The diagnostic accuracy of gray-scale ultrasonography compared with ERP and arteriography in the detection of pancreatic carcinoma.

Forty patients who underwent ultrasound, endoscopic retrograde pancreatography (ERP), and selective arteriography examinations and who had a finding of or were suspected of having pancreatic carcinoma in one or several of these examinations form the basis of the study. Nineteen pancreatic carcinomas were verified at operation. Eighteen of these could be detected with ERP, 18 with arteriography, and 15 with sonography. ERP had no, arteriography one and sonography one false-negative finding. All methods had difficulties in distinguishing carcinoma from other pancreatic diseases, and therefore there were many false-positive findings.