Abstract
The serum values of the tumour markers carcinoembryonic antigen (CEA), cancer-associated carboanhydrate antigens CA 50 and CA 242 were evaluated in 193 patients with hepatopancreato-biliary diseases by receiver operating characteristic (ROC) curve analysis in order to compare their diagnostic accuracy in pancreatic cancer (n = 26), and to define optimal cut-off levels for the serum values of these tumour markers in the diagnosis of pancreatic cancer. The ROC analysis showed that all marker tests are considerably sensitive (77-81%) at the specificity level of 80%. The CA 242 test was more sensitive than CEA and CA 50 at high specificity levels (> 0.90) but slightly less sensitive at low specificity levels (< 0.60). The CEA test and CA 50 test performed equally well at high and low specificity levels. According to this study, it would seem optimal to use the cut-off level of 4.1 ng ml-1 for CEA, and the level of 137 U ml-1 for CA 50, since they gave a sensitivity of 77% at the specificity levels of 83% and 84%, respectively. For CA 242 the optimal cut-off level was 21 U ml-1, which gave a sensitivity and specificity of 81%. In conclusion, the results of ROC curve analysis suggest that the CA 242 test has an advantage over CEA and CA 50 because of its higher specificity in pancreatic cancer. In addition, it would seem reasonable to use higher cut-off values than what has been recommended for CEA and CA 50 in the diagnosis of pancreatic cancer, but for CA 242 the recommended cut-off level of 20 U ml-1 seems appropriate.
Highlights
Clinical assessment, laboratory tests, and imaging methods were performed as previously decribed (Pasanen et al, 1992)
The diagnosis of chronic pancreatitis was based on histology in seven cases, on cytology in one, on secretin-cerulein test in six, on the imaging methods in 14 and on clinical course of the disease in six
The results showed that CA 242 was significantly more sensitive than carcinoembryonic antigen (CEA) and CA 50 at high (>0.90) specificity levels, but slightly less sensitive at low (
Summary
Laboratory tests, and imaging methods (ultrasound, computed tomography and endoscopic retrograde cholangio-pancreatography) were performed as previously decribed (Pasanen et al, 1992). Liver biopsy was obtained or secretin-cerulein test was performed if hepatocellular disease or chronic pancreatitis was suspected. The final diagnosis of a pancreatic cancer or cancer of the ampulla of Vater was based on histology in 16 cases, on cytology in three, on operative or endoscopic macroscopic morphologic findings in three, and on the imaging methods in four. The diagnosis of chronic pancreatitis was based on histology in seven cases, on cytology in one, on secretin-cerulein test in six, on the imaging methods in 14 and on clinical course of the disease in six
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
