Reappraisal of intraoperative ultrasonography for pancreatobiliary carcinomas: assessment of malignant portal venous invasion.

Abstract

Conventional preoperative imaging modalities are unreliable for assessing portal venous invasion by pancreatobiliary carcinoma. We evaluated the usefulness of intraoperative ultrasonography for detecting pancreatobiliary carcinoma and assessing portal venous invasion, compared with other imaging modalities. Ninety-one patients with pancreatic carcinoma (n = 66) or bile duct carcinoma (n = 25) underwent ultrasonography, computed tomography, angiography, and endoscopic ultrasonography preoperatively. All these patients underwent tumor resection, with (n = 23) or without (n = 68) portal vein resection, after intraoperative ultrasonography. Portal venous invasion was histologically examined in all patients. Intraoperative ultrasonography was significantly more sensitive (100%) than ultrasonography (79%), computed tomography (81%), and angiography (54%) for detecting carcinomas, especially bile duct carcinomas and small (< or = 2.0 cm) tumors. Portal venous invasion was confirmed histopathologically in 25 patients. For diagnosing portal venous invasion, intraoperative ultrasonography was more sensitive (92%) and specific (92%) than ultrasonography (56% and 73%), computed tomography (64% and 79%), and angiography (76% and 83%), respectively. Endoscopic ultrasonography showed a 95% detectability for carcinomas and a 92% accuracy for assessing portal venous invasion. Intraoperative ultrasonography is a simple and accurate procedure for detection of pancreatobiliary carcinomas and assessment of portal venous invasion.