Purpose: To compare the diagnostic performance (detection, local staging) of multiphasic 64-detector row computed tomography (CT) with that of gadobenate dimeglumine-enhanced 3.0-T magnetic resonance (MR) imaging in patients suspected of having pancreatic cancer.

Materials and methods: The institutional review board approved this prospective study, and all patients provided written informed consent. Multidetector CT and MR imaging were performed in 89 patients (48 women aged 46-89 years [mean, 65.6 years] and 41 men aged 46-86 years [mean, 65.3 years]) suspected of having pancreatic cancer on the basis of findings from clinical examination or previous imaging studies. Two readers independently assessed the images to characterize lesions and determine the presence of focal masses, vascular invasion, distant metastases, and resectability. Findings from surgery, biopsy, endosonography, or follow-up imaging were used as the standard of reference. Logistic regression, the McNemar test, and κ values were used for statistical analysis.

Results: Focal pancreatic masses were present in 63 patients; 43 patients had adenocarcinoma. For reader 1, the sensitivities and specificities in the detection of pancreatic adenocarcinoma were 98% (42 of 43 patients) and 96% (44 of 46 patients), respectively, for CT and 98% (42 of 43 patients) and 96% (44 of 46 patients) for MR imaging. For reader 2, the sensitivities and specificities were 93% (40 of 43 patients) and 96% (44 of 46 patients), respectively, for CT and 95% (41 of 43 patients) and 96% (44 of 46 patients) for MR imaging. Vessel infiltration was determined in 22 patients who underwent surgery, and reader 1 obtained sensitivities and specificities of 90% (nine of 10 vessels) and 98% (119 of 122 vessels), respectively, for CT and 80% (eight of 10 vessels) and 96% (117 of 122 vessels) for MR imaging; for reader 2, those values were 70% (seven of 10 vessels) and 98% (120 of 122 vessels) for CT and 50% (five of 10 vessels) and 98% (120 of 122 vessels) for MR imaging. Both readers correctly assessed resectability in 87% (13 of 15 patients) of cases with CT and 93% (14 of 15 patients) of cases with MR imaging. Nonresectability was assessed correctly with CT in 75% (six of eight patients) of cases by reader 1 and 63% (five of eight patients) of cases by reader 2; nonresectability was correctly assessed with MR imaging in 75% (six of eight patients) of cases by reader 1 and 50% (four of eight patients) of cases by reader 2. None of the differences between modalities and readers were statistically significant (P > .05).

Conclusion: Both CT and MR imaging are equally suited for detecting and staging pancreatic cancer.

Supplemental material: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.11101189/-/DC1.