Online resource for validation of brain segmentation methods

Privacy preserving image registration

Image registration is a key task in medical imaging applications, allowing to represent medical images in a common spatial reference frame. Current approaches to image registration are generally based on the assumption that the content of the images is usually accessible in clear form, from which the spatial transformation is subsequently estimated. This common assumption may not be met in practical applications, since the sensitive nature of medical images may ultimately require their analysis under privacy constraints, preventing to openly share the image content. In this work, we formulate the problem of image registration under a privacy preserving regime, where images are assumed to be confidential and cannot be disclosed in clear. We derive our privacy preserving image registration framework by extending classical registration paradigms to account for advanced cryptographic tools, such as secure multi-party computation and homomorphic encryption, that enable the execution of operations without leaking the underlying data. To overcome the problem of performance and scalability of cryptographic tools in high dimensions, we propose several techniques to optimize the image registration operations by using gradient approximations, and by revisiting the use of homomorphic encryption trough packing, to allow the efficient encryption and multiplication of large matrices. We focus on registration methods of increasing complexity, including rigid, affine, and non-linear registration based on cubic splines or diffeomorphisms parameterized by time-varying velocity fields. In all these settings, we demonstrate how the registration problem can be naturally adapted for accounting to privacy-preserving operations, and illustrate the effectiveness of PPIR on a variety of registration tasks.

Dimensionality Reduction Hybrid U-Net for Brain Extraction in Magnetic Resonance Imaging

In various applications, such as disease diagnosis, surgical navigation, human brain atlas analysis, and other neuroimage processing scenarios, brain extraction is typically regarded as the initial stage in MRI image processing. Whole-brain semantic segmentation algorithms, such as U-Net, have demonstrated the ability to achieve relatively satisfactory results even with a limited number of training samples. In order to enhance the precision of brain semantic segmentation, various frameworks have been developed, including 3D U-Net, slice U-Net, and auto-context U-Net. However, the processing methods employed in these models are relatively complex when applied to 3D data models. In this article, we aim to reduce the complexity of the model while maintaining appropriate performance. As an initial step to enhance segmentation accuracy, the preprocessing extraction of full-scale information from magnetic resonance images is performed with a cluster tool. Subsequently, three multi-input hybrid U-Net model frameworks are tested and compared. Finally, we propose utilizing a fusion of two-dimensional segmentation outcomes from different planes to attain improved results. The performance of the proposed framework was tested using publicly accessible benchmark datasets, namely LPBA40, in which we obtained Dice overlap coefficients of 98.05%. Improvement was achieved via our algorithm against several previous studies.

Single Ventricle Reconstruction III: Brain Connectome and Neurodevelopmental Outcomes: Design, Recruitment, and Technical Challenges of a Multicenter, Observational Neuroimaging Study

Patients with hypoplastic left heart syndrome who have been palliated with the Fontan procedure are at risk for adverse neurodevelopmental outcomes, lower quality of life, and reduced employability. We describe the methods (including quality assurance and quality control protocols) and challenges of a multi-center observational ancillary study, SVRIII (Single Ventricle Reconstruction Trial) Brain Connectome. Our original goal was to obtain advanced neuroimaging (Diffusion Tensor Imaging and Resting-BOLD) in 140 SVR III participants and 100 healthy controls for brain connectome analyses. Linear regression and mediation statistical methods will be used to analyze associations of brain connectome measures with neurocognitive measures and clinical risk factors. Initial recruitment challenges occurred that were related to difficulties with: (1) coordinating brain MRI for participants already undergoing extensive testing in the parent study, and (2) recruiting healthy control subjects. The COVID-19 pandemic negatively affected enrollment late in the study. Enrollment challenges were addressed by: (1) adding additional study sites, (2) increasing the frequency of meetings with site coordinators, and (3) developing additional healthy control recruitment strategies, including using research registries and advertising the study to community-based groups. Technical challenges that emerged early in the study were related to the acquisition, harmonization, and transfer of neuroimages. These hurdles were successfully overcome with protocol modifications and frequent site visits that involved human and synthetic phantoms.

Single Ventricle Reconstruction III: Brain Connectome and Neurodevelopmental Outcomes: Design, Recruitment, and Technical Challenges of a Multicenter, Observational Neuroimaging Study

Patients with hypoplastic left heart syndrome who have been palliated with the Fontan procedure are at risk for adverse neurodevelopmental outcomes, lower quality of life, and reduced employability. We describe the methods (including quality assurance and quality control protocols) and challenges of a multi-center observational ancillary study, SVRIII (Single Ventricle Reconstruction Trial) Brain Connectome. Our original goal was to obtain advanced neuroimaging (Diffusion Tensor Imaging and Resting-BOLD) in 140 SVR III participants and 100 healthy controls for brain connectome analyses. Linear regression and mediation statistical methods will be used to analyze associations of brain connectome measures with neurocognitive measures and clinical risk factors. Initial recruitment challenges occurred related to difficulties with: 1) coordinating brain MRI for participants already undergoing extensive testing in the parent study, and 2) recruiting healthy control subjects. The COVID-19 pandemic negatively affected enrollment late in the study. Enrollment challenges were addressed by 1) adding additional study sites, 2) increasing the frequency of meetings with site coordinators and 3) developing additional healthy control recruitment strategies, including using research registries and advertising the study to community-based groups. Technical challenges that emerged early in the study were related to the acquisition, harmonization, and transfer of neuroimages. These hurdles were successfully overcome with protocol modifications and frequent site visits that involved human and synthetic phantoms.

Trial registration number

ClinicalTrials.gov Registration Number: NCT02692443.

Methods of Brain Extraction from Magnetic Resonance Images of Human Head: A Review

Medical images are providing vital information to aid physicians in diagnosing a disease afflicting the organ of a human body. Magnetic resonance imaging is an important imaging modality in capturing the soft tissues of the brain. Segmenting and extracting the brain is essential in studying the structure and pathological condition of brain. There are several methods that are developed for this purpose. Researchers in brain extraction or segmentation need to know the current status of the work that have been done. Such an information is also important for improving the existing method to get more accurate results or to reduce the complexity of the algorithm. In this paper we review the classical methods and convolutional neural network-based deep learning brain extraction methods.

A dual-branch hybrid dilated CNN model for the AI-assisted segmentation of meningiomas in MR images

BACKGROUND AND OBJECTIVE

Treatment for meningiomas usually includes surgical removal, radiation therapy, and chemotherapy. Accurate segmentation of tumors significantly facilitates complete surgical resection and precise radiotherapy, thereby improving patient survival. In this paper, a deep learning model is constructed for magnetic resonance T1-weighted Contrast Enhancement (T1CE) images to develop an automatic processing scheme for accurate tumor segmentation.

METHODS

In this paper, a novel Convolutional Neural Network (CNN) model is proposed for the accurate meningioma segmentation in MR images. It can extract fused features in multi-scale receptive fields of the same feature map based on MR image characteristics of meningiomas. The attention mechanism is added as a helpful addition to the model to optimize the feature information transmission.

RESULTS AND CONCLUSIONS

The results were evaluated on two internal testing sets and one external testing set. Mean Dice Similarity Coefficient (DSC) values of 0.886, 0.851, and 0.874 are demonstrated, respectively. In this paper, a deep learning approach is proposed to segment tumors in T1CE images. Multi-center testing sets validated the effectiveness and generalization of the method. The proposed model demonstrates state-of-the-art tumor segmentation performance.

Automated Detection of Vascular Leakage in Fluorescein Angiography - A Proof of Concept

Purpose

The purpose of this paper was to develop a deep learning algorithm to detect retinal vascular leakage (leakage) in fluorescein angiography (FA) of patients with uveitis and use the trained algorithm to determine clinically notable leakage changes.

Methods

An algorithm was trained and tested to detect leakage on a set of 200 FA images (61 patients) and evaluated on a separate 50-image test set (21 patients). The ground truth was leakage segmentation by two clinicians. The Dice Similarity Coefficient (DSC) was used to measure concordance.

Results

During training, the algorithm achieved a best average DSC of 0.572 (95% confidence interval [CI] = 0.548–0.596). The trained algorithm achieved a DSC of 0.563 (95% CI = 0.543–0.582) when tested on an additional set of 50 images. The trained algorithm was then used to detect leakage on pairs of FA images from longitudinal patient visits. Longitudinal leakage follow-up showed a >2.21% change in the visible retina area covered by leakage (as detected by the algorithm) had a sensitivity and specificity of 90% (area under the curve [AUC] = 0.95) of detecting a clinically notable change compared to the gold standard, an expert clinician's assessment.

Conclusions

This deep learning algorithm showed modest concordance in identifying vascular leakage compared to ground truth but was able to aid in identifying vascular FA leakage changes over time.

Translational Relevance

This is a proof-of-concept study that vascular leakage can be detected in a more standardized way and that tools can be developed to help clinicians more objectively compare vascular leakage between FAs.

Weakly Supervised Skull Stripping of Magnetic Resonance Imaging of Brain Tumor Patients

Automatic brain tumor segmentation is particularly challenging on magnetic resonance imaging (MRI) with marked pathologies, such as brain tumors, which usually cause large displacement, abnormal appearance, and deformation of brain tissue. Despite an abundance of previous literature on learning-based methodologies for MRI segmentation, few works have focused on tackling MRI skull stripping of brain tumor patient data. This gap in literature can be associated with the lack of publicly available data (due to concerns about patient identification) and the labor-intensive nature of generating ground truth labels for model training. In this retrospective study, we assessed the performance of Dense-Vnet in skull stripping brain tumor patient MRI trained on our large multi-institutional brain tumor patient dataset. Our data included pretreatment MRI of 668 patients from our in-house institutional review board-approved multi-institutional brain tumor repository. Because of the absence of ground truth, we used imperfect automatically generated training labels using SPM12 software. We trained the network using common MRI sequences in oncology: T1-weighted with gadolinium contrast, T2-weighted fluid-attenuated inversion recovery, or both. We measured model performance against 30 independent brain tumor test cases with available manual brain masks. All images were harmonized for voxel spacing and volumetric dimensions before model training. Model training was performed using the modularly structured deep learning platform NiftyNet that is tailored toward simplifying medical image analysis. Our proposed approach showed the success of a weakly supervised deep learning approach in MRI brain extraction even in the presence of pathology. Our best model achieved an average Dice score, sensitivity, and specificity of, respectively, 94.5, 96.4, and 98.5% on the multi-institutional independent brain tumor test set. To further contextualize our results within existing literature on healthy brain segmentation, we tested the model against healthy subjects from the benchmark LBPA40 dataset. For this dataset, the model achieved an average Dice score, sensitivity, and specificity of 96.2, 96.6, and 99.2%, which are, although comparable to other publications, slightly lower than the performance of models trained on healthy patients. We associate this drop in performance with the use of brain tumor data for model training and its influence on brain appearance.

Convolutional Neural Net Learning Can Achieve Production-Level Brain Segmentation in Structural Magnetic Resonance Imaging

Deep learning implementations using convolutional neural nets have recently demonstrated promise in many areas of medical imaging. In this article we lay out the methods by which we have achieved consistently high quality, high throughput computation of intra-cranial segmentation from whole head magnetic resonance images, an essential but typically time-consuming bottleneck for brain image analysis. We refer to this output as “production-level” because it is suitable for routine use in processing pipelines. Training and testing with an extremely large archive of structural images, our segmentation algorithm performs uniformly well over a wide variety of separate national imaging cohorts, giving Dice metric scores exceeding those of other recent deep learning brain extractions. We describe the components involved to achieve this performance, including size, variety and quality of ground truth, and appropriate neural net architecture. We demonstrate the crucial role of appropriately large and varied datasets, suggesting a less prominent role for algorithm development beyond a threshold of capability.

DIKA-Nets: Domain-invariant knowledge-guided attention networks for brain skull stripping of early developing macaques

As non-human primates, macaques have a close phylogenetic relationship to human beings and have been proven to be a valuable and widely used animal model in human neuroscience research. Accurate skull stripping (aka. brain extraction) of brain magnetic resonance imaging (MRI) is a crucial prerequisite in neuroimaging analysis of macaques. Most of the current skull stripping methods can achieve satisfactory results for human brains, but when applied to macaque brains, especially during early brain development, the results are often unsatisfactory. In fact, the early dynamic, regionally-heterogeneous development of macaque brains, accompanied by poor and age-related contrast between different anatomical structures, poses significant challenges for accurate skull stripping. To overcome these challenges, we propose a fully-automated framework to effectively fuse the age-specific intensity information and domain-invariant prior knowledge as important guiding information for robust skull stripping of developing macaques from 0 to 36 months of age. Specifically, we generate Signed Distance Map (SDM) and Center of Gravity Distance Map (CGDM) based on the intermediate segmentation results as guidance. Instead of using local convolution, we fuse all information using the Dual Self-Attention Module (DSAM), which can capture global spatial and channel-dependent information of feature maps. To extensively evaluate the performance, we adopt two relatively-large challenging MRI datasets from rhesus macaques and cynomolgus macaques, respectively, with a total of 361 scans from two different scanners with different imaging protocols. We perform cross-validation by using one dataset for training and the other one for testing. Our method outperforms five popular brain extraction tools and three deep-learning-based methods on cross-source MRI datasets without any transfer learning.

Deep Learning in Selected Cancers' Image Analysis-A Survey

Deep learning algorithms have become the first choice as an approach to medical image analysis, face recognition, and emotion recognition. In this survey, several deep-learning-based approaches applied to breast cancer, cervical cancer, brain tumor, colon and lung cancers are studied and reviewed. Deep learning has been applied in almost all of the imaging modalities used for cervical and breast cancers and MRIs for the brain tumor. The result of the review process indicated that deep learning methods have achieved state-of-the-art in tumor detection, segmentation, feature extraction and classification. As presented in this paper, the deep learning approaches were used in three different modes that include training from scratch, transfer learning through freezing some layers of the deep learning network and modifying the architecture to reduce the number of parameters existing in the network. Moreover, the application of deep learning to imaging devices for the detection of various cancer cases has been studied by researchers affiliated to academic and medical institutes in economically developed countries; while, the study has not had much attention in Africa despite the dramatic soar of cancer risks in the continent.

Brain segmentation, spatial censoring, and averaging techniques for optical functional connectivity imaging in mice

Resting-state functional connectivity analysis using optical neuroimaging holds the potential to be a powerful bridge between mouse models of disease and clinical neurologic monitoring. However, analysis techniques specific to optical methods are rudimentary, and algorithms from magnetic resonance imaging are not always applicable to optics. We have developed visual processing tools to increase data quality, improve brain segmentation, and average across sessions with better field-of-view. We demonstrate improved performance using resting-state optical intrinsic signal from normal mice. The proposed methods increase the amount of usable data from neuroimaging studies, improve image fidelity, and should be translatable to human optical neuroimaging systems.

A new approach to automatic fetal brain extraction from MRI using a variational level set method

BACKGROUND AND PURPOSE

Appropriate images extracted from the MRI of mothers' wombs can be of great help in the medical diagnosis of fetal abnormalities. As maternal tissue may appear in such images, affecting visualization of myelination of the fetal brain, it is not possible to use methods routinely used for extraction of adult brains for fetal brains. The aim of the present study was to use a variational level set approach to extract fetal brain from T2-weighted MR images of the womb.

METHODS

Coronal T2-weighted images were acquired using fast MRI protocols (to avoid artifacts). The database includes 105 MR images from eight subjects. After correcting the inhomogeneity of the images, the fetal eyes were located, and from that information, the location of the fetus brain was automatically determined. Then, the variational level set was used for fetus brain extraction. The results were analyzed by a clinical specialist (radiologist) and the similarity (Dice and Jaccard coefficients), sensitivity and specificity were calculated.

RESULTS AND CONCLUSIONS

The means of the statistical analysis for the Dice and Jaccard coefficients, sensitivity and specificity, were 99.56%, 96.89%, 95.71%, and 97.96%, respectively. Thus, extraction of fetal brain from MR images was confirmed, both statistically and visually through cross-validation.

Post-acquisition processing confounds in brain volumetric quantification of white matter hyperintensities

BACKGROUND

Disparate research sites using identical or near-identical magnetic resonance imaging (MRI) acquisition techniques often produce results that demonstrate significant variability regarding volumetric quantification of white matter hyperintensities (WMH) in the aging population. The sources of such variability have not previously been fully explored.

NEW METHOD

3D FLAIR sequences from a group of randomly selected aged subjects were analyzed to identify sources-of-variability in post-acquisition processing that can be problematic when comparing WMH volumetric data across disparate sites. The methods developed focused on standardizing post-acquisition protocol processing methods to develop a protocol with less than 0.5% inter-rater variance.

RESULTS

A series of experiments using standard MRI acquisition sequences explored post-acquisition sources-of-variability in the quantification of WMH volumetric data. Sources-of-variability included: the choice of image center, software suite and version, thresholding selection, and manual editing procedures (when used). Controlling for the identified sources-of-variability led to a protocol with less than 0.5% variability between independent raters in post-acquisition WMH volumetric quantification.

COMPARISON WITH EXISTING METHOD(S)

Post-acquisition processing techniques can introduce an average variance approaching 15% in WMH volume quantification despite identical scan acquisitions. Understanding and controlling for such sources-of-variability can reduce post-acquisition quantitative image processing variance to less than 0.5%.

DISCUSSION

Considerations of potential sources-of-variability in MRI volume quantification techniques and reduction in such variability is imperative to allow for reliable cross-site and cross-study comparisons.

Realistic volumetric-approach to simulate transcranial electric stimulation-ROAST-a fully automated open-source pipeline

OBJECTIVE

Research in the area of transcranial electrical stimulation (TES) often relies on computational models of current flow in the brain. Models are built based on magnetic resonance images (MRI) of the human head to capture detailed individual anatomy. To simulate current flow on an individual, the subject's MRI is segmented, virtual electrodes are placed on this anatomical model, the volume is tessellated into a mesh, and a finite element model (FEM) is solved numerically to estimate the current flow. Various software tools are available for each of these steps, as well as processing pipelines that connect these tools for automated or semi-automated processing. The goal of the present tool-realistic volumetric-approach to simulate transcranial electric simulation (ROAST)-is to provide an end-to-end pipeline that can automatically process individual heads with realistic volumetric anatomy leveraging open-source software and custom scripts to improve segmentation and execute electrode placement.

APPROACH

ROAST combines the segmentation algorithm of SPM12, a Matlab script for touch-up and automatic electrode placement, the finite element mesher iso2mesh and the solver getDP. We compared its performance with commercial FEM software, and SimNIBS, a well-established open-source modeling pipeline.

MAIN RESULTS

The electric fields estimated with ROAST differ little from the results obtained with commercial meshing and FEM solving software. We also do not find large differences between the various automated segmentation methods used by ROAST and SimNIBS. We do find bigger differences when volumetric segmentation are converted into surfaces in SimNIBS. However, evaluation on intracranial recordings from human subjects suggests that ROAST and SimNIBS are not significantly different in predicting field distribution, provided that users have detailed knowledge of SimNIBS.

SIGNIFICANCE

We hope that the detailed comparisons presented here of various choices in this modeling pipeline can provide guidance for future tool development. We released ROAST as an open-source, easy-to-install and fully-automated pipeline for individualized TES modeling.

Proximal Support Vector Machine-Based Hybrid Approach for Edge Detection in Noisy Images

We propose a novel edge detector in the presence of Gaussian noise with the use of proximal support vector machine (PSVM). The edges of a noisy image are detected using a two-stage architecture: smoothing of image is first performed using regularized anisotropic diffusion, followed by the classification using PSVM, termed as regularized anisotropic diffusion-based PSVM (RAD-PSVM) method. In this process, a feature vector is formed for a pixel using the denoised coefficient’s class and the local orientations to detect edges in all possible directions in images. From the experiments, conducted on both synthetic and benchmark images, it is observed that our RAD-PSVM approach outperforms the other state-of-the-art edge detection approaches, both qualitatively and quantitatively.

Field of View Normalization in Multi-Site Brain MRI

Multi-site brain MRI analysis is needed in big data neuroimaging studies, but challenging. The challenges lie in almost every analysis step including skull stripping. The diversities in multi-site brain MR images make it difficult to tune parameters specific to subjects or imaging protocols. Alternatively, using constant parameter settings often leads to inaccurate, inconsistent and even failed skull stripping results. One reason is that images scanned at different sites, under different scanners or protocols, and/or by different technicians often have very different fields of view (FOVs). Normalizing FOV is currently done manually or using ad hoc pre-processing steps, which do not always generalize well to multi-site diverse images. In this paper, we show that (a) a generic FOV normalization approach is possible in multi-site diverse images; we show experiments on images acquired from Philips, GE, Siemens scanners, from 1.0T, 1.5T, 3.0T field of strengths, and from subjects 0-90 years of ages; and (b) generic FOV normalization improves skull stripping accuracy and consistency for multiple skull stripping algorithms; we show this effect for 5 skull stripping algorithms including FSL's BET, AFNI's 3dSkullStrip, FreeSurfer's HWA, BrainSuite's BSE, and MASS. We have released our FOV normalization software at http://www.nitrc.org/projects/normalizefov .

Bayesian convolutional neural network based MRI brain extraction on nonhuman primates

Brain extraction or skull stripping of magnetic resonance images (MRI) is an essential step in neuroimaging studies, the accuracy of which can severely affect subsequent image processing procedures. Current automatic brain extraction methods demonstrate good results on human brains, but are often far from satisfactory on nonhuman primates, which are a necessary part of neuroscience research. To overcome the challenges of brain extraction in nonhuman primates, we propose a fully-automated brain extraction pipeline combining deep Bayesian convolutional neural network (CNN) and fully connected three-dimensional (3D) conditional random field (CRF). The deep Bayesian CNN, Bayesian SegNet, is used as the core segmentation engine. As a probabilistic network, it is not only able to perform accurate high-resolution pixel-wise brain segmentation, but also capable of measuring the model uncertainty by Monte Carlo sampling with dropout in the testing stage. Then, fully connected 3D CRF is used to refine the probability result from Bayesian SegNet in the whole 3D context of the brain volume. The proposed method was evaluated with a manually brain-extracted dataset comprising T1w images of 100 nonhuman primates. Our method outperforms six popular publicly available brain extraction packages and three well-established deep learning based methods with a mean Dice coefficient of 0.985 and a mean average symmetric surface distance of 0.220 mm. A better performance against all the compared methods was verified by statistical tests (all p-values < 10-4, two-sided, Bonferroni corrected). The maximum uncertainty of the model on nonhuman primate brain extraction has a mean value of 0.116 across all the 100 subjects. The behavior of the uncertainty was also studied, which shows the uncertainty increases as the training set size decreases, the number of inconsistent labels in the training set increases, or the inconsistency between the training set and the testing set increases.

An efficient and accurate method for robust inter-dataset brain extraction and comparisons with 9 other methods

Brain extraction is an important first step in many magnetic resonance neuroimaging studies. Due to variability in brain morphology and in the appearance of the brain due to differences in scanner acquisition parameters, the development of a generally applicable brain extraction algorithm has proven challenging. Learning-based brain extraction algorithms in particular perform well when the target and training images are sufficiently similar, but often perform worse when this condition is not met. In this study, we propose a new patch-based multi-atlas segmentation method for brain extraction which is specifically developed for accurate and robust processing across datasets. Using a diverse collection of labeled images from 5 different datasets, extensive comparisons were made with 9 other commonly used brain extraction methods, both before and after applying error correction (a machine learning method for automatically correcting segmentation errors) to each method. The proposed method performed equal to or better than the other methods in each of two segmentation scenarios: a challenging inter-dataset segmentation scenario in which no dataset-specific atlases were used (mean Dice coefficient 98.57%, volumetric correlation 0.994 across datasets following error correction), and an intra-dataset segmentation scenario in which only dataset-specific atlases were used (mean Dice coefficient 99.02%, volumetric correlation 0.998 across datasets following error correction). Furthermore, combined with error correction, the proposed method runs in less than one-tenth of the time required by the other top-performing methods in the challenging inter-dataset comparisons. Validation on an independent multi-centre dataset also confirmed the excellent performance of the proposed method.

Auto-Context Convolutional Neural Network (Auto-Net) for Brain Extraction in Magnetic Resonance Imaging

Brain extraction or whole brain segmentation is an important first step in many of the neuroimage analysis pipelines. The accuracy and the robustness of brain extraction, therefore, are crucial for the accuracy of the entire brain analysis process. The state-of-the-art brain extraction techniques rely heavily on the accuracy of alignment or registration between brain atlases and query brain anatomy, and/or make assumptions about the image geometry, and therefore have limited success when these assumptions do not hold or image registration fails. With the aim of designing an accurate, learning-based, geometry-independent, and registration-free brain extraction tool, in this paper, we present a technique based on an auto-context convolutional neural network (CNN), in which intrinsic local and global image features are learned through 2-D patches of different window sizes. We consider two different architectures: 1) a voxelwise approach based on three parallel 2-D convolutional pathways for three different directions (axial, coronal, and sagittal) that implicitly learn 3-D image information without the need for computationally expensive 3-D convolutions and 2) a fully convolutional network based on the U-net architecture. Posterior probability maps generated by the networks are used iteratively as context information along with the original image patches to learn the local shape and connectedness of the brain to extract it from non-brain tissue. The brain extraction results we have obtained from our CNNs are superior to the recently reported results in the literature on two publicly available benchmark data sets, namely, LPBA40 and OASIS, in which we obtained the Dice overlap coefficients of 97.73% and 97.62%, respectively. Significant improvement was achieved via our auto-context algorithm. Furthermore, we evaluated the performance of our algorithm in the challenging problem of extracting arbitrarily oriented fetal brains in reconstructed fetal brain magnetic resonance imaging (MRI) data sets. In this application, our voxelwise auto-context CNN performed much better than the other methods (Dice coefficient: 95.97%), where the other methods performed poorly due to the non-standard orientation and geometry of the fetal brain in MRI. Through training, our method can provide accurate brain extraction in challenging applications. This, in turn, may reduce the problems associated with image registration in segmentation tasks.

ISLES 2015 - A public evaluation benchmark for ischemic stroke lesion segmentation from multispectral MRI

Ischemic stroke is the most common cerebrovascular disease, and its diagnosis, treatment, and study relies on non-invasive imaging. Algorithms for stroke lesion segmentation from magnetic resonance imaging (MRI) volumes are intensely researched, but the reported results are largely incomparable due to different datasets and evaluation schemes. We approached this urgent problem of comparability with the Ischemic Stroke Lesion Segmentation (ISLES) challenge organized in conjunction with the MICCAI 2015 conference. In this paper we propose a common evaluation framework, describe the publicly available datasets, and present the results of the two sub-challenges: Sub-Acute Stroke Lesion Segmentation (SISS) and Stroke Perfusion Estimation (SPES). A total of 16 research groups participated with a wide range of state-of-the-art automatic segmentation algorithms. A thorough analysis of the obtained data enables a critical evaluation of the current state-of-the-art, recommendations for further developments, and the identification of remaining challenges. The segmentation of acute perfusion lesions addressed in SPES was found to be feasible. However, algorithms applied to sub-acute lesion segmentation in SISS still lack accuracy. Overall, no algorithmic characteristic of any method was found to perform superior to the others. Instead, the characteristics of stroke lesion appearances, their evolution, and the observed challenges should be studied in detail. The annotated ISLES image datasets continue to be publicly available through an online evaluation system to serve as an ongoing benchmarking resource (www.isles-challenge.org).

Cloud-Based Evaluation of Anatomical Structure Segmentation and Landmark Detection Algorithms: VISCERAL Anatomy Benchmarks

Variations in the shape and appearance of anatomical structures in medical images are often relevant radiological signs of disease. Automatic tools can help automate parts of this manual process. A cloud-based evaluation framework is presented in this paper including results of benchmarking current state-of-the-art medical imaging algorithms for anatomical structure segmentation and landmark detection: the VISCERAL Anatomy benchmarks. The algorithms are implemented in virtual machines in the cloud where participants can only access the training data and can be run privately by the benchmark administrators to objectively compare their performance in an unseen common test set. Overall, 120 computed tomography and magnetic resonance patient volumes were manually annotated to create a standard Gold Corpus containing a total of 1295 structures and 1760 landmarks. Ten participants contributed with automatic algorithms for the organ segmentation task, and three for the landmark localization task. Different algorithms obtained the best scores in the four available imaging modalities and for subsets of anatomical structures. The annotation framework, resulting data set, evaluation setup, results and performance analysis from the three VISCERAL Anatomy benchmarks are presented in this article. Both the VISCERAL data set and Silver Corpus generated with the fusion of the participant algorithms on a larger set of non-manually-annotated medical images are available to the research community.

Person-Based Versus Generalized Impulsivity Disinhibition in Frontotemporal Dementia and Alzheimer Disease

BACKGROUND

While much disinhibition in dementia results from generalized impulsivity, in behavioral variant frontotemporal dementia (bvFTD) disinhibition may also result from impaired social cognition.

OBJECTIVE

To deconstruct disinhibition and its neural correlates in bvFTD vs. early-onset Alzheimer's disease (eAD).

METHODS

Caregivers of 16 bvFTD and 21 matched-eAD patients completed the Frontal Systems Behavior Scale disinhibition items. The disinhibition items were further categorized into (1) "person-based" subscale which predominantly associated with violating social propriety and personal boundary and (2) "generalized-impulsivity" subscale which included nonspecific impulsive acts. Subscale scores were correlated with grey matter volumes from tensor-based morphometry on magnetic resonance images.

RESULTS

In comparison to the eAD patients, the bvFTD patients developed greater person-based disinhibition ( P < 0.001) but comparable generalized impulsivity. Severity of person-based disinhibition significantly correlated with the left anterior superior temporal sulcus (STS), and generalized-impulsivity correlated with the right orbitofrontal cortex (OFC) and the left anterior temporal lobe (aTL).

CONCLUSIONS

Person-based disinhibition was predominant in bvFTD and correlated with the left STS. In both dementia, violations of social propriety and personal boundaries involved fronto-parieto-temporal network of Theory of Mind, whereas nonspecific disinhibition involved the OFC and aTL.

3D shape analysis of the brain's third ventricle using a midplane encoded symmetric template model

BACKGROUND

Structural changes of the brain's third ventricle have been acknowledged as an indicative measure of the brain atrophy progression in neurodegenerative and endocrinal diseases. To investigate the ventricular enlargement in relation to the atrophy of the surrounding structures, shape analysis is a promising approach. However, there are hurdles in modeling the third ventricle shape. First, it has topological variations across individuals due to the inter-thalamic adhesion. In addition, as an interhemispheric structure, it needs to be aligned to the midsagittal plane to assess its asymmetric and regional deformation.

METHOD

To address these issues, we propose a model-based shape assessment. Our template model of the third ventricle consists of a midplane and a symmetric mesh of generic shape. By mapping the template's midplane to the individuals' brain midsagittal plane, we align the symmetric mesh on the midline of the brain before quantifying the third ventricle shape. To build the vertex-wise correspondence between the individual third ventricle and the template mesh, we employ a minimal-distortion surface deformation framework. In addition, to account for topological variations, we implement geometric constraints guiding the template mesh to have zero width where the inter-thalamic adhesion passes through, preventing vertices crossing between left and right walls of the third ventricle. The individual shapes are compared using a vertex-wise deformity from the symmetric template.

RESULTS

Experiments on imaging and demographic data from a study of aging showed that our model was sensitive in assessing morphological differences between individuals in relation to brain volume (i.e. proxy for general brain atrophy), gender and the fluid intelligence at age 72. It also revealed that the proposed method can detect the regional and asymmetrical deformation unlike the conventional measures: volume (median 1.95ml, IQR 0.96ml) and width of the third ventricle. Similarity measures between binary masks and the shape model showed that the latter reconstructed shape details with high accuracy (Dice coefficient ≥0.9, mean distance 0.5mm and Hausdorff distance 2.7mm).

CONCLUSIONS

We have demonstrated that our approach is suitable to morphometrical analyses of the third ventricle, providing high accuracy and inter-subject consistency in the shape quantification. This shape modeling method with geometric constraints based on anatomical landmarks could be extended to other brain structures which require a consistent measurement basis in the morphometry.

Neuropsychological and Neuroanatomical Correlates of the Social Norms Questionnaire in Frontotemporal Dementia Versus Alzheimer's Disease

Traditional neuropsychological batteries may not distinguish early behavioral variant frontotemporal dementia (bvFTD) from Alzheimer's disease (AD) without the inclusion of a social behavioral measure. We compared 33 participants, 15 bvFTD, and 18 matched patients with early-onset AD (eAD), on the Social Norms Questionnaire (SNQ), neuropsychological tests and 3-dimensional T1-weighted magnetic resonance imaging (MRI). The analyses included correlations of SNQ results (total score, overendorsement or "overadhere" errors, and violations or "break" errors) with neuropsychological results and tensor-based morphometry regions of interest. Patients with BvFTD had significantly lower SNQ total scores and higher overadhere errors than patients with eAD. On neuropsychological measures, the SNQ total scores correlated significantly with semantic knowledge and the overadhere subscores with executive dysfunction. On MRI analysis, the break subscores significantly correlated with lower volume of lateral anterior temporal lobes (aTL). The results also suggest that endorsement of social norm violations corresponds to the role of the right aTL in social semantic knowledge.

Methods on Skull Stripping of MRI Head Scan Images-a Review

The high resolution magnetic resonance (MR) brain images contain some non-brain tissues such as skin, fat, muscle, neck, and eye balls compared to the functional images namely positron emission tomography (PET), single photon emission computed tomography (SPECT), and functional magnetic resonance imaging (fMRI) which usually contain relatively less non-brain tissues. The presence of these non-brain tissues is considered as a major obstacle for automatic brain image segmentation and analysis techniques. Therefore, quantitative morphometric studies of MR brain images often require a preliminary processing to isolate the brain from extra-cranial or non-brain tissues, commonly referred to as skull stripping. This paper describes the available methods on skull stripping and an exploratory review of recent literature on the existing skull stripping methods.

Accurate Learning with Few Atlases (ALFA): an algorithm for MRI neonatal brain extraction and comparison with 11 publicly available methods

Accurate whole-brain segmentation, or brain extraction, of magnetic resonance imaging (MRI) is a critical first step in most neuroimage analysis pipelines. The majority of brain extraction algorithms have been developed and evaluated for adult data and their validity for neonatal brain extraction, which presents age-specific challenges for this task, has not been established. We developed a novel method for brain extraction of multi-modal neonatal brain MR images, named ALFA (Accurate Learning with Few Atlases). The method uses a new sparsity-based atlas selection strategy that requires a very limited number of atlases 'uniformly' distributed in the low-dimensional data space, combined with a machine learning based label fusion technique. The performance of the method for brain extraction from multi-modal data of 50 newborns is evaluated and compared with results obtained using eleven publicly available brain extraction methods. ALFA outperformed the eleven compared methods providing robust and accurate brain extraction results across different modalities. As ALFA can learn from partially labelled datasets, it can be used to segment large-scale datasets efficiently. ALFA could also be applied to other imaging modalities and other stages across the life course.

Deep MRI brain extraction: A 3D convolutional neural network for skull stripping

Brain extraction from magnetic resonance imaging (MRI) is crucial for many neuroimaging workflows. Current methods demonstrate good results on non-enhanced T1-weighted images, but struggle when confronted with other modalities and pathologically altered tissue. In this paper we present a 3D convolutional deep learning architecture to address these shortcomings. In contrast to existing methods, we are not limited to non-enhanced T1w images. When trained appropriately, our approach handles an arbitrary number of modalities including contrast-enhanced scans. Its applicability to MRI data, comprising four channels: non-enhanced and contrast-enhanced T1w, T2w and FLAIR contrasts, is demonstrated on a challenging clinical data set containing brain tumors (N=53), where our approach significantly outperforms six commonly used tools with a mean Dice score of 95.19. Further, the proposed method at least matches state-of-the-art performance as demonstrated on three publicly available data sets: IBSR, LPBA40 and OASIS, totaling N=135 volumes. For the IBSR (96.32) and LPBA40 (96.96) data set the convolutional neuronal network (CNN) obtains the highest average Dice scores, albeit not being significantly different from the second best performing method. For the OASIS data the second best Dice (95.02) results are achieved, with no statistical difference in comparison to the best performing tool. For all data sets the highest average specificity measures are evaluated, whereas the sensitivity displays about average results. Adjusting the cut-off threshold for generating the binary masks from the CNN's probability output can be used to increase the sensitivity of the method. Of course, this comes at the cost of a decreased specificity and has to be decided application specific. Using an optimized GPU implementation predictions can be achieved in less than one minute. The proposed method may prove useful for large-scale studies and clinical trials.

MRBrainS Challenge: Online Evaluation Framework for Brain Image Segmentation in 3T MRI Scans

Many methods have been proposed for tissue segmentation in brain MRI scans. The multitude of methods proposed complicates the choice of one method above others. We have therefore established the MRBrainS online evaluation framework for evaluating (semi)automatic algorithms that segment gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) on 3T brain MRI scans of elderly subjects (65–80 y). Participants apply their algorithms to the provided data, after which their results are evaluated and ranked. Full manual segmentations of GM, WM, and CSF are available for all scans and used as the reference standard. Five datasets are provided for training and fifteen for testing. The evaluated methods are ranked based on their overall performance to segment GM, WM, and CSF and evaluated using three evaluation metrics (Dice, H95, and AVD) and the results are published on the MRBrainS13 website. We present the results of eleven segmentation algorithms that participated in the MRBrainS13 challenge workshop at MICCAI, where the framework was launched, and three commonly used freeware packages: FreeSurfer, FSL, and SPM. The MRBrainS evaluation framework provides an objective and direct comparison of all evaluated algorithms and can aid in selecting the best performing method for the segmentation goal at hand.

A comparison of FreeSurfer-generated data with and without manual intervention

This paper examined whether FreeSurfer-generated data differed between a fully-automated, unedited pipeline and an edited pipeline that included the application of control points to correct errors in white matter segmentation. In a sample of 30 individuals, we compared the summary statistics of surface area, white matter volumes, and cortical thickness derived from edited and unedited datasets for the 34 regions of interest (ROIs) that FreeSurfer (FS) generates. To determine whether applying control points would alter the detection of significant differences between patient and typical groups, effect sizes between edited and unedited conditions in individuals with the genetic disorder, 22q11.2 deletion syndrome (22q11DS) were compared to neurotypical controls. Analyses were conducted with data that were generated from both a 1.5 tesla and a 3 tesla scanner. For 1.5 tesla data, mean area, volume, and thickness measures did not differ significantly between edited and unedited regions, with the exception of rostral anterior cingulate thickness, lateral orbitofrontal white matter, superior parietal white matter, and precentral gyral thickness. Results were similar for surface area and white matter volumes generated from the 3 tesla scanner. For cortical thickness measures however, seven edited ROI measures, primarily in frontal and temporal regions, differed significantly from their unedited counterparts, and three additional ROI measures approached significance. Mean effect sizes for edited ROIs did not differ from most unedited ROIs for either 1.5 or 3 tesla data. Taken together, these results suggest that although the application of control points may increase the validity of intensity normalization and, ultimately, segmentation, it may not affect the final, extracted metrics that FS generates. Potential exceptions to and limitations of these conclusions are discussed.

The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)

In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) organized in conjunction with the MICCAI 2012 and 2013 conferences. Twenty state-of-the-art tumor segmentation algorithms were applied to a set of 65 multi-contrast MR scans of low- and high-grade glioma patients-manually annotated by up to four raters-and to 65 comparable scans generated using tumor image simulation software. Quantitative evaluations revealed considerable disagreement between the human raters in segmenting various tumor sub-regions (Dice scores in the range 74%-85%), illustrating the difficulty of this task. We found that different algorithms worked best for different sub-regions (reaching performance comparable to human inter-rater variability), but that no single algorithm ranked in the top for all sub-regions simultaneously. Fusing several good algorithms using a hierarchical majority vote yielded segmentations that consistently ranked above all individual algorithms, indicating remaining opportunities for further methodological improvements. The BRATS image data and manual annotations continue to be publicly available through an online evaluation system as an ongoing benchmarking resource.

Separation of overlapping dental arch objects using digital records of illuminated plaster casts

Background

Plaster casts of individual patients are important for orthodontic specialists during the treatment process and their analysis is still a standard diagnostical tool. But the growing capabilities of information technology enable their replacement by digital models obtained by complex scanning systems.

Method

This paper presents the possibility of using a digital camera as a simple instrument to obtain the set of digital images for analysis and evaluation of the treatment using appropriate mathematical tools of image processing. The methods studied in this paper include the segmentation of overlapping dental bodies and the use of different illumination sources to increase the reliability of the separation process. The circular Hough transform, region growing with multiple seed points, and the convex hull detection method are applied to the segmentation of orthodontic plaster cast images to identify dental arch objects and their sizes.

Results

The proposed algorithm presents the methodology of improving the accuracy of segmentation of dental arch components using combined illumination sources. Dental arch parameters and distances between the canines and premolars for different segmentation methods were used as a measure to compare the results obtained.

Conclusion

A new method of segmentation of overlapping dental arch components using digital records of illuminated plaster casts provides information with the precision required for orthodontic treatment. The distance between corresponding teeth was evaluated with a mean error of 1.38% and the Dice similarity coefficient of the evaluated dental bodies boundaries reached 0.9436 with a false positive rate \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$FPR=0.0381$$\end{document}FPR=0.0381 and false negative rate \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$FNR=0.0728$$\end{document}FNR=0.0728.

Brain Extraction Using Label Propagation and Group Agreement: Pincram

Accurately delineating the brain on magnetic resonance (MR) images of the head is a prerequisite for many neuroimaging methods. Most existing methods exhibit disadvantages in that they are laborious, yield inconsistent results, and/or require training data to closely match the data to be processed. Here, we present pincram, an automatic, versatile method for accurately labelling the adult brain on T1-weighted 3D MR head images. The method uses an iterative refinement approach to propagate labels from multiple atlases to a given target image using image registration. At each refinement level, a consensus label is generated. At the subsequent level, the search for the brain boundary is constrained to the neighbourhood of the boundary of this consensus label. The method achieves high accuracy (Jaccard coefficient > 0.95 on typical data, corresponding to a Dice similarity coefficient of > 0.97) and performs better than many state-of-the-art methods as evidenced by independent evaluation on the Segmentation Validation Engine. Via a novel self-monitoring feature, the program generates the "success index," a scalar metadatum indicative of the accuracy of the output label. Pincram is available as open source software.

Fuzzy Index to Evaluate Edge Detection in Digital Images

In literature, we can find different metrics to evaluate the detected edges in digital images, like Pratt's figure of merit (FOM), Jaccard’s index (JI) and Dice’s coefficient (DC). These metrics compare two images, the first one is the reference edges image, and the second one is the detected edges image. It is important to mention that all existing metrics must binarize images before their evaluation. Binarization step causes information to be lost because an incomplete image is being evaluated. In this paper, we propose a fuzzy index (FI) for edge evaluation that does not use a binarization step. In order to process all detected edges, images are represented in their fuzzy form and all calculations are made with fuzzy sets operators and fuzzy Euclidean distance between both images. Our proposed index is compared to the most used metrics using synthetic images, with good results.

Hippocampal Shape Modeling Based on a Progressive Template Surface Deformation and its Verification

Accurately recovering the hippocampal shapes against rough and noisy segmentations is as challenging as achieving good anatomical correspondence between the individual shapes. To address these issues, we propose a mesh-to-volume registration approach, characterized by a progressive model deformation. Our model implements flexible weighting scheme for model rigidity under a multi-level neighborhood for vertex connectivity. This method induces a large-to-small scale deformation of a template surface to build the pairwise correspondence by minimizing geometric distortion while robustly restoring the individuals' shape characteristics. We evaluated the proposed method's (1) accuracy and robustness in smooth surface reconstruction, (2) sensitivity in detecting significant shape differences between healthy control and disease groups (mild cognitive impairment and Alzheimer's disease), (3) robustness in constructing the anatomical correspondence between individual shape models, and (4) applicability in identifying subtle shape changes in relation to cognitive abilities in a healthy population. We compared the performance of the proposed method with other well-known methods--SPHARM-PDM, ShapeWorks and LDDMM volume registration with template injection--using various metrics of shape similarity, surface roughness, volume, and shape deformity. The experimental results showed that the proposed method generated smooth surfaces with less volume differences and better shape similarity to input volumes than others. The statistical analyses with clinical variables also showed that it was sensitive in detecting subtle shape changes of hippocampus.

Fully automated whole-head segmentation with improved smoothness and continuity, with theory reviewed

Individualized current-flow models are needed for precise targeting of brain structures using transcranial electrical or magnetic stimulation (TES/TMS). The same is true for current-source reconstruction in electroencephalography and magnetoencephalography (EEG/MEG). The first step in generating such models is to obtain an accurate segmentation of individual head anatomy, including not only brain but also cerebrospinal fluid (CSF), skull and soft tissues, with a field of view (FOV) that covers the whole head. Currently available automated segmentation tools only provide results for brain tissues, have a limited FOV, and do not guarantee continuity and smoothness of tissues, which is crucially important for accurate current-flow estimates. Here we present a tool that addresses these needs. It is based on a rigorous Bayesian inference framework that combines image intensity model, anatomical prior (atlas) and morphological constraints using Markov random fields (MRF). The method is evaluated on 20 simulated and 8 real head volumes acquired with magnetic resonance imaging (MRI) at 1 mm3 resolution. We find improved surface smoothness and continuity as compared to the segmentation algorithms currently implemented in Statistical Parametric Mapping (SPM). With this tool, accurate and morphologically correct modeling of the whole-head anatomy for individual subjects may now be feasible on a routine basis. Code and data are fully integrated into SPM software tool and are made publicly available. In addition, a review on the MRI segmentation using atlas and the MRF over the last 20 years is also provided, with the general mathematical framework clearly derived.

Level set method coupled with Energy Image features for brain MR image segmentation

Up until now, the noise and intensity inhomogeneity are considered one of the major drawbacks in the field of brain magnetic resonance (MR) image segmentation. This paper introduces the energy image feature approach for intensity inhomogeneity correction. Our approach of segmentation takes the advantage of image features and preserves the advantages of the level set methods in region-based active contours framework. The energy image feature represents a new image obtained from the original image when the pixels' values are replaced by local energy values computed in the 3×3 mask size. The performance and utility of the energy image features were tested and compared through two different variants of level set methods: one as the encompassed local and global intensity fitting method and the other as the selective binary and Gaussian filtering regularized level set method. The reported results demonstrate the flexibility of the energy image feature to adapt to level set segmentation framework and to perform the challenging task of brain lesion segmentation in a rather robust way.

Partial volume effect modeling for segmentation and tissue classification of brain magnetic resonance images: A review

Quantitative analysis of magnetic resonance (MR) brain images are facilitated by the development of automated segmentation algorithms. A single image voxel may contain of several types of tissues due to the finite spatial resolution of the imaging device. This phenomenon, termed partial volume effect (PVE), complicates the segmentation process, and, due to the complexity of human brain anatomy, the PVE is an important factor for accurate brain structure quantification. Partial volume estimation refers to a generalized segmentation task where the amount of each tissue type within each voxel is solved. This review aims to provide a systematic, tutorial-like overview and categorization of methods for partial volume estimation in brain MRI. The review concentrates on the statistically based approaches for partial volume estimation and also explains differences to other, similar image segmentation approaches.

Bidirectional iterative parcellation of diffusion weighted imaging data: separating cortical regions connected by the arcuate fasciculus and extreme capsule

This paper introduces a Bidirectional Iterative Parcellation (BIP) procedure designed to identify the location and size of connected cortical regions (parcellations) at both ends of a white matter tract in diffusion weighted images. The procedure applies the FSL option "probabilistic tracking with classification targets" in a bidirectional and iterative manner. To assess the utility of BIP, we applied the procedure to the problem of parcellating a limited set of well-established gray matter seed regions associated with the dorsal (arcuate fasciculus/superior longitudinal fasciculus) and ventral (extreme capsule fiber system) white matter tracts in the language networks of 97 participants. These left hemisphere seed regions and the two white matter tracts, along with their right hemisphere homologues, provided an excellent test case for BIP because the resulting parcellations overlap and their connectivity via the arcuate fasciculi and extreme capsule fiber systems are well studied. The procedure yielded both confirmatory and novel findings. Specifically, BIP confirmed that each tract connects within the seed regions in unique, but expected ways. Novel findings included increasingly left-lateralized parcellations associated with the arcuate fasciculus/superior longitudinal fasciculus as a function of age and education. These results demonstrate that BIP is an easily implemented technique that successfully confirmed cortical connectivity patterns predicted in the literature, and has the potential to provide new insights regarding the architecture of the brain.

Automated segmentation of multifocal basal ganglia T2*-weighted MRI hypointensities

Multifocal basal ganglia T2*-weighted (T2*w) hypointensities, which are believed to arise mainly from vascular mineralization, were recently proposed as a novel MRI biomarker for small vessel disease and ageing. These T2*w hypointensities are typically segmented semi-automatically, which is time consuming, associated with a high intra-rater variability and low inter-rater agreement. To address these limitations, we developed a fully automated, unsupervised segmentation method for basal ganglia T2*w hypointensities. This method requires conventional, co-registered T2*w and T1-weighted (T1w) volumes, as well as region-of-interest (ROI) masks for the basal ganglia and adjacent internal capsule generated automatically from T1w MRI. The basal ganglia T2*w hypointensities were then segmented with thresholds derived with an adaptive outlier detection method from respective bivariate T2*w/T1w intensity distributions in each ROI. Artefacts were reduced by filtering connected components in the initial masks based on their standardised T2*w intensity variance. The segmentation method was validated using a custom-built phantom containing mineral deposit models, i.e. gel beads doped with 3 different contrast agents in 7 different concentrations, as well as with MRI data from 98 community-dwelling older subjects in their seventies with a wide range of basal ganglia T2*w hypointensities. The method produced basal ganglia T2*w hypointensity masks that were in substantial volumetric and spatial agreement with those generated by an experienced rater (Jaccard index = 0.62 ± 0.40). These promising results suggest that this method may have use in automatic segmentation of basal ganglia T2*w hypointensities in studies of small vessel disease and ageing.

•

A novel method segmented focal T2*-weighted MRI hypointensities automatically.

•

The method was validated with MRI of a novel phantom and 98 elderly subjects.

•

The subject masks from the method and an experienced rater overlapped substantially.

•

The method is potentially useful for research into small vessel disease and ageing.

Nonlocal intracranial cavity extraction

Automatic and accurate methods to estimate normalized regional brain volumes from MRI data are valuable tools which may help to obtain an objective diagnosis and followup of many neurological diseases. To estimate such regional brain volumes, the intracranial cavity volume (ICV) is often used for normalization. However, the high variability of brain shape and size due to normal intersubject variability, normal changes occurring over the lifespan, and abnormal changes due to disease makes the ICV estimation problem challenging. In this paper, we present a new approach to perform ICV extraction based on the use of a library of prelabeled brain images to capture the large variability of brain shapes. To this end, an improved nonlocal label fusion scheme based on BEaST technique is proposed to increase the accuracy of the ICV estimation. The proposed method is compared with recent state-of-the-art methods and the results demonstrate an improved performance both in terms of accuracy and reproducibility while maintaining a reduced computational burden.

Skull-stripping with machine learning deformable organisms

BACKGROUND

Segmentation methods for medical images may not generalize well to new data sets or new tasks, hampering their utility. We attempt to remedy these issues using deformable organisms to create an easily customizable segmentation plan. We validate our framework by creating a plan to locate the brain in 3D magnetic resonance images of the head (skull-stripping).

NEW METHOD

Our method borrows ideas from artificial life to govern a set of deformable models. We use control processes such as sensing, proactive planning, reactive behavior, and knowledge representation to segment an image. The image may have landmarks and features specific to that dataset; these may be easily incorporated into the plan. In addition, we use a machine learning method to make our segmentation more accurate.

RESULTS

Our method had the least Hausdorff distance error, but included slightly less brain voxels (false negatives). It also had the lowest false positive error and performed on par to skull-stripping specific method on other metrics.

COMPARISON WITH EXISTING METHOD(S)

We tested our method on 838 T1-weighted images, evaluating results using distance and overlap error metrics based on expert gold standard segmentations. We evaluated the results before and after the learning step to quantify its benefit; we also compare our results to three other widely used methods: BSE, BET, and the Hybrid Watershed algorithm.

CONCLUSIONS

Our framework captures diverse categories of information needed for brain segmentation and will provide a foundation for tackling a wealth of segmentation problems.

MBIS: multivariate Bayesian image segmentation tool

We present MBIS (Multivariate Bayesian Image Segmentation tool), a clustering tool based on the mixture of multivariate normal distributions model. MBIS supports multichannel bias field correction based on a B-spline model. A second methodological novelty is the inclusion of graph-cuts optimization for the stationary anisotropic hidden Markov random field model. Along with MBIS, we release an evaluation framework that contains three different experiments on multi-site data. We first validate the accuracy of segmentation and the estimated bias field for each channel. MBIS outperforms a widely used segmentation tool in a cross-comparison evaluation. The second experiment demonstrates the robustness of results on atlas-free segmentation of two image sets from scan-rescan protocols on 21 healthy subjects. Multivariate segmentation is more replicable than the monospectral counterpart on T1-weighted images. Finally, we provide a third experiment to illustrate how MBIS can be used in a large-scale study of tissue volume change with increasing age in 584 healthy subjects. This last result is meaningful as multivariate segmentation performs robustly without the need for prior knowledge.

A pilot study evaluating presurgery neuroanatomical biomarkers for postoperative cognitive decline after total knee arthroplasty in older adults

BACKGROUND

Total knee arthroplasty improves quality of life but is associated with postoperative cognitive dysfunction in older adults. This prospective longitudinal pilot study with a parallel control group tested the hypotheses that (1) nondemented adults would exhibit primary memory and executive difficulties after total knee arthroplasty, and (2) reduced preoperative hippocampus/entorhinal volume would predict postoperative memory change, whereas preoperative leukoaraiosis and lacunae volumes would predict postoperative executive dysfunction.

METHODS

Surgery (n = 40) and age-education-matched controls with osteoarthritis (n = 15) completed pre- and postoperative (3 weeks, 3 months, and 1 yr) memory and cognitive testing. Hypothesized brain regions of interest were measured in patients completing preoperative magnetic resonance scans (surgery, n = 31; control, n = 12). Analyses used reliable change methods to identify the frequency of cognitive change at each time point.

RESULTS

The incidence of postoperative memory difficulties was shown with delay test indices (i.e., story memory test: 3 weeks = 17%, 3 months = 25%, 1 yr = 9%). Postoperative executive difficulty with measures of inhibitory function (i.e., Stroop Color Word: 3 weeks = 21%, 3 months = 22%, 1 yr = 9%). Hierarchical regression analysis assessing the predictive interaction of group (surgery, control) and preoperative neuroanatomical structures on decline showed that greater preoperative volumes of leukoaraiosis/lacunae were significantly contributed to postoperative executive (inhibitory) declines.

CONCLUSIONS

This pilot study suggests that executive and memory declines occur in nondemented adults undergoing orthopedic surgery. Severity of preoperative cerebrovascular disease may be relevant for understanding executive decline, in particular.

Automatic generation of statistical pose and shape models for articulated joints

Statistical analysis of motion patterns of body joints is potentially useful for detecting and quantifying pathologies. However, building a statistical motion model across different subjects remains a challenging task, especially for a complex joint like the wrist. We present a novel framework for simultaneous registration and segmentation of multiple 3-D (CT or MR) volumes of different subjects at various articulated positions. The framework starts with a pose model generated from 3-D volumes captured at different articulated positions of a single subject (template). This initial pose model is used to register the template volume to image volumes from new subjects. During this process, the Grow-Cut algorithm is used in an iterative refinement of the segmentation of the bone along with the pose parameters. As each new subject is registered and segmented, the pose model is updated, improving the accuracy of successive registrations. We applied the algorithm to CT images of the wrist from 25 subjects, each at five different wrist positions and demonstrated that it performed robustly and accurately. More importantly, the resulting segmentations allowed a statistical pose model of the carpal bones to be generated automatically without interaction. The evaluation results show that our proposed framework achieved accurate registration with an average mean target registration error of 0.34 ±0.27 mm. The automatic segmentation results also show high consistency with the ground truth obtained semi-automatically. Furthermore, we demonstrated the capability of the resulting statistical pose and shape models by using them to generate a measurement tool for scaphoid-lunate dissociation diagnosis, which achieved 90% sensitivity and specificity.

T1 magnetic resonance imaging head segmentation for diffuse optical tomography and electroencephalography

Accurate segmentation of structural magnetic resonance images is critical for creating subject-specific forward models for functional neuroimaging source localization. In this work, we present an innovative segmentation algorithm that generates accurate head tissue layer thicknesses that are needed for diffuse optical tomography (DOT) data analysis. The presented algorithm is compared against other publicly available head segmentation methods. The proposed algorithm has a root mean square scalp thickness error of 1.60 mm, skull thickness error of 1.96 mm, and summed scalp and skull error of 1.49 mm. We also introduce a segmentation evaluation metric that evaluates the accuracy of tissue layer thicknesses in regions of the head where optodes are typically placed. The presented segmentation algorithm and evaluation metric are tools for improving the localization accuracy of neuroimaging with DOT, and also multimodal neuroimaging such as combined electroencephalography and DOT.

An alternative approach to histopathological validation of PET imaging for radiation therapy image-guidance: a proof of concept

PURPOSE

In radiotherapy, PET images can be used to guide the delivery of selectively escalated doses to biologically relevant tumour subvolumes. Validation of PET for such applications requires demonstration of spatial coincidence between PET tracer uptake pattern and the histopathologically confirmed target. This study introduces a novel approach to histopathological validation of PET image segmentation for radiotherapy guidance.

METHODS AND MATERIALS

Sequential tissue sections from surgically excised whole-tumour specimens were used to acquire full 3D-sets of both histopathological images (microscopy) and PET tracer distribution images (autoradiography). After these datasets were accurately registered, a full 3D autoradiographic distribution of PET tracer was reconstructed and used to obtain synthetic PET images (sPET) by simulating the image deterioration induced by processes involved in PET image formation. To illustrate the method, sPET images were used in this study to investigate spatial coincidence between high FDG uptake areas and the distribution of viable tissue in two small animal tumour models.

RESULTS

The reconstructed 3D autoradiographic distribution of the PET tracer was spatially coherent, as indicated by the high average value of the normalised pixel-by-pixel correlation of intensities between successive slices (0.84 ± 0.05 and 0.94 ± 0.02). The loss of detail in the sPET images versus the 3D autoradiography was significant as indicated by Dice coefficient values corresponding to the two tumours (0 and 0.1 at 70% threshold). The maximum overlap between the FDG segmented volumes and the extent of the viable tissue as indicated by Dice coefficient values, was 0.8 for one tumour (for the image thresholded at 22% of max intensity) and 0.88 for the other (threshold of 14% of max intensity).

CONCLUSION

It was demonstrated that the use of synthetic PET images for histopathological validation allows for bypassing a technically challenging and error-prone step of registering non-invasive PET images with histopathology.