Reference tissue-based coupled linear separation of susceptibility sources in adolescent
Poster No:
1983
Submission Type:
Abstract Submission
Authors:
Joo-won Kim1, Dimitri Fiani1, Muhammad Haque2, Griselda Barba Villalobos1, Jacqueline Nguyen1, Stephanie Dinh1, Chadi Calarge1, Junqian Xu1
Institutions:
1Baylor College of Medicine, Houston, TX, 2UTHealth, Houston, TX
First Author:
Co-Author(s):
Introduction:
Paramagnetic Iron and diamagnetic myelin are the primary sources of susceptibility signal in the brain. Several approaches have sought to disentangle the two in QSM [1]. Recent efforts premises on the frequency shift dependence on the susceptibility sources. As such, quantitative T2 scans are acquired/approximated to reconstruct a magnitude decay kernel [2], [3]. These approaches have successfully estimated iron and myelin content, at levels validated by histology [2], [4]; however, they require additional scanning/assumptions. We introduce a simple REference TIssue-based Coupled Linear sEparation of susceptibility Sources (RETICLES) method that only requires a typical multi-echo GRE (mGRE) acquisition.
Methods:
Participants: 117 children and adolescents (7-18 years old, 77 females).
3T MRI acquisition:T1-weighted (T1w) image 1 mm isotropic. 3D mGRE: TR/4 TEs =30.4/7/14/21/28 ms, 0.7x0.7x1 mm.
QSM (χ) reconstruction: Phase unwrapping using ROMEO [5], phase reliability mask [6], background field removal and dipole inversion using QSMbox [7].
R2* values: mono-exponential fitting of mGRE image intensities.
Individual T1w image and QSM map were linearly combined into a hybrid image [8] to create a template using ANTs.
Manual delineation of subregions in the hybrid template space (Fig.1). Globus pallidus (GP): GPe/GPi (externa/interna), and wmGP (myelin-rich anterior caudal region adjacent to the Pu). Putamen (Pu): aPu (anterior), pPu (posterior), and wmPu (myelin-rich anterior medial wedge-shaped region separating the GP and Pu). Caudate (Cd): aCd (anterior), cCd (central), pCd (posterior), and vCd (vein-rich most ventromedial tip of the Cd). Posterior limb of the internal capsule (pLIC).
Assumptions: 1) Distribution of myelin is well-mixed with iron within a voxel. 2) Linear additive contribution from iron and myelin components. 3) The proportion of myelin component (f_myelin) or the proportion of iron component (f_iron) scales similarly for R2* and χ when referenced to a highly myelinated (pLIC) or iron-rich (GPe) structure, respectively. Solving the coupled equations below:
R2* = R2*_intrinsic + f_iron R2*_iron_Ref + f_myelin R2*_myelin_Ref
χ = χ_paramag + χ_diamag = f_iron χ_iron_Ref + fmyelin χ_myelin_Ref
where R2*_intrinsic = intrinsic R2*, assumed as a constant = 6 ms [9]. The reference values in the pLIC and GPe were obtained from the study group-average QSM or R2* values.
Summary statistics for each ROI, after transformation to the native QSM space, were used to compare χ, χ_paramag, and χ_diamag values (Fig.2, left).
3T MRI acquisition:T1-weighted (T1w) image 1 mm isotropic. 3D mGRE: TR/4 TEs =30.4/7/14/21/28 ms, 0.7x0.7x1 mm.
QSM (χ) reconstruction: Phase unwrapping using ROMEO [5], phase reliability mask [6], background field removal and dipole inversion using QSMbox [7].
R2* values: mono-exponential fitting of mGRE image intensities.
Individual T1w image and QSM map were linearly combined into a hybrid image [8] to create a template using ANTs.
Manual delineation of subregions in the hybrid template space (Fig.1). Globus pallidus (GP): GPe/GPi (externa/interna), and wmGP (myelin-rich anterior caudal region adjacent to the Pu). Putamen (Pu): aPu (anterior), pPu (posterior), and wmPu (myelin-rich anterior medial wedge-shaped region separating the GP and Pu). Caudate (Cd): aCd (anterior), cCd (central), pCd (posterior), and vCd (vein-rich most ventromedial tip of the Cd). Posterior limb of the internal capsule (pLIC).
Assumptions: 1) Distribution of myelin is well-mixed with iron within a voxel. 2) Linear additive contribution from iron and myelin components. 3) The proportion of myelin component (f_myelin) or the proportion of iron component (f_iron) scales similarly for R2* and χ when referenced to a highly myelinated (pLIC) or iron-rich (GPe) structure, respectively. Solving the coupled equations below:
R2* = R2*_intrinsic + f_iron R2*_iron_Ref + f_myelin R2*_myelin_Ref
χ = χ_paramag + χ_diamag = f_iron χ_iron_Ref + fmyelin χ_myelin_Ref
where R2*_intrinsic = intrinsic R2*, assumed as a constant = 6 ms [9]. The reference values in the pLIC and GPe were obtained from the study group-average QSM or R2* values.
Summary statistics for each ROI, after transformation to the native QSM space, were used to compare χ, χ_paramag, and χ_diamag values (Fig.2, left).
Results:
RETICLES-separated paramagnetic and diamagnetic images (Fig.1) show similar anatomical contrasts to those from other source-separation methods. In wmGP and wmPu with relatively high myelin content, correspondingly high χ_diamag are reflected in the group, as compared to all other subregions. The posterior putamen (pPu), which is particularly sensitive to neurodegeneration, shows nearly zero χ_diamag. The body of the caudate (encompassing both cCd and pCd), which receives more radial fibers from the internal capsule than the head of the caudate (aCd) [10], shows correspondingly higher χ_diamag.
Notably, after source-separation, BG subregions that typically have relatively low (median ~ 20 ppb) QSM values, such as the aCd, cCd, pCd, and aPu, have substantially (approximately 50%) expanded magnitude in their corresponding χ_paramag values.
Notably, after source-separation, BG subregions that typically have relatively low (median ~ 20 ppb) QSM values, such as the aCd, cCd, pCd, and aPu, have substantially (approximately 50%) expanded magnitude in their corresponding χ_paramag values.
Conclusions:
We propose a reference tissue-based QSM source separation method that only requires linear algebraic operations between a pair of coupled χ and R2* equations, and without additional T2 acquisition. It demonstrates reasonable anatomical contrasts and meaningful quantitative results in BG subregions. The expanded magnitude of the χ_paramag values in select BG subregions, presumably more accurately reflecting the underlying iron content, provides increased dynamic range and sensitivity to evaluate the potential impact of iron deficiency in adolescents.
Modeling and Analysis Methods:
Other Methods 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Subcortical Structures 2
Novel Imaging Acquisition Methods:
Imaging Methods Other
Keywords:
MRI
Other - QSM
1|2Indicates the priority used for review
Provide references using author date format
[2] H.-G. Shin et al., “χ-separation: Magnetic susceptibility source separation toward iron and myelin mapping in the brain,” NeuroImage, vol. 240, p. 118371, Oct. 2021, doi: 10.1016/j.neuroimage.2021.118371.
[3] Z. Zhang et al., “Blip up-down acquisition for spin- and gradient-echo imaging (BUDA-SAGE) with self-supervised denoising enables efficient T2, T2*, para- and dia-magnetic susceptibility mapping,” Magn. Reson. Med., vol. 88, no. 2, pp. 633–650, 2022, doi: 10.1002/mrm.29219.
[4] J. Emmerich, P. Bachert, M. E. Ladd, and S. Straub, “A novel phantom with dia- and paramagnetic substructure for quantitative susceptibility mapping and relaxometry,” Phys. Med., vol. 88, pp. 278–284, Aug. 2021, doi: 10.1016/j.ejmp.2021.07.015.
[5] B. Dymerska et al., “Phase unwrapping with a rapid opensource minimum spanning tree algorithm (ROMEO),” Magn. Reson. Med., vol. 85, no. 4, pp. 2294–2308, 2021, doi: 10.1002/mrm.28563.
[6] C. Wang et al., “Phenotypic and genetic associations of quantitative magnetic susceptibility in UK Biobank brain imaging,” Nat. Neurosci., vol. 25, no. 6, Art. no. 6, Jun. 2022, doi: 10.1038/s41593-022-01074-w.
[7] J. Acosta-Cabronero, C. Milovic, H. Mattern, C. Tejos, O. Speck, and M. F. Callaghan, “A robust multi-scale approach to quantitative susceptibility mapping,” NeuroImage, vol. 183, pp. 7–24, Dec. 2018, doi: 10.1016/j.neuroimage.2018.07.065.
[8] J. Hanspach et al., “Methods for the computation of templates from quantitative magnetic susceptibility maps (QSM) – toward improved atlas- and voxel-based analyses (VBA),” J. Magn. Reson. Imaging JMRI, vol. 46, no. 5, pp. 1474–1484, Nov. 2017, doi: 10.1002/jmri.25671.
[9] J. P. Wansapura, S. K. Holland, R. S. Dunn, and W. S. Ball Jr., “NMR relaxation times in the human brain at 3.0 tesla,” J. Magn. Reson. Imaging, vol. 9, no. 4, pp. 531–538, 1999, doi: 10.1002/(SICI)1522-2586(199904)9:4<531::AID-JMRI4>3.0.CO;2-L.
[10] S. A. Kotz, A. Anwander, H. Axer, and T. R. Knösche, “Beyond Cytoarchitectonics: The Internal and External Connectivity Structure of the Caudate Nucleus,” PLOS ONE, vol. 8, no. 7, p. e70141, Jul. 2013, doi: 10.1371/journal.pone.0070141.