Comparison of Phase Estimation Methods for Quantitative Susceptibility Mapping Using a Rotating-Tube Phantom

Table 2

All phase estimation methods. The first six methods were combined with other common phase processing techniques to process the multiecho data, as described in the Methods section.

Sequence	Method	Summary
SEGE, MEGE, sEPI	Unprocessed	Default channel combined output phase image produced by the scanner
	Phun	Region-based algorithm [46]
	GBC	Goldstein’s branch cut method [42]
	MEDI.RG	Region growing algorithm from MEDI Toolbox [45, 47]
	MEDI.LP	Laplacian algorithm from MEDI Toolbox [44, 45]
	Laplace	Laplacian algorithm [40]
MEGE	Slope	Computes the frequency offset from the slope of the complex data across echoes [41]
	Division	Computes the frequency offset from the complex division of the data at successive echoes [43]
	MAGPI-unopt	Applies the MAGPI postprocessing algorithm using a random (unoptimized) subset of echoes from the MEGE sequence. Since MAGPI requires uneven echo spacing, only four unequally spaced echoes were selected [37]
MAGPI	MAGPI	Maximum-likelihood estimate of phase from optimal echo spacing [37]