Skip to main content
NCBI home page
IBRO Neuroscience Reports logoLink to IBRO Neuroscience Reports
. 2026 Apr 1;20:540–547. doi: 10.1016/j.ibneur.2026.03.013

Prediction of early neurological deterioration using quantitative susceptibility mapping in small artery occlusion ischemic stroke

Xuelian Tang a,1, Yanhua Yu a,1, Yusheng Yu b,⁎, Jinan Chen a,⁎
PMCID: PMC13087643  PMID: 42004614

Abstract

Background

Approximately 20–30% of ischemic stroke patients with small artery occlusion (SAO) experience early neurological deterioration (END) after stroke onset, which is associated with a poor prognosis. Our aim was to determine the predictive ability of quantitative susceptibility mapping (QSM) for END in patients with SAO.

Methods

As neurological deficits, the National Institutes of Health Stroke Scale (NIHSS) was used. END was defined as an increase of ≥ 2 points on the NIHSS from baseline during the first week. 126 SAO stroke participants within 24 h of stroke onset were included in this study and divided into END group and nEND (no early neurological deterioration) group. QSM values were calculated from magnetic resonance imaging (MRI) data and clinical characteristics were collected. A cluster-based permutation test was conducted between groups and diagnostic efficacy of the QSM value was evaluated.

Results

As a result, END group showed a different cortical susceptibility pattern compared with nEND group. Specifically, END group showed higher QSM value in the left/right precentral gyrus, the middle frontal gyrus, the opercular part of the left/right inferior frontal gyrus, the triangular part of left inferior frontal gyrus and postcentral gyrus, while lower in the dorsolateral part of left superior frontal. The support vector machine (SVM) classifier’s analysis showed a high power for QSM diagnostic efficacy between END and nEND groups. In addition, homocysteine (Hcy), low-density lipoprotein (LDL), and total cholesterol (TC) were correlated with QSM values in some specific brain regions.

Conclusion

The cluster-based permutation test indicated the differences of brain QSM value distribution between END and nEND groups. The SVM classifier’s analysis showed a high power for QSM diagnostic efficacy between the two groups. Our study revealed the underlying predictive value of QSM in END of SAO stroke patients.

Keywords: Small artery occlusion, Quantitative susceptibility mapping, Neurological deficits, Early neurological deterioration

1. Introduction

Small artery occlusion (SAO) is one of the main subtypes of ischemic stroke (Adams et al., 1993). SAO is defined as subcortical infarcts stemming from occlusion of penetrating arteries (Sacco et al., 2006). Early neurological deterioration (END) occurs in approximately 20–30% of ischemic stroke patients and is often associated with poor functional outcomes (Sharma et al., 2024, Siegler et al., 2013). END can occur in all stroke subtypes, but the incidence has been reported to be higher in SAO stroke than other stroke etiological subtypes (Steinke and Ley, 2002). However, the pathogenesis of END in stroke is complex and we still lack effective treatment strategies for END (Liu et al., 2023). The consensus is that early prediction and prevention of END is one of the important therapeutic targets in acute ischemic stroke (AIS) management. However, sensitive methods predicting END have not been evaluated.

Quantitative Susceptibility Mapping (QSM) is an non-invasion MRI technique that allows quantification of local tissue magnetic susceptibility via complex post-processing of a multi-echo gradient-recalled echo(GRE) acquisition (Liu et al., 2015, Ruetten et al., 2019). QSM can provide valuable information regarding pathological processes such as iron deposition, hemorrhage, oxygen extraction fraction, calcification, and myelin (Deistung et al., 2017, Fushimi et al., 2024, Probst et al., 2021). Cerebral small vessel disease (SVD) refers to any pathologic process that damages small end arteries, arterioles, venules, and brain capillaries (Wardlaw et al., 2013). Multiple studies have shown that SVD is closely related to poor neurological prognosis in stroke patients (Guo et al., 2021, Wang et al., 2024). SVD causes lacunar stroke and intracerebral hemorrhage (Markus and de Leeuw, 2023). Studies have shown that iron homeostasis broke down on the onset of ischemic stroke which led to the abnormal accumulation of iron in brain tissue (DeGregorio-Rocasolano et al., 2019, Kataike et al., 2024, Stankiewicz et al., 2007). This iron accumulation usually contributed to worsen tissue damage and the severity of ischemic stroke (Selim and Ratan, 2004). As mentioned above, the abnormal iron deposition could be sensitively detected by QSM. Some studies have reported that QSM can assist in assessing the neurological function impairment of SVD patients (Wang et al., 2025). SAO is a manifestation of SVD (Munoz, 2003). However, the clinical application of QSM in END of SAO is still largely unexplored.

In this study we sought to employ QSM to quantify the distribution characteristics of magnetic susceptibility in the brain tissue of patients with SAO and explore its relationship with END in these patients. By characterizing these changes of magnetic susceptibility, this work aims to provide a deeper understanding of SAO and suggest potential focal treatment strategies for patients with SAO.

2. Materials and methods

2.1. Participants

One hundred and twenty-six consecutive patients were admitted from April 2023 to May 2024 at a tertiary care hospital, after our quality control, 111 patients were included. Patients were divided into two groups including END group and no early neurological deterioration (nEND) group, depending on if patients experienced an increase in NIHSS score of 2 points or more in the first 7 days from the onset of AIS. Neurological assessments using the NIHSS are routinely conducted at three critical time points: upon hospital admission, on the seventh day of hospitalization, and whenever clinical deterioration occurs within the first week. The highest NIHSS score recorded during this 7-day period is compared with the baseline admission score. According to the classification principles of the Trial of Org 10172 in Acute Stroke Treatment (TOAST), the detailed classification of SAO is described as follows: patients with a traditional clinical lacunar syndrome and brainstem or subcortical hemispheric lesion(s) with a diameter ≤ 1.5 cm on MRI. The inclusion criteria were as follows: (1) age ≥ 18 years old; (2) the etiologic classification of ischemic stroke was SAO subtype according to TOAST, which were confirmed by two neurological doctors who were blinded to the specific grouping information of the patients. (Adams et al., 1993); (3) cranial MRI examination completed within 24 h of first onset.

The exclusion criteria were as follows: (1) those receiving thrombolysis or endovascular treatment; (2) those with a history of major neurologic or psychiatric diseases or other clinically significant diseases; (3) those suffered from intracranial tumors and hemorrhagic strokes; (4) those could not cooperate or understand the study procedures. All the participants provided written informed consent. This study was in accordance with the Declaration of Helsinki and conducted after approval by medical ethics committee of nanjing jiangning hospital (2023–03–074-K01).

2.2. Clinical information

For all enrolled patients, neurological impairment was assessed using the NIHSS scores and the following clinical information was collected: age, sex, smoking, hypertension, diabetes mellitus, history of atrial fibrillation, NIHSS score at admission and 7 days after stroke onset. The laboratory data including the total cholesterol (TC), triglycerides (TG), homocysteine (Hcy), low-density lipoprotein (LDL), high-density lipoprotein (HDL), fibrinogen and D-Dimer were collected.

2.3. Image acquisition

MRI data were acquired using a 3.0 T magnetic resonance scanner (Siemens, Germany). T1 weighted images were acquired with a 3D fast spoiled gradient-echo sequence and the T2w sequence was acquired using a multi-echo 3D GRE. The scanning parameters for each sequence were as follows: 1)：T1w：TR = 1900 ms, TE = 2.01 ms, flip angle = 9°, field of view (FOV) = 256 mm × 256 mm, voxel size = 1 × 1 × 1 mm3, 176 slices, slice thickness = 1 mm. 2)：T2w：TR = 50 ms, first TE = 6.84 ms, last TE = 43.74 ms, number of echoes = 10, bandwidth = 350 kHz, flip angle = 15°, field of view= 195 mm × 256 mm, voxel size = 0.75 × 0.75 × 2 mm3.

2.4. QSM mapping preprocessing and voxel-based morphometry analysis

The standard pipeline in the SEPIA (SuscEptibility mapping PIpeline tool for phAse images) toolbox in MATLAB (Mathworks, Natick, MA) was used to generate the QSM maps from GRE data (Chan and Marques, 2021). We used the BET tool in FMRIB Software (FSL, https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL) from the MEDI toolbox to perform brain extraction on whole-brain magnitude data. The general process consists of three points: 1) total field recovery: the optimum weights method was used to perform echo phase combination (Robinson et al., 2017) and SEGUE was applied to perform phase unwrapping (Karsa and Shmueli, 2019). 2) background field removal: the background field was removed using the Laplacian boundary value (LBV) (Zhou et al., 2014). 3) QSM dipole inversion: magnetic susceptibility was quantitatively calculated using the streaking artifact reduction for QSM method and then QSM maps were generated (Wei et al., 2017). In addition to the core pipeline steps, we implemented rigorous quality control (QC) and artifact mitigation measures. Phase wrapping correction: The SEGUE algorithm applied in SEPIA was combined with post-unwrapping visual inspection to identify residual phase jumps (>π). Background field inhomogeneity removal: After LBV-based background suppression, a brain mask was applied to exclude extracranial signals, ensuring only intracranial phase contributions were retained. Air-tissue interface artifacts: Susceptibility maps were thresholded to mask voxels with values > 500 ppb (consistent with non-brain air/soft tissue), reducing blooming artifacts near sinuses and ear canals. QC metrics: Scans with > 5% artifact-contaminated voxels in target regions or SNR < 20 (magnitude images) were excluded. Supplementary Figure 1 showed representative QSM maps of one END patient and one nEND patient.

After getting the QSM maps, we used the voxel-based morphometry (VBM) analysis in the DPABI (Data Processing & Analysis for Brain Imaging) version V8.2, built on the MATLAB R2020b (MathWorks, Natick, MA, USA) to extract the mean QSM values of every cortical region (Chen et al., 2022a, Nemoto, 2017). DPABI integrates several core toolboxes essential for the workflow: SPM12 (Statistical Parametric Mapping version 12) (Wellcome Centre for Human Neuroimaging, London, UK) was employed for key preprocessing steps, including T1w image segmentation (via the "New Segment" module), spatial normalization using DARTEL, and modulation. SPM12’s diffeomorphic registration algorithms underpin the DARTEL framework, ensuring accurate alignment and template generation. Addressing Limitations of Spatial Normalization in Stroke Patients: To mitigate challenges posed by altered brain anatomy in stroke patients (e.g., infarction, edema, or atrophy), we implemented the following strategies: Robust Normalization with DARTEL: The VBM pipeline in DPABI uses the DARTEL (Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra) algorithm, which employs diffeomorphic transformations to align individual brains to the MNI template. Lesion Masking: For patients with visible infarcts on structural images, we manually delineated infarct regions using magnitude images and applied a lesion mask during normalization.

2.5. Statistical analysis

2.5.1. Distribution patterns of QSM in END and nEND

Firstly, we registered the QSM image to the AAL3v1 atlas and resampled it to a voxel size of 2 mm × 2 mm × 2 mm. To characterize the distribution patterns of QSM in END and nEND, we conducted a one-sample t-test on QSM maps in END and nEND within grey matter mask separately. The statistical threshold was set at p < 0.05 using a permutation test with threshold-free cluster enhancement (TFCE) as well as family-wise error (FWE) correction together with a cluster extent k > 200 voxels (1600 mm3). Age, sex, hypertension, atrial fibrillation and diabetes were included as covariates. Subsequently, masks were generated based on the group-specific brain regions with distinct QSM differences between the END and nEND cohorts, identified via a priori group-level QSM contrast analysis. Justification for 2 mm Isotropic Voxel Resampling: Resampling to 2 mm isotropic voxels was selected to balance spatial resolution and analytical robustness. Post-Normalization Quality Control: For post-normalization spatial alignment quality control, pre-defined target ROIs were extracted from the 1-mm isotropic MNI-152 standard anatomical template using the Automated Anatomical Labeling (AAL) atlas (no manual ROI delineation was performed on individual subject images), with ROIs selected for their relevance to post-stroke susceptibility and morphometric alterations. Following DARTEL-based normalization of individual images to the MNI template, the template-derived AAL ROIs were applied to normalized subject images, and spatial overlap between individual normalized target ROIs and template ROIs was quantified via the Dice Similarity Coefficient (DSC) in DPABI software (DSC ranges from 0 for no overlap to 1 for perfect congruence). A threshold of > 10% spatial misalignment was defined as a DSC corresponding to < 90% ROI overlap (i.e., >10% of voxels in the individual’s ROI lay outside the template ROI boundary). This criterion was established based on prior post-stroke image normalization quality control literature and our pilot data, which validated that < 90% overlap indicated unacceptably severe spatial distortion in regions of interest; subjects meeting this misalignment threshold were excluded to eliminate bias in subsequent group-level analyses.

2.5.2. Comparisons of QSM values between two groups based on whole-brain voxel-level

Further, voxel-wise comparisons of QSM were conducted between END and nEND within masks based on the QSM distribution pattern, while controlling for age some clinical information such as sex, atrial fibrillation, hypertension, and diabetes (with a TFCE-corrected p < 0.001 and a minimum cluster size of > 400 mm3).

2.5.3. Relationship between regional QSM values and clinical variables

To facilitate cross-variable comparisons and intuitive visualization of associations between clinical features and regional QSM values, all clinical characteristics of interest were standardized to z-scores prior to analysis. Z-score calculations were conducted using SPSS 22.0 (IBM Corp., Armonk, NY, USA), and the standardized z-scores were used for the correlation analysis.

As calculated from the previous step, brain regions which showed significantly different QSM values between the two groups were identified. Later, partial correlation analyses were performed to explore the associations between regional QSM values and clinical variables. To address the risk of false-positive results caused by multiple comparisons across brain regions and clinical factors, False Discovery Rate (FDR) correction (Benjamini-Hochberg procedure) was applied to all correlation test results, with a predefined significance level of α = 0.05. A correlation was considered statistically significant only if the original p-value was ≤ the corresponding critical FDR threshold.

2.5.4. Diagnostic efficacy of the QSM value

A support vector machine (SVM) approach was applied to distinguish END from nEND subjects. Subsequently, a leave-one-out cross-validation (LOOCV) strategy was used to assess the generalization of this SVM classifier. Meanwhile, LOOCV also be used to measure its accuracy, sensitivity, and specificity. We adopted receiver operating characteristic (ROC) to evaluate the diagnostic efficacy of the QSM values for separating the two groups and the diagnostic value was defined as the area under curve (AUC) > 0.7.

3. Results

3.1. Demographic information of all participants

One hundred and twenty-six patients with SAO were admitted, after our quality control, 111 patients were included. Across all patients, 72 (61%) were male, and 46(39%) were females, the mean (SD) age was 65.33(11.92) years. According to if the patients experienced an increment in NIHSS score of 2 points or more in the first 7 days from the onset of AIS, the patients were divided into two groups: END group and nEND group. Of all included patients, 28 (23.7%) patients suffered from END. The results showed there were no significant differences in terms of age, sex, vascular risk factors (hypertension, diabetes, atrial fibrillation), history of smoking and drinking, laboratory tests [i.e., Hcy, LDL, HDL, fibrinogen, D-Dimer]. However, there was a significant difference between the two groups with regard to triglyceride (P <0.05). The detailed clinical characteristics of all subjects are summarized in Table 1.

Table 1.

Demographic and clinical characteristics of all participants.

END (n = 28) nEND (n = 83) P Value
Age, mean ± SD 63.61(10.83) 65.93(12.28) 0.401
Sex, male (%) 19（67.86） 53（63.86） 0.690
AF, n (%) 2（7.14） 1（1.20） 0.177
Hypertensive, n (%) 20（71.43） 53（63.86） 0.598
Diabetes, n (%) 7（25） 31（37.35） 0.313
Smoker, n (%) 16（57.14）） 34（40.96） 0.331
TG, mean ± SD 1.47(0.50) 1.76(1.10) 0.033*
TC, mean ± SD 4.05(0.89) 4.49(0.88) 0.598
LDL-cholesterol, mean ± SD 2.61(0.85) 2.83(0.86) 0.634
HDL-cholesterol, mean ± SD 0.87(0.19) 0.96(0.21) 0.980
Fibrinogen, mean ± SD 3.29(0.61) 2.99(0.55) 0.398
D-Dimer, mean ± SD 0.66(0.81) 0.51(0.47) 0.162
Homocysteine, mean ± SD 18.93(9.90) 17.09(8.40) 0.297
Baseline NIHSS, mean ± SD 3.71(1.51) 2.89(2.12) 0.061
Highest NIHSS at admission 6 11
Open in a new tab

Abbreviations: AF: atrial fibrillation; TG: triglyceride; TC: total cholesterol; LDL: low-density lipoprotein; HDL: high-density lipoprotein; SD: standard deviation. *：P＜0.05

3.2. Differences in QSM values between END and nEND

As shown in Fig. 1, END and nEND showed different susceptibility related distribution patterns measured using QSM. Compared to nEND, END showed higher QSM value in some cortical regions, such as the left/right precentral gyrus, the middle frontal gyrus, the opercular part of the left/right inferior frontal gyrus, the triangular part of left inferior frontal gyrus and postcentral gyrus while lower in the dorsolateral part of left superior frontal.

Fig. 1.

Fig. 1

Open in a new tab

Differences in cortical grey matter susceptibilities measured using QSM between nEND and END. A) Voxel-wise one-sample t-test results of QSM in the nEND group. The colored clusters indicate brain regions where QSM values significantly deviated from the baseline (zero) within the nEND cohort; the color bar represents t-scores. B) Voxel-wise one-sample t-test results of QSM in the END group. Colored clusters represent brain regions with QSM values significantly deviating from the baseline within the END cohort; the color bar indicates t-scores. C) Voxel-wise two-sample t-test results of QSM between the nEND and END groups. The colored clusters highlight brain regions with significant QSM differences between the two groups; the color bar corresponds to the t-scores of the group comparison. D) Bar plot showing QSM values of the differential brain regions (identified in Panel C) in the nEND and END groups. Abbreviations: END, early neurological deterioration; nEND, no early neurological deterioration. PreCG.L,Precentral gyrus; PreCG.R, Precentral gyrus; SFGdor.L, Superior frontal gyrus, dorsolateral; MFG.L, Middle frontal gyrus; PoCG.L, Postcentral gyrus; IFGoperc.L, Inferior frontal gyrus, opercular part; IFGoperc.R, Inferior frontal gyrus, opercular part; IFGtriang.L, Inferior frontal gyrus, triangular part; R, right hemisphere. L, left hemisphere. P＜0.05.

3.3. Relationship between regional QSM values and clinical variables

As data in Table 1 showed, in terms of most previously mentioned clinical variables, no significant differences were observed between END and nEND. However, we found significant differences in terms of cortical susceptibilities via QSM in some specific brain regions between them, suggesting that there may be a relationship between regional QSM values and clinical variables. Therefore, we employed a partial correlation analysis to investigate the potential relationship between these regional QSM values and clinical variables. As shown in Fig. 2, QSM values in the opercular part of the right inferior frontal gyrus (r = 0.266, p = 0.019), right precentral gyrus (r = 0.266, p = 0.019) and triangular part of left inferior frontal gyrus (r = 0.300, p = 0.008) positively correlated with the Hcy value. Meanwhile, QSM values in the triangular part of left inferior frontal gyrus negatively were correlated with TC value (r = -0.257, p = 0.008), and LDL value (r = -0.247, p = 0.011).

Fig. 2.

Fig. 2

Open in a new tab

Correlations between z-scores of clinical characteristics and regional QSM values. Panels A–E display marginal scatter plots illustrating linear associations between z-scores of clinical characteristics (vertical axis) and QSM values in different brain regions (horizontal axis). Each panel integrates a scatter plot (core) with marginal distribution plots (top/right for the y-axis variable, bottom for the x-axis variable) to display the distribution of individual variables alongside their bivariate association. Blue points represent individual study samples; the solid blue line denotes the Pearson linear regression fit, and the shaded gray area indicates the 95% confidence interval around the regression line. The Pearson correlation coefficient (r) and corresponding p-value for each association are reported in the top-right corner of each panel. Abbreviations: IFGoperc.R, Inferior frontal gyrus, opercular part; PreCG.R, Precentral gyrus; IFGtriang.L, Inferior frontal gyrus, triangular part; R, right hemisphere. L, left hemisphere.

3.4. Distinguishing END from nEND subjects using QSM value

Inspired by the significantly different magnetic susceptibility of the brain cortical regions between END and nEND groups (Table 2 and Fig. 1), we further explore the potential of using QSM-value to differentiate END patients from nEND patients. As shown in Fig. 3, the SVM classifier's ROC curve displayed the ability to differentiate END from nEND on an individual subject basis, with an AUC of 70%, sensitivity of 60%, and specificity of 78%.

Table 2.

Comparisons of cortical grey matter susceptibilities measured using quantitative susceptibility mapping between END and nEND subjects.

Brain regions L/R MNI
 Tvalues Cluster size (mm3)
x y z
END > nEND
Precentral gyrus L -58 6 38 4.1983 621
Precentral gyrus R 60 8 16 4.075 106
Middle frontal gyrus L -46 22 36 3.7953 149
Postcentral gyrus L -60–6 26 3.8567 300
Inferior frontal gyrus, opercular part R 56 10 22 4.4249 172
Inferior frontal gyrus, opercular part L -54 8 14 3.4701 266
Inferior frontal gyrus, triangular part L -38 42 10 4.0057 181
END < nEND
Superior frontal gyrus, dorsolateral L -20 52 24 -4.6524 154
Open in a new tab

Abbreviations: END, early neurological deterioration; nEND, no early neurological deterioration; MNI, montreal neurological institute; L, left hemisphere; R, right hemisphere. Regions with significant voxel numbers≥ 100 with PTFCE correction < 0.005 were presented and defined from the regions of interest (ROIs) in AAL3v1 atlas.

Fig. 3.

Fig. 3

Open in a new tab

Classification of individuals as END and nEND by QSM-based“classifier”. ROC curve shows the classification power in QSM-based “classifier” of END from nEND. Abbreviations: END, early neurological deterioration; nEND, no early neurological deterioration.AUC, the area under the ROC curve; ACC, accuracy; Opt, optimum; ROC, receiver operating characteristic.

4. Discussion

The present study demonstrated the distribution patterns of the brain cortical magnetic susceptibility in END and nEND in SAO patients. The different susceptibility related distributed patterns acquired from QSM could accurately discriminate END from nEND in the early stages of SAO.

QSM has been proven useful for the evaluation of AIS in various ways (Haacke et al., 2010, Zhang et al., 2019). Recently, some studies have shown the potential value of magnetic susceptibility of thrombus in the occluded arteries. A study by Chen et al. showed that the thrombus susceptibility value could be used to discriminate the cardioembolic stroke from other stroke etiological subtypes in acute middle cerebral artery or basilar artery occluded patients (Chen et al., 2022b, Chen et al., 2022c). However, the application of QSM in the SAO stroke has not been explored at all. Our study firstly declared the characteristics of magnetic susceptibility in the cerebral cortex in SAO patients. We found the different magnetic susceptibility related patterns in END and nNED patients. Further, we found the different distribution patterns showed a power to discriminate END from nEND patients. We have not found similar results from other studies. However, the causal relationship between QSM findings and END remains unclear. The observed differences in cortical susceptibility could be either a cause or consequence of neurological deterioration.

Iron is a main contributor to the magnetic susceptibility signal (Langkammer et al., 2012, Ravanfar et al., 2021). Iron is a kind of paramagnetic material which means it has positive susceptibility while diamagnetic materials have negative susceptibility. Iron is necessary for many metabolic processes such as oxygen transport, DNA synthesis, myelin synthesis, and mitochondrial respiration (Gutteridge, 1992, Koeppen, 1995, Todorich et al., 2009). The effect of iron deposition on the ischemic stroke was getting more and more attention which may be stimulated by the development of novel MRI techniques. Numerous studies have reported that ischemic stroke could lead to the iron accumulation (van Etten et al., 2015, Wiegertjes et al., 2021). Yang et al. reported the susceptibilities of infarct regions in AIS patients were significantly higher than those in healthy controls via QSM (Yang et al., 2024). Another study by Zhou et al. also revealed increased iron concentrations in the GM nuclei of patients with AIS (Zhou et al., 2024). However, excess iron content was related to oxidative stress and aggravate neurological function (Carbonell and Rama, 2007, Connor et al., 1992, Ghassaban et al., 2019). The mechanism for iron accumulation during ischemic stroke are mainly included the breach of blood–brain barrier, ferritinophagy, iron export protein dysfunction (DeGregorio-Rocasolano et al., 2018, Li et al., 2021, Tuo et al., 2017). Previous studies have shown that QSM is a sensitive technique that can assess the magnetic properties of cerebral iron in vivo (Madden and Merenstein, 2023). In our study, we found QSM values of most of the cortical regions mentioned above of the END patients were significantly higher than nEND patients, which suggested that there were elevated iron concentrations in GM of END patients. Regrettably, we have found that these cortical regions with higher QSM values in END group mostly were not adjacent to the infarctions (such as the base of the pons or the posterior limb of the internal capsule). However, most of these cortical regions (such as the precentral gyrus, the middle frontal gyrus, the inferior frontal gyrus, the inferior frontal gyrus and postcentral gyrus) are key areas which control language, sensation and movement. Therefore, we propose two possible explanations for the iron deposition in the key cortical regions potentially leading to the progression of SAO: one is the iron deposition in the key regions triggers oxidative stress, directly damaging the functions of the cortical neurons in these areas and ultimately exacerbating neurological functions. Two is the changed distributed iron deposition may have affected the function of some connecting fiber tracts between these vital language or motor function cortical regions and the infarcted areas, ultimately leading to the progression of cerebral infarction. As far as we know, no previous studies have put forward this view. However, the exact molecular mechanism requires further prospective studies for confirmation. In conclusion, given our results, we supposed excess iron content could lead to worse neurofunction in ischemic stroke, which was consistent to many other studies (Armengou and Davalos, 2002, Carbonell and Rama, 2007). It is well-known that ischemic stroke is strongly associated with various risk factors such as hypertension, glucose metabolism disorder, high homocysteine, lipid metabolism disorder and atrial fibrillation (Migdady et al., 2021, Shi et al., 2021). Regretfully, our study still cannot essentially explain whether the END in SAO patients leads to iron deposition or whether iron deposition leads to END. The elucidation of causality needs a larger clinical cohort in the future.

On the other hand, our study showed there was no significant difference in the relevant clinical variables between END and nEND in this study except for TG. Many evidences in prospective cohort studies favor a positive association between TG levels and stroke (Labreuche et al., 2009). However, there are relatively few studies on the differences in the role of lipid metabolism in different types of ischemic stroke. Our study revealed that nEND exhibited marginally higher TG levels compared to END, a finding not previously documented in existing literature and seemingly inconsistent with conventional observations. Maybe the different medication history such as the use of statin is a possible reason to explain this result. It is pity that we did not collect data on the use of statins before the onset of the SAO. Furthermore, some studies showed lipids critically modulated the immunity and inflammation in ischemic stroke or some neurodegenerative diseases (Danik et al., 1999, Vassiliou and Farias-Pereira, 2023, Yan and Horng, 2020). However, the specific role of TG in SAO remains unclear at present. Further validation with a larger patient cohort is warranted to substantiate the reliability of these conclusions.

Besides，we found QSM values in specific brain regions were correlated with some clinical variables such as Hcy, TC and LDL. Rare studies have reported the relationships between brain susceptibilities and clinical variables. A previous research showed iron played an important role in cell-mediated LDL-oxidation (Yuan and Brunk, 1998), this result may explain the negative relationship between QSM and LDL in our study. In addition, we also speculated that dyslipidemia may affect cerebral blood flow perfusion through vascular lesions, leading to alterations in iron deposition. Joseph et al. reported that an elevated amount of non-protein-bound iron increases circulating Hcy (Baggott and Tamura, 2015), which helped us to explain our result about the positive relationship between Hcy and QSM values. These findings may contribute to elucidating the underlying mechanisms of QSM alterations in patients with SAO. The exact mechanism between the QSM and clinical variables needs more studies in the future.

To our knowledge, this is the first study on the predictive ability of QSM for END in patients with SAO. We hope our study could help to establish a predictive model which can assist clinicians in identifying SAO patients at an early stage who are likely to experience END, so that clinicians can provide these patients with more personalized medical strategies such as potent antiplatelet therapy or intensified statin treatment, ultimately improving their clinical prognosis. Of course, this requires further confirmation through larger-scale randomized controlled studies.

This study has limitations. Firstly, this was a single-center study with a limited number of patients included from 24-hour enrollment window, which may lead to potential selection bias. Secondly, besides the iron contents, many other components also affect susceptibility in the brain cortex in some ways, such as copper deposition, calcification, the concentration of paramagnetic deoxyhemoglobins and so on (Pauling, 1977, Skowrońska et al., 2013, Wang et al., 2017). Meanwhile, the anisotropic acquisition is not advisable for QSM which is also an inevitable problem in this study. Thirdly, we did not take the infarct location into consideration for some reasons in this study. Fourthly, this study could not answer the casual relationship between END and QSM susceptibility and to address these issues, further studies with a larger clinical cohort and comprehensive clinical records are needed. Fifthly, we still lack the histopathological validation for QSM-measured iron. Sixthly, we acknowledge that spatial normalization may still introduce minor biases in regions with anatomical distortion. However, the combination of DARTEL-based nonlinear registration, lesion masking, and rigorous QC have been shown in previous stroke imaging studies to effectively mitigate these issues. Future studies could employ subject-specific atlases or ROI-based analyses on native space images to further reduce normalization-related uncertainties. Lastly, in the SVM part, the overfitting problems should not be overlooked when we use machine learning on small simple sizes. We are going to plan to establish an external dataset by including more patients to solve the problem in the future work.

5. Conclusions

In conclusion, using QSM, we found different cortical susceptibility distributed patterns between END and nEND groups. Additionally, Hcy, LDL and TC were correlated with QSM values in some specific brain regions. The SVM classifier’s analysis showed a high power for QSM diagnostic efficacy to distinguish END from nEND. Therefore, QSM may provide a valuable method to predict END in the early stage of SAO.

CRediT authorship contribution statement

Jinan Chen: Writing – review & editing, Validation, Resources, Methodology, Funding acquisition, Conceptualization. Xuelian Tang: Writing – review & editing, Writing – original draft, Validation, Software, Investigation. Yusheng Yu: Supervision, Project administration, Data curation, Conceptualization. Yanhua Yu: Writing – original draft, Visualization, Software, Formal analysis, Data curation.

Consent for publication

Informed consent was obtained from all subjects involved in the study. Written informed consent was obtained from the patients to publish this paper.

Ethics approval and consent to participate

All the participants provided written informed consent. This study was in accordance with the Declaration of Helsinki and conducted after approval by medical ethics committee of nanjing jiangning hospital (2023–03–074-K01).

Fundings

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors declare no conflict of interest.

Acknowledgments

We thank all patients for granting permission to publish this information.

Footnotes

Appendix A

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ibneur.2026.03.013.

Contributor Information

Yusheng Yu, Email: yayiba@163.com.

Jinan Chen, Email: cjn1523@163.com.

Appendix A. Supplementary material

Supplementary material

mmc1.jpg (226.8KB, jpg)

Data availability

All data supporting the findings of this study are included in this published article and its supplementary information files.

References

  1. Adams H.P., Jr., Bendixen B.H., Kappelle L.J., Biller J., Love B.B., Gordon D.L., Marsh E.E., 3rd Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993;24:35–41. doi: 10.1161/01.str.24.1.35. [DOI] [PubMed] [Google Scholar]
  2. Armengou A., Davalos A. A review of the state of research into the role of iron in stroke. J. Nutr. Health Aging. 2002;6:207–208. [PubMed] [Google Scholar]
  3. Baggott J.E., Tamura T. Homocysteine, iron and cardiovascular disease: a hypothesis. Nutrients. 2015;7:1108–1118. doi: 10.3390/nu7021108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Carbonell T., Rama R. Iron, oxidative stress and early neurological deterioration in ischemic stroke. Curr. Med Chem. 2007;14:857–874. doi: 10.2174/092986707780363014. [DOI] [PubMed] [Google Scholar]
  5. Chan K.S., Marques J.P. SEPIA-Susceptibility mapping pipeline tool for phase images. Neuroimage. 2021;227 doi: 10.1016/j.neuroimage.2020.117611. [DOI] [PubMed] [Google Scholar]
  6. Chen J., Chen R., Xue C., Qi W., Hu G., Xu W., Chen S., Rao J., Zhang F., Zhang X. Hippocampal-subregion mechanisms of repetitive transcranial magnetic stimulation causally associated with amelioration of episodic memory in amnestic mild cognitive impairment. J. Alzheimers Dis. 2022;85:1329–1342. doi: 10.3233/jad-210661. [DOI] [PubMed] [Google Scholar]
  7. Chen J., Zhang Z., Nie X., Xu Y., Liu C., Zhao X., Wang Y. Basilar artery thrombus magnetic susceptibility for cardioembolic stroke identification. Quant. Imaging Med Surg. 2022;12:1579–1584. doi: 10.21037/qims-21-547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen J., Zhang Z., Nie X., Xu Y., Liu C., Zhao X., Wang Y. Predictive value of thrombus susceptibility for cardioembolic stroke by quantitative susceptibility mapping. Quant. Imaging Med Surg. 2022;12:550–557. doi: 10.21037/qims-21-235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Connor J.R., Snyder B.S., Beard J.L., Fine R.E., Mufson E.J. Regional distribution of iron and iron-regulatory proteins in the brain in aging and Alzheimer's disease. J. Neurosci. Res. 1992;31:327–335. doi: 10.1002/jnr.490310214. [DOI] [PubMed] [Google Scholar]
  10. Danik M., Champagne D., Petit-Turcotte C., Beffert U., Poirier J. Brain lipoprotein metabolism and its relation to neurodegenerative disease. Crit. Rev. Neurobiol. 1999;13:357–407. doi: 10.1615/critrevneurobiol.v13.i4.20. [DOI] [PubMed] [Google Scholar]
  11. DeGregorio-Rocasolano N., Martí-Sistac O., Gasull T. Deciphering the Iron Side of Stroke: Neurodegeneration at the Crossroads Between Iron Dyshomeostasis, Excitotoxicity, and Ferroptosis. Front Neurosci. 2019;13:85. doi: 10.3389/fnins.2019.00085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. DeGregorio-Rocasolano N., Martí-Sistac O., Ponce J., Castelló-Ruiz M., Millán M., Guirao V., García-Yébenes I., Salom J.B., Ramos-Cabrer P., Alborch E., Lizasoain I., Castillo J., Dávalos A., Gasull T. Iron-loaded transferrin (Tf) is detrimental whereas iron-free Tf confers protection against brain ischemia by modifying blood Tf saturation and subsequent neuronal damage. Redox Biol. 2018;15:143–158. doi: 10.1016/j.redox.2017.11.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Deistung A., Schweser F., Reichenbach J.R. Overview of quantitative susceptibility mapping. NMR Biomed. 2017;30 doi: 10.1002/nbm.3569. [DOI] [PubMed] [Google Scholar]
  14. Fushimi Y., Nakajima S., Sakata A., Okuchi S., Otani S., Nakamoto Y. Value of Quantitative Susceptibility Mapping in Clinical Neuroradiology. J. Magn. Reson Imaging. 2024;59:1914–1929. doi: 10.1002/jmri.29010. [DOI] [PubMed] [Google Scholar]
  15. Ghassaban K., He N., Sethi S.K., Huang P., Chen S., Yan F., Haacke E.M. Regional High Iron in the Substantia Nigra Differentiates Parkinson's Disease Patients From Healthy Controls. Front Aging Neurosci. 2019;11:106. doi: 10.3389/fnagi.2019.00106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Guo Y.K., Zhang M., Wang Q., Chen Z.Y., Dong G.Z., Yao T., Xu Y. Influence of neuroimaging markers of chronic cerebral small vessel disease on prognosis of patients with cardioembolic stroke] Zhonghua Yi Xue Za Zhi. 2021;101:2333–2336. doi: 10.3760/cma.j.cn112137-20201216-03370. [DOI] [PubMed] [Google Scholar]
  17. Gutteridge J.M. Iron and oxygen radicals in brain. Ann. Neurol. 1992;32:S16–S21. doi: 10.1002/ana.410320705. (Suppl) [DOI] [PubMed] [Google Scholar]
  18. Haacke E.M., Tang J., Neelavalli J., Cheng Y.C. Susceptibility mapping as a means to visualize veins and quantify oxygen saturation. J. Magn. Reson Imaging. 2010;32:663–676. doi: 10.1002/jmri.22276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Karsa A., Shmueli K. SEGUE: A Speedy rEgion-Growing Algorithm for Unwrapping Estimated Phase. IEEE Trans. Med Imaging. 2019;38:1347–1357. doi: 10.1109/tmi.2018.2884093. [DOI] [PubMed] [Google Scholar]
  20. Kataike V.M., Desmond P.M., Steward C., Mitchell P.J., Davey C., Yassi N., Bivard A., Parsons M.W., Campbell B.C.V., Ng F., Venkatraman V. Iron changes within infarct tissue in ischemic stroke patients after successful reperfusion quantified using QSM. Neuroradiology. 2024;66:2233–2242. doi: 10.1007/s00234-024-03444-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Koeppen A.H. The history of iron in the brain. J. Neurol. Sci. 1995;134:1–9. doi: 10.1016/0022-510x(95)00202-d. (Suppl) [DOI] [PubMed] [Google Scholar]
  22. Labreuche J., Touboul P.J., Amarenco P. Plasma triglyceride levels and risk of stroke and carotid atherosclerosis: a systematic review of the epidemiological studies. Atherosclerosis. 2009;203:331–345. doi: 10.1016/j.atherosclerosis.2008.08.040. [DOI] [PubMed] [Google Scholar]
  23. Langkammer C., Schweser F., Krebs N., Deistung A., Goessler W., Scheurer E., Sommer K., Reishofer G., Yen K., Fazekas F., Ropele S., Reichenbach J.R. Quantitative susceptibility mapping (QSM) as a means to measure brain iron? A post mortem validation study. Neuroimage. 2012;62:1593–1599. doi: 10.1016/j.neuroimage.2012.05.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li C., Sun G., Chen B., Xu L., Ye Y., He J., Bao Z., Zhao P., Miao Z., Zhao L., Hu J., You Y., Liu N., Chao H., Ji J. Nuclear receptor coactivator 4-mediated ferritinophagy contributes to cerebral ischemia-induced ferroptosis in ischemic stroke. Pharm. Res. 2021;174 doi: 10.1016/j.phrs.2021.105933. [DOI] [PubMed] [Google Scholar]
  25. Liu C., Li W., Tong K.A., Yeom K.W., Kuzminski S. Susceptibility-weighted imaging and quantitative susceptibility mapping in the brain. J. Magn. Reson Imaging. 2015;42:23–41. doi: 10.1002/jmri.24768. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Liu H., Liu K., Zhang K., Zong C., Yang H., Li Y., Li S., Wang X., Zhao J., Xia Z., Song B., Xu X., Gao Y. Early neurological deterioration in patients with acute ischemic stroke: a prospective multicenter cohort study. Ther. Adv. Neurol. Disord. 2023;16 doi: 10.1177/17562864221147743. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Madden D.J., Merenstein J.L. Quantitative susceptibility mapping of brain iron in healthy aging and cognition. Neuroimage. 2023;282 doi: 10.1016/j.neuroimage.2023.120401. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Markus H.S., de Leeuw F.E. Cerebral small vessel disease: Recent advances and future directions. Int J. Stroke. 2023;18:4–14. doi: 10.1177/17474930221144911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Migdady I., Russman A., Buletko A.B. Atrial Fibrillation and Ischemic Stroke: A Clinical Review. Semin Neurol. 2021;41:348–364. doi: 10.1055/s-0041-1726332. [DOI] [PubMed] [Google Scholar]
  30. Munoz D.G. Small vessel disease: neuropathology. Int Psychogeriatr. 2003;15(1):67–69. doi: 10.1017/s1041610203008986. [DOI] [PubMed] [Google Scholar]
  31. Nemoto K. [Understanding Voxel-Based Morphometry] Brain Nerve. 2017;69:505–511. doi: 10.11477/mf.1416200776. [DOI] [PubMed] [Google Scholar]
  32. Pauling L. Vol. 74. 1977. Magnetic properties and structure of oxyhemoglobin; pp. 2612–2613. (Proc Natl Acad Sci U S A). [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Probst J., Rohner M., Zahn M., Piccirelli M., Pangalu A., Luft A., Deistung A., Klohs J., Wegener S. Quantitative susceptibility mapping in ischemic stroke patients after successful recanalization. Sci. Rep. 2021;11 doi: 10.1038/s41598-021-95265-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ravanfar P., Loi S.M., Syeda W.T., Van Rheenen T.E., Bush A.I., Desmond P., Cropley V.L., Lane D.J.R., Opazo C.M., Moffat B.A., Velakoulis D., Pantelis C. Systematic Review: Quantitative Susceptibility Mapping (QSM) of Brain Iron Profile in Neurodegenerative Diseases. Front Neurosci. 2021;15 doi: 10.3389/fnins.2021.618435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Robinson S.D., Bredies K., Khabipova D., Dymerska B., Marques J.P., Schweser F. An illustrated comparison of processing methods for MR phase imaging and QSM: combining array coil signals and phase unwrapping. NMR Biomed. 2017;30 doi: 10.1002/nbm.3601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ruetten P.P.R., Gillard J.H., Graves M.J. Introduction to Quantitative Susceptibility Mapping and Susceptibility Weighted Imaging. Br. J. Radio. 2019;92 doi: 10.1259/bjr.20181016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sacco S., Marini C., Totaro R., Russo T., Cerone D., Carolei A. A population-based study of the incidence and prognosis of lacunar stroke. Neurology. 2006;66:1335–1338. doi: 10.1212/01.wnl.0000210457.89798.0e. [DOI] [PubMed] [Google Scholar]
  38. Selim M.H., Ratan R.R. The role of iron neurotoxicity in ischemic stroke. Ageing Res Rev. 2004;3:345–353. doi: 10.1016/j.arr.2004.04.001. [DOI] [PubMed] [Google Scholar]
  39. Sharma A., Pandit A.K., Mishra B., Srivastava M.V.P., Srivastava A.K., Vishnu V.Y., Singh R.K. Early neurological deterioration in acute ischemic stroke. Ir. J. Med Sci. 2024;193:949–955. doi: 10.1007/s11845-023-03485-5. [DOI] [PubMed] [Google Scholar]
  40. Shi Y., Guo L., Chen Y., Xie Q., Yan Z., Liu Y., Kang J., Li S. Risk factors for ischemic stroke: differences between cerebral small vessel and large artery atherosclerosis aetiologies. Folia Neuropathol. 2021;59:378–385. doi: 10.5114/fn.2021.112007. [DOI] [PubMed] [Google Scholar]
  41. Siegler J.E., Boehme A.K., Kumar A.D., Gillette M.A., Albright K.C., Martin-Schild S. What change in the National Institutes of Health Stroke Scale should define neurologic deterioration in acute ischemic stroke? J. Stroke Cereb. Dis. 2013;22:675–682. doi: 10.1016/j.jstrokecerebrovasdis.2012.04.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Skowrońska M., Litwin T., Dzieżyc K., Wierzchowska A., Członkowska A. Does brain degeneration in Wilson disease involve not only copper but also iron accumulation? Neurol. Neurochir. Pol. 2013;47(6):542. doi: 10.5114/ninp.2013.39071. [DOI] [PubMed] [Google Scholar]
  43. Stankiewicz J., Panter S.S., Neema M., Arora A., Batt C.E., Bakshi R. Iron in chronic brain disorders: imaging and neurotherapeutic implications. Neurotherapeutics. 2007;4:371–386. doi: 10.1016/j.nurt.2007.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Steinke W., Ley S.C. Lacunar stroke is the major cause of progressive motor deficits. Stroke. 2002;33:1510–1516. doi: 10.1161/01.str.0000016326.78014.fe. [DOI] [PubMed] [Google Scholar]
  45. Todorich B., Pasquini J.M., Garcia C.I., Paez P.M., Connor J.R. Oligodendrocytes and myelination: the role of iron. Glia. 2009;57:467–478. doi: 10.1002/glia.20784. [DOI] [PubMed] [Google Scholar]
  46. Tuo Q.Z., Lei P., Jackman K.A., Li X.L., Xiong H., Li X.L., Liuyang Z.Y., Roisman L., Zhang S.T., Ayton S., Wang Q., Crouch P.J., Ganio K., Wang X.C., Pei L., Adlard P.A., Lu Y.M., Cappai R., Wang J.Z., Liu R., Bush A.I. Tau-mediated iron export prevents ferroptotic damage after ischemic stroke. Mol. Psychiatry. 2017;22:1520–1530. doi: 10.1038/mp.2017.171. [DOI] [PubMed] [Google Scholar]
  47. van Etten E.S., van der Grond J., Dumas E.M., van den Bogaard S.J., van Buchem M.A., Wermer M.J. MRI Susceptibility Changes Suggestive of Iron Deposition in the Thalamus after Ischemic Stroke. Cereb. Dis. 2015;40:67–72. doi: 10.1159/000433560. [DOI] [PubMed] [Google Scholar]
  48. Vassiliou E., Farias-Pereira R. Impact of Lipid Metabolism on Macrophage Polarization: Implications for Inflammation and Tumor Immunity. Int J. Mol. Sci. 2023;24 doi: 10.3390/ijms241512032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Wang Y., Li H., Pan Y., Wang M., Liao X., Yang Y., Chen W., Meng X., Wang Y., Wang Y. Cerebral small vessel disease was associated with the prognosis in ischemic stroke with atrial fibrillation. CNS Neurosci. Ther. 2024;30 doi: 10.1111/cns.70052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang Y., Spincemaille P., Liu Z., Dimov A., Deh K., Li J., Zhang Y., Yao Y., Gillen K.M., Wilman A.H., Gupta A., Tsiouris A.J., Kovanlikaya I., Chiang G.C., Weinsaft J.W., Tanenbaum L., Chen W., Zhu W., Chang S., Lou M., Kopell B.H., Kaplitt M.G., Devos D., Hirai T., Huang X., Korogi Y., Shtilbans A., Jahng G.H., Pelletier D., Gauthier S.A., Pitt D., Bush A.I., Brittenham G.M., Prince M.R. Clinical quantitative susceptibility mapping (QSM): Biometal imaging and its emerging roles in patient care. J. Magn. Reson Imaging. 2017;46:951–971. doi: 10.1002/jmri.25693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wang Y., Ye C., Pan R., Tang B., Li C., Liu J., Tao W., Zhang X., Yang T., Yan Y., Jiang S., Lui S., Wu B. Cognitive implications and associated transcriptomic signatures of distinct regional iron depositions in cerebral small vessel disease. Alzheimers Dement. 2025;21 doi: 10.1002/alz.70196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wardlaw J.M., Smith E.E., Biessels G.J., Cordonnier C., Fazekas F., Frayne R., Lindley R.I., O'Brien J.T., Barkhof F., Benavente O.R., Black S.E., Brayne C., Breteler M., Chabriat H., Decarli C., de Leeuw F.E., Doubal F., Duering M., Fox N.C., Greenberg S., Hachinski V., Kilimann I., Mok V., Oostenbrugge R., Pantoni L., Speck O., Stephan B.C., Teipel S., Viswanathan A., Werring D., Chen C., Smith C., van Buchem M., Norrving B., Gorelick P.B., Dichgans M. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol. 2013;12:822–838. doi: 10.1016/s1474-4422(13)70124-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wei H., Dibb R., Decker K., Wang N., Zhang Y., Zong X., Lin W., Nissman D.B., Liu C. Investigating magnetic susceptibility of human knee joint at 7 Tesla. Magn. Reson Med. 2017;78:1933–1943. doi: 10.1002/mrm.26596. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wiegertjes K., Chan K.S., Telgte A.T., Gesierich B., Norris D.G., Klijn C.J., Duering M., Tuladhar A.M., Marques J.P., Leeuw F.E. Assessing cortical cerebral microinfarcts on iron-sensitive MRI in cerebral small vessel disease. J. Cereb. Blood Flow. Metab. 2021;41:3391–3399. doi: 10.1177/0271678x211039609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yan J., Horng T. Lipid Metabolism in Regulation of Macrophage Functions. Trends Cell Biol. 2020;30:979–989. doi: 10.1016/j.tcb.2020.09.006. [DOI] [PubMed] [Google Scholar]
  56. Yang J., Lv M., Han L., Li Y., Liu Y., Guo H., Feng H., Wu Y., Zhong J. Evaluation of brain iron deposition in different cerebral arteries of acute ischaemic stroke patients using quantitative susceptibility mapping. Clin. Radio. 2024;79:e592–e598. doi: 10.1016/j.crad.2024.01.007. [DOI] [PubMed] [Google Scholar]
  57. Yuan X.M., Brunk U.T. Iron and LDL-oxidation in atherogenesis. Apmis. 1998;106:825–842. doi: 10.1111/j.1699-0463.1998.tb00229.x. [DOI] [PubMed] [Google Scholar]
  58. Zhang S., Liu Z., Nguyen T.D., Yao Y., Gillen K.M., Spincemaille P., Kovanlikaya I., Gupta A., Wang Y. Clinical feasibility of brain quantitative susceptibility mapping. Magn. Reson Imaging. 2019;60:44–51. doi: 10.1016/j.mri.2019.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhou D., Liu T., Spincemaille P., Wang Y. Background field removal by solving the Laplacian boundary value problem. NMR Biomed. 2014;27:312–319. doi: 10.1002/nbm.3064. [DOI] [PubMed] [Google Scholar]
  60. Zhou L., Yang J., Zhang W., Han L., Zhou S., Zheng C., Feng H., Zhong J. Evaluating iron deposition in gray matter nuclei of patients with acute ischemic stroke using quantitative susceptibility mapping. Front Neurol. 2024;15 doi: 10.3389/fneur.2024.1518911. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material

mmc1.jpg (226.8KB, jpg)

Data Availability Statement

All data supporting the findings of this study are included in this published article and its supplementary information files.

Articles from IBRO Neuroscience Reports are provided here courtesy of Elsevier

RESOURCES