Comparison of multislice computed tomography and endoscopic ultrasonography with operative and histological findings in suspected pancreatic and periampullary malignancy.

BACKGROUND: This study compared multislice computed tomography (MSCT) with endoscopic ultrasonography (EUS) in the diagnosis and staging of pancreatic and periampullary malignancy.
METHODS: Data were collected prospectively on patients having MSCT and EUS for suspected pancreatic and periampullary malignancy.
RESULTS: Eighty-four patients had MSCT and EUS, of whom 35 underwent operative assessment (29 resections). In assessing malignancy, there was no significant difference between MSCT and EUS, and agreement was good (82 per cent, kappa = 0.49); the sensitivity and specificity of MSCT were 97 and 87 per cent, compared with 95 and 52 per cent respectively for EUS (P = 0.264). For portal vein/superior mesenteric vein invasion, MSCT was superior (P = 0.017) and agreement was moderate (72 per cent, kappa = 0.42); the sensitivity and specificity were 88 and 92 per cent for MSCT, and 50 and 83 per cent for EUS. For resectability, there was no significant difference and agreement was good (78 per cent, kappa = 0.51). EUS had an impact on the management of 14 patients in whom MSCT suggested benign disease or equivocal resectability.
CONCLUSION: MSCT is the imaging method of choice for pancreatic and periampullary tumours. Routine EUS should be reserved for those with borderline resectability on MSCT.