Rapid fully automatic segmentation of subcortical brain structures by shape-constrained surface adaptation

Brain MR image simulation for deep learning based medical image analysis networks

BACKGROUND AND OBJECTIVE

As large sets of annotated MRI data are needed for training and validating deep learning based medical image analysis algorithms, the lack of sufficient annotated data is a critical problem. A possible solution is the generation of artificial data by means of physics-based simulations. Existing brain simulation data is limited in terms of anatomical models, tissue classes, fixed tissue characteristics, MR sequences and overall realism.

METHODS

We propose a realistic simulation framework by incorporating patient-specific phantoms and Bloch equations-based analytical solutions for fast and accurate MRI simulations. A large number of labels are derived from open-source high-resolution T1w MRI data using a fully automated brain classification tool. The brain labels are taken as ground truth (GT) on which MR images are simulated using our framework. Moreover, we demonstrate that the T1w MR images generated from our framework along with GT annotations can be utilized directly to train a 3D brain segmentation network. To evaluate our model further on larger set of real multi-source MRI data without GT, we compared our model to existing brain segmentation tools, FSL-FAST and SynthSeg.

RESULTS

Our framework generates 3D brain MRI for variable anatomy, sequence, contrast, SNR and resolution. The brain segmentation network for WM/GM/CSF trained only on T1w simulated data shows promising results on real MRI data from MRBrainS18 challenge dataset with a Dice scores of 0.818/0.832/0.828. On OASIS data, our model exhibits a close performance to FSL, both qualitatively and quantitatively with a Dice scores of 0.901/0.939/0.937.

CONCLUSIONS

Our proposed simulation framework is the initial step towards achieving truly physics-based MRI image generation, providing flexibility to generate large sets of variable MRI data for desired anatomy, sequence, contrast, SNR, and resolution. Furthermore, the generated images can effectively train 3D brain segmentation networks, mitigating the reliance on real 3D annotated data.

Shape-constrained deformable brain segmentation: Methods and quantitative validation

MRI-guided neuro interventions require rapid, accurate, and reproducible segmentation of anatomical brain structures for identification of targets during surgical procedures and post-surgical evaluation of intervention efficiency. Segmentation algorithms must be validated and cleared for clinical use. This work introduces a methodology for shape-constrained deformable brain segmentation, describes the quantitative validation used for its clinical clearance, and presents a comparison with manual expert segmentation and FreeSurfer, an open source software for neuroimaging data analysis. ClearPoint Maestro is software for fully-automatic brain segmentation from T1-weighted MRI that combines a shape-constrained deformable brain model with voxel-wise tissue segmentation within the cerebral hemispheres and the cerebellum. The performance of the segmentation was validated in terms of accuracy and reproducibility. Segmentation accuracy was evaluated with respect to training data and independently traced ground truth. Segmentation reproducibility was quantified and compared with manual expert segmentation and FreeSurfer. Quantitative reproducibility analysis indicates superior performance compared to both manual expert segmentation and FreeSurfer. The shape-constrained methodology results in accurate and highly reproducible segmentation. Inherent point based-correspondence provides consistent target identification ideal for MRI-guided neuro interventions.

Multi-modal brain tumor segmentation via conditional synthesis with Fourier domain adaptation

Accurate brain tumor segmentation is critical for diagnosis and treatment planning, whereby multi-modal magnetic resonance imaging (MRI) is typically used for analysis. However, obtaining all required sequences and expertly labeled data for training is challenging and can result in decreased quality of segmentation models developed through automated algorithms. In this work, we examine the possibility of employing a conditional generative adversarial network (GAN) approach for synthesizing multi-modal images to train deep learning-based neural networks aimed at high-grade glioma (HGG) segmentation. The proposed GAN is conditioned on auxiliary brain tissue and tumor segmentation masks, allowing us to attain better accuracy and control of tissue appearance during synthesis. To reduce the domain shift between synthetic and real MR images, we additionally adapt the low-frequency Fourier space components of synthetic data, reflecting the style of the image, to those of real data. We demonstrate the impact of Fourier domain adaptation (FDA) on the training of 3D segmentation networks and attain significant improvements in both the segmentation performance and prediction confidence. Similar outcomes are seen when such data is used as a training augmentation alongside the available real images. In fact, experiments on the BraTS2020 dataset reveal that models trained solely with synthetic data exhibit an improvement of up to 4% in Dice score when using FDA, while training with both real and FDA-processed synthetic data through augmentation results in an improvement of up to 5% in Dice compared to using real data alone. This study highlights the importance of considering image frequency in generative approaches for medical image synthesis and offers a promising approach to address data scarcity in medical imaging segmentation.

A deep learning toolbox for automatic segmentation of subcortical limbic structures from MRI images

A tool was developed to automatically segment several subcortical limbic structures (nucleus accumbens, basal forebrain, septal nuclei, hypothalamus without mammillary bodies, the mammillary bodies, and fornix) using only a T1-weighted MRI as input. This tool fills an unmet need as there are few, if any, publicly available tools to segment these clinically relevant structures. A U-Net with spatial, intensity, contrast, and noise augmentation was trained using 39 manually labeled MRI data sets. In general, the Dice scores, true positive rates, false discovery rates, and manual-automatic volume correlation were very good relative to comparable tools for other structures. A diverse data set of 698 subjects were segmented using the tool; evaluation of the resulting labelings showed that the tool failed in less than 1% of cases. Test-retest reliability of the tool was excellent. The automatically segmented volume of all structures except mammillary bodies showed effectiveness at detecting either clinical AD effects, age effects, or both. This tool will be publicly released with FreeSurfer (surfer.nmr.mgh.harvard.edu/fswiki/ScLimbic). Together with the other cortical and subcortical limbic segmentations, this tool will allow FreeSurfer to provide a comprehensive view of the limbic system in an automated way.

Differences Between MR Brain Region Segmentation Methods: Impact on Single-Subject Analysis

For the segmentation of magnetic resonance brain images into anatomical regions, numerous fully automated methods have been proposed and compared to reference segmentations obtained manually. However, systematic differences might exist between the resulting segmentations, depending on the segmentation method and underlying brain atlas. This potentially results in sensitivity differences to disease and can further complicate the comparison of individual patients to normative data. In this study, we aim to answer two research questions: 1) to what extent are methods interchangeable, as long as the same method is being used for computing normative volume distributions and patient-specific volumes? and 2) can different methods be used for computing normative volume distributions and assessing patient-specific volumes? To answer these questions, we compared volumes of six brain regions calculated by five state-of-the-art segmentation methods: Erasmus MC (EMC), FreeSurfer (FS), geodesic information flows (GIF), multi-atlas label propagation with expectation–maximization (MALP-EM), and model-based brain segmentation (MBS). We applied the methods on 988 non-demented (ND) subjects and computed the correlation (PCC-v) and absolute agreement (ICC-v) on the volumes. For most regions, the PCC-v was good (>0.75), indicating that volume differences between methods in ND subjects are mainly due to systematic differences. The ICC-v was generally lower, especially for the smaller regions, indicating that it is essential that the same method is used to generate normative and patient data. To evaluate the impact on single-subject analysis, we also applied the methods to 42 patients with Alzheimer’s disease (AD). In the case where the normative distributions and the patient-specific volumes were calculated by the same method, the patient’s distance to the normative distribution was assessed with the z-score. We determined the diagnostic value of this z-score, which showed to be consistent across methods. The absolute agreement on the AD patients’ z-scores was high for regions of thalamus and putamen. This is encouraging as it indicates that the studied methods are interchangeable for these regions. For regions such as the hippocampus, amygdala, caudate nucleus and accumbens, and globus pallidus, not all method combinations showed a high ICC-z. Whether two methods are indeed interchangeable should be confirmed for the specific application and dataset of interest.

Morphological analysis of subcortical structures for assessment of cognitive dysfunction in Parkinson's disease using multi-atlas based segmentation

Cognitive impairment in Parkinson's Disease (PD) is the most prevalent non-motor symptom that requires analysis of anatomical associations to cognitive decline in PD. The objective of this study is to analyse the morphological variations of the subcortical structures to assess cognitive dysfunction in PD. In this study, T1 MR images of 58 Healthy Control (HC) and 135 PD subjects categorised as 91 Cognitively normal PD (NC-PD), 25 PD with Mild Cognitive Impairment (PD-MCI) and 19 PD with Dementia (PD-D) subjects, based on cognitive scores are utilised. The 132 anatomical regions are segmented using spatially localized multi-atlas model and volumetric analysis is carried out. The morphological alterations through textural features are captured to differentiate among the HC and PD subjects under different cognitive domains. The volumetric differences in the segmented subcortical structures of accumbens, amygdala, caudate, putamen and thalamus are able to predict cognitive impairment in PD. The volumetric distribution of the subcortical structures in PD-MCI subjects exhibit an overlap with the HC group due to lack of spatial specificity in their atrophy levels. The 3D GLCM features extracted from the significant subcortical structures could discriminate HC, NC-PD, PD-MCI and PD-D subjects with better classification accuracies. The disease related atrophy levels of the subcortical structures captured through morphological analysis provide sensitive evaluation of cognitive impairment in PD.

Brain Volumetric Alterations in Preclinical HIV-Associated Neurocognitive Disorder Using Automatic Brain Quantification and Segmentation Tool

Purpose

This study aimed to determine if people living with HIV (PLWH) in preclinical human immunodeficiency virus (HIV)-associated neurocognitive disorder (HAND), with no clinical symptoms and without decreased daily functioning, suffer from brain volumetric alterations and its patterns.

Method

Fifty-nine male PLWH at the HAND preclinical stage were evaluated, including 19 subjects with asymptomatic neurocognitive impairment (ANI), 17 subjects with cognitive abnormality that does not reach ANI (Not reach ANI), and 23 subjects with cognitive integrity. Moreover, 23 healthy volunteers were set as the seronegative normal controls (NCs). These individuals underwent sagittal three-dimensional T₁-weighted imaging (3D T₁WI). Quantified data and volumetric measures of brain structures were automatically segmented and extracted using AccuBrain^®. In addition, the multiple linear regression analysis was performed to analyze the relationship of volumes of brain structures and clinical variables in preclinical HAND, and the correlations of the brain volume parameters with different cognitive function states were assessed by Pearson's correlation analysis.

Results

The significant difference was shown in the relative volumes of the ventricular system, bilateral lateral ventricle, thalamus, caudate, and left parietal lobe gray matter between the preclinical HAND and NCs. Furthermore, the relative volumes of the bilateral thalamus in preclinical HAND were negatively correlated with attention/working memory (left: r = -0.271, p = 0.042; right: r = -0.273, p = 0.040). Higher age was associated with increased relative volumes of the bilateral lateral ventricle and ventricular system and reduced relative volumes of the left thalamus and parietal lobe gray matter. The lower CD4⁺/CD8⁺ ratio was associated with increased relative volumes of the left lateral ventricle and ventricular system. Longer disease course was associated with increased relative volumes of the bilateral thalamus. No significant difference was found among preclinical HAND subgroups in all indices, and the difference between the individual groups (Not reach ANI and Cognitive integrity groups) and NCs was also insignificant. However, there was a significant difference between ANI and NCs in the relative volumes of the bilateral caudate and lateral ventricle.

Conclusion

Male PLWH at the HAND preclinical stage suffer from brain volumetric alterations. AccuBrain^® provides potential value in evaluating HIV-related neurocognitive dysfunction.

QSM reconstruction challenge 2.0: A realistic in silico head phantom for MRI data simulation and evaluation of susceptibility mapping procedures

PURPOSE

To create a realistic in silico head phantom for the second QSM reconstruction challenge and for future evaluations of processing algorithms for QSM.

METHODS

We created a digital whole-head tissue property phantom by segmenting and postprocessing high-resolution (0.64 mm isotropic), multiparametric MRI data acquired at 7 T from a healthy volunteer. We simulated the steady-state magnetization at 7 T using a Bloch simulator and mimicked a Cartesian sampling scheme through Fourier-based processing. Computer code for generating the phantom and performing the MR simulation was designed to facilitate flexible modifications of the phantom in the future, such as the inclusion of pathologies as well as the simulation of a wide range of acquisition protocols. Specifically, the following parameters and effects were implemented: TR and TE, voxel size, background fields, and RF phase biases. Diffusion-weighted imaging phantom data are provided, allowing future investigations of tissue-microstructure effects in phase and QSM algorithms.

RESULTS

The brain part of the phantom featured realistic morphology with spatial variations in relaxation and susceptibility values similar to the in vivo setting. We demonstrated some of the phantom's properties, including the possibility of generating phase data with nonlinear evolution over TE due to partial-volume effects or complex distributions of frequency shifts within the voxel.

CONCLUSION

The presented phantom and computer programs are publicly available and may serve as a ground truth in future assessments of the faithfulness of quantitative susceptibility reconstruction algorithms.

Quantitative comparison of subcortical and ventricular volumetry derived from MPRAGE and MP2RAGE images using different brain morphometry software

OBJECTIVE

In brain volume assessment with MR imaging, it is of interest to know the effects of the pulse sequence and software used, to determine whether they provide equivalent data. The aim of this study was to compare cross-sectional volumes of subcortical and ventricular structures and their repeatability derived from MP2RAGE and MPRAGE images using MorphoBox, and FIRST or ALVIN.

MATERIALS AND METHODS

MPRAGE and MP2RAGE T1-weighted images were obtained from 24 healthy volunteers. Back-to-back scans were performed in 12 of them. Volumes, coefficients of variation, concordance, and correlations were determined.

RESULTS

Significant differences were found for volumes derived from MorphoBox and FIRST. Ventricular volumes determined by MorphoBox and ALVIN were similar. Differences between volumes obtained using MPRAGE and MP2RAGE were significant for a few regions. Coefficients of variation, ranged from 0.2 to 9.1%, showed a significant inverse correlation with the mean volume. There was a correlation between volume measures, but agreement was rated as poor for most regions.

CONCLUSION

MP2RAGE sequences and MorphoBox are valid options for assessing subcortical and ventricular volumes, in the same way as MPRAGE and FIRST or ALVIN, accepted tools for clinical research. However, caution is needed when comparing volumes obtained with different tools.

An MRI-based strategy for differentiation of frontotemporal dementia and Alzheimer's disease

BACKGROUND

The differential diagnosis of frontotemporal dementia (FTD) and Alzheimer's disease (AD) is difficult due to the overlaps of clinical symptoms. Structural magnetic resonance imaging (sMRI) presents distinct brain atrophy and potentially helps in their differentiation. In this study, we aim at deriving a novel integrated index by leveraging the volumetric measures in brain regions with significant difference between AD and FTD and developing an MRI-based strategy for the differentiation of FTD and AD.

METHODS

In this study, the data were acquired from three different databases, including 47 subjects with FTD, 47 subjects with AD, and 47 normal controls in the NACC database; 50 subjects with AD in the ADNI database; and 50 subjects with FTD in the FTLDNI database. The MR images of all subjects were automatically segmented, and the brain atrophy, including the AD resemblance atrophy index (AD-RAI), was quantified using AccuBrain®. A novel MRI index, named the frontotemporal dementia index (FTDI), was derived as the ratio between the weighted sum of the volumetric indexes in "FTD dominant" structures over that obtained from "AD dominant" structures. The weights and the identification of "FTD/AD dominant" structures were acquired from the statistical analysis of NACC data. The differentiation performance of FTDI was validated using independent data from ADNI and FTLDNI databases.

RESULTS

AD-RAI is a proven imaging biomarker to identify AD and FTD from NC with significantly higher values (p < 0.001 and AUC = 0.88) as we reported before, while no significant difference was found between AD and FTD (p = 0.647). FTDI showed excellent accuracy in identifying FTD from AD (AUC = 0.90; SEN = 89%, SPE = 75% with threshold value = 1.08). The validation using independent data from ADNI and FTLDNI datasets also confirmed the efficacy of FTDI (AUC = 0.93; SEN = 96%, SPE = 70% with threshold value = 1.08).

CONCLUSIONS

Brain atrophy in AD, FTD, and normal elderly shows distinct patterns. In addition to AD-RAI that is designed to detect abnormal brain atrophy in dementia, a novel index specific to FTD is proposed and validated. By combining AD-RAI and FTDI, an MRI-based decision strategy was further proposed as a promising solution for the differential diagnosis of AD and FTD in clinical practice.

Patient-Specific Sensor Registration for Electrical Source Imaging Using a Deformable Head Model

OBJECTIVE

Electrical source imaging of brain activity is most accurate when using individualized bioelectric head models. Constructing these models requires identifying electrode positions on the scalp surface. Current methods such as photogrammetry involve significant user interaction that limits integration in clinical workflows. This work introduces and validates a new, fully-automatic method for sensor registration.

METHODS

Average electrode coordinates are registered to the mean scalp mesh of a shape-constrained deformable head model used for tissue segmentation. Patient-specific electrode positions can be identified on the deformed scalp surface using point-based correspondence after model adaptation.

RESULTS

The performance of the proposed method for sensor registration is evaluated with simulated and real data. Electrode variability is quantified for a photogrammetry-based solution and compared against the proposed sensor registration.

CONCLUSION

A fully-automated model-based approach can identify electrode locations with similar accuracy as a current state-of-the-art photogrammetry system.

SIGNIFICANCE

The new method for sensor registration presented in this work is rapid and fully automatic. It eliminates any user dependent inaccuracy introduced in sensor registration and ensures reproducible results. More importantly, it can more easily be integrated in clinical workflows, enabling broader adoption of electrical source imaging technologies.

Quantifying deep grey matter atrophy using automated segmentation approaches: A systematic review of structural MRI studies

The deep grey matter (DGM) nuclei of the brain play a crucial role in learning, behaviour, cognition, movement and memory. Although automated segmentation strategies can provide insight into the impact of multiple neurological conditions affecting these structures, such as Multiple Sclerosis (MS), Huntington's disease (HD), Alzheimer's disease (AD), Parkinson's disease (PD) and Cerebral Palsy (CP), there are a number of technical challenges limiting an accurate automated segmentation of the DGM. Namely, the insufficient contrast of T1 sequences to completely identify the boundaries of these structures, as well as the presence of iso-intense white matter lesions or extensive tissue loss caused by brain injury. Therefore in this systematic review, 269 eligible studies were analysed and compared to determine the optimal approaches for addressing these technical challenges. The automated approaches used among the reviewed studies fall into three broad categories, atlas-based approaches focusing on the accurate alignment of atlas priors, algorithmic approaches which utilise intensity information to a greater extent, and learning-based approaches that require an annotated training set. Studies that utilise freely available software packages such as FIRST, FreeSurfer and LesionTOADS were also eligible, and their performance compared. Overall, deep learning approaches achieved the best overall performance, however these strategies are currently hampered by the lack of large-scale annotated data. Improving model generalisability to new datasets could be achieved in future studies with data augmentation and transfer learning. Multi-atlas approaches provided the second-best performance overall, and may be utilised to construct a "silver standard" annotated training set for deep learning. To address the technical challenges, providing robustness to injury can be improved by using multiple channels, highly elastic diffeomorphic transformations such as LDDMM, and by following atlas-based approaches with an intensity driven refinement of the segmentation, which has been done with the Expectation Maximisation (EM) and level sets methods. Accounting for potential lesions should be achieved with a separate lesion segmentation approach, as in LesionTOADS. Finally, to address the issue of limited contrast, R2*, T2* and QSM sequences could be used to better highlight the DGM due to its higher iron content. Future studies could look to additionally acquire these sequences by retaining the phase information from standard structural scans, or alternatively acquiring these sequences for only a training set, allowing models to learn the "improved" segmentation from T1-sequences alone.

Automatic Pathological Lung Segmentation in Low-Dose CT Image Using Eigenspace Sparse Shape Composition

The segmentation of lungs with severe pathology is a nontrivial problem in the clinical application. Due to complex structures, pathological changes, individual differences, and low image quality, accurate lung segmentation in clinical 3-D computed tomography (CT) images is still a challenging task. To overcome these problems, a novel dictionary-based approach is introduced to automatically segment pathological lungs in 3-D low-dose CT images. Sparse shape composition is integrated with the eigenvector space shape prior model, called eigenspace sparse shape composition, to reduce local shape reconstruction error caused by the weak and misleading appearance prior information. To initialize the shape model, a landmark recognition method based on discriminative appearance dictionary is introduced to handle lesions and local details. Furthermore, a new vertex search strategy based on the gradient vector flow field is also proposed to drive the shape deformation to the target boundary. The proposed algorithm is tested on 78 3-D low-dose CT images with lung tumors. Compared to the state-of-the-art methods, the proposed approach can robustly and accurately detect pathological lung surface.

Data consistency-driven determination of B 0 -fluctuations in gradient-echo MRI

PURPOSE

Introduce a method to estimate B 0 -fluctuations based on the analysis of raw k-space data, without sequence modifications or external hardware, and correct for their detrimental effects in gradient-echo MRI.

THEORY

Inconsistencies in multi-channel raw k-space data can be used to estimate B 0 -fluctuations by exploiting coil-sensitivity information.

METHODS

The proposed method, dubbed consistency navigation, is used to extract B 0 -fluctuations from T 2 * -weighted 3D gradient-echo data. These results are compared with the results from an MR phase navigator and respiratory bellows. The spatial variation of the B 0 -fluctuation amplitude is derived using the sensitivity maps of the coil array and compared with direct measurements based on dynamic 2D gradient-echo data.

RESULTS

B 0 -fluctuations derived from the consistency navigator and MR phase navigator are highly correlated. Images corrected for these fluctuations show marked improvements in homogeneity and tissue delineation. The spatial variation of the B 0 -fluctuation amplitude follows closely the variation directly measured from time-resolved 2D scans.

CONCLUSIONS

Based on the consistency navigator, an accurate estimation of the spatiotemporal characteristics of B 0 -fluctuations and correction of T 2 * -weighted images has been demonstrated.

Motion artifacts in standard clinical setting obscure disease-specific differences in quantitative susceptibility mapping

As quantitative susceptibility mapping (QSM) is maturing, more clinical applications are being explored. With this comes the question whether QSM is sufficiently robust and reproducible to be directly used in a clinical setting where patients are possibly not cooperative and/or unable to suppress involuntary movements sufficiently. Twenty-nine patients with Alzheimer's disease, 31 patients with mild cognitive impairment and 41 healthy controls were scanned on a 3 T scanner, including a multi-echo gradient-echo sequence for QSM and an inversion-prepared segmented gradient-echo sequence (T1-TFE, MPRAGE). The severity of motion artifacts (excessive/strong/noticeable/invisible) was categorized via visual inspection by two independent raters. Quantitative susceptibility was reconstructed using 'joint background-field removal and segmentation-enhanced dipole inversion', based on segmented subcortical gray-matter regions, as well as using 'morphology enabled dipole inversion'. Statistical analysis of the susceptibility maps was performed per region. A large fraction of the data showed motion artifacts, visible in both magnitude images and susceptibility maps. No statistically significant susceptibility differences were found between groups including motion-affected data. Considering only subjects without visible motion, significant susceptibility differences were observed in caudate nucleus as well as in putamen. Motion-effects can obscure statistically significant differences in QSM between patients and controls. Additional measures to restrict and/or compensate for subject motion should be taken for QSM in standard clinical settings to avoid risk of false findings.