What Tumor Markers in Blood Tests Can (and Can't) Tell You

A tumor marker is usually a protein that shows up in the blood in larger amounts than normal. Sometimes cancer cells make it directly. Other times it appears because the body is reacting to a tumor's presence. Either way, the basic idea behind tumor marker testing is simple: measure the substance, and use the level to learn something about cancer.

That simplicity is appealing, which is part of why tumor marker tests get so much attention. But clinical guidelines are blunt about a key limitation. No blood-based tumor marker available today is accurate enough, on its own, to diagnose cancer. Confirming a cancer diagnosis still requires imaging and a tissue biopsy, where a small sample of suspicious tissue is examined directly. A marker level can raise suspicion or help guide what happens next, but it cannot stand in for that direct confirmation.

Why a High Number Doesn't Automatically Mean Cancer

To understand why, it helps to think about how any test can fail in two different directions. A test can miss people who truly have the disease, or it can flag people who don't. Statisticians call these two properties sensitivity, how well a test catches real cases, and specificity, how well it avoids false alarms in healthy people. A good screening test needs to be strong on both counts.

Most tumor markers fall short on at least one side. A protein that rises with cancer often rises for other reasons too: inflammation, infection, benign growths, or even normal variation between individuals. When that happens, a healthy person can get a positive-looking result. This is called a false positive, and it matters because it can trigger more testing, anxiety, and sometimes invasive procedures for something that was never cancer at all.

There's a flip side as well. Some early cancers don't produce enough of a marker to register as abnormal, so a normal result offers no guarantee. A tumor marker test is a bit like a car alarm wired to detect vibration. It will catch a break-in sometimes, but it will also go off for a passing truck, and it can stay silent during a quiet, careful theft. Neither a high reading nor a normal one tells the whole story by itself.

The Prostate Cancer Lesson: What PSA Screening Trials Revealed

Prostate-specific antigen, or PSA, is a protein made by the prostate gland, and it remains one of the most studied tumor markers in medicine. Two major randomized trials, one run in the United States and one across several European countries, followed large groups of men for years to see whether routine PSA screening actually saved lives. The results were mixed and, at best, modest: screening found more early-stage cancers, but the effect on deaths from prostate cancer was small and inconsistent between the two trials.

The harder problem was overdiagnosis, detecting cancers so slow-growing that they would never have caused symptoms or shortened a man's life. Finding those cancers led to real treatment, including surgery and radiation, with real side effects, for disease that may never have needed treating at all. The PSA story is now the clearest cautionary example in oncology of how a useful biological signal can still make a poor general screening tool.

Why CA-125 Hasn't Become an Ovarian Cancer Screening Test

CA-125 is a protein that often rises in ovarian cancer, which makes it a natural candidate for early detection. A large UK-based randomized trial tested this idea directly, screening tens of thousands of women over many years using CA-125 measurements and ultrasound. Despite that scale and duration, the trial did not show a clear reduction in ovarian cancer deaths among women who were screened.

CA-125 also rises in many ordinary, noncancerous conditions, including endometriosis, fibroids, and even menstruation itself. In a screening context, that lack of specificity translates into false alarms, and false alarms in ovarian cancer screening are not trivial. They can lead to unnecessary surgery to remove ovaries that turn out to be healthy. The CA-125 trial result reinforced a pattern already visible with PSA: a marker tied to a real biological process in cancer does not automatically make a workable population screening test.

Where These Markers Do Earn Their Place

None of this means tumor markers are useless. Their real value shows up after a cancer diagnosis has already been made, not before. Carcinoembryonic antigen, or CEA, is a good example. It isn't used to screen for or diagnose colorectal cancer, but once someone is being treated for it, tracking CEA over time helps show whether treatment is working and can offer an early signal if the cancer has returned.

Alpha-fetoprotein, or AFP, plays a similar supporting role in liver cancer and in certain tumors arising from reproductive cells, called germ cell tumors. It's used mainly for monitoring, often alongside imaging, rather than catching early disease on its own; a meaningful share of early liver cancers don't raise AFP enough to be caught by the marker alone. A related marker, CA 19-9, is sometimes used in a similar monitoring role for pancreatic cancer, with the same caveat that non-cancerous conditions affecting the bile ducts or pancreas can push its levels up too.

The common thread is context. These markers work best as a way to watch a known cancer over time, comparing a person's own numbers against their own baseline, rather than as a yes-or-no verdict applied to someone with no prior diagnosis.

A Newer Approach: Liquid Biopsy and Multi-Analyte Testing

In recent years, researchers have tried to get around the limits of single-protein markers by testing for several things at once. One notable study combined measurements of multiple proteins with fragments of circulating tumor DNA, small pieces of genetic material shed into the blood by tumor cells, a strategy sometimes called liquid biopsy. The 2018 study, led by a research team including Joshua Cohen and colleagues, showed that this combined approach could detect and roughly localize several types of surgically treatable cancers in blood samples.

It's an encouraging direction, but an early one. Combining several imperfect signals can improve overall accuracy, yet it does not automatically solve the underlying problems of false positives and missed cases. These multi-analyte and liquid biopsy approaches remain investigational for most uses. They are not yet established replacements for imaging and tissue biopsy in diagnosing cancer, though they represent an active and evolving area of research.

Putting a Number in Its Proper Place

The common thread running through PSA, CA-125, CEA, AFP, and the newer blood-based tests is that a tumor marker is a clue, not a verdict. An elevated result can be worth investigating further. A normal result doesn't rule cancer out. Decisions about whether to test, and how to interpret a result, depend on a person's individual history, symptoms, and other findings, not the marker number alone.

That nuance is easy to lose in casual conversation, where a lab result over the normal range can feel like a diagnosis in itself. The research record says otherwise: these tests gained their clinical value by being used carefully, in specific situations, alongside imaging and tissue confirmation, rather than as a stand-alone cancer check. Whether the newer combined blood tests will eventually change that balance is still an open question the field is working through.