The objective of this study was to evaluate the diagnostic benefit of an intravoxel incoherent motion (IVIM) model–based characterization of pancreatic masses from diffusion-weighted imaging (DWI) with 3 b values.
This retrospective study had an approval from the institutional review board, and informed patient consent was waived. The 1.5-T DWI data of 42 patients with or without pancreatic disease, acquired by a respiratory-gated spin-echo echo-planar imaging sequence with 3 b values (0, 50, 800 s/mm2), were retrospectively analyzed. The IVIM-related parameters D′, which is the apparent diffusion coefficient [ADC(50,800)], and f′, as well as ADC(0,50), and conventional ADC(0,800) were calculated voxelwise. Regions of interest were analyzed in pancreatic adenocarcinomas (CAs, n = 12), neuroendocrine pancreatic tumors (NETs, n = 9), and chronic pancreatitis (CPs, n = 11), not affected tissue of each pathologic group, and in the head, body, and tail of the healthy pancreas (n = 10).
By ADC(0,800) and D′, CAs could hardly be distinguished from neuroendocrine pancreatic tumors and chronic pancreatitis. However, CAs revealed very low ADC(0,50) and f′ values, which differed significantly from all other groups. In the healthy pancreas, ADC(0,800) and D′ values were significantly higher for the head than for the body and tail, but no significant differences were found for ADC(0,50) and f′.
The determination of IVIM-based microcirculation-sensitive parameter maps from DWI with 3 b values significantly improved the discrimination of CAs from NETs, CPs, and the healthy tissue.
From the *Department of Radiology, University of Bonn, Bonn; †Institute of Medical Engineering, Ruhr-University Bochum, Bochum; ‡Department of Internal Medicine 3, University of Bonn, Bonn, Germany; and §Philips Healthcare, Best, The Netherlands.
Received for publication May 16, 2013; and accepted for publication, after revision, July 26, 2013.
Conflicts of interest and sources of funding: none declared.
Reprints: Petra Mürtz, PhD, Radiologische Klinik der Universität Bonn, Sigmund-Freud-Straße 25, 53105 Bonn, Germany. E-mail: firstname.lastname@example.org.