Fifty-three patients underwent endoscopic retrograde cholangiopancreatography (ERCP). Thirty-eight patients had pancreatic disease, diagnosed on the basis of surgical and/or instrumental-clinical findings, and consisted of 17 cases of carcinoma (group A), 10 cases of chronic pancreatitis with benign stenosis of the pancreatic duct (group B), and 11 cases of chronic pancreatitis without stenosis (group C). Linear discriminant analysis was employed to evaluate differential diagnosis data relative to secondary pancreatic ducts. With this method, 16/17 patients (94%) were correctly assigned to group A (high sensitivity), 7/8 patients (88%) to group B, and 7/10 patients (70%) to group C. A readily found radiological sign, such as abrupt and/or irregular stenosis of the pancreatic duct enables the correct diagnosis of carcinoma, but while this criterion is highly specific, it is poorly sensitive. If a clearly neoplastic stenosis of the main duct is not present, an analysis of the secondary ducts is determinant.