OncoFirm™ / Cancer Atlas 2026 / Tumor-associated antigens

Cancer Atlas 2026 · Tumor-associated antigens

Tumor-Associated Antigens: Classes, Glycan Antigens and Their Role in Cancer Diagnostics (2026)

Tumor-associated antigens are molecules that tumors display or release in unusual amounts or unusual forms. Almost every blood tumor marker in clinical use is one of them, which explains both their value and their limits. This guide covers the main antigen classes, the sugar-based antigens TF, Tn, sialyl-Tn and sialyl-Lewis a, and what is established in diagnostics versus still in research.

By OncoFirm™ Scientific TeamPublished 1 Oct 2026Updated 1 Oct 2026Sources peer-reviewed, guidelines, regulators, linked

At a glance

Topic

Tumor-associated antigens (TAAs) and tumor-associated carbohydrate antigens

Key idea

TAAs are abnormal in amount or form, but not absent from normal tissue

Established

CA 19-9, CA-125, CA 15-3, CEA, AFP and AFP-L3% for defined uses, not screening

Research stage

Circulating TF and Tn, anti-TF antibodies, autoantibody panels

OncoFirm™ status

TF assay concepts are concept-stage; research use only; nothing cleared or for sale

On this page

Key numbers

Tumor-associated antigens in three numbers

0.9%positive predictive value of CA 19-9 in people without symptoms, in one systematic review
1,028women in the phase III sialyl-Tn vaccine trial, which showed no benefit
2015first FDA approval of a drug aimed at a glycan antigen (dinutuximab, anti-GD2)

Sources: Lee et al., World Journal of Gastrointestinal Surgery 2020; Miles et al., The Oncologist 2011; Berois et al., Cancers 2022.

Definitions

Tumor-associated and tumor-specific antigens

An antigen is a molecule that antibodies or immune cells can recognize. Tumor-associated antigens (TAAs) are normal, non-mutated molecules that tumors express abnormally: in excess, at the wrong stage of life, or in an altered form. Their expression in normal tissue is limited but not zero. Tumor-specific antigens, also called neoantigens, arise from mutations in the tumor’s DNA and are absent from normal cells.

The line between the two is not sharp, but the practical point is clear. CEA, AFP, PSA, CA-125, CA 15-3 and CA 19-9 are all tumor-associated, not tumor-specific. That is the root of their limited specificity, and the reason the National Cancer Institute (NCI) lists them mainly for following treatment and recurrence. Those uses are covered in Cancer biomarkers explained, the companion chapter of the OncoFirm™ Cancer Atlas 2026.

Classes

The main classes of tumor antigens

Classes of tumor antigens and tumor-associated glycansGrid of eight cells: oncofetal antigens (CEA, AFP); differentiation antigens (PSA, tyrosinase); overexpressed antigens (HER2, MUC1); cancer-testis antigens (NY-ESO-1, MAGE-A1); viral antigens (HPV E6 and E7); neoantigens created by tumor mutations; truncated O-glycans (Tn, sialyl-Tn, TF); and other glycans (sialyl-Lewis a, ganglioside GD2).TUMOR ANTIGEN CLASSES AND GLYCAN ANTIGENSOncofetalMade in fetal life andre-expressed by tumors.Examples: CEA, AFP.DifferentiationLimited to one tissuelineage. Examples: PSA,tyrosinase in melanoma.OverexpressedNormal proteins made inexcess. Examples: HER2,MUC1.Cancer-testisNormally limited to germcells. Examples: NY-ESO-1,MAGE-A1.ViralEncoded by cancer-causingviruses. Examples: HPV E6and E7.NeoantigensCreated by tumor mutations;absent from normal cells(tumor-specific).Truncated O-glycansTn, sialyl-Tn and TF: shortsugar chains, normallyextended or masked.Other glycansSialyl-Lewis a (the CA 19-9epitope); gangliosides suchas GD2.All classes except neoantigens and viral antigens are tumor-associated, not tumor-specific.

Six classes of tumor antigens and two groups of tumor-associated glycans. Sources: Xie et al. 2023; Rajesh and Radhakrishnan 2023; Berois et al. 2022.

Six classes are defined by where the antigen comes from. The last two cells show glycan (sugar) antigens, which cut across the protein classes: the antigen is an abnormal sugar chain on an otherwise normal protein or lipid.

Viral antigens are a large class: NCI attributes about 690,000 cancers a year worldwide to human papillomavirus (HPV). Because TAAs are the body’s own molecules, the immune system partly tolerates them. That is one reason T-cell responses against TAAs have often been weak in trials.

Carbohydrate antigens

Carbohydrate antigens: why sugars change in cancer

Mucins are heavily sugar-coated proteins, such as MUC1. Their sugar chains, called O-glycans, are built one sugar at a time. In many carcinomas (cancers of lining tissues) building stops early, and short chains that are normally extended or masked become exposed. These are tumor-associated carbohydrate antigens (TACAs).

AntigenStructureNote
TnOne sugar (GalNAc) on a serine or threonineFirst step of O-glycan building
Sialyl-TnTn capped with a sialic acidMade by the enzyme ST6GalNAc-I
TF (Thomsen-Friedenreich)Two sugars (Gal-GalNAc), called core 1Made from Tn by T-synthase (C1GALT1)
Sialyl-Lewis aA four-sugar chainThe epitope measured by CA 19-9

Several defects produce these short chains. T-synthase needs a helper protein called Cosmc to fold correctly, and Cosmc can be lost through mutation or silenced by DNA methylation. Other routes include altered levels of the enzymes that start or cap the chain and a shortage of sugar building blocks.

How common are these antigens? Early histochemical and serological work (Springer, Science 1984) reported unmasked T and Tn antigens in most carcinomas, and later reviews commonly cite about 90%. That figure is a literature estimate, not a validated epidemiological prevalence. Measured frequency varies with tumor type and with the reagent used (peanut agglutinin versus monoclonal antibodies). In breast cancer, for example, reported sialyl-Tn positivity ranges from 20% to 80% depending on the detection method.

Glycan antigens are not passive flags. In cell and animal studies the protein galectin-3 binds TF on MUC1 and helps tumor cells clump and attach to blood vessel walls, and Tn and sialyl-Tn dampen anti-tumor immunity. None of this is yet a validated clinical test or treatment. Background: what raises Thomsen-Friedenreich antigen.

Established

What is established in diagnostics

Several TAA assays are routine, each for a defined job in people already under care. None is recommended for screening the general population.

AssayAntigenEstablished useMain limits
CA 19-9Sialyl-Lewis a glycanMonitoring treatment of pancreatic and biliary cancersRaised in pancreatitis and bile duct disease
CA-125Epitope on the MUC16 mucinOvarian cancer: aid to diagnosis, response, recurrenceScreening did not reduce deaths (UKCTOCS)
CA 15-3, CA 27.29Epitopes on MUC1Breast cancer: response and recurrenceRoutine use to find recurrence early is disputed
CEAOncofetal glycoproteinColorectal cancer: treatment effect and recurrenceNot a screening test
AFP-L3%Lectin-reactive AFP glycoformLiver cancer risk assessment in chronic liver diseaseAn adjunct; not for diagnosis or screening

CA 19-9 shows the pattern. In people with symptoms, one systematic review reported 79% sensitivity and 82% specificity for pancreatic cancer. In people without symptoms the positive predictive value (the share of positive results that are true cancers) was 0.9%. The same review reports that about 6% of Caucasian and 22% of non-Caucasian people are Lewis-negative and cannot make the antigen. See the pancreatic cancer chapter.

AFP-L3% is the clearest example of a cleared glycan-defined test. It measures the share of AFP that carries a fucose sugar and binds a lentil lectin (a sugar-binding protein). The FDA cleared it through the 510(k) route on 23 February 2011 (K100464) for hepatocellular carcinoma risk assessment. More is in the liver cancer chapter and our guides to AFP and CA-125.

Research stage

What is still research

  • Circulating TF and Tn antigen. In the sources reviewed for this Atlas, no cleared or CE-marked assay measures TF or Tn antigen in blood, and no quantitative clinical studies were found.
  • Anti-TF antibodies. People naturally carry antibodies to TF, possibly induced by gut bacteria. A 2020 review reported lower anti-TF IgM in cancer patients than in controls, and exploratory accuracy of up to 80% for gastric cancer using antibody glycosylation markers. It also lists obstacles: levels depend on blood group, and cause and effect are unclear. Multicenter validation was not found.
  • Autoantibody panels. The immune system can make antibodies against TAAs. In the Scottish ECLS trial of about 12,200 people at high risk of lung cancer, an autoantibody blood test followed by imaging lowered the share of cancers found at stage III or IV at 2 years (58.9% versus 73.2%; hazard ratio 0.64, 95% CI 0.41 to 0.99). Deaths from lung cancer and from all causes did not differ significantly. In people with lung nodules, a pooled analysis found 22% sensitivity and 92% specificity, which its authors called poor.

The lesson recurs across the Atlas: finding cancer at an earlier stage is not the same as fewer deaths. Panel design is covered in multi-biomarker analysis, newer platforms in emerging diagnostic technologies, and trial endpoints in our early detection evidence hub.

Therapy context

Antigens as treatment targets

  • Anti-GD2 antibodies. GD2 is a ganglioside, a sugar-bearing lipid. Dinutuximab was approved by the FDA in 2015 for high-risk neuroblastoma in children, the first approved agent aimed at a TACA. Naxitamab received FDA accelerated approval on 25 November 2020.
  • The sialyl-Tn vaccine. Theratope was tested in 1,028 women with metastatic breast cancer. Time to progression was 3.4 versus 3.0 months and overall survival 23.1 versus 22.3 months, with no significant difference.
  • Tn-MUC1 CAR T cells. A first-in-human phase 1 trial (NCT04025216) enrolled 16 participants and was terminated by its sponsor, which judged the balance of risk and benefit unfavorable.
  • Antibody-drug conjugates (ADCs). ADCs carry a drug to cells that display a target antigen. Industry lists of FDA-approved ADCs include targets such as HER2, TROP2, nectin-4, folate receptor alpha, tissue factor and c-Met. None of these is tumor-specific.

Two cautions follow. The ADC target tissue factor is a protein and should not be confused with the TF (Thomsen-Friedenreich) glycan, despite the shared abbreviation. And no approved therapy targets TF, Tn or sialyl-Tn in the sources reviewed. Success as a drug target does not make an antigen a good diagnostic: a test must separate cancer from benign disease. HER2 testing is covered in the breast cancer chapter.

Pitfalls

Pitfalls: benign disease, reagents and standards

  • Benign disease. The same antigens rise in inflammation and liver or bile duct disease. CA 19-9 is raised in pancreatitis, pancreatic cysts, diabetes and benign cholestatic disease.
  • Lectin specificity. Peanut agglutinin (PNA) is one of the most used tools for detecting TF, but it also binds other sugar structures, with lower affinity. A PNA-based assay therefore measures PNA-reactive glycans, not TF alone. See the role of peanut agglutinin.
  • Standardization. Many tumor markers lack agreed international reference preparations, so kits can give different results for the same sample. No reference materials or agreed cut-offs for TF assays were found.
  • Optimistic early studies. Accuracy in case-control sets often falls in the intended-use population, as the lung nodule data show.

A credible TF or TACA assay would still need to show:

  1. Reagent specificity, tested against related glycans.
  2. A defined analyte: free glycan, glycan on a carrier such as MUC1, or antibody.
  3. Reference materials and cut-offs.
  4. Performance in people with inflammatory and benign disease, not only healthy donors.
  5. Prospective validation in the intended-use population, including the effect of blood group and Lewis type.
  6. Clinical utility. A shift to earlier stage alone is not a mortality benefit.

Outlook

What to watch

  • Results of the phase III GLORIA trial of a Globo H vaccine in triple-negative breast cancer (575 enrolled; no results were located).
  • Longer ECLS follow-up for lung cancer deaths; see the lung cancer chapter.
  • Multicenter, prospective studies of TF, Tn or anti-TF antibody assays that include benign-disease controls.
  • Regulatory status: as of 1 October 2026, no TF-based test was found to be FDA cleared or approved in the sources reviewed.

OncoFirm™ roadmap

Where OncoFirm™ fits

OncoFirm™ is developing quantitative fluorescent lateral flow assays for two tumor-associated antigens, CEA and PSA. OncoFirm™ makes no screening or multi-cancer claims. Its assays, reader and software are in development, are not cleared or approved by the FDA or any other regulatory authority, are for research use only and are not for sale.

CEA and PSA (in development)

Designed ranges are CEA 1 to 100 ng/mL and PSA 0.5 to 50 ng/mL, with planned traceability to WHO standards 73/601 and 17/100. See the CEA guide.

AFP (future pipeline candidate)

AFP is an oncofetal antigen used with ultrasound in people at high risk of liver cancer. See causes of high AFP.

TF antigen (concept-stage)

The reagent specificity, reference material and benign-control questions above all apply. Research partners can use the collaboration inquiry.

Platform

One reader and strip format, with a development target of a result in 20 minutes or less. See the fluorescent lateral flow platform.

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 are not for sale. Designed ranges and result times are development targets, not validated performance claims.

FAQ

Frequently asked questions

What is a tumor-associated antigen?

A tumor-associated antigen is a normal, non-mutated molecule that tumors express abnormally, for example in excess or in an altered form. It is not strictly absent from normal tissue. CEA, AFP, PSA, CA-125 and CA 19-9 are all tumor-associated antigens.

How does a tumor-associated antigen differ from a neoantigen?

A neoantigen is created by a mutation in the tumor's DNA and is absent from normal cells, so it is tumor-specific. A tumor-associated antigen is a normal molecule expressed abnormally. This difference explains why blood tumor markers can be raised in benign disease.

Is the TF antigen present in 90% of cancers?

Early work by Springer in 1984 reported unmasked T and Tn antigens in most carcinomas, and later reviews commonly cite about 90%. That figure is a literature estimate, not a validated epidemiological prevalence. Measured frequency varies with tumor type and with the reagent used.

Why is CA 19-9 not used to screen for pancreatic cancer?

CA 19-9 rises in benign conditions such as pancreatitis and bile duct disease, and some people cannot make the antigen at all. In people without symptoms its positive predictive value was 0.9% in one systematic review. It is used mainly to monitor treatment.

Does OncoFirm™ offer a TF antigen test?

No. OncoFirm™ TF antigen assay concepts are concept-stage. OncoFirm™ assays are in development, are not cleared or approved by the FDA or any other regulatory authority, are for research use only and are not for sale.

Sources

References

  1. Xie et al. Review of neoantigens and tumor antigens as targets for cancer therapy. Signal Transduction and Targeted Therapy. 2023;8:9. www.nature.com/articles/s41392-022-01270-x
  2. National Cancer Institute. Tumor Marker Tests in Common Use. Reviewed 7 December 2023. www.cancer.gov/about-cancer/diagnosis-staging/diagnosis/tumor-markers-list
  3. National Cancer Institute. HPV and Cancer. Updated 9 May 2025. www.cancer.gov/about-cancer/causes-prevention/risk/infectious-agents/hpv-and-cancer
  4. Springer GF. T and Tn, general carcinoma autoantigens. Science. 1984;224(4654):1198-1206. pubmed.ncbi.nlm.nih.gov/6729450/
  5. Sindrewicz, Lian, Yu. Review of galectin-3 and the Thomsen-Friedenreich antigen in cancer. Frontiers in Oncology. 2016;6:79. www.frontiersin.org/articles/10.3389/fonc.2016.00079/full
  6. Rajesh, Radhakrishnan. Review of Tn and sialyl-Tn antigens in cancer. Frontiers in Oncology. 2023;12:1093496. www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.1093496/full
  7. Cazet et al. Review of tumour-associated carbohydrate antigens in breast cancer. Breast Cancer Research. 2010;12:204. link.springer.com/article/10.1186/bcr2577
  8. Berois, Pittini, Osinaga. Review of therapies targeting tumor-associated carbohydrate antigens. Cancers. 2022;14:645. www.mdpi.com/2072-6694/14/3/645
  9. Lee, Teng, Shelat. Review of carbohydrate antigen 19-9. World Journal of Gastrointestinal Surgery. 2020;12:468-490. www.wjgnet.com/1948-9366/full/v12/i12/468.htm
  10. US Food and Drug Administration. 510(k) K100464: AFP-L3% test for hepatocellular carcinoma risk assessment. Decision 23 February 2011. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K100464
  11. Mayo Clinic Laboratories. Alpha-fetoprotein L3 percent (test L3AFP). www.mayocliniclabs.com/test-catalog/overview/88878
  12. Menon et al. UKCTOCS long-term mortality results. The Lancet. 2021;397(10290):2182-2193. www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)00731-5/fulltext
  13. Duffy MJ. Tumor markers in clinical practice: a review. Medical Principles and Practice. 2013;22:4. karger.com/mpp/article/22/1/4/203271/Tumor-Markers-in-Clinical-Practice-A-Review
  14. Duffy MJ, McGing P. Guidelines for the Use of Tumour Markers, 5th edition. Association of Clinical Biochemists in Ireland; 2018. acbi.ie/wp-content/uploads/2022/12/1644913336-1602832758-Tumour-markers-5th.pdf
  15. Duarte et al. Evidence review of the EarlyCDT Lung autoantibody test. Health Technology Assessment. 2022;26(49). www.ncbi.nlm.nih.gov/books/NBK587810/
  16. Kurtenkov. Review of naturally occurring antibodies to the Thomsen-Friedenreich antigen. BioMed Research International. 2020;2020:9747040. www.hindawi.com/journals/bmri/2020/9747040/
  17. Miles et al. Phase III trial of the sialyl-Tn (STn-KLH) vaccine in metastatic breast cancer. The Oncologist. 2011;16(8):1092-1100. academic.oup.com/oncolo/article/16/8/1092/6401188
  18. ClinicalTrials.gov. NCT04025216: phase 1 study of TnMUC1-targeted CAR T cells (terminated). clinicaltrials.gov/study/NCT04025216
  19. ClinicalTrials.gov. NCT03562637: phase III GLORIA study of adagloxad simolenin in Globo H-positive triple-negative breast cancer. clinicaltrials.gov/study/NCT03562637
  20. US Food and Drug Administration. FDA grants accelerated approval to naxitamab for high-risk neuroblastoma in bone or bone marrow. 25 November 2020. www.fda.gov/drugs/resources-information-approved-drugs/fda-grants-accelerated-approval-naxitamab-high-risk-neuroblastoma-bone-or-bone-marrow
  21. National Cancer Institute. Dinutuximab (drug information). Updated 14 September 2022. www.cancer.gov/about-cancer/treatment/drugs/dinutuximab
  22. BioPharmaSpec. List of FDA-approved antibody-drug conjugates (industry source). biopharmaspec.com/blog/fda-approved-adc-drugs/

About this article. Sourced from peer-reviewed reviews, NCI and FDA pages and trial registry records; some primary papers were read only as abstracts, and prevalence figures for glycan antigens are literature estimates. Part of the OncoFirm™ Cancer Atlas 2026, written by the OncoFirm™ Scientific Team from the sources linked above and reflecting public information as of 1 October 2026. It is for education and is not medical advice. OncoFirm™ has no affiliation with the companies or tests named. OncoFirm™ assays are in development, have not been cleared or approved by the FDA and are not available for sale.

Collaborate on the next edition

Clinicians, laboratories, researchers and industry partners can contribute data, corrections or validation studies. Collaboration inquiry · Investor inquiry

[email protected]+1 516.900.2606Port Washington, NY