We evaluated the clinical utility of a new tumor marker tissue polypeptide specific antigen (TPS) in the diagnosis of pancreatic carcinoma. Serum concentrations were determined in 113 patients with jaundice, in 18 patients with laboratory values suggesting cholestasis and in 60 patients with suspicion of chronic pancreatitis or pancreatic tumor. Twenty-four of these 191 patients had pancreatic carcinoma and 2 patients had carcinoma of the papilla of Vater. The highest median serum TPS value was detected in patients with malignant liver disease, but high median values were also measured in patients with pancreatic cancer, bile duct cancer or benign liver disease. The sensitivity of TPS was 50.0%, with a specificity of 73% and an efficiency of 70%. In comparison with carcinoembryonic antigen (CEA), CA 50 and CA 242, the TPS test showed lower sensitivity, but the differences in specificity and negative predictive value were considerably smaller. The utility of TPS as a complementary test was also analyzed. When TPS was combined with other marker tests, their specificities clearly improved, being highest in the combination of TPS and CA 242 (92.5%). In this combination, efficiency and positive likelihood ratio were also clearly better (85% and 5.6) than those of the marker tests alone. In conclusion, TPS seems less accurate than CEA, CA 50 or CA 242 in the diagnosis of pancreatic cancer, but because of its different nature it may be considered to be used as a complementary test.