Benjamin Henninger1, Heinz Zoller2, Stephan Kannengiesser3, Xiaodong Zhong4, Werner Jaschke1, Christian Kremser1. 1. Department of Radiology, Medical University of Innsbruck, Innsbruck, Austria. 2. Department of Internal Medicine, Medical University of Innsbruck, Innsbruck, Austria. 3. MR Applications Development, Siemens, Healthcare Sector, Erlangen, Germany. 4. MR R&D Collaborations, Siemens Healthcare, Atlanta, Georgia, USA.
Abstract
PURPOSE: To prospectively evaluate a new 3D-multiecho-Dixon (3D-ME-Dixon) sequence for the quantification of hepatic iron and fat in a clinical setting. MATERIALS AND METHODS: In all, 120 patients underwent 1.5T magnetic resonance imaging of the liver between December 2013 and June 2015 including the following three sequences: 3D-ME-Dixon with inline calculation of R2* and proton-density fat-fraction (PDFF) maps, single-voxel-spectroscopy (SVS), 2D multigradient-echo sequence (2D-ME-GRE). SVS and 2D-ME-GRE were used as reference for PDFF and R2*, respectively. R2*- and PDFF-values from 3D-ME-Dixon were compared with those of the reference. Linear regression analysis, Bland-Altman plots, and agreement parameters were calculated. RESULTS: In total, 103 patients were finally included (87 men and 16 women; mean age, 50.51 years); 17/120 were excluded due to fat/water-swaps or R2*-values exceeding the constraint of 400 1/s for 3D-ME-Dixon. A strong correlation (r = 0.992, P < 0.001) between R2* of 3D-ME-Dixon and the reference 2D-ME-GRE was found. Bland-Altman analysis revealed systematically lower values for 3D-ME-Dixon (16.499%). Using an adapted threshold of 57 1/s, 3D-ME-Dixon obtained a positive/negative percentage agreement (PPA/NPA) of 84.4%/91.4% for detecting hepatic iron overload. For hepatic fat the correlation between 3D-ME-Dixon and the reference SVS was strong (r = 0.957, P < 0.001); PPA/NPA was 88.3%/91.4%. CONCLUSION: The 3D-ME-Dixon sequence is a valuable tool for the evaluation of hepatic iron and fat in a clinical setting. Fat/water-swaps remain a drawback requiring improvements to the implementation and making it necessary to have proven conventional sequences at hand in case of an eventual occurrence. LEVEL OF EVIDENCE: 1. Technical Efficacy: Stage 2 J. MAGN. RESON. IMAGING 2017;46:793-800.
PURPOSE: To prospectively evaluate a new 3D-multiecho-Dixon (3D-ME-Dixon) sequence for the quantification of hepatic iron and fat in a clinical setting. MATERIALS AND METHODS: In all, 120 patients underwent 1.5T magnetic resonance imaging of the liver between December 2013 and June 2015 including the following three sequences: 3D-ME-Dixon with inline calculation of R2* and proton-density fat-fraction (PDFF) maps, single-voxel-spectroscopy (SVS), 2D multigradient-echo sequence (2D-ME-GRE). SVS and 2D-ME-GRE were used as reference for PDFF and R2*, respectively. R2*- and PDFF-values from 3D-ME-Dixon were compared with those of the reference. Linear regression analysis, Bland-Altman plots, and agreement parameters were calculated. RESULTS: In total, 103 patients were finally included (87 men and 16 women; mean age, 50.51 years); 17/120 were excluded due to fat/water-swaps or R2*-values exceeding the constraint of 400 1/s for 3D-ME-Dixon. A strong correlation (r = 0.992, P < 0.001) between R2* of 3D-ME-Dixon and the reference 2D-ME-GRE was found. Bland-Altman analysis revealed systematically lower values for 3D-ME-Dixon (16.499%). Using an adapted threshold of 57 1/s, 3D-ME-Dixon obtained a positive/negative percentage agreement (PPA/NPA) of 84.4%/91.4% for detecting hepatic iron overload. For hepatic fat the correlation between 3D-ME-Dixon and the reference SVS was strong (r = 0.957, P < 0.001); PPA/NPA was 88.3%/91.4%. CONCLUSION: The 3D-ME-Dixon sequence is a valuable tool for the evaluation of hepatic iron and fat in a clinical setting. Fat/water-swaps remain a drawback requiring improvements to the implementation and making it necessary to have proven conventional sequences at hand in case of an eventual occurrence. LEVEL OF EVIDENCE: 1. Technical Efficacy: Stage 2 J. MAGN. RESON. IMAGING 2017;46:793-800.
Authors: Andy McKay; Henry R Wilman; Andrea Dennis; Matt Kelly; Michael L Gyngell; Stefan Neubauer; Jimmy D Bell; Rajarshi Banerjee; E Louise Thomas Journal: PLoS One Date: 2018-12-21 Impact factor: 3.240