The aim of this study was to evaluate the diagnostic accuracy of: a) three imaging techniques: ultrasonography (US), computed tomography (CT) and endoscopic retrograde cholangiopancreatography (ERCP), and b) the serum tumor markers CEA, CA 19-9 and CA 125, in the diagnosis of pancreatic cancer.
A total of 137 patients were prospectively evaluated. Pancreatic cancer was diagnosed in 25 patients; chronic pancreatitis in 24; acute pancreatitis in 22; extrapancreatic malignancies in 24, and benign digestive disease in 42. The reference interval for each marker was determined in 36 healthy volunteers. Diagnostic accuracy was determined using receiver operating characteristic (ROC) curves. Each diagnostic test was used in order: a) to diagnose pancreatic cancer when the disease was clinically suspected, and b) to differentiate between pancreatic cancer and chronic pancreatitis.
CT (0.976-0.888) and US (0.857) diagnostic accuracy resulted greater than the best serum tumor marker CA 19-9 (0.755-0.786) (p = NS), in both diagnostic conditions. To obtain a diagnostic specificity of 90% in pancreatic cancer, the CA 19-9 cutoff level should increase up to 3 and 7.5 folds being in this case better than CEA and CA 125 (p < 0.05).
CT scan offers the greatest accuracy in the diagnosis of PC. Among the tumor markers, CA 19-9 offers the best accuracy in the diagnosis of pancreatic cancer.