test for early cancer detection
    Margaret McCartney, Deborah Cohen. The BMJ, 8 Aug 2024. Accessed 21 Oct 2026.

Revision history

3 changes since this profile was created.

Changes on 21 Oct 2026
Published on owner instruction (content loop iteration 8)950bcd7
Sources and claims re-checked against originalsfbb434c
Initial profile drafted from primary sources7e66249