Patients with pancreatic cancer usually lack signs and symptoms in the early course of the disease. Even when malignancy is suspected, differential diagnosis between benign and malignant pancreatic disorders may be difficult with current methods. An increasing interest has been focused on the utility of immunological tumour markers. CEA has been widely used since the early seventies, but the results in diagnosis of pancreatic cancer have been disappointing. Tumour marker tests for CA 19-9 and CA 50 are based on monoclonal antibodies to colonic carcinoma cell lines. CA 19-9 and CA 50 are strongly expressed in most tissue specimens from pancreatic carcinomas, but are also found in normal pancreas and benign pancreatic diseases. The CA 19-9 and CA 50 antigens are shed or released into the circulation, and are found in increased concentrations in 70-80% of patients with pancreatic cancer. Also 50-65% of patients with small resectable carcinomas have elevated CA 19-9 and CA 50 levels, although very high serum concentrations usually indicate advanced disease. Slightly elevated serum CA 19-9 and CA 50 levels are seen in some patients with benign pancreatic diseases, more often in acute than in chronic pancreatitis. Elevated values are often observed in patients with benign obstruction of the common bile duct, particularly in patients with cholangitis. In patients with jaundice of hepatocellular origin, the CA 19-9 and CA 50 levels are lower than in extrahepatic cholestasis. CA 19-9 and CA 50 have better diagnostic accuracy for pancreatic cancer than CEA, CA 125, DU-PAN-2, TPA and PSTI/TATI. However, the sensitivities and specificities of CA 19-9 and CA 50 are too low for screening of an asymptomatic population. Nevertheless, CA 19-9 and CA 50 have in our experience shown to be useful complements to other diagnostic methods in symptomatic patients with suspicion of pancreatic cancer. Combinations of different markers improve the sensitivity only slightly compared to the use of CA 19-9 or CA 50 alone. Follow-up using CA 19-9 and CA 50 is a simple and sensitive way of monitoring the postoperative course of patients with pancreatic cancer, and may give a lead time of several months for a recurrence compared to conventional methods.