QSMxT: Robust Masking and Artefact Reduction for Quantitative Susceptibility Mapping

Abstract

Purpose: Quantitative Susceptibility Mapping (QSM) is a post-processing technique applied to gradient-echo phase data. QSM algorithms require a signal mask to delineate regions with reliable phase signal for subsequent susceptibility estimation. Existing masking techniques used in QSM have limitations that introduce artefacts, exclude anatomical detail, and rely on parameter tuning and anatomical priors that narrow their application. Here, a robust masking and reconstruction procedure is presented to overcome these limitations and enable automated QSM processing for a wide range of use-cases implemented in an open-source software framework: QSMxT. Methods: A robust masking technique that automatically separates reliable from less reliable phase regions was developed and combined with a two-pass reconstruction procedure that operates on the separated sources before combination, extracting more information while reducing the influence of artefacts. Results: Compared with standard masking and reconstruction procedures, the two-pass inversion reduces streaking artefacts caused by unreliable phase and high dynamic ranges of susceptibility sources. QSMxT is robust across a range of datasets at 3 T in healthy volunteers and phantoms, at 7 T in tumour patients, and in the QSM challenge 2.0 simulated brain dataset, with significant artefact and error reductions, greater anatomical detail, and minimal parameter tuning. Conclusion: QSMxT generates masks for QSM that separate reliable from less reliable phase regions, enables a more accurate QSM reconstruction that mitigates artefacts, operates without anatomical priors, and requires minimal parameter tuning. QSMxT makes QSM processing more accessible, reliable and reproducible.

QSMxT: Robust masking and artifact reduction for quantitative susceptibility mapping

“…The new, two‐pass QSM approach was implemented as part of the QSMxT framework within a modular reconstruction and analysis ecosystem 30 . The first module consists of a continuous integration system using GitHub actions to automatically build software containers, including the container required for the QSMxT framework.…”

“…The first module consists of a continuous integration system using GitHub actions to automatically build software containers, including the container required for the QSMxT framework. The QSMxT container incorporates the FMRIB Software Library (FSL) v6.0.4, 36 FastSurfer, 37 ANTS 2.3.4, 38 Julia v1.5.3, 39 Bidscoin v2.3, 40 Python v3.8, 41 TGV‐QSM v1.0.0, 42,43 the Python packages pydicom, 44 nibabel, 45 seaborn 46,47 and NiPype, 48 the Julia packages ArgParse 49 and MriResearchTools, 50 as well as the QSMxT NiPype workflow 30 . QSMxT runs natively on Linux operating systems and high‐performance clusters (HPCs), only requiring Singularity 51 as a dependency.…”

Ironsmith: An automated pipeline for QSM-based data analyses

“…The first publicly available software package for QSM reconstruction was MEDI Toolbox ( De Rochefort et al, 2010 ; J. Liu et al, 2012 ; Liu et al, 2011a , 2011b ) which uses a Morphology Enabled Dipole Inversion approach (MEDI) for dipole deconvolution. The MEDI Toolbox remains the predominant method for reconstructing QSM maps in the literature, although other software toolboxes are now publicly available, including COSMOS ( Liu et al, 2009 ), FANSI toolbox ( Bilgic et al, 2015 ; Milovic et al, 2018 ), QSM Toolbox ( Kames et al, 2018 ), QSMxT ( Stewart et al, 2021 ), SEPIA ( Chan and Marques, 2021 ) and STI suite ( Liu et al, 2015 ). In addition, automated, hardware-based QSM reconstruction solutions are also becoming available.…”

“…To our knowledge, few QSM toolboxes are currently available for extracting QSM-based brain iron concentrations from anatomical ROIs. Two examples we are aware of are the Johns Hopkins University (JHU) multi-atlas tool ( Li et al, 2019 ) and the QSMxT toolbox ( Stewart et al, 2021 ). Both are valuable, however, neither are fully automated.…”

Quantitative susceptibility mapping of the head‐and‐neck using SMURF fat‐water imaging with chemical shift and relaxation rate corrections

“…This, and the fact that magnitudebased masks include voxels with high magnitude signal but unreliable phase (e.g., veins) have led to the suggestion to generate masks using phase image. 46 The aim of this study was to address the challenges of QSM outside the brain and, using simultaneous fat-water imaging with SMURF, to generate chemical shift and relaxation rate bias-free, high CNR susceptibility maps of the entire head-and-neck region. Since QSM generally benefits from the use of higher field, we extended SMURF to 7 T, although the need for different RF excitation pulses and the modified field variations this a non-trivial translation.…”