Tumor Markers: What They Are, the List & What High Means

Tumor markers are substances — usually proteins — measured in the blood, whose level can rise when certain cancers are present. The name is frightening, so let's say the most important thing calmly and up front: a high tumor marker is not a diagnosis of cancer. With a few debated exceptions, these tests are used to monitor a cancer that is already known — to judge how it responds to treatment and to watch for recurrence — and they are not meant to screen healthy people without symptoms. Many benign conditions push them up (smoking, inflammation, liver disease, menstruation, pregnancy, a gallstone), and some cancers do not raise them at all — so both false positives and false negatives are common. This hub explains, plainly, what a tumor marker is, what it is used for (and what it is not), and walks through the seven you are most likely to meet — PSA, CEA, CA-125, CA 19-9, CA 15-3, AFP, and beta-hCG — each with its own detailed guide. The interpretation always belongs to your clinician.

Key takeaways

• A tumor marker is a substance in the blood that can rise with cancer — but is also present in healthy tissue and in many benign conditions.¹²
• Their main job is monitoring a cancer that is already diagnosed (response to treatment, recurrence) — not mass screening of people without symptoms.¹²
• A high value is not cancer. False positives from benign causes are frequent, and false negatives happen too (some cancers raise no marker).¹³
• The seven most common: PSA (prostate), CEA (colorectal), CA-125 (ovarian), CA 19-9 (pancreatic/biliary), CA 15-3 (breast), AFP (liver and germ-cell), beta-hCG (germ-cell and trophoblastic) — each has a dedicated guide.¹
• No cutoff is a verdict. What counts is less an isolated number than its trend over time and the clinical context, and values vary by laboratory.²⁴
• Marker-based screening does not save lives for ovarian cancer (CA-125) and gives only a modest benefit, at the cost of overdiagnosis, for prostate cancer (PSA).⁵⁶

What are tumor markers?

A tumor marker is a substance — most often a protein, or a sugar carried on a protein — that the body's cells make and that can be measured in the blood (and sometimes in urine or tissue).¹ The word "tumor" comes from the fact that, for some cancers, the cells produce more of it, which raises the blood level. But it is a misleading label: the very same substance is usually present in healthy tissue and in a long list of benign conditions.²

That single fact governs everything else. A tumor marker is neither specific to one cancer (many non-cancer causes raise it) nor reliably sensitive (some cancers do not secrete it). A raised number is therefore a signal to interpret in context, never proof of disease.¹² Most of these tests are targeted — ordered in a defined clinical situation, not printed on a routine blood panel. When one does turn up unexpectedly on a broad panel, that is a reason to put it back in context with a clinician, not a reason to panic.

The U.S. National Cancer Institute frames it clearly: tumor markers are used mostly to help guide treatment and monitor cancer, and — apart from a small number of debated exceptions — are not recommended to screen people who have no symptoms.¹⁴

The cardinal rule: monitoring, not screening

This is the thesis of the whole page, and it is the source of most of the worry these tests create. With few exceptions, tumor markers are not screening tests for healthy people. They earn their place after a cancer is diagnosed, as a way to follow it over time.¹²

Two ideas do most of the work:

• A high value does not mean cancer. Benign causes are common, so most mildly raised markers in the general population are false alarms, not tumors.¹³
• A normal value does not rule cancer out. Some cancers raise no marker at all, so a reassuring number can miss real disease.²³

Because a screening test needs to be both sensitive (catch real disease) and specific (not flag healthy people), a test that fails on both counts cannot reliably find cancer in someone without symptoms — it generates far more false positives than true early cancers, driving anxiety, extra scans, and unnecessary biopsies.²¹ That is why, for the cancers with proven screening programs, screening rests on other tools: mammography for breast cancer, colonoscopy or a stool test (FIT) for colorectal cancer, and cervical screening for cervical cancer — not on blood markers.⁴¹

The nuance sits with a few markers. PSA is the clearest partial exception: used for individual prostate-cancer screening it gives a modest reduction in prostate-cancer deaths, but at the price of overdiagnosis and overtreatment, which is why it is a shared decision rather than an automatic test.⁶ For ovarian cancer, by contrast, a very large trial (UKCTOCS) showed that screening with CA-125 does not reduce deaths.⁵ The takeaway is consistent: outside a specific, doctor-defined context, a tumor marker is not a "cancer check."

How tumor markers are used

Used well, tumor markers are genuinely valuable; used as a general "cancer screen," they mostly create alarm and unnecessary tests.² Within a context set by your clinician, they serve four main purposes:¹⁷

• Supporting a diagnosis. Faced with an abnormality already found another way — a mass on imaging, a suspicious symptom — a marker can, alongside imaging and often a biopsy, help characterize the situation. It never makes the diagnosis alone.¹²
• Staging and prognosis. The level of certain markers at diagnosis carries prognostic information and can help estimate how advanced a disease is, guiding the specialist's plan.²⁸
• Monitoring treatment response. In someone with a known cancer, the marker is tracked over time: a fall during surgery or chemotherapy is generally reassuring, while a plateau or rise prompts the team to reconsider.²⁹
• Detecting recurrence. After treatment, a gradual rise can be an early signal that a cancer is coming back — triggering imaging — as with CEA after colorectal cancer.⁹⁸

In every case the marker is an input, read alongside imaging and the clinical picture, and interpreted by the treating team — not a stand-alone answer.¹²

The main tumor markers

Here are the seven tumor markers you are most likely to encounter. Each has its own dedicated guide with reference ranges, causes of elevation, myths, and FAQs — follow the links for the detail.

PSA — prostate

Prostate-specific antigen is made by the prostate, not by tumors specifically. It is the central tool in prostate-cancer screening (the most debated in U.S. medicine), but it also rises with age, with an enlarged prostate (BPH), and with infection or prostatitis — so a high PSA far more often signals something benign than cancer. The 4–10 ng/mL "gray zone" carries only about a 1-in-4 chance of cancer and is refined with percent-free PSA, PSA density, and an MRI before any biopsy.⁶ See the full PSA test guide.

CEA — colorectal cancer

Carcinoembryonic antigen is used mainly to monitor a known colorectal cancer — a rising CEA can flag recurrence early, a falling one suggests treatment is working. It is not a screening test (colonoscopy and FIT are). Its most common benign cause is smoking, which raises it in a dose-dependent way — hence the higher "normal" ceiling for smokers — along with inflammatory bowel disease, liver and lung disease.⁹⁸ See the full CEA test guide.

CA-125 — ovarian cancer

Cancer antigen 125 is best known as an ovarian-cancer marker, used to monitor known disease and, with the marker HE4 (the ROMA score) and ultrasound, to help evaluate a pelvic mass. In women it rises very commonly for benign reasons — menstruation, pregnancy, endometriosis, ovarian cysts, fibroids — and it is more reliable after menopause. Population screening with CA-125 does not save lives.⁵ See the full CA-125 test guide.

CA 19-9 — pancreatic and biliary cancers

CA 19-9 is tied mainly to pancreatic and bile-duct cancers, used for monitoring and prognosis, not screening. Two limits define it: false positives from cholestasis or obstructive jaundice — a simple gallstone can send it very high with no tumor, and the level often falls once the blockage is cleared — and false negatives in the 5–10% of people who are "Lewis-negative" and make little or no CA 19-9, so a normal result never rules cancer out.¹⁰⁸ See the full CA 19-9 test guide.

CA 15-3 — breast cancer

CA 15-3, a fragment of the MUC1 protein, is used to monitor advanced or metastatic breast cancer — not to screen or diagnose (mammography does the screening). Its sensitivity is low in early, localized disease, so a normal value does not reassure, and benign causes include liver disease, benign breast conditions, and pregnancy or lactation.⁸² See the full CA 15-3 test guide.

AFP — liver and germ-cell cancers

Alpha-fetoprotein is a fetal protein that is very low in adults. As a tumor marker it is used for surveillance of liver cancer (hepatocellular carcinoma) in at-risk people — always with ultrasound, never alone — and to diagnose and monitor germ-cell tumors of the testicle and ovary (with beta-hCG and LDH). Chronic liver disease, hepatitis, cirrhosis, and pregnancy all raise it benignly.²¹ See the full AFP test guide.

Beta-hCG — germ-cell and trophoblastic tumors

Beta-hCG is best known as the pregnancy hormone, but it doubles as a tumor marker: for germ-cell tumors of the testicle and ovary (used with AFP and LDH) and, above all, for gestational trophoblastic disease (molar pregnancy), where serial hCG is the key marker followed to confirm cure. Low-level positives outside pregnancy usually have benign explanations (pituitary hCG, assay interference).¹ See the full beta-hCG test guide.

A note on reference values. The cutoffs used for these markers are indicative and vary by laboratory and assay. The value that counts is the one printed on your report, and in monitoring it is the trend — ideally compared within the same lab — that carries the information, far more than a single isolated number.²⁹

Why a high tumor marker usually isn't cancer

This is the most important, and most reassuring, message on this page. A marker above the reference range worries people, but it should be read carefully, because two symmetrical traps explain most results.

1. False positives: many benign conditions raise markers. In the general population this is actually the usual reason a marker is high. Concrete examples:²¹

• smoking raises CEA in a dose-dependent way, with no cancer;
• menstruation, pregnancy, endometriosis, cysts, and fibroids raise CA-125 in women;
• cholestasis or a simple gallstone can send CA 19-9 very high;
• hepatitis, cirrhosis, or a regenerating liver raise AFP;
• an enlarged prostate, prostatitis, or a urinary infection raise PSA;
• more broadly, inflammation, liver disease, diabetes, and infection can all contribute to non-cancer elevations.²³

2. False negatives: a normal marker does not rule cancer out. Some tumors simply do not produce the marker. The clearest case is CA 19-9: about 5–10% of people are "Lewis-negative" and make none of it, so their result stays low even with pancreatic cancer.¹⁰ Likewise, CA 15-3 is often normal in early breast cancer, and AFP can be normal while an early liver cancer is developing.²

These two traps are exactly why a tumor marker is never read alone and does not work as a screening test. Faced with a raised value, the right response is neither panic nor self-diagnosis: it is to discuss it calmly with a clinician, who places the number in context — your history, smoking, liver, cycle, symptoms, and imaging — and, if needed, rechecks it later.¹²

Limitations: sensitivity, specificity, and false results

The value of any test comes down to two properties. Sensitivity is its ability to catch real disease (few false negatives); specificity is its ability to stay negative in healthy people (few false positives). Tumor markers tend to fall short on both, which is precisely why they make poor screening tests but useful monitoring tools.²¹

Low specificity is the false-positive problem: because the same substances appear in healthy tissue and benign disease, a raised marker often points nowhere in particular. Low sensitivity is the false-negative problem: a normal marker cannot guarantee the absence of cancer. Add the lab-to-lab variability of assays — the reason serial monitoring should stay in the same laboratory — and it becomes clear why no single number is ever a verdict.⁴²

There is one more limitation worth stating plainly: markers are often drawn in combination (for example CEA with CA 19-9). But different markers move for different reasons, so one can be high while another is normal without either pointing to a specific disease. None is interpreted in isolation; it is the clinician who synthesizes the markers with imaging and the clinical picture.⁸²

Recent research

According to recent PubMed publications, the field is moving beyond classic protein markers — while newer blood tests remain, honestly, promising but unproven for the biggest question, early detection in healthy people:

• Circulating tumor DNA (ctDNA) — finer than protein markers. "Liquid biopsies" that detect tumor DNA in the blood can, in colorectal cancer for example, flag minimal residual disease after surgery and catch recurrence earlier and more precisely than protein markers like CEA. It is a fast-moving field, still being folded into routine care rather than fully established.¹¹ (Loft M et al., Lancet Gastroenterol Hepatol, 2023.)
• Multi-cancer early detection (MCED) blood tests. Tests such as Galleri, built on cell-free DNA, aim to detect a signal from several cancers from a single blood draw. In the prospective PATHFINDER study, an MCED test returned a cancer signal in about 1.4% of adults, and roughly 38% of those signals were true cancers — meaning the majority of positives were false alarms, and the test found some cancers but missed many others.¹² (Schrag D et al., Lancet, 2023.)
• Framed honestly: not yet standard of care. These MCED tests are not approved for routine screening and have not been shown to reduce cancer mortality; large randomized trials (including the NHS-Galleri trial) are still underway to answer that question. Reviews stress the real risks — false positives, overdiagnosis, and follow-up burden — alongside the promise.¹³ (Wan JCM et al., Nat Rev Clin Oncol, 2025.) For now, they do not replace the proven screening programs.

These findings concern research and monitoring; they do not authorize self-medication, and interpretation thresholds must be confirmed with your physician.

Get your tumor markers interpreted by AI DiagMe

A tumor marker is never read alone: its meaning depends on your context (smoking, liver, cycle, history), its trend over time, the imaging, and your other tests. A high value is not a diagnosis of cancer, and a normal one does not reassure by itself. That cross-referencing is what gives the result its real value.

👉 AI DiagMe interprets your lab results — blood, urine, or stool — in plain language, taking your whole profile into account. An informational service that does not provide a diagnosis and complements, never replaces, your physician.

Frequently asked questions

Bottom line

Tumor markers — PSA, CEA, CA-125, CA 19-9, CA 15-3, AFP, and beta-hCG — are blood substances useful mainly to monitor a cancer that is already known (treatment response, recurrence), and, with few exceptions, not to screen healthy people without symptoms. The central, reassuring message is that a high value is not a diagnosis of cancer: false positives from benign causes (smoking, inflammation, liver disease, menstruation, pregnancy, gallstones) are common, and false negatives exist (a normal marker does not rule cancer out, as in Lewis-negative people for CA 19-9). No cutoff is a verdict, values vary by laboratory, and it is the trend over time — crossed with imaging and your context, and interpreted by your clinician — that matters. Newer blood tests (ctDNA, MCED) are promising but not yet proven to save lives and do not replace established screening. No value is read alone: it is your full set of markers and your profile that gives a result meaning — which is what AI DiagMe provides, alongside your physician.

Sources

Official sources and peer-reviewed publications (PubMed) used for this guide: