Brain tissue segmentation in neurosurgery: a systematic analysis for quantitative tractography approaches

Diffusion magnetic resonance imaging (dMRI) is a cutting-edge imaging method that provides a macro-scale in vivo map of the white matter pathways in the brain. The measurement of brain microstructure and the enhancement of tractography rely heavily on dMRI tissue segmentation. Anatomical MRI technique (e.g., T1- and T2-weighted imaging) is the most widely used method for segmentation in dMRI. In comparison to anatomical MRI, dMRI suffers from higher image distortions, lower image quality, and making inter-modality registration more difficult. The dMRI tractography study of brain connectivity has become a major part of the neuroimaging landscape in recent years. In this research, we provide a high-level overview of the methods used to segment several brain tissues types, including grey and white matter and cerebrospinal fluid, to enable quantitative studies of structural connectivity in the brain in health and illness. In the first part of our review, we discuss the three main phases in the quantitative analysis of tractography, which are correction, segmentation, and quantification. Methodological possibilities are described for each phase, along with their popularity and potential benefits and drawbacks. After that, we will look at research that used quantitative tractography approaches to examine the white and grey matter of the brain, with an emphasis on neurodevelopment, ageing, neurological illnesses, mental disorders, and neurosurgery as possible applications. Even though there have been substantial advancements in methodological technology and the spectrum of applications, there is still no consensus regarding the "optimal" approach in the quantitative analysis of tractography. As a result, researchers should tread carefully when interpreting the findings of quantitative analysis of tractography.

Estimation of undistorted images in brain echo-planar images with distortions using the conjugate gradient method with anatomical regularization

PURPOSE

Although echo-planar imaging (EPI) is widely used for diffusion magnetic resonance (MR) imaging, EPI images suffer from susceptibility-induced geometric distortions. We herein propose a new estimation method for undistorted EPI images using anatomical T₁ -weighted images (T₁ WIs) based on the physics of MR imaging.

METHODS

Our proposed method estimates the undistorted EPI image in the image domain while estimating the magnetic field inhomogeneity map using the conjugate gradient method with anatomical regularization. Our method synthesizes the distorted image to match the measured EPI image containing geometric distortions by alternately updating the undistorted EPI image and the magnetic field inhomogeneity map. We evaluated our proposed method and compared it with a nonrigid registration-based distortion correction method using simulated data and using real data. In the evaluation of the estimation of the magnetic field inhomogeneity map, we used the normalized root-mean-squared error (NRMSE) between the estimated results and the ground truth. In the evaluation of the estimation of undistorted images, we used mutual information (MI) between the undistorted EPI image and the anatomical T₁ WI.

RESULTS

Using the simulated data, the means and standard deviations of the NRMSE values in the nonrigid registration-based method and proposed method were 1.29 ± 0.63 and 0.64 ± 0.30, respectively. The MI values in the proposed method were larger than those in the nonrigid registration-based method in all evaluated slices. For the real data, the proposed method improved the distortion, and the MI values in the proposed method were larger than those in the nonrigid registration-based method. In the estimation of the magnetic field inhomogeneity map, the NRMSE values in our method were smaller than those in the nonrigid registration-based method.

CONCLUSIONS

We demonstrated that our proposed method can estimate the regions with compressed distortions that are not well represented by the nonrigid registration-based methods. The results suggest that the proposed method could be useful in analyses combining EPI images with T₁ WIs.

Deep learning based segmentation of brain tissue from diffusion MRI

Segmentation of brain tissue types from diffusion MRI (dMRI) is an important task, required for quantification of brain microstructure and for improving tractography. Current dMRI segmentation is mostly based on anatomical MRI (e.g., T1- and T2-weighted) segmentation that is registered to the dMRI space. However, such inter-modality registration is challenging due to more image distortions and lower image resolution in dMRI as compared with anatomical MRI. In this study, we present a deep learning method for diffusion MRI segmentation, which we refer to as DDSeg. Our proposed method learns tissue segmentation from high-quality imaging data from the Human Connectome Project (HCP), where registration of anatomical MRI to dMRI is more precise. The method is then able to predict a tissue segmentation directly from new dMRI data, including data collected with different acquisition protocols, without requiring anatomical data and inter-modality registration. We train a convolutional neural network (CNN) to learn a tissue segmentation model using a novel augmented target loss function designed to improve accuracy in regions of tissue boundary. To further improve accuracy, our method adds diffusion kurtosis imaging (DKI) parameters that characterize non-Gaussian water molecule diffusion to the conventional diffusion tensor imaging parameters. The DKI parameters are calculated from the recently proposed mean-kurtosis-curve method that corrects implausible DKI parameter values and provides additional features that discriminate between tissue types. We demonstrate high tissue segmentation accuracy on HCP data, and also when applying the HCP-trained model on dMRI data from other acquisitions with lower resolution and fewer gradient directions.

Objective Ventricle Segmentation in Brain CT with Ischemic Stroke Based on Anatomical Knowledge

Ventricle segmentation is a challenging technique for the development of detection system of ischemic stroke in computed tomography (CT), as ischemic stroke regions are adjacent to the brain ventricle with similar intensity. To address this problem, we developed an objective segmentation system of brain ventricle in CT. The intensity distribution of the ventricle was estimated based on clustering technique, connectivity, and domain knowledge, and the initial ventricle segmentation results were then obtained. To exclude the stroke regions from initial segmentation, a combined segmentation strategy was proposed, which is composed of three different schemes: (1) the largest three-dimensional (3D) connected component was considered as the ventricular region; (2) the big stroke areas were removed by the image difference methods based on searching optimal threshold values; (3) the small stroke regions were excluded by the adaptive template algorithm. The proposed method was evaluated on 50 cases of patients with ischemic stroke. The mean Dice, sensitivity, specificity, and root mean squared error were 0.9447, 0.969, 0.998, and 0.219 mm, respectively. This system can offer a desirable performance. Therefore, the proposed system is expected to bring insights into clinic research and the development of detection system of ischemic stroke in CT.

Functional consequences of neurite orientation dispersion and density in humans across the adult lifespan

As humans age, a characteristic pattern of widespread neocortical dendritic disruption coupled with compensatory effects in hippocampus and other subcortical structures is shown in postmortem investigations. It is now possible to address age-related effects on gray matter (GM) neuritic organization and density in humans using multishell diffusion-weighted MRI and the neurite-orientation dispersion and density imaging (NODDI) model. In 45 healthy individuals across the adult lifespan (21-84 years), we used a multishell diffusion imaging and the NODDI model to assess the intraneurite volume fraction and neurite orientation-dispersion index (ODI) in GM tissues. We also determined the functional correlates of variations in GM microstructure by obtaining resting-state fMRI and behavioral data. We found a significant age-related deficit in neocortical ODI (most prominently in frontoparietal regions), whereas increased ODI was observed in hippocampus and cerebellum with advancing age. Neocortical ODI outperformed cortical thickness and white matter fractional anisotropy for the prediction of chronological age in the same individuals. Higher GM ODI sampled from resting-state networks with known age-related susceptibility (default mode and visual association networks) was associated with increased functional connectivity of these networks, whereas the task-positive networks tended to show no association or even decreased connectivity. Frontal pole ODI mediated the negative relationship of age with executive function, whereas hippocampal ODI mediated the positive relationship of age with executive function. Our in vivo findings align very closely with the postmortem data and provide evidence for vulnerability and compensatory neural mechanisms of aging in GM microstructure that have functional and cognitive impact in vivo.

Improvement of partial volume segmentation for brain tissue on diffusion tensor images using multiple-tensor estimation

To improve evaluations of cortical and subcortical diffusivity in neurological diseases, it is necessary to improve the accuracy of brain diffusion tensor imaging (DTI) data segmentation. The conventional partial volume segmentation method fails to classify voxels with multiple white matter (WM) fiber orientations such as fiber-crossing regions. Our purpose was to improve the performance of segmentation by taking into account the partial volume effects due to both multiple tissue types and multiple WM fiber orientations. We quantitatively evaluated the overall performance of the proposed method using digital DTI phantom data. Moreover, we applied our method to human DTI data, and compared our results with those of a conventional method. In the phantom experiments, the conventional method and proposed method yielded almost the same root mean square error (RMSE) for gray matter (GM) and cerebrospinal fluid (CSF), while the RMSE in the proposed method was smaller than that in the conventional method for WM. The volume overlap measures between our segmentation results and the ground truth of the digital phantom were more than 0.8 in all three tissue types, and were greater than those in the conventional method. In visual comparisons for human data, the WM/GM/CSF regions obtained using our method were in better agreement with the corresponding regions depicted in the structural image than those obtained using the conventional method. The results of the digital phantom experiment and human data demonstrated that our method improved accuracy in the segmentation of brain tissue data on DTI compared to the conventional method.

Integration of sparse multi-modality representation and anatomical constraint for isointense infant brain MR image segmentation

Segmentation of infant brain MR images is challenging due to poor spatial resolution, severe partial volume effect, and the ongoing maturation and myelination processes. During the first year of life, the brain image contrast between white and gray matters undergoes dramatic changes. In particular, the image contrast inverses around 6-8months of age, where the white and gray matter tissues are isointense in T1 and T2 weighted images and hence exhibit the extremely low tissue contrast, posing significant challenges for automated segmentation. In this paper, we propose a general framework that adopts sparse representation to fuse the multi-modality image information and further incorporate the anatomical constraints for brain tissue segmentation. Specifically, we first derive an initial segmentation from a library of aligned images with ground-truth segmentations by using sparse representation in a patch-based fashion for the multi-modality T1, T2 and FA images. The segmentation result is further iteratively refined by integration of the anatomical constraint. The proposed method was evaluated on 22 infant brain MR images acquired at around 6months of age by using a leave-one-out cross-validation, as well as other 10 unseen testing subjects. Our method achieved a high accuracy for the Dice ratios that measure the volume overlap between automated and manual segmentations, i.e., 0.889±0.008 for white matter and 0.870±0.006 for gray matter.

An active contour model for medical image segmentation with application to brain CT image

PURPOSE

Cerebrospinal fluid (CSF) segmentation in computed tomography (CT) is a key step in computer-aided detection (CAD) of acute ischemic stroke. Because of image noise, low contrast and intensity inhomogeneity, CSF segmentation has been a challenging task. A region-based active contour model, which is insensitive to contour initialization and robust to intensity inhomogeneity, was developed for segmenting CSF in brain CT images.

METHODS

The energy function of the region-based active contour model is composed of a range domain kernel function, a space domain kernel function, and an edge indicator function. By minimizing the energy function, the region of edge elements of the target could be automatically identified in images with less dependence on initial contours. The energy function was optimized by means of the deepest descent method with a level set framework. An overlap rate between segmentation results and the reference standard was used to assess the segmentation accuracy. The authors evaluated the performance of the proposed method on both synthetic data and real brain CT images. They also compared the performance level of our method to those of region-scalable fitting (RSF) and global convex segment (GCS) models.

RESULTS

For the experiment of CSF segmentation in 67 brain CT images, their method achieved an average overlap rate of 66% compared to the average overlap rates of 16% and 46% from the RSF model and the GCS model, respectively.

CONCLUSIONS

Their region-based active contour model has the ability to achieve accurate segmentation results in images with high noise level and intensity inhomogeneity. Therefore, their method has great potential in the segmentation of medical images and would be useful for developing CAD schemes for acute ischemic stroke in brain CT images.

Diffusion abnormalities in temporal lobes of children with temporal lobe epilepsy: a preliminary diffusional kurtosis imaging study and comparison with diffusion tensor imaging

In this preliminary study, we aimed to investigate the abnormalities of water diffusion in children with temporal lobe epilepsy (TLE). Eight children with unilateral TLE (according to electroencephalography, EEG) and eight age- and sex-matched controls were recruited. Diffusion tensor imaging (DTI)/diffusional kurtosis imaging (DKI) acquisitions were performed. Radial diffusivity (λ(⊥)), axial diffusivity (λ(∥)), mean diffusivity (MD) and fractional anisotropy (FA) maps were calculated for both DTI and DKI, and radial kurtosis (K(⊥)), axial kurtosis (K(∥)) and mean kurtosis (MK) maps were calculated for DKI only. Mann-Whitney test showed that, for white matter in the temporal lobe, DKI-derived λ(∥) , MD and K(∥) were significantly different in bilateral temporal lobes and EEG-abnormal and EEG-normal sides of the temporal lobe between patients and controls, whereas DTI showed no abnormalities. For gray matter, DKI detected significantly higher MD and MK in the same three comparisons, whereas DTI detected abnormalities only in the comparison between bilateral temporal lobes and between EEG-normal sides in cases and left-right matched sides in controls. No significant difference was observed between EEG-abnormal and EEG-normal sides in cases. These preliminary results indicate that DKI is more sensitive than DTI for the detection of diffusion abnormalities in the temporal lobes of children with TLE, even when EEG signals are normal. These findings pave the way for the application of DKI for in-depth studies on TLE in children.

Diffusion MR imaging: basic principles

From their origin as simple techniques primarily used for detecting acute cerebral ischemia, diffusion MR imaging techniques have rapidly evolved into a versatile set of tools that provide the only noninvasive means of characterizing brain microstructure and connectivity, becoming a mainstay of both clinical and investigational brain MR imaging. In this article, the basic principles required for understanding diffusion MR imaging techniques are reviewed with clinical neuroradiologists in mind.