CA 19-9: A Pancreatic Cancer Monitoring Tool with Big Caveats

CA 19-9 (also called sialyl Lewis a) is a carbohydrate antigen sitting on glycoproteins and glycolipids in the pancreaticobiliary tree, the GI tract, and other epithelial tissues. Pancreatic adenocarcinoma cells often overexpress it. So do cells under stress from cholestasis, inflammation, or infection.

Synthesis depends on a working fucosyltransferase encoded by the FUT3 (Lewis) gene. People homozygous for inactivating FUT3 alleles, the Lewis-negative phenotype, cannot make CA 19-9 at all. About 5 to 10 percent of people of European ancestry, with variation by population. If you are Lewis-negative, your CA 19-9 result is uninformative. Lewis-status testing is worth considering before relying on CA 19-9 for surveillance.

At a 37 U/mL cutoff in symptomatic patients with imaging findings, sensitivity and specificity for pancreatic cancer are each around 80 percent. In screening unselected populations, the positive predictive value collapses, which is why no major society recommends it for that purpose.

The usual cutoff is below 37 U/mL, with small variation by lab:

• Below 37 U/mL: "normal" range. Does not rule out cancer, especially in Lewis-negative individuals.
• 37 to 100 U/mL: mild elevation. Most often benign: pancreatitis, cholestasis, smoking, recent biliary instrumentation.
• 100 to 1000 U/mL: moderate elevation. Cancer is on the differential, but cholestasis and chronic biliary disease still account for many cases. Imaging is the next step.
• Above 1000 U/mL: high probability of pancreatic, biliary, or another GI malignancy if cholestasis has been excluded. Markedly elevated values often track with locally advanced or metastatic disease.
• Above 100 U/mL after biliary obstruction has been relieved: persistent elevation despite drainage shifts the probability strongly toward malignancy.

One pattern matters more than any cutoff. A CA 19-9 that does not normalize after biliary drainage is much more concerning than the same number while stones or stricture are still in place.

The benign list is long, and several entries are common:

• Cholestasis from any cause. Gallstones with biliary obstruction, common bile duct stricture, primary sclerosing cholangitis, primary biliary cholangitis. Cholestasis alone can drive CA 19-9 into the thousands. The first interpretation question is always, "Is the bile flowing?"
• Acute and chronic pancreatitis.
• Cirrhosis and chronic hepatitis.
• Cystic fibrosis. Chronic CA 19-9 elevation is common.
• Diabetes mellitus. Modestly elevated baselines.
• Smoking.
• Endometriosis, ovarian cysts, hydronephrosis, thyroid disease, rheumatoid arthritis. Less common but documented.
• Other GI cancers. Cholangiocarcinoma, gallbladder, gastric, colorectal, hepatocellular. Some non-GI cancers (lung, gynecologic) occasionally do the same.

The right way to read an elevated CA 19-9:

1. Check liver function tests. If alkaline phosphatase, GGT, and bilirubin point to cholestasis, the elevation is at least partly explained. Relieve the obstruction, then recheck.
2. Get imaging. CT or MRI of the abdomen with pancreas protocol. CA 19-9 without imaging in a symptomatic patient is an incomplete workup.
3. If imaging is negative and cholestasis is excluded, repeat CA 19-9 in 4 to 6 weeks. Mild benign elevations often drift back down. Rising values warrant endoscopic ultrasound and continued workup.
4. Consider Lewis-status if a known pancreatic cancer patient has a stubbornly normal CA 19-9.

A low CA 19-9 in a treated pancreatic cancer patient is reassuring and consistent with treatment response or remission. There is no clinically meaningful "too low" value.

An undetectable CA 19-9 in someone with an obvious pancreatic mass on imaging should raise the Lewis-negative question. About 5 to 10 percent of pancreatic cancer patients cannot produce CA 19-9 because of FUT3 genotype. In them, CA 19-9 cannot be used for surveillance. CEA, CA-125, and serial imaging are used instead.

CA 19-9 has substantial day-to-day variability and is acutely sensitive to bile flow. The clinical signal lives in the trend, especially in three contexts:

• After biliary drainage. CA 19-9 should fall substantially within 1 to 2 weeks once cholestasis is relieved. A persistent post-drainage elevation strongly suggests underlying malignancy.
• After resection of pancreatic cancer. CA 19-9 should fall to within normal range within 4 to 8 weeks. Failure to normalize predicts early recurrence and worse survival. A subsequent rise above the post-op nadir is a sensitive recurrence signal, often before imaging changes.
• During chemotherapy for advanced disease. A drop of more than 50 percent from baseline correlates with treatment response in metastatic pancreatic cancer. Rising values during therapy suggest progression.

Outside known disease, the trend matters less than the explanation. An isolated CA 19-9 of 80 in someone with gallstones is a different problem from a CA 19-9 of 80 in someone with a pancreatic mass.

• Markedly elevated CA 19-9 (above 1000 U/mL) without cholestasis. High probability of malignancy. Prioritize cross-sectional imaging.
• Persistently elevated CA 19-9 after biliary drainage. The single most useful diagnostic pattern this test produces.
• Rising CA 19-9 after pancreatic cancer surgery. Often the earliest recurrence signal, sometimes months before imaging changes.
• Failure to normalize CA 19-9 within 4 to 8 weeks of resection. Predicts early recurrence and may influence adjuvant therapy.
• New, otherwise unexplained CA 19-9 elevation in someone with new abdominal symptoms, weight loss, or jaundice. Workup with cross-sectional imaging and consideration of EUS.

What does not warrant action: an isolated mild CA 19-9 elevation ordered as part of a "tumor marker panel" in an asymptomatic person with normal liver tests and no imaging findings. The most likely outcome of further workup is a normal scan, an unexplained number, and lasting anxiety.

• Alkaline phosphatase, GGT, bilirubin, ALT/AST. Define whether cholestasis is the explanation for an elevated CA 19-9.
• CT or MRI abdomen with pancreas protocol. The diagnostic test for any concerning CA 19-9. Imaging dominates this decision space.
• Endoscopic ultrasound (EUS). Best test for small pancreatic lesions; high-yield in high-risk surveillance and in patients with elevated CA 19-9 and ambiguous imaging.
• CEA. Sometimes elevated alongside CA 19-9 in pancreatic, gastric, and colorectal cancers; complementary in monitoring.
• CA-125. Occasionally helpful in mucinous pancreatic neoplasms.
• Lewis blood group genotype/phenotype. Worth considering if CA 19-9 is being used for surveillance and the value is suspiciously low.