Multimodal MRI Research Articles

Assessment of hypoxia and oxidative-related changes in a lung-derived brain metastasis model by [64Cu][Cu(ATSM)] PET and proteomic studies

BackgroundBrain metastases (BM) are the most frequent malignant brain tumors. The aim of this study was to characterize the tumor microenvironment (TME) of BM and particularly hypoxia and redox state, known to play a role in tumor growth and treatment resistance with multimodal PET and MRI imaging, immunohistochemical and proteomic approaches in a human lung cancer (H2030-BrM3)-derived BM model in rats.ResultsFirst, in vitro studies confirmed that H2030-BrM3 cells respond to hypoxia with increasing expression of HIF-1, HIF-2 and their target genes. Proteomic analyses revealed, among expression changes, proteins associated with metabolism, oxidative stress, metal response and hypoxia signaling in particular in cortical BM. [64Cu][Cu(ATSM)] PET revealed a significant uptake by cortical BM (p < 0.01), while no uptake is observed in striatal BM 23 days after tumor implantation. Pimonidazole, HIF-1α, HIF-2α, CA-IX as well as GFAP, CTR1 and DMT1 immunostainings are positive in both BM.ConclusionOverall, [64Cu][Cu(ATSM)] imaging and proteomic results showed the presence of hypoxia and protein expression changes linked to hypoxia and oxidative stress in BM, which are more pronounced in cortical BM compared to striatal BM. Moreover, it emphasized the interest of [64Cu][Cu(ATSM)] PET to characterize TME of BM and depict inter-metastasis heterogeneity that could be useful to guide treatments.

Prediction of Fuhrman grade of renal clear cell carcinoma by multimodal MRI radiomics: a retrospective study

To explore the value of multimodal magnetic resonance imaging (MRI) radiomics combined with traditional radiologist-defined semantic characteristics and conventional (cMRI) and functional MRI (fMRI) texture features in predicting Fuhrman grade of clear cell renal cell carcinoma (ccRCC). The data of 89 patients with histopathologically proven ccRCC (low-grade, 54; high-grade, 35) were collected. Texture features were extracted from cMRI (T1- and T2-weighted imaging) and fMRI (Dixon-MRI; blood-oxygen-level dependent [BOLD]-MRI; and susceptibility-weighted imaging [SWI]) images, and the traditional characteristics (TC) were evaluated. Logistic regression analysis was performed to develop models based on TC, cMRI, and fMRI texture features for grading. Receiver operating characteristic (ROC) curve analysis and leave-group-out cross-validation (LGOCV) were performed to test the reliability of combined models. Two T2-weighted imaging-based, two Dixon_W-based, one Dixon_F-based, one BOLD-based, and three SWI-based texture features, and three TC were extracted for feature selection. TC, cMRI, fMRI, cMRI+fMRI, cMRI+TC, fMRI+TC, and cMRI+fMRI+TC models were constructed. The AUC of the cMRI+fMRI+TC model for differentiating high- from low-grade ccRCC was 0.74, with 81.42% accuracy, 75.93% sensitivity, and 91.43% specificity. The fMRI+TC model exhibited a performance similar to that of the cMRI+fMRI+TC model (p>0.05). The areas under the curve (AUCs) of the fMRI+TC and cMRI+fMRI+TC models were significantly higher than those of the other five models (all p<0.05). For the cMRI+fMRI+TC model, the mean accuracy was 85.40% after 100 LGOCV for the test sets. Multimodal MRI radiomics combined with TC, cMRI, and fMRI texture features may be a reliable quantitative approach for differentiating high-grade ccRCC from low-grade ccRCC.

FastSurfer-HypVINN: Automated sub-segmentation of the hypothalamus and adjacent structures on high-resolutional brain MRI

Abstract The hypothalamus plays a crucial role in the regulation of a broad range of physiological, behavioral, and cognitive functions. However, despite its importance, only a few small-scale neuroimaging studies have investigated its substructures, likely due to the lack of fully automated segmentation tools to address scalability and reproducibility issues of manual segmentation. While the only previous attempt to automatically sub-segment the hypothalamus with a neural network showed promise for 1.0 mm isotropic T1-weighted (T1w) magnetic resonance imaging (MRI), there is a need for an automated tool to sub-segment also high-resolutional (HiRes) MR scans, as they are becoming widely available, and include structural detail also from multi-modal MRI. We, therefore, introduce a novel, fast, and fully automated deep-learning method named HypVINN for sub-segmentation of the hypothalamus and adjacent structures on 0.8 mm isotropic T1w and T2w brain MR images that is robust to missing modalities. We extensively validate our model with respect to segmentation accuracy, generalizability, in-session test-retest reliability, and sensitivity to replicate hypothalamic volume effects (e.g., sex differences). The proposed method exhibits high segmentation performance both for standalone T1w images as well as for T1w/T2w image pairs. Even with the additional capability to accept flexible inputs, our model matches or exceeds the performance of state-of-the-art methods with fixed inputs. We, further, demonstrate the generalizability of our method in experiments with 1.0 mm MR scans from both the Rhineland Study and the UK Biobank—an independent dataset never encountered during training with different acquisition parameters and demographics. Finally, HypVINN can perform the segmentation in less than a minute (graphical processing unit [GPU]) and will be available in the open source FastSurfer neuroimaging software suite, offering a validated, efficient, and scalable solution for evaluating imaging-derived phenotypes of the hypothalamus.

Multi-modal MRI of hippocampal morphometry and connectivity after pediatric severe TBI

This investigation explores memory performance using the California Verbal Learning Test in relation to morphometric and connectivity measures of the memory network in severe traumatic brain injury. Twenty-two adolescents with severe traumatic brain injury were recruited for multimodal MRI scanning 1–2 years post-injury at 13 participating sites. Analyses included hippocampal volume derived from anatomical T1-weighted imaging, fornix white matter microstructure from diffusion tensor imaging, and hippocampal resting-state functional magnetic resonance imaging connectivity as well as diffusion-based structural connectivity. A typically developing control cohort of forty-nine age-matched children also underwent scanning and neurocognitive assessment. Results showed hippocampus volume was decreased in traumatic brain injury with respect to controls. Further, hippocampal volume loss was associated with worse performance on memory and learning in traumatic brain injury subjects. Similarly, hippocampal fornix fractional anisotropy was reduced in traumatic brain injury with respect to controls, while decreased fractional anisotropy in the hippocampal fornix also was associated with worse performance on memory and learning in traumatic brain injury subjects. Additionally, reduced structural connectivity of left hippocampus to thalamus and calcarine sulcus was associated with memory and learning in traumatic brain injury subjects. Functional connectivity in the left hippocampal network was also associated with memory and learning in traumatic brain injury subjects. These regional findings from a multi-modal neuroimaging approach should not only be useful for gaining valuable insight into traumatic brain injury induced memory and learning disfunction, but may also be informative for monitoring injury progression, recovery, and for developing rehabilitation as well as therapy strategies.

NIMG-30. DEEP LEARNING SEGMENTATION OF INFILTRATIVE AND ENHANCING CELLULAR TUMOR IN POST-TREATMENT GLIOBLASTOMA PATIENTS WITH MULTI-SHELL DIFFUSION MRI

Abstract Multimodal MRI is used for the evaluation of tumor burden in glioblastoma (GBM) patients after surgical resection, radiation, and chemotherapy. However, it can be challenging to identify non-enhancing infiltrative tumor and differentiate recurrent enhancing tumor from post-treatment changes. Restriction spectrum imaging (RSI), an advanced multishell diffusion technique, has shown promise in distinguishing cellular tumor from edema and enhancing post-treatment changes including pseudoprogression and radiation necrosis. We developed a convolutional network (nnUNet) that incorporates traditional multimodal MRI with DSC perfusion and RSI to detect and quantify enhancing and nonenhancing cellular tumor in post-treatment glioblastoma patients. This cellular tumor segmentation network was evaluated in a cohort of 186 post-treatment GBM patients (55.5 +/- 13.7 years, 158 male) with 243 timepoints using 5-fold cross validation. Cellular tumor volumes were then used to predict overall survival (OS) and progression free survival (PFS) in 94 patients with imaging within 3 months of initial surgery. Accuracy for segmenting cellular tumor as measured by the Dice-similarity coefficient was 0.77 (IQR 0.55-0.87) and the AUC of the ROC curve for detecting residual/recurrent tumor compared to post-treatment changes was 0.86. The predicted cellular tumor volumes were an independent hazard over age, sex, GTR status and conventional enhancing residual tumor for both OS (p=0.0003) and PFS (p=0.0004). The tumor segmentation network was validated in an external dataset of 22 GBM patients (55 timepoints) with a median Dice of 0.73 and AUC of 0.91. Survival prediction was validated in a different external dataset of 70 post-treatment GBM patients, with cellular tumor volumes being an independent predictor of OS (p=0.03). In conclusion, deep learning models incorporating advanced imaging can accurately segment enhancing and nonenhancing cellular tumor, which can help distinguish recurrent/residual tumor from post-treatment changes and predict OS and PFS.

Changes in brain perfusion with training-related visuomotor improvement in MS.

Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system. A better understanding of the mechanisms supporting brain plasticity in MS would help to develop targeted interventions to promote recovery. A total of 29 MS patients and 19 healthy volunteers underwent clinical assessment and multi-modal MRI acquisition [fMRI during serial reaction time task (SRT), DWI, T1w structural scans and ASL of resting perfusion] at baseline and after 4-weeks of SRT training. Reduction of functional hyperactivation was observed in MS patients following the training, shown by the stronger reduction of the BOLD response during task execution compared to healthy volunteers. The functional reorganization was accompanied by a positive correlation between improvements in task accuracy and the change in resting perfusion after 4 weeks' training in right angular and supramarginal gyri in MS patients. No longitudinal changes in WM and GM measures and no correlation between task performance improvements and brain structure were observed in MS patients. Our results highlight a potential role for CBF as an early marker of plasticity, in terms of functional (cortical reorganization) and behavioral (performance improvement) changes in MS patients that may help to guide future interventions that exploit preserved plasticity mechanisms.

Segmentation of glioblastomas in early post-operative multi-modal MRI with deep neural networks

Extent of resection after surgery is one of the main prognostic factors for patients diagnosed with glioblastoma. To achieve this, accurate segmentation and classification of residual tumor from post-operative MR images is essential. The current standard method for estimating it is subject to high inter- and intra-rater variability, and an automated method for segmentation of residual tumor in early post-operative MRI could lead to a more accurate estimation of extent of resection. In this study, two state-of-the-art neural network architectures for pre-operative segmentation were trained for the task. The models were extensively validated on a multicenter dataset with nearly 1000 patients, from 12 hospitals in Europe and the United States. The best performance achieved was a 61% Dice score, and the best classification performance was about 80% balanced accuracy, with a demonstrated ability to generalize across hospitals. In addition, the segmentation performance of the best models was on par with human expert raters. The predicted segmentations can be used to accurately classify the patients into those with residual tumor, and those with gross total resection.

2 Higher White Matter Hyperintensity Load Adversely Affects Pre-Post Proximal Cognitive Training Performance in Healthy Older Adults

Objective:Cognitive training has shown promise for improving cognition in older adults. Aging involves a variety of neuroanatomical changes that may affect response to cognitive training. White matter hyperintensities (WMH) are one common age-related brain change, as evidenced by T2-weighted and Fluid Attenuated Inversion Recovery (FLAIR) MRI. WMH are associated with older age, suggestive of cerebral small vessel disease, and reflect decreased white matter integrity. Higher WMH load associates with reduced threshold for clinical expression of cognitive impairment and dementia. The effects of WMH on response to cognitive training interventions are relatively unknown. The current study assessed (a) proximal cognitive training performance following a 3-month randomized control trial and (b) the contribution of baseline whole-brain WMH load, defined as total lesion volume (TLV), on pre-post proximal training change.Participants and Methods:Sixty-two healthy older adults ages 65-84 completed either adaptive cognitive training (CT; n=31) or educational training control (ET; n=31) interventions. Participants assigned to CT completed 20 hours of attention/processing speed training and 20 hours of working memory training delivered through commercially-available Posit Science BrainHQ. ET participants completed 40 hours of educational videos. All participants also underwent sham or active transcranial direct current stimulation (tDCS) as an adjunctive intervention, although not a variable of interest in the current study. Multimodal MRI scans were acquired during the baseline visit. T1- and T2-weighted FLAIR images were processed using the Lesion Segmentation Tool (LST) for SPM12. The Lesion Prediction Algorithm of LST automatically segmented brain tissue and calculated lesion maps. A lesion threshold of 0.30 was applied to calculate TLV. A log transformation was applied to TLV to normalize the distribution of WMH. Repeated-measures analysis of covariance (RM-ANCOVA) assessed pre/post change in proximal composite (Total Training Composite) and sub-composite (Processing Speed Training Composite, Working Memory Training Composite) measures in the CT group compared to their ET counterparts, controlling for age, sex, years of education and tDCS group. Linear regression assessed the effect of TLV on post-intervention proximal composite and sub-composite, controlling for baseline performance, intervention assignment, age, sex, years of education, multisite scanner differences, estimated total intracranial volume, and binarized cardiovascular disease risk.Results:RM-ANCOVA revealed two-way group*time interactions such that those assigned cognitive training demonstrated greater improvement on proximal composite (Total Training Composite) and sub-composite (Processing Speed Training Composite, Working Memory Training Composite) measures compared to their ET counterparts. Multiple linear regression showed higher baseline TLV associated with lower pre-post change on Processing Speed Training sub-composite (ß = -0.19, p = 0.04) but not other composite measures.Conclusions:These findings demonstrate the utility of cognitive training for improving postintervention proximal performance in older adults. Additionally, pre-post proximal processing speed training change appear to be particularly sensitive to white matter hyperintensity load versus working memory training change. These data suggest that TLV may serve as an important factor for consideration when planning processing speed-based cognitive training interventions for remediation of cognitive decline in older adults.

The alteration of cortical microstructure similarity in drug-resistant epilepsy correlated with mTOR pathway genes

Drug-resistant epilepsy (DRE) is associated with distributed laminar disruptions due to cytoarchitectonic pathologies, which may be characterized by multimodal MRI approaches such as morphometric similarity networks (MSNs). However, the genetic and histological underpinning of MSN alterations in DRE remains poorly understood, hampering its clinical application. We enrolled 60 patients with DRE and 23 controls, acquiring T1 and diffusion spectrum imaging data with a 3.0T GE SIGNA Premier scanner. Morphometric similarity networks (MSNs) were constructed and analyzed to identify microstructure similarity differences between patients and controls. Subsequently, patient-specific MSN alteration patterns were associated with gene expression using the GAMBA tool, and layer-specific neuronal signature mapping were also applied. During these analyses, sex and age were adjusted as covariates and multiple comparisons corrections were applied when appropriate. We observed widespread MSN changes in patients with DRE and identified five distinct MSN alteration patterns. Major patterns presented pattern-specific associations with expressions of epilepsy-related genes, particularly involving the mTOR pathway. Histological analysis confirmed the presence of cortical microstructure changes in areas with MSN alterations and revealed cellular abnormalities matching the aforementioned genetic risks. Our findings highlight the potential of quantifying laminar-related microstructure integrity using MSN to uncover the cytoarchitectonic changes in the pathophysiology of DRE. This approach may facilitate the identification of genetic and histological underpinnings of MSN alterations in DRE, ultimately aiding in the development of targeted therapeutic strategies. The National Natural Science Foundation of China, the Ministry of Science and Technology of the People's Republic of China, and the Beijing Municipal Health Commission.

BTSC-TNAS: A neural architecture search-based transformer for brain tumor segmentation and classification

Glioblastoma (GBM), isolated brain metastasis (SBM), and primary central nervous system lymphoma (PCNSL) possess a high level of similarity in histomorphology and clinical manifestations on multimodal MRI. Such similarities have led to challenges in the clinical diagnosis of these three malignant tumors. However, many existing models solely focus on either the task of segmentation or classification, which limits the application of computer-aided diagnosis in clinical diagnosis and treatment. To solve this problem, we propose a multi-task learning transformer with neural architecture search (NAS) for brain tumor segmentation and classification (BTSC-TNAS). In the segmentation stage, we use a nested transformer U-shape network (NTU-NAS) with NAS to directly predict brain tumor masks from multi-modal MRI images. In the tumor classification stage, we use the multiscale features obtained from the encoder of NTU-NAS as the input features of the classification network (MSC-NET), which are integrated and corrected by the classification feature correction enhancement (CFCE) block to improve the accuracy of classification. The proposed BTSC-TNAS achieves an average Dice coefficient of 80.86% and 87.12% for the segmentation of tumor region and the maximum abnormal region in clinical data respectively. The model achieves a classification accuracy of 0.941. The experiments performed on the BraTS 2019 dataset show that the proposed BTSC-TNAS has excellent generalizability and can provide support for some challenging tasks in the diagnosis and treatment of brain tumors.

Application value of multimodal MRI combined with PET metabolic parameters in temporal lobe epilepsy with dual pathology

ObjectivesTo investigate the application value of multimodal MRI combined with PET metabolic parameters in detecting temporal lobe epilepsy (TLE) with dual pathology (DP) and the prediction effect of post-surgical outcomes in these patients. MethodsWe retrospectively reviewed 50 patients with TLE-DP who underwent surgery at our hospital between January 2016 and December 2021 and collected the demographics, clinical characteristics, video-electroencephalography (v-EEG), neuroimaging, and surgical data. Seizure outcome data were collected during a regular follow-up of at least 12 months and were graded using Engel scores. Fisher's exact test was used to compare the differences in DP detection rates of various diagnostic modalities. Univariate and multivariate analyses were performed to explore the prognostic factors for predicting seizure outcomes post-surgery. ResultsOf the 50 patients, 20 were males. The median age was 30, the median age at first seizure was 14, and the median duration was ten years. Voxel-based morphometry-PET statistical parametric mapping-PET/MRI (VBM-PSPM-PET/MRI) had the highest detection rate, followed by PET/MRI, VBM analysis, and PET-SPM. Regardless of follow-up duration, v-EEG, PET, image post-processing methods, and VBM-PSPM-PET/MRI statistically correlated with seizure outcomes using the log-rank test in the Kaplan-Meier analysis. Multivariate analysis showed that VBM-PSPM-PET/MRI was an independent predictor of TLE-DP (hazard ratio (HR) = 15.674, 95 % CI = 0.002–0.122, P < 0.00 1). ConclusionsOur study illustrates that VBM-PSPM-PET/MRI has the highest detection value in patients with TLE-DP and can provide independent prognostic information for patients who undergo surgery. This approach has the most substantial potential for the selection of candidates for patients who undergo surgical treatment and for prognostic stratification.

Functional-structural decoupling in visual network is associated with cognitive decline in patients with type 2 diabetes mellitus: evidence from a multimodal MRI analysis.

Type 2 diabetes mellitus (T2DM) and cognitive dysfunction are highly prevalent disorders worldwide. Although visual network (VN) alteration and functional-structural coupling are potential warning factors for mild cognitive impairment (MCI) in T2DM patients, the relationship between the three in T2DM without MCI is unclear. Thirty T2DM patients without MCI and twenty-nine healthy controls (HC) were prospectively enrolled. Visual components (VC) were estimated by independent component analysis (ICA). Degree centrality (DC), amplitude of low frequency fluctuation (ALFF) and fractional anisotropy (FA) were established to reflect functional and structural characteristics in these VCs respectively. Functional-structural coupling coefficients were further evaluated using combined FA and DC or ALFF. Partial correlations were performed among neuroimaging indicators and neuropsychological scores and clinical variables. Three VCs were selected using group ICA. Deteriorated DC, ALFF and DC-FA coefficients in the VC1 were observed in the T2DM group compared with the HC group, while FA and ALFF-FA coefficients in these three VCs showed no significant differences. In the T2DM group, DC in the VC1 positively correlated with 2 dimensions in the California Verbal Learning Test, including Trial 4 and Total trial 1-5. The impaired DC-FA coefficients in the VC1 markedly affected the Total perseverative responses % of the Wisconsin Card Sorting Test. These findings indicate that DC and DC-FA coefficients in VN may be potential imaging biomarkers revealing early cognitive deficits in T2DM.

Decreased DTI-ALPS and choroid plexus enlargement in fibromyalgia: a preliminary multimodal MRI study.

The glymphatic system is a fluid exchange pathway that clears waste products that is crucial for the maintenance of brain homeostasis. However, the exact role it plays in the emergence of fibromyalgia (FM) is still not fully understood. Here, we explored the changes in non-invasive MRI proxy probably related to the glymphatic function in FM patients, and explored brain-behavior relationships. A total of 40 participants, consisting of 20 individuals with FM and 20 healthy controls (HCs), were included in the study. The participants underwent structural T1-weighted MRI, diffusion tensor imaging (DTI), and clinical assessment. The data was obtained from an open access dataset. The study compared non-invasive MRI indices, including choroid plexus (CP) volume and DTI analysis along the perivascular space (ALPS), between the FM and HC groups. Furthermore, correlation analysis was conducted to determine the correlation between clinical parameters and both CP volume and DTI-ALPS index. Patients with FM had significantly higher CP volume and a lower DTI-ALPS index than HCs adjusting for age and intracranial volume. Higher CP volume was associated with lower DTI-ALPS index, and longer disease duration. Our findings demonstrate aberrant glymphatic function in FM, and that dysfunction in the brain glymphatic system may play a role in the neural mechanisms underlying FM.

Multiencoder‐based federated intelligent deep learning model for brain tumor segmentation

AbstractGlioma, a primary tumor derived from brain glial cells, is around 45% of all intracranial tumors. Magnetic resonance imaging's (MRI's) precise glioma segmentation is crucial for clinical purposes. This article presents a novel automatic brain tumor segmentation approach based on a multi‐encoder‐based federated intelligent deep learning framework. The suggested method uses a U‐shaped network design that multiplies the single contraction path into several paths to explore semantic information modalities deeply. The basic convolutional layer uses an Inception module and dilated convolutions to extract multi‐scale features from the images using artificial intelligent. To emphasize segmentation‐related information while ignoring redundant channel dimension information and improving the accuracy of network segmentation, lightweight channel attention efficient channel attention (ECA) modules are inserted into the bottleneck layer and decoder. The collection of data for the 2018 Brain Tumor Segmentation Challenge (BraTS 2018) is used to test the effectiveness of the suggested structure, and the findings indicate that the growth core, for the entire tumor and the augmented tumor regions, respectively, the average Dice coefficients are 0.880, 0.784, and 0.757. These findings support the proposed algorithm's ability to accurately and successfully segregate multimodal MRI brain tumors.

Brain-derived neurotrophic factor gene polymorphism affects cognitive function and neurofilament light chain level in patients with subcortical ischaemic vascular dementia.

To investigate the effects of brain-derived neurotrophic factor (BDNF) gene polymorphism on cognitive function, neuroimaging and blood biological markers in patients with subcortical ischaemic vascular dementia (SIVD). A total of 81 patients with SIVD were included. According to their BDNF gene polymorphism, the participants were divided into the Val/Val (n = 26), Val/Met (n = 35), and Met/Met (n = 20) groups. A comprehensive neuropsychological evaluation and multimodal brain MRI scan were performed. MRI markers for small vessel disease were visually rated or quantitatively analysed. Moreover, 52 patients were further evaluated with blood marker assays, including amyloid beta (Aβ), phosphorylated tau at threonine-181 (P-tau181), glial fibrillary acidic protein (GFAP), total tau (T-tau) and neurofilament light chain (NfL). There were no significant differences in demographics, disease duration or MRI markers of small vessel disease between the three groups. Compared with the Val/Val and Val/Met groups, the Met/Met group showed worse performance in the verbal fluency test and higher levels of plasma NfL. The rs6265 polymorphism of the BDNF gene is associated with semantic language fluency in patients with SIVD. The Met genotype may be a risk factor for cognitive impairment and neuronal injury.

Diagnostic Value of Radiomics Analysis Based on Multimodal MRI for Advanced Liver Fibrosis in Patients with Hepatitis B.

The goal was to explore the value of using radiomics analysis based on multimodal MRI for evaluating the advanced fibrosis in patients with hepatitis B. One hundred and forty-three patients with hepatitis B fibrosis were randomly divided into training and validation cohorts in a 2:1 ratio. In the training cohort, a clinical model was established with logistic regression, a radiomics signature based on multimodal MRI was established with support vector machine (SVM), and a nomogram integrated the radiomics signature and clinical factors. The value of three models was assessed by ROC analysis in the training and validation cohorts. The nomogram demonstrated the largest area under the ROC curve. The nomogram presented good agreement in the prediction probability of advanced liver fibrosis in two cohorts. Radiomics analysis has good diagnostic value for advanced liver fibrosis and the nomogram can enhance the diagnostic value.

Multimodal MRI study of the relationship between plaque characteristics and hypoperfusion in patients with transient ischemic attack.

Transient ischemic attack is a significant risk factor for acute cerebral infarction. Previous studies have demonstrated that hypoperfusion in patients with transient ischemic attack was associated with the recurrence of transient ischemic attack, stroke, and persistent worsening of neurological symptoms. Moreover, transient ischemic attack patients classified as high-risk group according to the ABCD2 score have a higher incidence of stroke. Therefore, the objective of this study was to investigate the plaque characteristics of transient ischemic attack patients with concomitant cerebral hypoperfusion using multimodal MRI, as well as hemodynamic changes in the high-risk group with transient ischemic attack patients. A total of 151 patients with transient ischemic attack were prospectively recruited for this study. All enrolled patients underwent multimodal MRI, including DWI, TOF-MRA, HR-VWI, and DSC-PWI. Finally, 56 patients met the inclusion criteria. Based on DSC-PWI images, patients were divided into two groups: hypoperfusion (n = 41) and non-hypoperfusion (n = 15). Clinical baseline characteristics and plaque characteristics were analyzed between the two groups. Furthermore, within the hypoperfusion group, patients were further classified into low-risk (n = 11) and high-risk (n = 30) subgroups based on the ABCD2 score. Hemodynamic differences between these subgroups were also analyzed. Compared with the non-hypoperfusion group, the hypoperfusion group had a significantly higher prevalence of hypertension (68.3% vs. 33.3%, p = 0.019) and hyperhomocysteinemia (65.9% vs. 33.3%, p = 0.029). Moreover, the hypoperfusion group exhibited more significant luminal stenosis degree [41.79 31.36 vs. 17.6213.62, p = 0.006] and greater NWI (57.1%20.47% vs. 40.21% 21.56%, p = 0.009) compared to the non-hypoperfusion group. In addition, the high-risk group identified by the ABCD2 score had a higher rMTT [117.6(109.31-128.14) vs. 108.36(100.67-119.92), p = 0.037]. Transient ischemic attack patients with hypoperfusion exhibited a higher prevalence of hypertension and hyperhomocysteinemia, as well as higher luminal stenosis degree, and greater NWI. Furthermore, Transient ischemic attack patients in the high-risk group demonstrated higher MTT.

Stroke Mimics: Insights from a Retrospective Neuroimaging Study

Objectives: The study’s goals are to evaluate the management of Stroke Mimics (SMs), conditions with stroke-like symptoms but non-vascular origins. It seeks to assess the effect of unnecessary intravenous thrombolysis on therapy delays and determine the best SMs diagnosis approach. Materials and methods: A review was conducted of all patients admitted to the Emergency Department under a “stroke code” from January 1, 2018, to January 31, 2019. Anamnestic and clinical data, along with information on neuroimaging protocols and findings, were collected. Advanced MRI sequences, such as Arterial Spin Labeling (ASL) MR perfusion and MR spectroscopy, were revised to confirm the diagnoses. Results: 513 stroke codes were found; a neurovascular disease was diagnosed in 414 cases: 282 ischemic strokes and 73 intracranial hemorrages and 59 TIAs. The 99 SMs included, 13 infections, 12 syncopes, 11 epileptic seizures, 11 hemodinamic conditions, 10 tumors, 9 metabolic disorders, 9 diziness, 7 migraines, 4 drug/alcohol intoxication, 3 functional disorders, 3 acute hydrocefalus, 2 multiple sclerosis, 2 arteriovenous malformations, 1 spinal cord compression and 2 unexplained conditions. Specific neuroimaging findings were evaluated for all cases. Positive and negative predictive values of clinical diagnosis of SM were respectively 0.23 and 0.35. 125 SM patients underwent MRI examination, 40 of which within 3 hours from the activation of stroke code. Advanced MRI sequences as ASL, MR perfusion, MR spectroscopy were used to reach the correct diagnosis. Conclusions: Advanced multimodal MRI can be a valuable tool in the assessment of, particularly in cases where conventional imaging techniques such as DWI-FLAIR mismatch are inconclusive. The novelty of this retrospective study is to demonstrate that the consistent use of arterial spin labeling perfusion in cases with stroke code leads to a rapid and accurate diagnosis of SMs. The implementation of an MRI-based pathway can expedite the diagnosis and treatment of underlying non-vascular causes such as SMs.

CNN-Res: deep learning framework for segmentation of acute ischemic stroke lesions on multimodal MRI images

Background Accurate segmentation of stroke lesions on MRI images is very important for neurologists in the planning of post-stroke care. Segmentation helps clinicians to better diagnose and evaluation of any treatment risks. However, manual segmentation of brain lesions relies on the experience of neurologists and is also a very tedious and time-consuming process. So, in this study, we proposed a novel deep convolutional neural network (CNN-Res) that automatically performs the segmentation of ischemic stroke lesions from multimodal MRIs.MethodsCNN-Res used a U-shaped structure, so the network has encryption and decryption paths. The residual units are embedded in the encoder path. In this model, to reduce gradient descent, the residual units were used, and to extract more complex information in images, multimodal MRI data were applied. In the link between the encryption and decryption subnets, the bottleneck strategy was used, which reduced the number of parameters and training time compared to similar research.ResultsCNN-Res was evaluated on two distinct datasets. First, it was examined on a dataset collected from the Neuroscience Center of Tabriz University of Medical Sciences, where the average Dice coefficient was equal to 85.43%. Then, to compare the efficiency and performance of the model with other similar works, CNN-Res was evaluated on the popular SPES 2015 competition dataset where the average Dice coefficient was 79.23%.ConclusionThis study presented a new and accurate method for the segmentation of MRI medical images using a deep convolutional neural network called CNN-Res, which directly predicts segment maps from raw input pixels.

Stronger influence of systemic than local hemodynamic-vascular factors on resting-state BOLD functional connectivity

Correlated fluctuations in the blood oxygenation level dependent (BOLD) signal of resting-state functional MRI (i.e., BOLD-functional connectivity, BOLD-FC) reflect a spectrum of neuronal and non-neuronal processes. In particular, there are multiple hemodynamic-vascular influences on BOLD-FC on both systemic (e.g., perfusion delay) and local levels (e.g., neurovascular coupling). While the influence of individual factors has been studied extensively, combined and comparative studies of systemic and local hemodynamic-vascular factors on BOLD-FC are scarce, notably in humans. We employed a multi-modal MRI approach to investigate and compare distinct hemodynamic-vascular processes and their impact on homotopic BOLD-FC in healthy controls and patients with unilateral asymptomatic internal carotid artery stenosis (ICAS). Asymptomatic ICAS is a cerebrovascular disorder, in which neuronal functioning is largely preserved but hemodynamic-vascular processes are impaired, mostly on the side of stenosis. Investigated indicators for local hemodynamic-vascular processes comprise capillary transit time heterogeneity (CTH) and cerebral blood volume (CBV) from dynamic susceptibility contrast (DSC) MRI, and cerebral blood flow (CBF) from pseudo-continuous arterial spin labeling (pCASL). Indicators for systemic processes are time-to-peak (TTP) from DSC MRI and BOLD lags from functional MRI. For each of these parameters, their influence on BOLD-FC was estimated by a comprehensive linear mixed model. Equally across groups, we found that individual mean BOLD-FC, local (CTH, CBV, and CBF) and systemic (TTP and BOLD lag) hemodynamic-vascular factors together explain 40.7% of BOLD-FC variance, with 20% of BOLD-FC variance explained by hemodynamic-vascular factors, with an about two-times larger contribution of systemic versus local factors. We conclude that regional differences in blood supply, i.e., systemic perfusion delays, exert a stronger influence on BOLD-FC than impairments in local neurovascular coupling.