Cancer blood tests in 2026: what Galleri, Shield and liquid biopsy can and cannot do

A single blood draw that looks for dozens of cancers sounds like the end of screening as we know it. The evidence says it is a useful addition, not a replacement.

How a blood test can find a tumour

Cells throughout the body constantly die and release small fragments of DNA into the bloodstream. When a tumour is present, some of those fragments come from cancer cells. This is called circulating tumour DNA. A liquid biopsy tries to find that faint signal against a much larger background of normal DNA.

Tests use several kinds of clues:

• Methylation patterns: chemical tags on DNA that differ between cancer and normal cells, and also reveal which organ the DNA came from. Galleri relies mainly on this.
• Mutations: cancer-related changes in genes such as KRAS or TP53.
• Fragment patterns: cancer DNA breaks into pieces of different lengths. DELFI's FirstLook Lung uses this approach with machine learning.
• Proteins: some tests, such as Cancerguard, combine DNA signals with known blood markers like CA-125 and CEA.

Machine learning is what makes this workable. Each sample produces an enormous amount of data, and algorithms trained on thousands of patients separate cancer signals from noise, including age-related blood cell mutations that can otherwise cause false alarms.

The key limitation: small tumours shed little DNA

The amount of tumour DNA in the blood rises with the size and aggressiveness of the cancer. Large, advanced cancers are found reliably. Small stage I cancers, which are the most curable, often shed too little DNA to detect. This single fact explains most of the results below.

Which tests exist and where they stand

What the evidence shows

Colorectal blood tests

In the ECLIPSE study of about 7,900 people published in the New England Journal of Medicine in 2024, Shield detected 83% of colorectal cancers compared with colonoscopy, with about 90% specificity. But it found only 13% of advanced pre-cancerous polyps. SimpleScreen showed a similar pattern in its PREEMPT CRC study. Colonoscopy can find and remove nearly all such polyps, and stool DNA tests detect more of them than blood tests do.

Galleri in the PATHFINDER 2 study

Results presented at ESMO in October 2025, from more than 23,000 people aged 50 and over, showed very few false positives (specificity 99.6%). About six in ten positive results turned out to be cancer, and the test predicted the organ of origin correctly in about 92% of cases. Overall it detected roughly 40% of cancers diagnosed within a year, and much more of the most deadly types than of slow-growing ones. The FDA's own analysis gave a somewhat lower sensitivity and noted that the true figure is uncertain.

The NHS-Galleri trial

This randomised trial in England enrolled about 142,000 people aged 50 to 77, who gave blood once a year for three years. Full results were presented at ASCO in June 2026. The trial did not significantly reduce the main target, the number of cancers diagnosed at stage III or IV combined. GRAIL highlighted a reduction in stage IV diagnoses as a secondary finding. Whether the test saves lives will depend on longer follow-up.

Cost and access

• United States: Medicare covers Shield for eligible average-risk adults. Galleri costs about $949 and Cancerguard $689, both self-pay, though eligible for health savings accounts. A 2026 US law created a route for Medicare to cover multi-cancer tests once they are FDA-approved, but not immediately.
• United Kingdom: the NHS does not provide Galleri outside research; private clinics sell it. Bowel screening on the NHS uses stool tests (FIT).
• European Union: no multi-cancer blood test is certified for population screening under the EU diagnostics regulation (IVDR). Access is limited to private clinics.
• Turkey: we found no screening blood test approved by the medicines agency (TİTCK) or reimbursed by SGK. Private oncology centres use liquid biopsy mainly to choose treatment for people already diagnosed with advanced cancer.

Risks you should know about

• False alarms lead to investigations. Even with 99.5% specificity, testing 100,000 healthy people produces several hundred false positives, each followed by scans, endoscopies or biopsies.
• False reassurance. The biggest danger is skipping colonoscopy, mammography or smear tests because a blood test was negative.
• Overdiagnosis. Some slow-growing cancers found by testing would never have caused harm, but may still be treated.
• No proven survival benefit yet. Finding cancer earlier does not automatically mean living longer. Mortality data from randomised trials is still awaited.

Trials to watch

• NHS-Galleri (NCT05611632): ongoing follow-up of cancer deaths.
• US National Cancer Institute Vanguard Study: a feasibility trial of about 24,000 people, preparing a much larger trial to test whether multi-cancer screening reduces deaths.
• CRANE: Exact Sciences' real-world study of Cancerguard.
• FIND-CRC (Natera): another colorectal blood test aiming for FDA submission.

Questions patients ask