Objective: To assess the accuracy of ultrasonography (US), computed tomography (CT), and endoscopic retrograde cholangiopancreatography (ERCP) in distinguishing between benign and malignant causes of jaundice and cholestasis without jaundice.
Design: Prospective study.
Setting: University Hospital in Finland.
Subjects: A consecutive series of patients with jaundice (n = 187) or cholestasis without jaundice (n = 33).
Main outcome measures: Correlation between diagnosis on imaging and final diagnosis at follow up 6 months later. Final diagnoses made on histology (n = 79), cytology (n = 5), operative or endoscopic findings (n = 96), and clinical course or serology (n = 40).
Results: The most common benign disease was choledocholithiasis (n = 83) and the most common malignant disease was carcinoma of pancreas (n = 33). The benign nature of the extrahepatic obstruction was correctly defined by US, CT, and ERCP in 53%, 53%, and 90% of patients, respectively, and the corresponding figures for choledocholithiasis were 22%, 25%, and 79% (ERCP compared with each of the other techniques, p < 0.0001). Intrahepatic benign diseases were diagnosed by US and CT in a third of cases. Malignant extrahepatic obstruction was correctly diagnosed in 57%, 80%, and 83%, respectively and the corresponding figures for pancreatic cancer were 60%, 97%, and 89% (US compared with CT, p < 0.01, and with ERCP, p < 0.05). Intrahepatic malignant lesions were diagnosed by US, CT, and ERCP in 100%, 77%, and 60% of patients, respectively.
Conclusions: When the obstruction was benign and extrahepatic ERCP was the most accurate, but when it was malignant CT was comparable. Intrahepatic disease was best diagnosed by US and CT. The results emphasise that the three methods of imaging are complementary.