Hilar Cholangiocarcinoma: Diagnosis and Evaluation of Resectability with the Three-Dimensional Thin-section Contrast-enhanced Dynamic MR Imaging Sequence
Longlin Yin,Bin Song,Juan Xu,Yingchun Li,Yang Yang,Xian Chen,Zhenlin Li,Changxian Li,Kexiang Zhong,Jiayu Sun
2007-01-01
Abstract:Hilar cholangiocarcinoma is a malignant tumour with high mortality, thus early diagnosis of tumour and accurate staging are crucial for planning of treatment and survival improvement. The purpose of this study was to evaluate the diagnostic value of the three-dimensional (3D) T1-weighted thin-section dynamic enhanced MR imaging sequence in the preoperative evaluation of hilar cholangiocarcinoma. Thirty-one patients with hilar cholangiocarcinomas confirmed by surgery and pathology were performed preoperative MR imaging examination, sequences including two-dimensional (2D) T1- and T2-weighted plain scanning, 2D MR cholangiopancreatography (MRCP), 3D T1-weighted thin-section dynamic enhanced tri-phasic scanning in the early arterial, late arterial and portal venous phases, and 2D T1-weighted enhanced scanning in the equilibrium phase. Imaging data of 3D T1-weighted enhanced sequence, 2D T1-weighted enhanced sequence, and 2D MRCP were interpreted by two abdominal radiologists through consensus reading in blind manner, focusing on the assessment of the morphological type, the longitudinal extent of tumour infiltration in the bile ducts and the involvement of neighboring blood vessels. The accuracy between 3D T1-weighted and 2D T1-weighted enhanced sequences in assessing the tumour resectability was compared. 3D T1-weighted enhanced images directly displayed hilar tumour in 31 patients and accurately classified their morphological types, whereas 2D T1-weighted enhanced images missed 8 patients of periductal-infiltrating tumours (8/31, 25.8%). Using the Bismuth-Corlette classification, 3D T1-weighted enhanced sequence was closed to MRCP (28/31, 90.3%) in delineating the intraductal extent of tumour infiltration, but 2D T1-weighted enhanced sequence (10/31, 32.3%) obviously underestimated the extent. Difference between the two MR sequences had statistical significance (P<0.05). Involvement of the hepatic artery, the portal venous trunk and their branches was shown more frequently on 3D T1-weighted enhanced images than on 2D T1-weighted enhanced images. The positive predictive value and accuracy of 3D T1-weighted sequence (84.2%, 90.3%) for assessing tumour resectability were higher than those of 2D T1-weighted enhanced sequence (64.0%, 71.0%). Therefore, 3D T1-weighted thin-section dynamic enhanced sequence is better than 2D T1-weighted enhanced sequence in the preoperative assessment of the morphological type, the intraductal infiltrating extent and the tumour resectability of hilar cholangiocarcinomas.