1Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, 2Leonard Wolfson Experimental Neurology Centre, UCL Institute of Neurology, University College London, London, United Kingdom
In Susceptibility Mapping (SM)
using multi-echo gradient-echo phase data, unwrapping and/or background-field
removal is often performed using Laplacian-based methods. However, SM pipelines
in the literature have applied these methods at different stages. Here, using
simulated and acquired images, we compared the performance of three pipelines
that apply Laplacian-based methods at different stages. We showed that
Laplacian-based methods alter the linearity of the phase over time. We
demonstrated that only a processing pipeline that takes this into account, i.e.
by fitting the multi-echo data over time to correctly estimate a field map
before applying Laplacian-based methods, gives accurate susceptibility values.