DeepQSMSeg: A Deep Learning-based Sub-cortical Nucleus Segmentation Tool for Quantitative Susceptibility Mapping.

BackgroundCerebral microbleeds (CMBs) are small hypointense round lesions that indicate leakage of blood products from cerebral vessels damaged by β‐amyloid‐40 (Aβ) and typically are detected by T2*‐weighted and susceptibility weighted imaging (SWI) on MRI. They are indicators of cerebral small vessel diseases, especially cerebral amyloid angiopathy (CAA), affecting cortical small arteries. Quantitative susceptibility mapping (QSM) is an advanced MRI imaging technique used to quantify the magnetic susceptibility of tissues in the human body. In this study, we utilized the QSM technique to analyze venous structures at ultra‐high resolution in a cohort of patients with AD and age‐matched control subjects using 7T MRI. Our aim is to evaluate the feasibility of using ultra‐high resolution QSM at 7T MRI to find a link between small veins and CMBs in individuals with AD.MethodIn‐vivo human imaging was performed using a 7T MRI scanner and were analyzed by two neuroradiologists. CMBs were defined as hyperintense, round lesions in the QSM images. For verification, their appearance was matched on the SWI images, where CMBs are hypointense and associated with. CMBs with an immediate adjacent to a vein were assessed in the QSM images. They were defined as CMBs with small veins connection.ResultThis high‐resolution 7T neuroimaging study investigated whether a spatial relationship between small veins and CMBs could be detected in a cohort of AD patients. QSM enabled depiction of cerebral CMBs with a direct connection to a vein and their localization. Our data support the notion that CMBs relate anatomically to veins. Understanding the role of vein pathology in CAA and its role in neurodegenerative pathologies might open novel insights in the pathophysiology of AD.ConclusionA key strength of our study lies in the utilization of high‐resolution QSM at 7T MRI. High‐resolution MRI enhances the visibility of small structures, and when maintained at the same resolution, imaging at a higher field MRI increases the signal‐to‐noise ratio. QSM elevates CMB detection sensitivity while facilitating precise visualization of the brain's venous vasculature. Notably, the postprocessing algorithm ensures a high level of specificity for venous blood susceptibilities, minimizing the likelihood of measuring arterioles or slow‐flow small arteries.

Quantitative susceptibility mapping (QSM) is a magnetic resonance imaging (MRI) technique that can measure the magnetic susceptibility of tissues, which can reflect their iron content. QSM has been used to detect iron accumulation in cortical and subcortical brain regions. However, its application in subcortical regions such as the basal ganglia, particularly the putamen, is rare in patients with amyotrophic lateral sclerosis (ALS). We present the case of a 40-year-old male patient with ALS who underwent an MRI for QSM. We compared his QSM images with those of a control subject and performed a quantitative analysis of the magnetic susceptibility values in the putamen regions. We also reviewed the literature on previous QSM studies in ALS and summarized their methods and findings. Our QSM analysis revealed increased magnetic susceptibility values in the bilateral putamen of the ALS patient compared to controls, indicating iron overload. This finding is consistent with previous studies reporting iron dysregulation in subcortical nuclei in ALS. We also discussed the QSM processing techniques used in our study and in the literature, highlighting their advantages and limitations. This case report demonstrates the potential of QSM as a sensitive MRI biomarker for evaluating iron levels in subcortical regions of ALS patients. QSM can provide quantitative information on iron deposition patterns in both motor and extra-motor areas of ALS patients, which may help understand the pathophysiology of ALS and monitor disease progression. Further studies with larger samples are needed to validate these results and explore the clinical implications of QSM in ALS.

Voxel-based morphometry (VBM) is widely used to quantify the progression of Alzheimer's disease (AD), but improvement is still needed for accurate early diagnosis. We evaluated the feasibility of a novel diagnosis index for early diagnosis of AD based on quantitative susceptibility mapping (QSM) and VBM. Thirty-seven patients with AD, 24 patients with mild cognitive impairment (MCI) due to AD, and 36 cognitively normal (NC) subjects from four centers were included. A hybrid sequence was performed by using 3-T MRI with a 3D multi-echo GRE sequence to obtain both a T1-weighted image for VBM and phase images for QSM. The index was calculated from specific voxels in QSM and VBM images by using a linear support vector machine. The method of voxel extraction was optimized to maximize diagnostic accuracy, and the optimized index was compared with the conventional VBM-based index using receiver operating characteristic analysis. The index was optimal when voxels were extracted as increased susceptibility (AD > NC) in the parietal lobe and decreased gray matter volume (AD < NC) in the limbic system. The optimized proposed index showed excellent performance for discrimination between AD and NC (AUC = 0.94, p = 1.1 × 10-10) and good performance for MCI and NC (AUC = 0.87, p = 1.8 × 10-6), but poor performance for AD and MCI (AUC = 0.68, p = 0.018). Compared with the conventional index, AUCs were improved for all cases, especially for MCI and NC (p < 0.05). In this preliminary study, the proposed index based on QSM and VBM improved the diagnostic performance between MCI and NC groups compared with the VBM-based index. • We developed a novel diagnostic index for Alzheimer's disease based on quantitative susceptibility mapping (QSM) and voxel-based morphometry (VBM). • QSM and VBM images can be acquired simultaneously in a single sequence with little increasing scan time. • In this preliminary study, the proposed diagnostic index improved the discriminative performance between mild cognitive impairment and normal control groups compared with the conventional VBM-based index.

Standard functional MRI (fMRI), which includes resting-state or paradigm-driven designs, is widely used in studies of brain function, aging, and disease. These fMRI studies typically use two-dimensional gradient echo-planar imaging, which inherently contains phase data that enables quantitative susceptibility mapping (QSM). This work focuses on the dual value of QSM within fMRI studies, by providing both a localized analysis of functional changes in activated tissue, and iron-sensitive structural maps in deep grey matter (DGM). Using a visual paradigm fMRI study on healthy volunteers at clinical (1.5 T) and high field strength (4.7 T), we perform functional analysis of magnitude and QSM time series, and at the same time harness structural QSM of iron-rich DGM, including globus pallidus, putamen, caudate head, substantia nigra, and red nucleus. The effects of fMRI spatial resolution and time series variation on structural DGM QSM are investigated. Our results indicate that structural DGM QSM is feasible within existing fMRI studies, provided that the voxel dimensions are equal to or less than 3 mm, with higher resolutions preferred. The mean DGM QSM values were about 40 to 220 ppb, while the interquartile ranges of the DGM QSM time series varied from about 3 to 9 ppb, depending on structure and resolution. In contrast, the peak voxel functional QSM (fQSM) changes in activated visual cortex ranged from about -10 to -30 ppb, and functional clusters were consistently smaller on QSM than magnitude fMRI. Mean-level DGM QSM of the time series was successfully extracted in all cases, while fQSM results were more prone to residual background fields and showed less functional change compared with standard magnitude fMRI. Under the conditions prescribed, standard fMRI studies may be used for robust mean-level DGM QSM, enabling study of DGM iron accumulation, in addition to functional analysis. Copyright © 2016 John Wiley & Sons, Ltd.

Iron quantification in Parkinson's disease using an age-based threshold on susceptibility maps: The advantage of local versus entire structure iron content measurements

Quantitative susceptibility mapping (QSM) technique had been used to measure the magnetic susceptibility of brain tissue in clinical practice. However, QSM presented echo-time (TE) dependence, and an appropriate number of echo-times (nTEs) for QSM became more important to obtain the reliable susceptibility value. The aim of the study was to explore the optimal nTEs for quantitative susceptibility mapping (QSM) measurements of basal ganglia nuclei in the healthy brain. 3D multi-echo enhanced gradient recalled echo T2 star weighted angiography (ESWAN) sequence was acquired on a 3.0T MR scanner for QSM analysis. Regions of interests (ROIs) were drawn along the margin of the head of the caudate nucleus (HCN), putamen (Pu) and globus pallidus (GP). The mean susceptibility value and standard deviation of the ROIs were derived from the pixels within each region. CV analysis demonstrated that TE6, TE8 and TE14 ESWAN sequences presented consistent lower CV value (< 1) for QSM measure of HCN, Pu and GP. ANOVA identified that susceptibility value showed no significant difference between TE6 and TE8 in HCN, Pu and GP (P > 0.05). ICC analysis demonstrated that the susceptibility value of TE6-TE8 had the highest ICC value as compared with TE6-TE14 and TE8-TE14 in HCN, Pu and GP. Combined with the timeefficiency of MRI scanning, TE6 sequence could not only provide the reliable QSM measurement but also short imaging time. The current study identified that the optimal nTEs of ESWAN were 6 TEs (2.9ms ~ 80.9ms) for QSM measurement of basal ganglia nuclei in the healthy brain.

Quantitative susceptibility mapping has been proposed to assess intraplaque hemorrhage (IPH) in the carotid artery. The purpose of this study was to compare the diagnostic accuracy of preoperative quantitative susceptibility mapping with that of the conventional T1-weighed 3D-FSE sequence for detecting IPH in cervical ICA stenosis in patients undergoing carotid endarterectomy by using histology as the reference standard. Carotid T1-weighted 3D-FSE and QSM images were obtained from 16 patients with cervical ICA stenosis before carotid endarterectomy. Relative signal intensity and susceptibility of the ICA were measured on 3 axial images, including the location of most severe stenosis on T1-weighted 3D-FSE and quantitative susceptibility mapping images, respectively. Three transverse sections of carotid plaques excised by carotid endarterectomy, which corresponded with images on MR imaging, were stained with H&E, antibody against glycophorin A, and Prussian blue, and the relative area of histologic IPH was calculated. The correlation coefficient was significantly greater between susceptibility and relative area-histologic IPH (ρ = 0.691) than between relative signal intensity and relative area-histologic IPH (ρ = 0.413; P = .0259). The areas under the receiver operating characteristic curves for detecting histologic sections consisting primarily of IPH (relative area-histologic IPH > 40.7%) tended to be greater for susceptibility (0.964) than for T1WI FSE-relative signal intensity (0.811). Marginal homogeneity was observed between susceptibility and histologic sections consisting primarily of IPH (P = .0412), but not between T1-weighted FSE-relative signal intensity and histologic sections consisting primarily of IPH (P = .1824). Pre-carotid endarterectomy quantitative susceptibility mapping detects histologic IPH in cervical ICA stenosis more accurately than preoperative T1-weighted 3D-FSE imaging.

ObjectiveIn dynamic susceptibility contrast MRI (DSC-MRI), an arterial input function (AIF) is required to quantify perfusion. However, estimation of the concentration of contrast agent (CA) from magnitude MRI signal data is challenging. A reasonable alternative would be to quantify CA concentration using quantitative susceptibility mapping (QSM), as the CA alters the magnetic susceptibility in proportion to its concentration.Material and methodsAIFs with reasonable appearance, selected on the basis of conventional criteria related to timing, shape, and peak concentration, were registered from both ΔR2* and QSM images and mutually compared by visual inspection. Both ΔR2*- and QSM-based AIFs were used for perfusion calculations based on tissue concentration data from ΔR2*as well as QSM images.ResultsAIFs based on ΔR2* and QSM data showed very similar shapes and the estimated cerebral blood flow values and mean transit times were similar. Analysis of corresponding ΔR2* versus QSM-based concentration estimates yielded a transverse relaxivity estimate of 89 s−1 mM−1, for voxels identified as useful AIF candidate in ΔR2* images according to the conventional criteria.DiscussionInterestingly, arterial concentration time curves based on ΔR2* versus QSM data, for a standard DSC-MRI experiment, were generally very similar in shape, and the relaxivity obtained in voxels representing blood was similar to tissue relaxivity obtained in previous studies.

Dynamic changes in cerebrovascular reactivity after acetazolamide administration vary markedly among patients with major cerebral arterial steno-occlusive disease. MR quantitative susceptibility mapping can dynamically quantify the cerebral magnetic susceptibility. The purpose of this study was to determine whether dynamic changes in susceptibility after administration of acetazolamide on 7T quantitative susceptibility mapping are associated with pre-existing states of CBV and the cerebral metabolic rate of oxygen in the cerebral hemispheres with major cerebral arterial steno-occlusive disease. Sixty-five patients underwent 7T MR imaging at baseline and at 5, 10, 15, and 20 minutes after acetazolamide administration. Differences between the susceptibility of venous structures and surrounding brain tissue were calculated in the quantitative susceptibility mapping images. Susceptibility differences at 5, 10, 15, and 20 minutes after acetazolamide administration relative to baseline were calculated in 97 cerebral hemispheres with major cerebral arterial steno-occlusive disease. CBV and the cerebral metabolic rate of oxygen were also calculated using 15O-gas PET in the resting state. Dynamic changes of susceptibility after acetazolamide administration were classified into 3 patterns: abnormally increasing 5 or 10 minutes after acetazolamide administration; abnormally decreasing within 20 minutes after acetazolamide administration; and remaining unchanged after acetazolamide administration. CBV was significantly greater in the first pattern than in the latter 2. The cerebral metabolic rate of oxygen differed significantly in descending order from the first to middle to last pattern. Dynamic changes of susceptibility after acetazolamide administration on 7T MR quantitative susceptibility mapping are associated with pre-existing states of CBV and the cerebral metabolic rate of oxygen in major cerebral arterial steno-occlusive disease.

Chronic drug doping has a great influence on the functional and morphological properties of brain tissues. However, the underlying mechanism related to these variations remains unclear. Since there are evidences showing that iron concentration affects the synthesis of dopamine and monoamine metabolism, in the present work, we propose to use quantitative susceptibility mapping (QSM) to investigate the iron distribution in the brain tissues. To achieve this, 9 heroin users (HUs) and 2 controls (CLs) were recruited for MRI scanning. QSM images were obtained firstly with STI-suite and then registered and segmented with a QSM atlas and parcellation maps. The results demonstrated that, the HUs experienced higher iron accumulation in putamen, globus pallidus, caudate nucleus and thalamus than CLs. Our findings showed that there is a link between the morphological changes and the iron concentration variations in the brain of Hus. This would be helpful for understanding the underlying mechanism by which drug addiction affects the brain and providing the guidance for effective treatment strategies.

The aims of this study were to investigate and validate quantitative susceptibility mapping (QSM) for lesional iron quantification in cerebral cavernous malformations (CCMs). Magnetic resonance imaging studies were performed in phantoms and 16 patients on a 3-T scanner. Susceptibility weighted imaging, QSM, and R2* maps were reconstructed from in vivo data acquired with a 3-dimensional, multi-echo, and T2*-weighted gradient echo sequence. Magnetic susceptibility measurements were correlated to susceptibility weighted imaging and R2* results. In addition, iron concentrations from surgically excised CCM lesion specimens were determined using inductively coupled plasma mass spectrometry and correlated with QSM measurements. The QSM images demonstrated excellent image quality for depicting CCM lesions in both sporadic and familial cases. Susceptibility measurements revealed a positive linear correlation with R2* values (R(2) = 0.99 for total, R(2) = 0.69 for mean; P < 0.01). Quantitative susceptibility mapping values of known iron-rich brain regions matched closely with those of previous studies and in interobserver consistency. A strong correlation was found between QSM and the concentration of iron phantoms (0.925; P < 0.01), as well as between QSM and mass spectroscopy estimation of iron deposition (0.999 for total iron, 0.86 for iron concentration; P < 0.01) in 18 fragments of 4 excised human CCM lesion specimens. The ability of QSM to evaluate iron deposition in CCM lesions was illustrated via phantom, in vivo, and ex vivo validation studies. Quantitative susceptibility mapping may be a potential biomarker for monitoring CCM disease activity and response to treatments.

Post-carotid endarterectomy (CEA) cerebral hyperperfusion (CH) can cause intracerebral hemorrhage and cognitive decline. Alterations in susceptibility in response to acetazolamide (ACZ) on 7T MRI quantitative susceptibility mapping (QSM) detect elevated CBV occurring due to impaired cerebrovascular autoregulation. We explored preoperative relative susceptibility changes on 7T MRI QSM in response to ACZ and their ability to predict CH following CEA. Sixty-three patients with uni- or bilateral cervical ICA stenosis of ≥70% underwent 7T MRI at baseline and at 5, 10, 15, and 20 minutes after ACZ administration before surgery. The difference between the susceptibility of venous structures and surrounding brain parenchyma at each time point after ACZ administration relative to the difference at baseline (relative susceptibility difference [RSD]) on QSM images was calculated in the cerebral hemisphere ipsilateral to the surgery site. Brain perfusion SPECT was conducted preoperatively and immediately following CEA to detect postoperative CH (≥100% rise in CBF postoperatively). In 9 patients with postoperative CH, the RSD was significantly increased at 5 or 10 minutes following ACZ administration (P < .05) but was reduced at 15 and 20 minutes (P < .05). In 54 patients without postoperative CH, the SD at all 4 time points after ACZ administration was significantly lower than the baseline value (P < .05). The area under the receiver operating characteristic curve to predict postoperative CH was significantly greater in RSD5 (0.981; 95% CI, 0.910-0.999) than in RSD15 (0.872; 95% CI, 0.764-0.943) (P < .05) or RSD20 (0.780; 95% CI, 0.658-0.874) (P < .01). Sensitivity, specificity, and positive and negative predictive values for RSD5 at a cutoff near the left upper corner of the curve were 100%, 89%, 60%, and 100%, respectively. Logistic regression analysis revealed that only RSD5 significantly predicted postoperative CH (95% CI, 455.9-4043.6; P < .05). Changes in susceptibility on preoperative 7T MRI QSM following ACZ administration predict CH following CEA. Patients with increased RSD5 on pre-CEA 7T MRI QSM following ACZ administration should undergo brain perfusion imaging immediately after surgery. Detection of CH on postoperative brain perfusion imaging warrants intensive blood pressure control.

Myotonic dystrophy type 1 is a dominantly inherited disorder, affecting musculoskeletal and central nervous systems and mainly characterized by progressive muscular atrophy and multisystemic damages including cardiac, respiratory and sleep dysfunctions. Neuroimaging studies conducted in myotonic dystrophy type 1 patients have documented widespread cerebral alterations encompassing structural, microstructural, functional and metabolic aspects of the brain, while comparatively few studies have investigated the role of iron concentration in the pathophysiology of central nervous system impairment. We report here the use of quantitative susceptibility (χ) mapping to explore iron content of both cortical and subcortical structures in myotonic dystrophy type 1 patients and to assess its possible clinical relevance, combining imaging and clinical data. Thirty-four myotonic dystrophy type 1 participants (20 females, 46.8 ± 12.0 years old) and 35 age- and sex- matched healthy controls (20 females, 50.5 ± 17.3 years old) were included in the study. All participants underwent MRI examinations in the same 3-Tesla scanner. The MRI protocol included 3D morphological T1-weighted magnetization prepared rapid gradient echo and T2*weighted multi-echo gradient echo for quantitative susceptibility mapping reconstruction. Cortical and subcortical structures were automatically segmented, and a volume of interest–based analysis was performed; χ distributions were compared between the two groups and myotonic dystrophy type 1 χ values were correlated with clinical and laboratory data. In the myotonic dystrophy type 1 group, a significant increase of χ was found in almost all cortical gyri, as a non-specific sign of neurodegeneration. Among subcortical structures, χ was significantly higher in myotonic dystrophy type 1 group in both thalamus (ventral and pulvinar nuclei) and brainstem (pons and medulla), compared to healthy controls. Additionally, correlation analysis showed some links between χ in subcortical structures and clinical signs, suggesting greater iron concentration with deterioration of clinical conditions. Thalamic χ values were associated with cardiological parameters and disability scores and, as with brainstem χ, they were also positively correlated with the number of central apnoeas; finally, thalamic and brainstem χ were negatively correlated with the age of onset. This study showed a correlation between autonomic dysfunction related to certain subcortical structures and their χ; higher values of χ correlated with greater functional impairment, suggesting iron accumulation detected by the quantitative susceptibility mapping technique is a possible biomarker of disease progression.

Background: Mild traumatic brain injury (mTBI) is a common neurological condition characterized by subtle brain dysfunction. Although generally considered less severe than moderate and severe TBI, it can have significant consequences on cognitive and motor function. Iron deposition in the basal ganglia nuclei, particularly in individuals with mTBI, has been reported as a potential biomarker for neurodegenerative and neurocognitive changes following injury. While previous studies have explored the relationship between iron deposition and mTBI, few human studies have examined these changes using quantitative susceptibility mapping (QSM). The QSM is a powerful technique that enables the non-invasive measurement of iron content in deep brain structures, providing valuable insights into the pathophysiology of mTBI. Objectives: This study aimed to compare magnetic susceptibility values (iron content) in the deep gray matter (DGM) nuclei regions of patients with mTBI and healthy controls (HCs). Patients and Methods: In this case-control study, we enrolled 10 acute mTBI patients and 10 age-matched HCs who underwent imaging with a 3 Tesla Prisma scanner magnetic resonance imaging (MRI), including T1 weighted magnetization-prepared rapid gradient echo (T1-MPRAGE) and multi-echo 3D gradient-recalled echo (GRE) sequences for QSM reconstruction. Inclusion criteria included patients aged 16 - 50 years, a Glasgow Coma Scale (GCS) score of 13 - 15 upon emergency department arrival, and a history of loss of consciousness for less than 30 minutes. Exclusion criteria included prior brain injuries, neurological disorders, psychiatric conditions, or substance abuse. The magnetic susceptibility of the DGM nuclei was calculated using the streaking artifact reduction for QSM (STAR-QSM) technique. Group comparisons were performed using independent t-tests. Results: Significantly higher magnetic susceptibility values were observed in the DGM nuclei regions, specifically in the right caudate (P &lt; 0.001), left caudate (P = 0.002), right thalamus (P = 0.004), and right hippocampus (P = 0.002), in mTBI patients compared to HCs. Conclusion: These findings suggest that mTBI may be associated with alterations in iron content or magnetic susceptibility within the DGM nuclei regions compared to HCs, potentially reflecting microstructural changes. This study provides preliminary evidence for the potential of QSM as a tool for investigating tissue susceptibility in acute mTBI. Future research should focus on longitudinal studies with larger sample sizes to examine the temporal evolution of iron accumulation and its potential relationship with functional outcomes in mTBI patients.

Quantitative susceptibility mapping (QSM) on MRI quantifies tissue magnetic susceptibility, which increases with iron accumulation, myelin loss, and neuroinflammation. Elevated QSM in the substantia nigra (SN) has been reported in Lewy body disease and other parkinsonian disorders, but from existing literature it remains unclear whether these findings are driven by neurodegeneration-related iron deposition or other neuropathologic features. We studied 59 autopsied participants who underwent antemortem 3T MRI with QSM (median age at death, 78.5years; MRI-to-death interval, 2.0years), including clinical diagnoses of 18 with Alzheimer's-type dementia, 15 cognitively unimpaired, 9 with mild cognitive impairment, and 9 with dementia with Lewy bodies. A machine learning-incorporated digital histopathology pipeline quantified tau burden, iron deposition, and neuronal densities. The SN was divided into geometric quadrants, and QSM values were analyzed in relation to corresponding neuropathologic measures within each quadrant. Iron deposition correlated with QSM in all quadrants (ρ = 0.41-0.56, all P < 0.005). Tau burden correlated with QSM in the ventromedial (VM) quadrant (ρ = 0.45, P = 0.002), whereas lower pigmented neuron density was associated with higher QSM in the dorsomedial quadrant (ρ = -0.35, P = 0.007). Rank regression analysis confirmed iron as the strongest predictor of QSM across all quadrants (β = 0.35-1.06, P ≤ 0.026), with tau independently associated with QSM in the VM (β = 0.45, P = 0.015). Mediation analysis demonstrated that tau exerted direct (0.45, P = 0.018) and indirect effects via iron (0.12, P = 0.046) on QSM in the VM, with 80% of the effect being direct. These findings underscore the contributions of tau pathology, pigmented neuron density, and iron deposition to nigral magnetic susceptibility and highlight the potential for QSM to serve as a sensitive biomarker for diverse neuropathologies.