BACKGROUND: Early detection and accurate staging are crucial for planning treatment and improving survival rate of hilar cholangiocarcinomas. This study investigated the diagnostic value of a three dimensional, spoiled gradient echo, T1-weighted magnetic resonance (MR) imaging sequence (3D volumetric interpolated breath-hold examination, 3D-VIBE) in the preoperative evaluation of hilar cholangiocarcinoma. METHODS: Thirty-one patients with surgically and histologically confirmed hilar cholangiocarcinomas underwent preoperative MR imaging examination. Unenhanced two-dimensional T1- and T2-weighted images, 2D MR cholangiopancreatographs (MRCP), gadolinium enhanced 3D-VIBE images in the early arterial, late arterial and portal venous phases followed by 2D T1-weighted images in the equilibrium phase were acquired. Images from 3D-VIBE, 2D T1-weighted enhanced sequences and 2D MRCP were interpreted by two abdominal radiologists through consensus reading in blind manner, focussing on the assessment of the morphological type, the longitudinal extent of tumor infiltration in the bile ducts and the involvement of neighbouring blood vessels. The accuracy of 3D-VIBE and 2D T1-weighted enhanced sequences in assessing the tumor resectability was compared. RESULTS: All the 31 tumors were directly displayed and accurately classified on 3D-VIBE images whereas 8 periductal infiltrating tumors (8/31, 25.8%) were not depicted on 2D T1-weighted enhanced images. Using the Bismuth Corlette classification, 3D-VIBE was closer to MRCP in delineating the intraductal extent of tumor infiltration than 2D T1-weighted enhanced (28/31, 90.3%; 10/31, 32.3%; chi2 = 22.0, P < 0.05). Involvement of the hepatic artery, the portal venous trunk and their branches was shown more frequently on 3D VIBE than 2D T1-weighted enhanced images. The positive predictive value and accuracy of 3D-VIBE (84.2%; 90.3%) for assessing tumor resectability were higher than those of 2D T1-weighted enhanced images (64.0%; 71.0%, all P < 0.05). CONCLUSION: Gadolinium enhanced 3D-VIBE is better than 2D T1-weighted enhanced sequence in the preoperative assessment of the morphologicalal type, the intraductal infiltrating extent and the tumor resectability of hilar cholangiocarcinomas.
BACKGROUND: Early detection and accurate staging are crucial for planning treatment and improving survival rate of hilar cholangiocarcinomas. This study investigated the diagnostic value of a three dimensional, spoiled gradient echo, T1-weighted magnetic resonance (MR) imaging sequence (3D volumetric interpolated breath-hold examination, 3D-VIBE) in the preoperative evaluation of hilar cholangiocarcinoma. METHODS: Thirty-one patients with surgically and histologically confirmed hilar cholangiocarcinomas underwent preoperative MR imaging examination. Unenhanced two-dimensional T1- and T2-weighted images, 2D MR cholangiopancreatographs (MRCP), gadolinium enhanced 3D-VIBE images in the early arterial, late arterial and portal venous phases followed by 2D T1-weighted images in the equilibrium phase were acquired. Images from 3D-VIBE, 2D T1-weighted enhanced sequences and 2D MRCP were interpreted by two abdominal radiologists through consensus reading in blind manner, focussing on the assessment of the morphological type, the longitudinal extent of tumor infiltration in the bile ducts and the involvement of neighbouring blood vessels. The accuracy of 3D-VIBE and 2D T1-weighted enhanced sequences in assessing the tumor resectability was compared. RESULTS: All the 31 tumors were directly displayed and accurately classified on 3D-VIBE images whereas 8 periductal infiltrating tumors (8/31, 25.8%) were not depicted on 2D T1-weighted enhanced images. Using the Bismuth Corlette classification, 3D-VIBE was closer to MRCP in delineating the intraductal extent of tumor infiltration than 2D T1-weighted enhanced (28/31, 90.3%; 10/31, 32.3%; chi2 = 22.0, P < 0.05). Involvement of the hepatic artery, the portal venous trunk and their branches was shown more frequently on 3D VIBE than 2D T1-weighted enhanced images. The positive predictive value and accuracy of 3D-VIBE (84.2%; 90.3%) for assessing tumor resectability were higher than those of 2D T1-weighted enhanced images (64.0%; 71.0%, all P < 0.05). CONCLUSION:Gadolinium enhanced 3D-VIBE is better than 2D T1-weighted enhanced sequence in the preoperative assessment of the morphologicalal type, the intraductal infiltrating extent and the tumor resectability of hilar cholangiocarcinomas.