Premium
Inhomogeneous static magnetic field‐induced distortion correction applied to diffusion weighted MRI of the breast at 3T
Author(s) -
Teruel Jose R.,
Fjøsne Hans E.,
Østlie Agnes,
Holland Dominic,
Dale Anders M.,
Bathen Tone F.,
Goa Pål E.
Publication year - 2015
Publication title -
magnetic resonance in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.696
H-Index - 225
eISSN - 1522-2594
pISSN - 0740-3194
DOI - 10.1002/mrm.25489
Subject(s) - diffusion mri , distortion (music) , encoding (memory) , nuclear magnetic resonance , mathematics , mutual information , nuclear medicine , similarity (geometry) , norm (philosophy) , physics , computer science , algorithm , magnetic resonance imaging , artificial intelligence , medicine , image (mathematics) , radiology , amplifier , political science , law , optoelectronics , cmos
Purpose To evaluate the performance of an advanced method for correction of inhomogeneous static magnetic field induced distortion in echo‐planar imaging (EPI), applied to diffusion‐weighted MRI (DWI) of the breast. Methods An algorithm for distortion correction based on the symmetry of the distortion induced by static field inhomogeneity when the phase encoding polarity is reversed was evaluated in 36 data sets of patients who received an MRI examination that included DWI (b = 0 and 700 s/mm 2 ) and an extra b = 0 s/mm 2 sequence with opposite phase encoding polarity. The decrease of the L 2 ‐square norm after correction between opposed phase encoding b = 0 images was calculated. Mattes mutual information between b = 0 images and fat‐suppressed T 2 ‐weighted images was calculated before and after correction. Results The L 2 ‐square norm between different phase encoding polarities for b = 0 images was reduced 94.3% on average after distortion correction. Furthermore, Mattes mutual information between b = 0 images and fat‐suppressed T 2 ‐weighted images increased significantly after correction for all cases ( P < 0.001). Conclusion Geometric distortion correction in DWI of the breast results in higher similarity of DWI to anatomical non‐EPI T 2 ‐weighted images and would potentially allow for a more reliable lesion segmentation mapping among different MRI modalities. Magn Reson Med 74:1138–1144, 2015. © 2014 Wiley Periodicals, Inc.