Tumor Markers (CEA, CA-125, AFP): What They Screen and What They Cannot Prove

It’s 11:47 p.m., and someone is staring at a patient portal. One line glows on the screen: CA-125: 41 U/mL (H). That little “H” — high — sends them straight to a search bar, where the first result mentions ovarian cancer and the fourth mentions menstruation. Both, it turns out, are telling part of the truth.

Blood tests for tumor markers occupy a strange space in modern medicine. They sound definitive — a cancer number, measured to the decimal. Yet oncologists treat them more like weather vanes than verdicts: useful for direction, unreliable in isolation, and easily spun by an unrelated breeze.

Here is what CEA, CA-125, and AFP genuinely measure, what a high value does and doesn’t mean, and why the most honest sentence in this article is also the shortest: no tumor marker, by itself, diagnoses cancer.

What exactly is a tumor marker?

A tumor marker is a substance — most often a protein — that can be measured in blood, urine, or tissue and tends to appear in larger amounts when certain cancers are present. Some markers are made by cancer cells themselves; others are produced by healthy tissue reacting to a tumor, which is one early clue that these tests measure association, not proof.

The idea is older than most people realize. Carcinoembryonic antigen (CEA) was identified in 1965 in colon cancer tissue, and researchers initially hoped it would become a simple blood test to catch cancer early in anyone. That hope collapsed within a decade, for a reason that still defines the field: CEA also rises with smoking, inflammatory bowel disease, liver disease, and ordinary lung conditions. The marker was real. Its exclusivity was not.

Modern medicine has settled into a more modest, more useful role for these tests. According to MedlinePlus, tumor markers today help doctors monitor how a known cancer responds to treatment, watch for recurrence after therapy ends, estimate prognosis, and occasionally sharpen a diagnosis that imaging has already suggested. Screening healthy people is, with narrow exceptions we’ll get to, not on that list.

Think of a marker as a smoke detector wired to several rooms at once. When it sounds, something may be burning — or someone may simply be making toast. The alarm tells you to investigate. It never tells you what you’ll find.

What are the common tumor markers?

Dozens of markers exist, but a handful account for most tests ordered in everyday practice. Three headline this article; a few others deserve a seat at the table because patients encounter them constantly.

Marker	Most associated cancers	Noncancer causes of elevation	Main clinical use
CEA	Colorectal; also lung, breast, pancreatic	Smoking, IBD, cirrhosis, pancreatitis, lung disease	Monitoring colorectal cancer after treatment
CA-125	Ovarian; some uterine and peritoneal	Endometriosis, fibroids, menstruation, pregnancy, pelvic infection, liver disease	Tracking ovarian cancer treatment; evaluating a pelvic mass
AFP	Liver (hepatocellular); certain testicular and ovarian germ cell tumors	Hepatitis, cirrhosis, pregnancy	Liver cancer surveillance in high-risk people; germ cell tumor staging
PSA	Prostate	Benign prostate enlargement, prostatitis, recent ejaculation	Prostate cancer detection and monitoring, via shared decision-making
CA 19-9	Pancreatic, bile duct	Gallstones, pancreatitis, bile duct blockage	Monitoring pancreatic cancer treatment
hCG	Gestational trophoblastic disease, some germ cell tumors	Pregnancy (its normal job)	Diagnosis and monitoring of trophoblastic disease

Read down the third column and a pattern emerges: nearly every marker on this list can be raised by something entirely benign — often something as common as smoking, a period, or gallstones. That column, more than any other, explains why interpretation requires a clinician rather than a search engine.

Why aren't tumor markers used to screen healthy people?

Two words: false alarms. A screening test applied to millions of healthy people needs both high sensitivity (it catches most cancers) and high specificity (it rarely flags people without cancer). Most tumor markers fail on both counts, and the arithmetic is unforgiving.

Consider CA-125. Roughly 80% of women with advanced ovarian cancer have elevated levels — but only about half of those with stage I disease do, according to Mayo Clinic. So the test misses many of the very cancers screening exists to catch early. Meanwhile, benign conditions like endometriosis and fibroids raise CA-125 in women who will never develop ovarian cancer, sending them toward ultrasounds, anxiety, and sometimes surgery they didn’t need.

This isn’t theoretical. A landmark UK trial followed more than 200,000 women for over 15 years, comparing CA-125-based screening strategies against no screening. The screened groups had some cancers detected earlier — yet deaths from ovarian cancer did not meaningfully decline. Earlier detection on paper did not translate into lives saved, which is the only outcome that justifies screening a healthy population.

Contrast that with screening tests that earned their place: colonoscopy, mammography, cervical screening. Each demonstrated, in large trials, that finding disease earlier changed outcomes. Tumor markers, tested the same way, largely have not. The honest conclusion isn’t that these blood tests are worthless — it’s that they were built for a different job, and asking them to screen the healthy is like asking a thermometer to predict the flu season.

CEA: the colorectal marker that a cigarette can raise

Carcinoembryonic antigen is a protein the fetal gut produces in abundance and the adult body mostly retires. Colorectal cancers frequently reactivate it, which is why CEA became the workhorse marker of gastrointestinal oncology.

The reference range itself tells a story about specificity. Many labs consider values up to about 3 ng/mL normal for nonsmokers — but extend that ceiling to roughly 5 ng/mL for smokers, because tobacco alone reliably nudges CEA upward. When a routine habit can move a “cancer number,” you begin to see why no one is diagnosed from it.

Where CEA shines is after a colorectal cancer diagnosis, in three specific moments:

• Before surgery, a baseline level helps estimate prognosis and gives future tests a point of comparison.
• After surgery, CEA has a half-life of a few days, so a successful operation should send an elevated level drifting back toward normal within weeks. A level that refuses to fall raises the question of remaining disease.
• During follow-up, a steady climb across serial tests — not one wobbly reading — often prompts imaging to look for recurrence, sometimes before symptoms appear.

Even in this best-case role, CEA has blind spots. Some colorectal tumors never secrete it, so a normal level after treatment is reassuring but not a guarantee. And an isolated bump can reflect a chest infection or a flare of liver inflammation. The trend is the message; the single value is just noise until proven otherwise.

Is CA-125 a tumor marker? Yes — with unusually loud caveats

CA-125 is absolutely a tumor marker, and probably the one most often misread by anxious portal-checkers. It’s a protein found on many ovarian cancer cells, and the conventional cutoff — about 35 U/mL — sits atop one of the noisiest backgrounds in laboratory medicine.

The tissue lining the abdomen and pelvis makes CA-125 whenever it’s irritated, and irritation has many innocent sources. Menstruation can raise it. So can pregnancy, endometriosis, uterine fibroids, pelvic inflammatory disease, and liver conditions. Cleveland Clinic notes that premenopausal women are especially prone to benign elevations, which is precisely why the test isn’t recommended as a general screen for them.

Where CA-125 earns real trust is in two settings. First, monitoring: in a woman being treated for ovarian cancer, a falling CA-125 suggests the treatment is working, while a sustained rise after therapy can flag recurrence months before scans show it. Second, evaluating a pelvic mass that imaging has already found: combined with ultrasound features and menopausal status, the level helps clinicians judge how urgently a specialist should be involved. In a postmenopausal woman with a suspicious mass, a markedly elevated CA-125 shifts the odds considerably — though the diagnosis still comes from tissue, not blood.

The number 41 from our opening scene? In a menstruating 32-year-old with known endometriosis, it may mean nothing at all. In a 68-year-old with new bloating and an ovarian cyst on ultrasound, it means a prompt referral. Same number, different worlds — context is the entire test.

AFP: a liver marker with a fetal backstory

Alpha-fetoprotein is essentially the fetal version of albumin, churned out by a developing baby’s liver and yolk sac. After birth, production plummets, and healthy adults typically carry less than about 10 ng/mL. Two situations bring it roaring back: pregnancy, where it’s expected, and certain cancers, where it isn’t.

Hepatocellular carcinoma — the most common form of primary liver cancer — often produces AFP, sometimes in the hundreds or thousands of nanograms per milliliter. Germ cell tumors of the testicle or ovary can do the same. That makes AFP genuinely useful in two roles, per MedlinePlus:

• Surveillance in high-risk livers. People with cirrhosis or chronic hepatitis B face a substantially elevated liver cancer risk, and many clinicians pair AFP with an abdominal ultrasound roughly every six months in these patients. This is targeted surveillance of a known high-risk group — a very different proposition from screening the general public.
• Staging and monitoring germ cell tumors. In testicular cancer, AFP joins other blood markers as a formal part of staging, and its trajectory after surgery helps guide further treatment.

The familiar caveat applies with force here: the same liver inflammation that puts someone at risk for cancer — hepatitis, active cirrhosis — also raises AFP on its own. A mildly elevated level in a person with hepatitis C may reflect the hepatitis, the beginnings of a tumor, or both. That ambiguity is exactly why AFP is interpreted alongside imaging rather than instead of it.

What happens if tumor markers are high?

Less than most people fear, and more methodically than the internet suggests. A single elevated marker sets a process in motion, not a diagnosis.

The first step is usually to look backward before looking forward. Your clinician reviews why the test was ordered, what your baseline has been, and what else could explain the number — smoking status for CEA, menstrual timing or endometriosis for CA-125, liver inflammation for AFP. Often the explanation is sitting in plain sight in the chart.

Next comes repetition. Because lab variation and transient benign causes are common, a modestly elevated result is frequently rechecked after several weeks, ideally at the same laboratory (different assay methods can produce genuinely different numbers from the same blood). A value that returns to normal usually ends the story. A value that climbs steadily changes it.

If the elevation persists or the clinical picture is concerning, imaging follows — ultrasound, CT, or MRI aimed at the organ the marker is most associated with. Only if imaging finds something suspicious does the conversation turn to biopsy, because tissue under a microscope remains the only way to actually diagnose cancer.

Notice what’s absent from this sequence: panic, and treatment based on a blood number alone. Oncologists have watched thousands of elevated markers resolve into gallstones, cysts, and cigarette habits. The number opens an investigation. The investigation — not the number — delivers the answer.

What counts as a high number for tumor markers?

People understandably want a bright line: below this, safe; above it, danger. Laboratories do publish cutoffs — commonly around 3 ng/mL for CEA in nonsmokers (about 5 for smokers), 35 U/mL for CA-125, and 10 ng/mL for AFP — but three realities complicate any simple reading.

Magnitude matters more than the threshold. A CA-125 of 42 is a raised eyebrow; a CA-125 of 900 is a different conversation. An AFP of 14 in someone with hepatitis may be the hepatitis talking, while an AFP in the thousands points strongly toward a tumor. Clinicians weigh how far above the line a value sits, not merely whether it crossed.

Direction matters more than either. A CEA that reads 6, then 6, then 5 over nine months tells a calmer story than one that reads 3, then 7, then 15. Serial measurements at the same lab turn a snapshot into a film, and the film is what guides decisions.

Cutoffs are statistical, not biological. The 35 U/mL line for CA-125 was drawn where it usefully separated groups of patients in studies — it is not a cliff edge in your body. Healthy people live above these lines; people with cancer live below them.

So the honest answer to “what is a high number?” is: high relative to your own baseline, sustained across repeat tests, and interpreted against your history. Any framing simpler than that is selling certainty the biology doesn’t offer.

What tumor markers can actually do well

After all these caveats, it’s worth being equally clear about where these tests genuinely earn their keep — because in the right role, they’re quietly excellent.

Tracking treatment response. During chemotherapy for ovarian or colorectal cancer, a marker that falls with each cycle offers near-real-time evidence the treatment is biting, often faster and more cheaply than repeated scans. A marker that plateaus or rises can prompt a change in strategy sooner.

Watching for recurrence. After treatment ends, rising CEA or CA-125 can flag returning disease months before it’s visible on imaging or felt as symptoms. That early warning lets doctors aim scans precisely rather than searching blindly.

Sharpening a diagnosis already in motion. When imaging has found a pelvic mass or a liver lesion, marker levels help stratify how suspicious it is and how quickly a specialist should act. Blood plus imaging outperforms either alone.

Informing prognosis and staging. In testicular cancer, marker levels are written directly into the official staging system — a rare case where the blood test carries formal diagnostic weight.

And one genuine near-exception to the “markers never diagnose” rule: hCG in gestational trophoblastic disease. In the right clinical setting, an hCG pattern is so characteristic that it drives diagnosis and treatment monitoring almost single-handedly. It’s the exception that proves the rule — one marker, one rare disease, decades of validation. The other markers never reached that bar, which is precisely why they’re handled with more humility.

What tumor markers cannot prove — the honest limits

Every limitation below trips up real patients weekly, so they deserve plain statement rather than fine print.

• A high marker cannot prove cancer. Benign conditions raise every marker discussed here. Elevated CEA has been traced to smoking; elevated CA-125 to fibroids; elevated AFP to hepatitis. Diagnosis requires tissue.
• A normal marker cannot rule cancer out. Some tumors simply don’t secrete their “assigned” marker — these are called nonsecretors. About half of stage I ovarian cancers produce a normal CA-125. A reassuring number is evidence, not exoneration.
• A marker cannot locate anything. CEA is “associated with” colorectal cancer, but breast, lung, and pancreatic tumors can raise it too. An elevated value points in a general direction at best; imaging does the actual finding.
• A marker cannot distinguish benign from malignant. The protein doesn’t know why it’s elevated. Only the clinical workup can sort a cyst from a tumor.
• A marker cannot predict an individual’s future. Population studies link higher levels with worse average prognosis, but no blood number forecasts one person’s course.

There’s a deeper point beneath the list. Tumor markers measure a downstream consequence of disease, not the disease itself — like judging a fire by its smoke. Smoke can billow from a small blaze or barely rise from a large one; wind, weather, and toast all interfere. Firefighters still go look. So do oncologists.

When tumor markers genuinely belong in screening

“Not for screening healthy people” comes with a crucial qualifier: average-risk healthy people. For specific high-risk groups, the math changes, because cancer is common enough among them that a flawed test still catches enough disease to matter.

Liver cancer surveillance is the clearest case. In people with cirrhosis or chronic hepatitis B, liver cancer risk is high enough that many clinicians recommend an abdominal ultrasound roughly every six months, often paired with an AFP blood test. Neither test is perfect alone; together they catch more early tumors, and early-stage liver cancer has meaningfully more treatment options. This is surveillance of a defined risk group, guided by a specialist — a world away from ordering AFP on a healthy 40-year-old.

High genetic risk for ovarian cancer occupies murkier ground. Women carrying certain inherited gene changes face substantially elevated ovarian cancer risk, and some are offered periodic CA-125 testing with pelvic ultrasound. Here honesty requires saying what the evidence actually shows: this monitoring has not been proven to save lives, and guidelines treat it as an interim measure for women who haven’t yet pursued risk-reducing surgery, not as a reliable safety net. It may find some cancers earlier; it has not demonstrated the mortality benefit that would make it true screening.

PSA for prostate cancer sits in its own category — the one marker widely used in screening conversations for average-risk men, and even there, major organizations frame it as a shared decision between patient and clinician after weighing benefits against the harms of overdiagnosis. The debate itself is instructive: even the most screening-adjacent tumor marker remains a judgment call, not a default.

What the test itself is like

The logistics, at least, are refreshingly simple. Tumor marker tests are ordinary blood draws — a needle, a vial, a few minutes. Fasting usually isn’t required, though your care team will say if a specific panel needs it. Results typically return within a few days.

Two practical details are worth knowing before the needle, because they prevent most of the confusion afterward.

Stick with one laboratory when possible. Different manufacturers’ assays can yield noticeably different values from identical blood. A CA-125 of 30 at one lab and 38 at another may represent zero biological change. When a marker is being followed over time, consistency of lab matters more than any single result.

Expect the portal to be blunt. Patient portals often release results before your clinician has reviewed them, flagged with an “H” and stripped of context. That flag reflects a statistical reference range, not a judgment about you. If you tend toward late-night result-checking — and our opening reader is hardly alone — it can help to decide in advance that you’ll note the number, write down your questions, and wait for the interpretive phone call rather than outsourcing your 2 a.m. anxiety to a search engine.

One more quiet truth: doctors rarely order these tests casually. If a marker appears on your lab slip, it usually has a specific job — establishing a baseline, tracking a treatment, or checking a trend. Asking “what question is this test answering?” is the single best way to understand your own result.

When to see a doctor

Tumor markers should never be the reason you delay acting on your own body’s signals — and never the reassurance that lets you ignore them. Two situations warrant a timely appointment.

If you have persistent symptoms, see a doctor regardless of any blood test. Warning signs that deserve evaluation include:

• Bloating, pelvic pain, feeling full quickly, or urinary changes lasting more than a few weeks
• Blood in the stool, a persistent change in bowel habits, or unexplained iron-deficiency anemia
• Yellowing of the skin or eyes, dark urine, or new pain under the right ribs
• Unexplained weight loss, drenching night sweats, or a lump anywhere — including the testicle
• Fatigue that is new, profound, and unexplained

Most of these symptoms trace back to benign causes. But each is common enough in early cancer that mainstream guidance — from the NHS to Mayo Clinic — urges evaluation rather than watchful worrying, and a normal tumor marker does not substitute for that evaluation. Remember: half of early ovarian cancers travel with a normal CA-125.

If you’ve received an abnormal marker result, book the follow-up rather than the search spiral. Bring three questions: Why was this test ordered? What benign causes fit my situation? What is the plan — repeat testing, imaging, or watchful waiting, and on what timeline? A ten-minute conversation with someone who knows your history will do what no article, including this one, can: interpret your number in the context of you.

Frequently asked questions

What are the common tumor markers?

The most frequently ordered tumor markers are CEA (associated with colorectal cancer), CA-125 (ovarian), AFP (liver and germ cell tumors), PSA (prostate), CA 19-9 (pancreatic), and hCG (gestational trophoblastic disease and some germ cell tumors). Each is measured with a simple blood draw, and each can also be elevated by benign conditions — smoking, endometriosis, hepatitis, and gallstones among them — which is why results are interpreted alongside imaging and clinical history.

What happens if tumor markers are high?

An elevated tumor marker triggers investigation, not diagnosis. Your clinician will typically review benign explanations first — smoking, menstrual timing, liver inflammation — then repeat the test after a few weeks at the same laboratory. If the level stays high or climbs, imaging of the relevant organ usually follows. Biopsy is considered only if imaging finds something suspicious, because tissue examination remains the only way to actually confirm cancer.

Is CA125 a tumor marker?

Yes. CA-125 is a protein found on many ovarian cancer cells and is one of the most widely used tumor markers. Its main roles are tracking ovarian cancer during and after treatment and helping evaluate a pelvic mass found on imaging. It is not a reliable screening test, because endometriosis, fibroids, menstruation, pregnancy, and pelvic infections can all raise it, while about half of early-stage ovarian cancers leave it normal.

What is a high number for tumor markers?

Common laboratory cutoffs are about 3 ng/mL for CEA in nonsmokers (roughly 5 for smokers), 35 U/mL for CA-125, and 10 ng/mL for AFP — but crossing a cutoff is only the starting point. Clinicians weigh how far above the line a value sits, whether it rises across repeat tests, and what benign causes fit your history. A slightly elevated, stable value often means far less than a lower value that keeps climbing.

Can a tumor marker test detect cancer early in healthy people?

Generally, no. Large studies show most tumor markers miss many early cancers while flagging many people who don’t have cancer, and a major UK trial found CA-125-based screening did not reduce ovarian cancer deaths. The exceptions are targeted: AFP paired with ultrasound is used for liver cancer surveillance in people with cirrhosis or chronic hepatitis B, and PSA testing is offered to men through shared decision-making rather than as a universal screen.

Can tumor markers be normal even if you have cancer?

Yes, and this happens often enough to matter. Some tumors are “nonsecretors” that simply don’t produce their associated marker — about half of stage I ovarian cancers, for example, come with a normal CA-125, and some colorectal cancers never raise CEA. This is a core reason doctors never use a normal marker result to rule out cancer, and why persistent symptoms warrant evaluation regardless of what a blood test shows.

Do I need to fast before a tumor marker blood test?

Usually not. Most tumor marker tests, including CEA, CA-125, and AFP, are standard blood draws that require no special preparation, according to MedlinePlus. If your marker is being ordered as part of a larger panel that includes fasting tests, your care team will tell you. One preparation step that genuinely helps: when a marker will be tracked over time, try to use the same laboratory, since different assays can produce different values from identical blood.

Why does my doctor keep checking CEA after colon cancer treatment?

Serial CEA testing is one of the best-validated uses of any tumor marker. After successful colorectal cancer surgery, an elevated CEA should fall toward normal within weeks. During follow-up, a level that climbs steadily across several tests can signal recurrence months before scans or symptoms would reveal it, letting your team investigate early with targeted imaging. A single mildly elevated reading matters far less than the trend across visits.

Can endometriosis or fibroids raise CA-125?

Yes, both commonly do. The tissue lining the pelvis releases CA-125 when irritated, so endometriosis, uterine fibroids, pelvic inflammatory disease, menstruation, and pregnancy can all push levels above the 35 U/mL cutoff without any cancer present. This background noise is strongest in premenopausal women, which is why an isolated elevation in a younger woman is often rechecked and interpreted against her gynecologic history rather than treated as an alarm.

Is the AFP test in pregnancy the same as the AFP tumor marker test?

It measures the same protein but answers a different question. During pregnancy, AFP made by the fetus is checked as part of prenatal screening to assess fetal development. Outside pregnancy, elevated AFP in an adult can point toward liver cancer or certain germ cell tumors — though hepatitis and cirrhosis raise it too. Because pregnancy naturally elevates AFP, the tumor marker interpretation doesn’t apply to pregnant patients.

References

• CEA Test — MedlinePlus
• Alpha-Fetoprotein (AFP) Tumor Marker Test — MedlinePlus

This article is for general information only and is not a substitute for professional medical advice. Please consult a qualified doctor about your individual situation.