Rapid calculations of susceptibility-induced magnetostatic field perturbations for in vivo magnetic resonance

• Published in: Physics in medicine & biology Vol. 51; no. 24; pp. 6381 - 6402
• Main Authors: Koch, Kevin M, Papademetris, Xenophon, Rothman, Douglas L, de Graaf, Robin A
• Format: Journal Article
• Language: English
• Published: England IOP Publishing 21.12.2006
• Subjects: Air; Algorithms; Brain - pathology; Fourier Analysis; Humans; Image Processing, Computer-Assisted; Magnetic Resonance Imaging - methods; Magnetics; Models, Statistical; Models, Theoretical; Phantoms, Imaging; Software; Tomography, X-Ray Computed - methods; Water - chemistry; Whole Body Imaging
• ISSN: 0031-9155; 1361-6560
• DOI: 10.1088/0031-9155/51/24/007

Static magnetic field perturbations generated by variations of magnetic susceptibility within samples reduce the quality and integrity of magnetic resonance measurements. These perturbations are difficult to predict in vivo where wide variations of internal magnetic susceptibility distributions are common. Recent developments have provided rapid computational means of estimating static field inhomogeneity within the small susceptibility limits of materials typically studied using magnetic resonance. Such a predictive mechanism could be a valuable tool for sequence simulation, field shimming and post-acquisition image correction. Here, we explore this calculation protocol and demonstrate its predictive power in estimating in vivo inhomogeneity within the human brain. Furthermore, we quantitatively explore the predictive limits of the computation. For in vivo comparison, a method of magnetic susceptibility registration using MRI and CT data is presented and utilized to carry out subject-specific inhomogeneity estimation. Using this algorithm, direct comparisons in human brain and phantoms are made between field map acquisitions and calculated inhomogeneity. Distortion correction in echo-planar images due to static field inhomogeneity is also demonstrated using the computed field maps.