OncoFirm™ / Research / Early detection
Pillar guide · Early detection
AI-Driven Biomarkers and Early Cancer Detection: Where Multi-Cancer Blood Tests Stand in 2026
For decades, early cancer detection meant a handful of single-cancer screening tests. Today, AI-driven biomarkers are making it possible to look for many cancers in one blood sample. 2026 has been a turning point: large trials have reported, and the first multi-cancer test has reached an FDA advisory committee. This guide explains how these tests work, what the evidence shows, what is still unproven and where quantitative protein measurement fits.
At a glance
Topic
AI-driven biomarkers for early, multi-cancer detection
Biggest 2026 news
FDA advisory committee backed Galleri on 23 Sep 2026; FDA decision pending
Best evidence
PATHFINDER 2 (35,878 people) and NHS-Galleri (142,000 people)
Still unproven
Whether multi-cancer blood tests reduce cancer deaths
OncoFirm™ status
Quantitative protein assays in development; no multi-cancer claims
On this page
Key points
Early cancer detection in 2026, in brief
60.3%of positive Galleri results were confirmed cancer in PATHFINDER 2, with 99.64% specificity
142,000people in NHS-Galleri; primary stage III–IV endpoint not met, but stage IV fell over 20% in later rounds
7–2FDA advisory committee vote on Galleri's benefit-risk profile, 23 September 2026
Figures are from the PATHFINDER 2 publication and GRAIL’s trial and regulatory announcements listed in the references.
Why early detection
Why finding cancer earlier matters
Cancer caused an estimated 9.7 million deaths worldwide in 2022, with about 20 million new cases (GLOBOCAN). For most cancers, survival is far higher when the disease is found while it is still localized than after it has spread. Yet recommended population screening exists for only a few cancers, such as breast, cervical, colorectal and, for high-risk people, lung cancer. Most cancer deaths come from cancers with no recommended screening test.
That gap is what multi-cancer early detection (MCED) tests aim to close, by looking for many cancers at once from a single blood draw. See why early detection matters and the future of cancer screening.
How it works
How AI turns biomarkers into a cancer signal
MCED tests combine one or more biomarker classes with a trained classifier. The output is a yes-or-no cancer signal and, for some tests, a predicted tissue of origin that guides follow-up.
- DNA methylation: cancers leave distinctive methylation patterns on cell-free DNA. Machine-learning classifiers trained on thousands of samples recognize these patterns and can predict the likely organ of origin.
- Fragmentomics and mutations: the size and end patterns of DNA fragments, chromosome changes and mutations add further signal. Our explainer on cell-free DNA and the MIGHT studies covers why inflammation can mimic these signals.
- Proteins: tumor-associated proteins such as CEA and CA-125 add independent information and can catch tumors that shed little DNA.
- The AI layer: the classifier must be locked, validated on independent data and tuned to very high specificity, because even a small false-positive rate means many unnecessary workups when testing healthy people.
The 2026 landscape
Multi-cancer and blood-based screening tests to know
| Test (developer) | Biomarkers | Scope | U.S. status, Oct 2026 |
|---|---|---|---|
| Galleri (GRAIL) | cfDNA methylation + AI classifier | More than 50 cancer types, with predicted origin | First MCED premarket approval application; advisory committee voted in favor 23 Sep 2026; decision pending |
| Shield (Guardant Health) | cfDNA | Colorectal cancer only | FDA approved July 2024 for average-risk adults 45+ |
| Shield MCD (Guardant Health) | cfDNA | Multiple cancers | Breakthrough Device designation; selected for NCI Vanguard study |
| Avantect (ClearNote Health) | cfDNA epigenomic signals | Multiple cancers | Selected for NCI Vanguard study |
| Cancerguard (Exact Sciences, now Abbott) | DNA methylation + proteins | Multiple cancers | Launched 2025 as a laboratory test; not FDA approved |
Status reflects public announcements as of 1 October 2026 and may change. OncoFirm™ has no affiliation with these companies.
What the trials show
The evidence so far
- PATHFINDER 2 (Nature Medicine, 2026; 35,878 enrolled): cancer signal detected in 0.54% of people; positive predictive value 60.3%; specificity 99.64%; predicted origin correct in 91.3%; 53% of detected new cancers were stage I–II. Episode sensitivity was 39.3% for all cancers and 69.8% for 12 prespecified deadly cancers. Only 0.6% of participants had an invasive procedure after a positive test.
- NHS-Galleri (about 142,000 people aged 50–77, results February 2026): adding annual Galleri to standard screening found about four times more cancers. The primary endpoint, a statistically significant reduction in combined stage III–IV cancers, was not met, although stage IV diagnoses of 12 deadly cancers fell by more than 20% in the second and third rounds.
- FDA advisory committee (23 September 2026): 7–2 with one abstention on benefit-risk, unanimous on safety and 6–4 on effectiveness for adults 50 and older, as a supplement to recommended screening.
- NCI Vanguard study: enrolling up to 24,000 people with two selected assays to test the feasibility of a much larger randomized trial designed to measure effects on cancer deaths.
Reading the numbers: specificity tells you how rarely healthy people test positive; positive predictive value tells you how often a positive result is real cancer; episode sensitivity tells you how many cancers diagnosed within a year the test caught. All three matter, and none alone shows that lives are saved.
Open questions
What is still unproven
- Mortality benefit: no completed randomized trial has yet shown that MCED screening reduces cancer deaths.
- Missed cancers: sensitivity for early-stage cancers remains limited, so a negative result can be falsely reassuring.
- False positives and overdiagnosis: even with high specificity, testing millions of people produces many false alarms and may find slow cancers that would never have caused harm.
- Confounders: inflammation, clonal blood-cell changes and benign disease can mimic cancer signals.
- Cost and equity: current tests are expensive and laboratory-based, and access is uneven within and between countries.
- Guidelines: no U.S. guideline body currently recommends MCED testing for routine screening.
Where proteins fit
Proteins, point of care and the access gap
The original CancerSEEK study at Johns Hopkins combined DNA mutations with eight proteins because the two signal types catch different tumors. Its successor, Cancerguard, is described by its developer as the only MCED test on the market that includes protein biomarkers. Proteins such as CEA, PSA, AFP and CA-125 are inexpensive to measure and well understood.
Meanwhile, the Lancet Commission on diagnostics estimates that 47% of the world’s population has little or no access to diagnostics. Sequencing-based tests in central laboratories will not close that gap on their own. Quantitative protein measurement near the patient is one route to wider access, as discussed in AI multi-biomarker blood tests for triage and the global need for affordable diagnostics.
OncoFirm™ roadmap
Our role: accessible, quantitative protein data
OncoFirm™ is not developing a sequencing-based multi-cancer test. Our goal is complementary: reliable, quantitative protein biomarker measurement at the point of care, so that validated panels and algorithms can one day reach people who will never see a sequencing laboratory.
Quantitative protein foundation
Fluorescent assays designed to report CEA (1–100 ng/mL) and PSA (0.5–50 ng/mL) with planned traceability to WHO standards. Platform
Panels and algorithms
Research on multi-marker panels read on one strip, interpreted by locked, validated software. OncoFirm™ MultiDx · AI-driven biomarkers
Specificity by design
Validation plans that include inflammation and benign disease, the main sources of false positives. Clinical collaborations
Access first
A handheld reader and low-cost strip intended for settings without a central laboratory. Point-of-care oncology
OncoFirm™ assays, reader and software are in development, have not been cleared or approved by the FDA or any other regulatory authority, are for research use only and make no multi-cancer detection claims.
FAQ
Frequently asked questions
What is a multi-cancer early detection (MCED) test?
An MCED test is a blood test designed to look for signals from many types of cancer at once in people without symptoms. It measures biomarkers such as cell-free DNA methylation, DNA fragments or proteins, and uses a trained algorithm to decide whether a cancer signal is present and where in the body it probably comes from.
Is Galleri FDA approved?
As of 1 October 2026, not yet. GRAIL submitted the first premarket approval application for an MCED test, and on 23 September 2026 an FDA advisory committee voted in favor of its benefit-risk profile. The FDA's decision was still pending at the time of writing.
Does a negative MCED result mean I do not have cancer?
No. Current MCED tests miss a substantial share of cancers, especially early ones. A negative result does not rule out cancer, and recommended screening such as mammography, colonoscopy or stool tests, cervical screening and lung CT should continue.
What happens after a positive MCED result?
A positive result means a cancer signal was detected, not that cancer is confirmed. Clinicians use the predicted origin to guide imaging and other tests, and only a biopsy or other diagnostic procedure can confirm cancer. In PATHFINDER 2, about 60% of positive results were confirmed as cancer.
Can AI-driven blood tests replace standard cancer screening?
No. MCED tests are being developed to complement, not replace, guideline-recommended screening. Whether they reduce cancer deaths has not yet been shown in a completed randomized trial.
Does OncoFirm™ offer a multi-cancer test?
No. OncoFirm™ is developing quantitative fluorescent protein assays, starting with CEA and PSA, and researching multi-marker panels. These are in development, have not been cleared or approved by the FDA or any other regulatory authority, and make no multi-cancer detection claims.
Keep reading
Related guides
Sources
References
- Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229–263. doi.org/10.3322/caac.21834
- Nabavizadeh N, et al. Performance and safety of a multi-cancer early detection test: the PATHFINDER 2 study. Nature Medicine. 2026. www.nature.com/articles/s41591-026-04618-w
- GRAIL. Landmark NHS-Galleri trial results. News release, February 19, 2026. grail.com/press-releases/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. FDA advisory committee votes in favor of approval of the Galleri multi-cancer early detection test. News release, September 23, 2026. grail.com/press-releases/fda-advisory-committee-votes-in-favor-of-approval-of-grails-galleri-multi-cancer-early-detection-test/
- National Cancer Institute, Division of Cancer Prevention. NCI selects two assays for Vanguard Study of multi-cancer detection tests. January 7, 2025. prevention.cancer.gov/news-and-events/news/nci-selects-two-assays-vanguard-study-multi-cancer-detection-tests
- The ASCO Post. Guardant Health's Shield blood test approved by the FDA as a primary screening option for colorectal cancer. July 2024. ascopost.com/news/july-2024/guardant-health-s-shield-blood-test-approved-by-the-fda-as-a-primary-screening-option-for-colorectal-cancer/
- Cancerguard. The science behind Cancerguard (DNA methylation and protein biomarkers). www.cancerguard.com/science
- Cohen JD, Li L, Wang Y, et al. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science. 2018;359:926–930. doi.org/10.1126/science.aar3247
- Fleming KA, Horton S, Wilson ML, et al. The Lancet Commission on diagnostics: transforming access to diagnostics. The Lancet. 2021;398:1997–2050. www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)00673-5/fulltext
About this article. Written by the OncoFirm™ Scientific Team from peer-reviewed publications, regulatory announcements and company releases, all linked above. OncoFirm™ has no affiliation with the companies or tests named. Regulatory status reflects public information as of 1 October 2026. This article is for education and is not medical advice. OncoFirm™ assays are in development, have not been cleared or approved by the FDA and are not available for sale.
Partner on accessible early detection
Clinicians, researchers and investors interested in quantitative protein biomarkers at the point of care can reach the OncoFirm™ team. Collaboration inquiry · Investor inquiry