• Giang-Chau Ngo, Berkin Bilgic, Borjan A Gagoski, and Bradley P Sutton.
• Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, Illinois.
• Magn Reson Med. 2019 Mar 1; 81 (3): 1645-1658.

PurposeQuantitative susceptibility mapping (QSM) provides information about tissue composition and organization. However, current acquisition methods can be quite time consuming, limiting the use of QSM in studies and in monitoring dynamic events. In this work, time efficient spiral acquisitions are combined with a model-based image reconstruction approach. Magnetic field inhomogeneity artifacts are corrected to obtain full brain susceptibility maps.MethodsMagnetic field maps, capturing the magnetic field inhomogeneity distribution due to non-tissue sources (background field), were estimated from a calibration scan. Tissue phase information present in the background field maps was identified using variable spherical mean value filtering and removed from the estimated field maps. The resulting background field maps were then used in the image reconstruction to correct the magnetic field inhomogeneity artifacts. Images were acquired on a 3 T system with a 3D spiral-in acquisition based on a rotated stack of spirals. For comparison, standard gradient echo images were also acquired. Susceptibility maps were obtained for both acquisitions using a dipole inversion algorithm based on a compressed sensing compensated formulation with wavelet and total variation penalties.ResultsSusceptibility maps obtained with the spiral acquisition and the model-based reconstruction agree with the ones obtained with the spin-warp gradient echo acquisition. Using the 3D spiral acquisition, full brain susceptibility maps at the resolution of 1 mm isotropic are obtained in 23 s CONCLUSIONS: Image distortions and blurring due to magnetic field inhomogeneity are removed while maintaining tissue magnetic susceptibility information, resulting in QSM maps in much shorter acquisition time.© 2018 International Society for Magnetic Resonance in Medicine.

26 keywords...

hide…