Automated Ischemic Lesion Segmentation in MRI Mouse Brain Data after Transient Middle Cerebral Artery Occlusion

Estimation of the Ischemic Lesion in the Experimental Stroke Studies Using Magnetic Resonance Imaging (Review)

In translational animal study aimed at evaluation of the effectiveness of innovative methods for treating cerebral stroke, including regenerative cell technologies, of particular importance is evaluation of the dynamics of changes in the volume of the cerebral infarction in response to therapy. Among the methods for assessing the focus of infarction, MRI is the most effective and convenient tool for use in preclinical studies. This review provides a description of MR pulse sequences used to visualize cerebral ischemia at various stages of its development, and a detailed description of the MR semiotics of cerebral infarction. A comparison of various methods for morphometric analysis of the focus of a cerebral infarction, including systems based on artificial intelligence for a more objective measurement of the volume of the lesion, is also presented.

Deep learning-based automated lesion segmentation on mouse stroke magnetic resonance images

Magnetic resonance imaging (MRI) is widely used for ischemic stroke lesion detection in mice. A challenge is that lesion segmentation often relies on manual tracing by trained experts, which is labor-intensive, time-consuming, and prone to inter- and intra-rater variability. Here, we present a fully automated ischemic stroke lesion segmentation method for mouse T2-weighted MRI data. As an end-to-end deep learning approach, the automated lesion segmentation requires very little preprocessing and works directly on the raw MRI scans. We randomly split a large dataset of 382 MRI scans into a subset (n = 293) to train the automated lesion segmentation and a subset (n = 89) to evaluate its performance. We compared Dice coefficients and accuracy of lesion volume against manual segmentation, as well as its performance on an independent dataset from an open repository with different imaging characteristics. The automated lesion segmentation produced segmentation masks with a smooth, compact, and realistic appearance that are in high agreement with manual segmentation. We report dice scores higher than the agreement between two human raters reported in previous studies, highlighting the ability to remove individual human bias and standardize the process across research studies and centers.

Automated Stroke Lesion Segmentation in Rat Brain MR Images Using an Encoder-Decoder Framework

Stroke is a leading cause of serious long-term disability and the major cause of mortality worldwide. Experimental ischemic stroke models play an important role in realizing the mechanism of cerebral ischemia and evaluating the development of pathological extent. An accurate and reliable image segmentation tool to automatically identify the stroke lesion is important in the subsequent processes. However, the intensity distribution of the infarct region in the diffusion weighted imaging (DWI) images is usually nonuniform with blurred boundaries. A deep learning-based infarct region segmentation framework is developed in this paper to address the segmentation difficulties. The proposed solution is an encoder-decoder network that includes a hybrid block model for efficient multiscale feature extraction. An in-house DWI image dataset was created to evaluate this automated stroke lesion segmentation scheme. Through massive experiments, accurate segmentation results were obtained, which outperformed many competitive methods both qualitatively and quantitatively. Our stroke lesion segmentation system is potential in providing a decent tool to facilitate preclinical stroke investigation using DWI images.Clinical Relevance- This facilitates neuroscientists the investigation of a new scoring system with less examination time and better inter-rater reliability, which helps to understand the function of specific brain areas underlying neuroimaging signatures clinically.

Reactivation of corticogenesis-related transcriptional factors BCL11B and SATB2 after ischemic lesion of the adult mouse brain

The aim of this study was to characterize expression of corticogenesis-related transcription factors BCL11B and SATB2 after brain ischemic lesion in the adult mice, and to analyze their correlation to the subsequent brain recovery. Ischemic brain lesion was induced by transient middle cerebral artery occlusion followed by reperfusion, and the animals with ischemic lesion were compared to the sham controls. Progression of the brain damage and subsequent recovery was longitudinally monitored structurally, by magnetic resonance imaging, and functionally, by neurological deficit assessment. Seven days after the ischemic injury the brains were isolated and analyzed by immunohistochemistry. The results showed higher expression in the brain of both, BCL11B and SATB2 in the animals with ischemic lesion compared to the sham controls. The co-expression of both markers, BCL11B and SATB2, increased in the ischemic brains, as well as the co-expression of BCL11B with the beneficial transcriptional factor ATF3 but not its co-expression with detrimental HDAC2. BCL11B was mainly implicated in the ipsilateral and SATB2 in the contralateral brain hemisphere, and their level in these regions correlated with the functional recovery rate. The results indicate that the reactivation of corticogenesis-related transcription factors BCL11B and SATB2 is beneficial after brain ischemic lesion.

RU-Net: skull stripping in rat brain MR images after ischemic stroke with rat U-Net

BACKGROUND

Experimental ischemic stroke models play a fundamental role in interpreting the mechanism of cerebral ischemia and appraising the development of pathological extent. An accurate and automatic skull stripping tool for rat brain image volumes with magnetic resonance imaging (MRI) are crucial in experimental stroke analysis. Due to the deficiency of reliable rat brain segmentation methods and motivated by the demand for preclinical studies, this paper develops a new skull stripping algorithm to extract the rat brain region in MR images after stroke, which is named Rat U-Net (RU-Net).

METHODS

Based on a U-shape like deep learning architecture, the proposed framework integrates batch normalization with the residual network to achieve efficient end-to-end segmentation. A pooling index transmission mechanism between the encoder and decoder is exploited to reinforce the spatial correlation. Two different modalities of diffusion-weighted imaging (DWI) and T2-weighted MRI (T2WI) corresponding to two in-house datasets with each consisting of 55 subjects were employed to evaluate the performance of the proposed RU-Net.

RESULTS

Extensive experiments indicated great segmentation accuracy across diversified rat brain MR images. It was suggested that our rat skull stripping network outperformed several state-of-the-art methods and achieved the highest average Dice scores of 98.04% (p < 0.001) and 97.67% (p < 0.001) in the DWI and T2WI image datasets, respectively.

CONCLUSION

The proposed RU-Net is believed to be potential for advancing preclinical stroke investigation and providing an efficient tool for pathological rat brain image extraction, where accurate segmentation of the rat brain region is fundamental.

Automatic Cerebral Hemisphere Segmentation in Rat MRI with Ischemic Lesions via Attention-based Convolutional Neural Networks

We present MedicDeepLabv3+, a convolutional neural network that is the first completely automatic method to segment cerebral hemispheres in magnetic resonance (MR) volumes of rats with ischemic lesions. MedicDeepLabv3+ improves the state-of-the-art DeepLabv3+ with an advanced decoder, incorporating spatial attention layers and additional skip connections that, as we show in our experiments, lead to more precise segmentations. MedicDeepLabv3+ requires no MR image preprocessing, such as bias-field correction or registration to a template, produces segmentations in less than a second, and its GPU memory requirements can be adjusted based on the available resources. We optimized MedicDeepLabv3+ and six other state-of-the-art convolutional neural networks (DeepLabv3+, UNet, HighRes3DNet, V-Net, VoxResNet, Demon) on a heterogeneous training set comprised by MR volumes from 11 cohorts acquired at different lesion stages. Then, we evaluated the trained models and two approaches specifically designed for rodent MRI skull stripping (RATS and RBET) on a large dataset of 655 MR rat brain volumes. In our experiments, MedicDeepLabv3+ outperformed the other methods, yielding an average Dice coefficient of 0.952 and 0.944 in the brain and contralateral hemisphere regions. Additionally, we show that despite limiting the GPU memory and the training data, our MedicDeepLabv3+ also provided satisfactory segmentations. In conclusion, our method, publicly available at https://github.com/jmlipman/MedicDeepLabv3Plus , yielded excellent results in multiple scenarios, demonstrating its capability to reduce human workload in rat neuroimaging studies.

Serum Exosomal microRNA-27-3p Aggravates Cerebral Injury and Inflammation in Patients with Acute Cerebral Infarction by Targeting PPARγ

Acute cerebral infarction (ACI) possesses high mortality. Exosomes present in serum have potential application value in ACI diagnosis. This study investigated the mechanism of serum exosomes in ACI. Serum exosomes isolated from ACI patients and normal people were identified and then injected into the established middle cerebral artery occlusion (MCAO) rat model to evaluate cerebral injury and inflammation. Exosomal microRNA (miR)-27-3p expression was detected and interfered to analyze rat cerebral inflammation. The binding relationship between miR-27-3p and PPARγ was predicted and verified. The lipopolysaccharide (LPS)-treated microglia model was established and intervened with miR-27-3p to detect PPARγ, Iba-1, and inflammation-related factor expressions. After overexpressing PPARγ, rat cerebral inflammation was evaluated. The clinical significance of serum exosomal miR-27-3p in ACI was evaluated. Serum exosomes from ACI patients caused exacerbated MCAO rat cerebral injury and poor behavior recovery, as well as promoted cerebral inflammation. Serum exosomal miR-27-3p deepened rat brain inflammation. miR-27-3p targeted PPARγ to promote microglia activation and inflammation-related factor expressions in MCAO rats, and overexpressing PPARγ attenuated MCAO rat cerebral inflammation. Serum exosomal miR-27-3p promised to be a biomarker for ACI. We proved that serum exosomes from ACI patients aggravated ACI patient cerebral inflammation via the miR-27-3p/PPARγ axis.

Impact of hydroxytyrosol on stroke: tracking therapy response on neuroinflammation and cerebrovascular parameters using PET-MR imaging and on functional outcomes

Immune cells have been implicated in influencing stroke outcomes depending on their temporal dynamics, number, and spatial distribution after ischemia. Depending on their activation status, immune cells can have detrimental and beneficial properties on tissue outcome after stroke, highlighting the need to modulate inflammation towards beneficial and restorative immune responses. Novel dietary therapies may promote modulation of pro- and anti-inflammatory immune cell functions. Among the dietary interventions inspired by the Mediterranean diet, hydroxytyrosol (HT), the main phenolic component of the extra virgin olive oil (EVOO), has been suggested to have antioxidant and anti-inflammatory properties in vitro. However, immunomodulatory effects of HT have not yet been studied in vivo after stroke. The aim of this project is therefore to monitor the therapeutic effect of a HT-enriched diet in an experimental stroke model using non-invasive in vivo multimodal imaging, behavioural phenotyping and cross-correlation with ex vivo parameters. Methods: A total of N = 22 male C57BL/6 mice were fed with either a standard chow (n = 11) or a HT enriched diet (n = 11) for 35 days, following a 30 min transient middle cerebral artery occlusion (tMCAo). T₂-weighted (lesion) and perfusion (cerebral blood flow)-/diffusion (cellular density)-weighted MR images were acquired at days 1, 3, 7, 14, 21 and 30 post ischemia. [¹⁸F]DPA-714 (TSPO, neuroinflammation marker) PET-CT scans were acquired at days 7, 14, 21 and 30 post ischemia. Infarct volume (mm³), cerebral blood flow (mL/100g/min), apparent diffusion coefficient (10^-4·mm²/s) and percentage of injected tracer dose (%ID/mL) were assessed. Behavioural tests (grip test, rotarod, open field, pole test) were performed prior and after ischemia to access therapy effects on sensorimotor functions. Ex vivo analyses (IHC, IF, WB) were performed to quantify TSPO expression, immune cells including microglia/macrophages (Iba-1, F4/80), astrocytes (GFAP) and peripheral markers in serum such as thiobarbituric acid reactive substances (TBARS) and nitric oxide (NO) 35 days post ischemia. Additionally, gene expression of pro- and anti-inflammatory markers were assessed by rt-qPCR, including tspo, cd163, arg1, tnf and Il-1β. Results: No treatment effect was observed on temporal [¹⁸F]DPA-714 uptake within the ischemic and contralateral region (two-way RM ANOVA, p = 0.71). Quantification of the percentage of TSPO⁺ area by immunoreactivity indicated a slight 2-fold increase in TSPO expression within the infarct region in HT-fed mice at day 35 post ischemia (p = 0.011) correlating with a 2-3 fold increase in Iba-1⁺ cell population expressing CD163 as anti-inflammatory marker (R² = 0.80). Most of the GFAP⁺ cells were TSPO^-. Only few F4/80⁺ cells were observed at day 35 post ischemia in both groups. No significant treatment effect was observed on global ADC and CBF within the infarct and the contralateral region over time. Behavioural tests indicated improved strength of the forepaws at day 14 post ischemia (p = 0.031). Conclusion: An HT-enriched diet significantly increased the number of Iba-1⁺ microglia/macrophages in the post-ischemic area, inducing higher expression of anti-inflammatory markers while no clear-cut effect was observed. Also, HT did not affect recovery of the cerebrovascular parameters, including ADC and CBF. Altogether, our data indicated that a prolonged dietary intervention with HT, as a single component of the Mediterranean diet, induces molecular changes that may improve stroke outcomes. Therefore, we support the use of the Mediterranean diet as a multicomponent therapy approach after stroke.

Machine learning identifies stroke features between species

Identification and localization of ischemic stroke (IS) lesions is routinely performed to confirm diagnosis, assess stroke severity, predict disability and plan rehabilitation strategies using magnetic resonance imaging (MRI). In basic research, stroke lesion segmentation is necessary to study complex peri-infarction tissue changes. Moreover, final stroke volume is a critical outcome evaluated in clinical and preclinical experiments to determine therapy or intervention success. Manual segmentations are performed but they require a specialized skill set, are prone to inter-observer variation, are not entirely objective and are often not supported by histology. The task is even more challenging when dealing with large multi-center datasets, multiple experimenters or large animal cohorts. On the other hand, current automatized segmentation approaches often lack histological validation, are not entirely user independent, are often based on single parameters, or in the case of complex machine learning methods, require vast training datasets and are prone to a lack of model interpretation.

Methods: We induced IS using the middle cerebral artery occlusion model on two rat cohorts. We acquired apparent diffusion coefficient (ADC) and T2-weighted (T2W) images at 24 h and 1-week after IS induction. Subsets of the animals at 24 h and 1-week post IS were evaluated using histology and immunohistochemistry. Using a Gaussian mixture model, we segmented voxel-wise interactions between ADC and T2W parameters at 24 h using one of the rat cohorts. We then used these segmentation results to train a random forest classifier, which we applied to the second rat cohort. The algorithms' stroke segmentations were compared to manual stroke delineations, T2W and ADC thresholding methods and the final stroke segmentation at 1-week. Volume correlations to histology were also performed for every segmentation method. Metrics of success were calculated with respect to the final stroke volume. Finally, the trained random forest classifier was tested on a human dataset with a similar temporal stroke on-set. Manual segmentations, ADC and T2W thresholds were again used to evaluate and perform comparisons with the proposed algorithms' output.

Results: In preclinical rat data our framework significantly outperformed commonly applied automatized thresholding approaches and segmented stroke regions similarly to manual delineation. The framework predicted the localization of final stroke regions in 1-week post-stroke MRI with a median Dice similarity coefficient of 0.86, Matthew's correlation coefficient of 0.80 and false positive rate of 0.04. The predicted stroke volumes also strongly correlated with final histological stroke regions (Pearson correlation = 0.88, P < 0.0001). Lastly, the stroke region characteristics identified by our framework in rats also identified stroke lesions in human brains, largely outperforming thresholding approaches in stroke volume prediction (P<0.01).

Conclusion: Our findings reveal that the segmentation produced by our proposed framework using 24 h MRI rat data strongly correlated with the final stroke volume, denoting a predictive effect. In addition, we show for the first time that the stroke imaging features can be directly translated between species, allowing identification of acute stroke in humans using the model trained on animal data. This discovery reduces the gap between the clinical and preclinical fields, unveiling a novel approach to directly co-analyze clinical and preclinical data. Such methods can provide further biological insights into human stroke and highlight the differences between species in order to help improve the experimental setups and animal models of the disease.

RatLesNetv2: A Fully Convolutional Network for Rodent Brain Lesion Segmentation

We present a fully convolutional neural network (ConvNet), named RatLesNetv2, for segmenting lesions in rodent magnetic resonance (MR) brain images. RatLesNetv2 architecture resembles an autoencoder and it incorporates residual blocks that facilitate its optimization. RatLesNetv2 is trained end to end on three-dimensional images and it requires no preprocessing. We evaluated RatLesNetv2 on an exceptionally large dataset composed of 916 T2-weighted rat brain MRI scans of 671 rats at nine different lesion stages that were used to study focal cerebral ischemia for drug development. In addition, we compared its performance with three other ConvNets specifically designed for medical image segmentation. RatLesNetv2 obtained similar to higher Dice coefficient values than the other ConvNets and it produced much more realistic and compact segmentations with notably fewer holes and lower Hausdorff distance. The Dice scores of RatLesNetv2 segmentations also exceeded inter-rater agreement of manual segmentations. In conclusion, RatLesNetv2 could be used for automated lesion segmentation, reducing human workload and improving reproducibility. RatLesNetv2 is publicly available at https://github.com/jmlipman/RatLesNetv2.

Individual in vivo Profiles of Microglia Polarization After Stroke, Represented by the Genes iNOS and Ym1

Microglia are the brain-innate immune cells which actively surveil their environment and mediate multiple aspects of neuroinflammation, due to their ability to acquire diverse activation states and phenotypes. Simplified, M1-like microglia are defined as pro-inflammatory cells, while the alternative M2-like cells promote neuroprotection. The modulation of microglia polarization is an appealing neurotherapeutic strategy for stroke and other brain lesions, as well as neurodegenerative diseases. However, the activation profile and change of phenotype during experimental stroke is not well understood. With a combined magnetic resonance imaging (MRI) and optical imaging approach and genetic targeting of two key genes of the M1- and M2-like phenotypes, iNOS and Ym1, we were able to monitor in vivo the dynamic adaption of the microglia phenotype in response to experimental stroke.

A Heavy Tailed Expectation Maximization Hidden Markov Random Field Model with Applications to Segmentation of MRI

A wide range of segmentation approaches assumes that intensity histograms extracted from magnetic resonance images (MRI) have a distribution for each brain tissue that can be modeled by a Gaussian distribution or a mixture of them. Nevertheless, intensity histograms of White Matter and Gray Matter are not symmetric and they exhibit heavy tails. In this work, we present a hidden Markov random field model with expectation maximization (EM-HMRF) modeling the components using the α-stable distribution. The proposed model is a generalization of the widely used EM-HMRF algorithm with Gaussian distributions. We test the α-stable EM-HMRF model in synthetic data and brain MRI data. The proposed methodology presents two main advantages: Firstly, it is more robust to outliers. Secondly, we obtain similar results than using Gaussian when the Gaussian assumption holds. This approach is able to model the spatial dependence between neighboring voxels in tomographic brain MRI.