The Value of Adding Conventional MR Imaging to MR Cholangiopancreatography in Differentiation of Benign and Malignant Causes of Postoperative Disorders
Ya Xiong
2003-01-01
Abstract:Objective To determine the value of conventional T 1- and T 2-weighted images and gadolinium-enhanced magnetic resonance (MR) images as a supplement to MR cholangiopancreatography(MRCP) in differentiation of benign from malignant causes of postoperative disorders in the biliary ductal system. Methods Sixty-one patients with postoperative disorders in the biliary ductal system with proved causes underwent MRCP, conventional T 1- and less heavily T 2-weighted images, as well as gadolinium-enhanced images. Two radiologists independently reviewed MRCP images alone, MRCP plus nonenhanced T 1- and T 2-weighted images, and MRCP plus nonenhanced and gadolinium-enhanced images. The results of MR findings were compared with that of the surgical findings and the pathology. Results For the diagnosis of postoperative disorders only with MRCP images, the sensitivity, specificity, and accuracy was 42.1%, 80.9%, and 68.9% for radiologist 1 and 47.4%, 85.7%, and 73.8% for radiologist 2, respectively. When MRCP images were interpreted with T 1- and T 2-weighted images, the sensitivity, specificity, and accuracy was 78.9%, 92.9%, and 88.5% for radiologist 1 and, 78.9%, 95.2%, and 90.2% for radiologist 2, respectively. When MRCP images were combined with both nonenhanced T 1- and T 2-weighted images and enhanced MR images, the sensitivity, specificity, and accuracy was 84.2%, 95.2%, and 91.8% for radiologist 1 and 84.2%, 97.6%, and 93.4% for radiologist 2, respectively. There was no significant difference between the 2 readers (P0.05). For differentiation of benign from malignant causes of postoperative disorders, the area under the receiver operating characteristic curve (Az) was significantly larger for MRCP images interpreted with T 1- and T 2-weighted images (0.907 for reader 1,0.920 for reader 2) than for MRCP images alone (0.682 reader 1,0.714 for reader 2) (P0.05). The addition of gadolinium-enhanced MR images to MRCP images with nonenhanced T 1- and T 2-weighted images did not significantly increase the accuracy (Az =0.948 for reader 1,0.944 for reader 2) (P0.05). Conclusion The usage of nonenhanced T 1- and T 2-weighted images and gadolinium-enhanced MR images with MRCP significantly improved the diagnostic accuracy of MR examinations in postoperative disorders in the biliary ductal system.