Differential diagnosis of benign and malignant distal biliary strictures: Value of adding diffusion-weighted imaging to conventional magnetic resonance cholangiopancreatography

Abstract

To determine the value of adding diffusion-weighted imaging (DWI) to conventional magnetic resonance cholangiopancreatography (MRCP) for differentiating benign from malignant distal biliary strictures. Two independent readers reviewed three image sets (1: MRCP alone; 2: MRCP and DWI combined; and 3: MRCP, DWI, and contrast-enhanced T1-weighted imaging [T1WI] combined) of 60 patients with suspected distal biliary strictures and rated the probability of malignancy. Diagnostic performance and accuracy were compared using the receiver operating characteristic (ROC) curves and McNemar two-tailed test. κ coefficients were calculated to assess the interobserver agreement. The Az value and accuracy improved significantly after additional review of DWI for both readers: Az = 0.780 vs. 0.916 (P = 0.003) for Reader 1 and 0.784 vs. 0.853 (P = 0.037) for Reader 2; accuracy = 69% vs. 93% for Reader 1 (P < 0.001) and 57% vs. 85% for Reader 2 (P = 0.002). No significant difference in the Az values and accuracy was found between MRCP and DWI combined, and MRCP, DWI, and contrast-enhanced T1WI combined (P > 0.050). There was substantial interobserver agreement in all three image sets (κ = 0.695-0.732). The addition of DWI to MRCP significantly improved diagnostic accuracy in the characterization of distal biliary strictures.