MR cholangiography versus cholangioscopy for evaluation of longitudinal extension of hilar cholangiocarcinoma

Background: The utility of magnetic resonance cholangiography for assessment of longitudinal tumor extension of hilar cholangiocarcinoma was investigated with reference to findings by percutaneous transhepatic cholangioscopy. Methods: Ninety-nine patients with hilar cholangiocarcinoma underwent both magnetic resonance cholangiography and percutaneous transhepatic cholangioscopy. Longitudinal tumor extension was described with the Bismuth-Corlette classification. Hilar cholangiocarcinoma was classified morphologically into stenotic, diffuse sclerosing, and polypoid types based on selective cholangiographic findings obtained during percutaneous transhepatic cholangioscopy. Agreement between percutaneous transhepatic cholangioscopy and magnetic resonance cholangiography according to the Bismuth-Corlette classification was compared. The degree of agreement between magnetic resonance cholangiography and percutaneous transhepatic cholangioscopy according to each morphologic type was also compared in each subgroup without reference to Bismuth-Corlette type. Results: The overall agreement between magnetic resonance cholangiography and percutaneous transhepatic cholangioscopy with regard to Bismuth-Corlette types was 87.9% (kappa = 0.832, p < 0.01). The agreement of magnetic resonance cholangiography for each Bismuth-Corlette type with reference to percutaneous transhepatic cholangioscopy was as follows: type I (n = 18), 16/18 (88.9%); type II (n = 16), 14/16 (87.5%); type IIIa (n = 23), 19/23 (82.6%); type IIIb (n = 14), 14/14 (100%); and type IV (n = 28), 24/28 (85.7%). The overall agreement between magnetic resonance cholangiography and percutaneous transhepatic cholangioscopy for Bismuth-Corlette type according to selective cholangiographic findings was as follows: stenotic type, 58/61 (95.1%, kappa = 0.929, p < 0.01); diffuse sclerosing type, 12/16 (75%, kappa = 0.619, p < 0.01); and polypoid type, 17/22 (77.3%, kappa = 0.696, p < 0.01). Conclusion: There is good overall agreement between magnetic resonance cholangiography and percutaneous transhepatic cholangioscopy on longitudinal extension of hilar cholangiocarcinoma. Especially for the stenotic type of hilar cholangiocarcinoma (based on selective cholangiographic findings), magnetic resonance cholangiography may replace percutaneous transhepatic cholangioscopy in the determination of longitudinal tumor extension. For polypoid or diffuse sclerosing types, however, percutaneous transhepatic cholangioscopy is