Purpose: The aim of this study was to compare diffusion-weighted magnetic resonance imaging (DWI) and multidetector-row computed tomography (MDCT) for detection of primary pancreatic cancer by reviewing images of patients at high risk for pancreatic cancer with main pancreatic duct (MPD) dilatation shown by magnetic resonance cholangiopancreatography (MRCP).
Methods: From October 2007 to September 2009, 83 patients who had undergone both DWI and MDCT with MPD dilatation were identified and were reviewed by four readers (2 radiologists and 2 gastroenterology fellows). Diagnostic performance in pancreatic cancer detection was evaluated with 95% confidence intervals. Statistically significant differences in the detection of pancreatic cancer between DWI and MDCT were compared by receiver operating characteristics and the confidence of the diagnosis by the paired t test.
Results: Thirty-two of 83 patients were diagnosed with primary pancreatic cancer by histological evaluation of 15 surgical and 2 endoscopic ultrasound-guided fine needle aspiration samples, and by the clinical course for 15 lesions. Overall average accuracies of pancreatic cancer detection by the four readers were 84% with DWI and 86% with MDCT.
Conclusion: Performance of DWI and MDCT was equivocal for detection of pancreatic cancer in a high-risk population with MPD dilatation. The combination of MRCP and DWI for detection of pancreatic cancer allowed identification of a high-risk population and tumor detection with a single imaging modality with no need for contrast medium.