Objective:

To determine the added value of gadolinium-enhanced dynamic magnetic resonance (MR) images compared with unenhanced and MR cholangiography (MRC) images, to evaluate the longitudinal extension of bile duct cancer.

Materials and Method:

Thirty-three patients with hilar cholangiocarcinoma or common duct cancer who had undergone MRC, unenhanced, and gadolinium-enhanced dynamic MR images and surgery were included in this study. Two experienced radiologists independently reviewed 2 image sets in 2 steps, that is, the MRC set (unenhanced and MRC) and the combined image set (MRC set with dynamic images). At each step, the readers determined the tumor status according to the Bismuth-Corlette classification. The readers assigned their confidence levels on a 5-point scale regarding whether the tumor involved the secondary confluence of the bile duct and the intrapancreatic common bile duct. The radiologists' diagnostic confidence of the 2 image sets was analyzed using receiver operating characteristic analysis.

Results:

Receiver operating characteristic analysis showed higher areas under the curve values when the combined image set was interpreted (0.990±0.017 for reader 1 and 0.951±0.027 for reader 2) than when the MRC set was interpreted (0.982±0.017 for reader 1 and 0.902±0.038 for reader 2); however, the difference was not statistically significant for either reader (P> 0.05). In addition, regarding evaluation of the tumor status according to the Bismuth-Corlette classification, the overall accuracy was higher for the combined image set than for the MRC set alone, but the difference was not significant (P> 0.05). When dynamic images were added to the MRC images, interobserver agreement improved from 0.72 to 0.84.

Conclusions:

The addition of contrast-enhanced dynamic images to unenhanced and MRC images did not significantly improve the diagnostic accuracy for assessment of the longitudinal extent of bile duct cancer.