An illustrated comparison of processing methods for phase MRI and QSM: removal of background field contributions from sources outside the region of interest

Schweser, Ferdinand; Robinson, Simon; Rochefort, Ludovic de; Li, Wei; Bredies, Kristian

doi:10.1002/nbm.3604

Cited by 131 publications

(211 citation statements)

References 43 publications

Supporting

Mentioning

205

Contrasting

Order By: Relevance

“…More accurate BFR could potentially overcome this problem. Here, however, we used 2 state‐of‐the‐art techniques with similar results, and it has been shown that all known BFR techniques lead to errors toward the mask edges . The error introduced by the SC step is probably because of regions of the local field map being cut off at reduced coverage (Figure B, red arrows).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

The effect of low resolution and coverage on the accuracy of susceptibility mapping

Karsa

Punwani

Shmueli

2018

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose Quantitative susceptibility mapping (QSM) has found increasing clinical applications. However, to reduce scan time, clinical acquisitions often use reduced resolution and coverage, particularly in the through‐slice dimension. The effect of these factors on QSM has begun to be assessed using only balloon phantoms and downsampled brain images. Here, we investigate the effects (and their sources) of low resolution or coverage on QSM using both simulated and acquired images. Methods Brain images were acquired at 1 mm isotropic resolution and full brain coverage, and low resolution (up to 6 mm slice thickness) or coverage (down to 20 mm) in 5 healthy volunteers. Images at reduced resolution or coverage were also simulated in these volunteers and in a new, anthropomorphic, numerical phantom. Mean susceptibilities in 5 brain regions, including white matter, were investigated over varying resolution and coverage. Results The susceptibility map contrast decreased with increasing slice thickness and spacing, and with decreasing coverage below ~40 mm for 2 different QSM pipelines. Our simulations showed that calculated susceptibility values were erroneous at low resolution or very low coverage, because of insufficient sampling and overattenuation of the susceptibility‐induced field perturbations. Susceptibility maps calculated from simulated and acquired images showed similar behavior. Conclusions Both low resolution and low coverage lead to loss of contrast and errors in susceptibility maps. The widespread clinical practice of using low resolution and coverage does not provide accurate susceptibility maps. Simulations in images of healthy volunteers and in a new, anthropomorphic numerical phantom were able to accurately model low‐resolution and low‐coverage acquisitions.

show abstract

Section: Discussionmentioning

confidence: 99%

“…This was repeated with the above QSM pipeline but with LBV as the background field removal (BFR) step to examine how the results are affected by different BFR techniques. LBV was chosen because it was shown to perform similarly well to PDF in the brain while introducing slightly different error patterns near the tissue edges …”

Section: Methodsmentioning

confidence: 99%

The effect of low resolution and coverage on the accuracy of susceptibility mapping

Karsa

Punwani

Shmueli

2018

Magnetic Resonance in Med

View full text Add to dashboard Cite

show abstract

“…A mask delineating the region of the local effects of interest has to be defined. Many background-correction methods presented in literature will either end up eroding this mask (losing relevant information on outer parts of the brain) or giving values that are unreliable close to the boundary, for an illustrated comparison of many of these methods refer to (Schweser et al, 2016). The second step is the calculation of the QSM from the measured field maps (after background field removal).…”

Section: Introductionmentioning

confidence: 99%

Studying cyto and myeloarchitecture of the human cortex at ultra-high field with quantitative imaging: R1, R2* and magnetic susceptibility

2017

View full text Add to dashboard Cite

A B S T R A C TIn this manuscript, the use of quantitative imaging at ultra-high field is evaluated as a mean to study cyto and myelo-architecture of the cortex. The quantitative contrasts used are the longitudinal relaxation rate (R 1 ), apparent transverse relaxation rate (R 2 *) and quantitative susceptibility mapping (QSM).The quantitative contrasts were computed using high resolution in-vivo (0.65mm isotropic) brain data acquired at 7 T.The performance of the different quantitative approaches was evaluated by visualizing the contrast between known highly myelinated primary sensory cortex regions and the neighbouring cortex. The transition from the inner layers to the outer layers (from white matter to the pial surface) of the human cortex, which is known to have varying cyto-and myelo architecture, was evaluated.The across cortex and through depth behaviour observed for the different quantitative maps was in good agreement between the different subjects, clearly allowing the differentiation between different Brodmann regions, suggesting these features could be used for individual cortical brain parcellation. While both R 1 and R 2 * maps decrease monotonically from the white matter to the pial surface due to the decrease of myelin and iron between these regions, magnetic susceptibility maps have a more complex behaviour reflecting its opposing sensitivity to myelin and iron concentration.

show abstract

“…This is commonly referred as the “shadowing” artifact. Past studies have shown that background‐removal (and single‐step) algorithms can effectively eliminate large‐scale (externally sourced) fields that are severely truncated by information loss due to low SNR or FOV definition, hence simplifying the problem into inverting local dipole fields that are only partially truncated by ROI masking. However, due to these and other confounding factors (e.g., nonzero transmit‐driven initial phases, unwrapping errors), the efficiency of background‐removal algorithms is often impaired leading to excessive local‐field contamination (Figure ).…”

Section: Discussionmentioning

confidence: 99%

Weak‐harmonic regularization for quantitative susceptibility mapping

Milovic

Bilgic̦

Zhao

et al. 2018

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose Background‐field removal is a crucial preprocessing step for quantitative susceptibility mapping (QSM). Remnants from this step often contaminate the estimated local field, which in turn leads to erroneous tissue‐susceptibility reconstructions. The present work aimed to mitigate this undesirable behavior with the development of a new approach that simultaneously decouples background contributions and local susceptibility sources on QSM inversion. Methods Input phase data for QSM can be seen as a composite scalar field of local effects and residual background components. We developed a new weak‐harmonic regularizer to constrain the latter and to separate the 2 components. The resulting optimization problem was solved with the alternating directions of multipliers method framework to achieve fast convergence. In addition, for convenience, a new alternating directions of multipliers method–based preconditioned nonlinear projection onto dipole fields solver was developed to enable initializations with wrapped‐phase distributions. Weak‐harmonic QSM, with and without nonlinear projection onto dipole fields preconditioning, was compared with the original (alternating directions of multipliers method–based) total variation QSM algorithm in phantom and in vivo experiments. Results Weak‐harmonic QSM returned improved reconstructions regardless of the method used for background‐field removal, although the proposed nonlinear projection onto dipole fields method often obtained better results. Streaking and shadowing artifacts were substantially suppressed, and residual background components were effectively removed. Conclusion Weak‐harmonic QSM with field preconditioning is a robust dipole inversion technique and has the potential to be extended as a single‐step formulation for initialization with uncombined multi‐echo data.

show abstract

An illustrated comparison of processing methods for phase MRI and QSM: removal of background field contributions from sources outside the region of interest

Cited by 131 publications

References 43 publications

The effect of low resolution and coverage on the accuracy of susceptibility mapping

The effect of low resolution and coverage on the accuracy of susceptibility mapping

Studying cyto and myeloarchitecture of the human cortex at ultra-high field with quantitative imaging: R1, R2* and magnetic susceptibility

Weak‐harmonic regularization for quantitative susceptibility mapping

Contact Info

Product

Resources

About