What Are Cancer Antigen Levels and What Do They Mean?

Cancer antigen levels are measurements of specific proteins or other substances that tumors shed into the bloodstream, and doctors use them primarily to track how a known cancer is responding to treatment or to watch for signs of recurrence. They are not, for the most part, reliable tools for diagnosing cancer in the first place. That distinction trips up a lot of people, and it matters enormously for understanding what a test result actually tells you. The biology behind these markers is messier than a simple “high means cancer, low means clear” reading, and the practical reality of how oncologists interpret the numbers involves context that rarely makes it onto the lab report.

What These Tests Actually Measure

Cancer antigen tests measure the concentration of particular proteins circulating in your blood. These proteins are produced by cancer cells, by normal cells reacting to a tumor, or sometimes by perfectly healthy tissue under stress. The word “antigen” in this context is a bit of a historical artifact; it refers to the fact that these molecules were originally identified because the immune system could recognize them. In practice, they are simply biological markers whose blood levels tend to rise when certain cancers are present or growing.

Each marker is associated with one or a few cancer types. Some of the most commonly ordered ones include:

PSA (prostate-specific antigen): linked to prostate cancer, though the prostate also produces it normally
CA-125: associated mainly with ovarian cancer
CA 19-9: tied primarily to pancreatic cancer
CEA (carcinoembryonic antigen): used in colorectal cancer surveillance
CA 15-3: associated with breast cancer

The numbers on your lab report represent how much of that protein was found per unit of blood, usually in units per milliliter (U/mL) or nanograms per milliliter (ng/mL). Each marker has a reference range that defines what is considered “normal,” but those cutoffs are not clean dividing lines between healthy and sick. They are statistical thresholds chosen to balance the trade-off between catching real cancers and flagging too many false alarms.

Why Doctors Rarely Use These Tests for Screening

If cancer antigens go up when cancer is present, you might wonder why doctors do not just test everyone regularly and catch cancers early. The answer is that these markers are not specific or sensitive enough for that job in most situations. A screening test needs to be accurate in a population where cancer is rare, and that is where antigen tests fall apart.

CA 19-9 illustrates the problem clearly. In people who already have symptoms pointing toward pancreatic cancer, it has reasonable accuracy, with sensitivity around 79 to 81 percent and specificity around 82 to 90 percent. But when you try to use it as a screening tool in the general population, the positive predictive value drops to about 0.5 to 0.9 percent. That means if you tested a thousand healthy people and some came back positive, fewer than one in a hundred of those positives would actually have pancreatic cancer.¹ The rest would be chasing false alarms through invasive follow-up procedures.

CA-125 faces the same limitation with ovarian cancer. Despite being the primary ovarian cancer marker for four decades, no professional society recommends using it to screen average-risk women who have no symptoms. It simply has not improved survival when used that way.² PSA screening for prostate cancer has been the notable exception where population screening was widely adopted, but even that remains contentious because of how many men end up biopsied or treated for cancers that would never have caused harm.

Where Cancer Antigens Genuinely Shine

The real value of these tests is in monitoring people who have already been diagnosed. Once a cancer is identified and treatment begins, serial measurements of the relevant marker give oncologists a running gauge of whether things are heading in the right direction.

In ovarian cancer, for example, a significant drop in CA-125 within the first three months of chemotherapy is associated with better overall survival. Patients whose CA-125 does not decline meaningfully during that window tend to have a particularly poor prognosis, and that information can help doctors decide whether to adjust the treatment plan.³

CEA serves a similar role in colorectal cancer. After surgical removal of a colon tumor, regular CEA blood draws are part of standard follow-up to catch recurrence early.⁴ A rising CEA after surgery does not prove cancer has returned, but it triggers further investigation with imaging. A stable, low CEA is reassuring.

In prostate cancer, the speed at which PSA rises after treatment can itself be informative. Researchers have found that the PSA doubling time, meaning how quickly the level doubles, correlates with tumor aggressiveness. Patients whose PSA doubling time is longer than about 80 days after relapse tend to have a better prognosis than those whose levels climb faster.⁵ This kind of kinetic information, tracking the rate of change rather than a single snapshot, often matters more than any individual number.

What a High Level Tells You About Prognosis

Beyond monitoring treatment response, the absolute level of a cancer antigen at the time of diagnosis often carries prognostic weight. Higher levels tend to mean more advanced disease, though the relationship is not perfectly linear.

With CA 19-9 and pancreatic cancer, preoperative levels below the standard cutoff of 37 U/mL are associated with a median survival of roughly 32 to 36 months. Patients with elevated levels face a median survival of about 12 to 15 months. And levels above 100 U/mL suggest the tumor may not be surgically removable or has already spread.¹ A meta-analysis looking specifically at the question of predicting metastatic disease found that a CA 19-9 cutoff of 336 U/mL could identify metastases with about 90 percent sensitivity and 80 percent specificity, and that patients with metastatic disease had dramatically higher mean levels than those without.⁶

In breast cancer, the marker CA 15-3 behaves similarly. A large study of patients with early-stage breast cancer who started with normal CA 15-3 found that those whose levels rose during follow-up had roughly 76 percent higher recurrence risk compared to patients with stable levels. When the rise was substantial, at least one standard deviation above baseline, the risk jumped even more steeply.⁷ Interestingly, the CA 15-3 elevation typically preceded the actual clinical diagnosis of recurrence by a median of about ten months, suggesting the blood test can pick up signals the imaging has not caught yet.

In ovarian cancer, tumor subtype matters. Type II ovarian cancers, which are generally more aggressive and diagnosed at later stages, tend to produce significantly higher CA-125 levels than the slower-growing type I tumors.² So the same marker can indirectly hint at biology, not just burden.

When Elevated Levels Have Nothing to Do With Cancer

One of the most anxiety-producing realities of cancer antigen testing is that benign conditions can push levels up. Every major marker has a list of non-cancerous causes for elevation, and if you are not aware of them, an abnormal result can feel like a death sentence when it is actually something far less dire.

CA 19-9 rises in the presence of obstructive jaundice, a blockage of the bile duct that can happen from gallstones or inflammation. Between 10 and 60 percent of patients with obstructive jaundice will show elevated CA 19-9 without any cancer at all.¹ CA-125 can be elevated by endometriosis, liver disease, and even menstruation. PSA rises with benign prostatic enlargement, prostatitis, and even vigorous exercise or recent ejaculation. CEA goes up in smokers and people with inflammatory bowel disease.

Even more specialized markers are not immune to this problem. S-100, a protein used to track melanoma, showed mildly elevated levels in a quarter of patients with entirely benign diseases in one study. The culprits were mainly liver cirrhosis and kidney failure, both of which interfered with the body’s ability to clear the protein from the blood.⁸ This is a recurring theme across tumor markers: organ dysfunction that affects how the body metabolizes or excretes these proteins can mimic the pattern of cancer.

The PSA Gray Zone

PSA deserves its own discussion because it is the cancer antigen test most people will encounter in routine care, and also the one surrounded by the most confusion. A total PSA between about 4 and 10 ng/mL sits in what urologists call the “gray zone,” where both cancer and benign enlargement of the prostate are common explanations. A biopsy has traditionally been the next step, but a large share of those biopsies come back negative.

One refinement that helps is the percentage of “free” PSA. PSA circulates in the blood in two forms: bound to other proteins, and free. Cancer tends to produce more of the bound form, so a lower percentage of free PSA relative to total PSA raises suspicion. A multicenter trial found that in men with total PSA between 4 and 10 ng/mL, the median free PSA percentage was 12 percent in those who turned out to have cancer compared to 18 percent in those with benign disease. Using a free PSA cutoff of 25 percent could have avoided about one in five unnecessary biopsies while still catching 95 percent of cancers.⁹ Combining the free-to-total PSA ratio with PSA density, which accounts for prostate size, further improves the ability to distinguish cancer from benign enlargement.¹⁰

The broader lesson here applies to all cancer antigens: a single number in isolation rarely tells a clear story. Context, trend, and supplementary tests matter far more than any one reading.

Temporary Spikes During Treatment

If you are undergoing cancer treatment and your marker suddenly jumps, the instinct is to panic. But short-lived spikes in tumor marker levels can actually be a sign that therapy is working. When cancer cells are killed rapidly by chemotherapy or radiation, they release their contents into the bloodstream, including whatever antigen they were producing. This phenomenon, sometimes called tumor lysis, can cause a temporary surge in marker levels shortly after treatment begins.¹¹ The key word is temporary. If levels spike and then drop below where they started, the trajectory is favorable. If they spike and keep climbing, that is a different conversation.

This is one reason oncologists emphasize trends over snapshots. A single elevated reading mid-treatment does not mean much by itself. What matters is the direction over weeks and months.

Why Some People Cannot Produce Certain Markers

Here is a biological quirk that most patients never hear about: roughly 5 to 10 percent of the population is genetically unable to produce meaningful amounts of CA 19-9, regardless of whether they have pancreatic cancer or not. These individuals are “Lewis antigen negative,” meaning they lack the enzyme needed to attach a specific sugar molecule to the CA 19-9 protein.¹² In these patients, CA 19-9 will stay low or undetectable even in the presence of advanced cancer, leading to false-negative results that could delay diagnosis or give a misleading picture of treatment response.

The Lewis blood group system directly governs this. CA 19-9 expression occurs mainly in people who carry certain Lewis antigen phenotypes, while it is largely absent in those who lack them entirely.¹³ If you are one of these individuals and you develop pancreatic or biliary tract cancer, your doctor needs to know that CA 19-9 cannot be relied on and should use alternative markers or imaging for monitoring.

This is not unique to CA 19-9 in principle. Other markers have their own biological blind spots tied to genetics, organ function, or tumor biology. Some tumors simply do not secrete much of the protein associated with their cancer type, which means a normal marker level never fully rules out disease.

Circulating Tumor DNA and the Next Generation of Blood Tests

Traditional cancer antigens are proteins, and they have served oncology for decades. But the field is shifting toward a different kind of blood-based marker: circulating tumor DNA, or ctDNA. These are tiny fragments of DNA that tumors shed into the bloodstream, and they carry the genetic mutations specific to the cancer.

Compared to protein markers, ctDNA offers some meaningful advantages. It tends to be more cancer-specific, meaning fewer false positives from benign conditions, and more sensitive at detecting early recurrence or residual disease after surgery. In colorectal cancer, multiple studies have found ctDNA to be superior to CEA for picking up disease that remains after surgery. In breast cancer, ctDNA has outperformed CA 15-3 in detecting early recurrences. ctDNA also has a shorter half-life in the blood, so levels respond faster to changes in disease status, and it can sometimes predict which therapies a tumor is likely to respond to or reveal how it is developing resistance.¹⁴

The trade-offs are practical rather than scientific. ctDNA tests are more expensive, less widely available, take longer to report results, and are not as well standardized across laboratories as protein antigen tests.¹⁴ For now, protein markers remain the workhorse in most oncology clinics, with ctDNA increasingly used as a complement rather than a replacement.

An even more ambitious development is the multi-cancer early detection (MCED) test, a single blood draw that uses cell-free DNA along with machine learning algorithms to look for a shared cancer signal across more than 50 cancer types and predict where in the body the signal is coming from.¹⁵ These tests are still being evaluated in large clinical studies, and they are not yet part of standard care. But they represent a fundamentally different approach from the one-marker-one-cancer model that has defined the field for decades.

How to Think About Your Own Results

If you have been handed a lab report with a cancer antigen level on it, the most important thing to understand is that the number only means something in context. A slightly elevated CA-125 in a premenopausal woman with no other symptoms is far less alarming than the same level in a postmenopausal woman with a pelvic mass on imaging. A PSA of 6 ng/mL in a man with a known enlarged prostate and a stable trend over years is a different situation from a PSA of 6 that jumped from 2 in the past twelve months.

For people being monitored after cancer treatment, the trend line is almost always more informative than any single reading. A level that is slowly creeping up over several measurements tells a clearer story than one that bounced once. And if your levels are being tracked, try not to obsess over small fluctuations between individual draws. Laboratory assays have inherent variability, and your body’s production of these proteins fluctuates day to day for reasons that have nothing to do with cancer.

It is also worth asking your oncologist which marker they are using and why, especially if you have a cancer type where the marker has known limitations. If you happen to be Lewis antigen negative, for example, CA 19-9 should not be the primary tool guiding your treatment decisions for pancreatic cancer. And if your marker rises during the first cycle of chemotherapy, ask whether a treatment-related spike is a plausible explanation before assuming the worst.

The era of relying on a single protein measurement to make major cancer decisions is gradually giving way to panels of markers, genomic tests, and imaging combined. But protein antigens remain deeply embedded in oncology practice, and understanding what they can and cannot do puts you in a much better position to interpret the numbers that show up on your chart.