3D Magnetic Resonance Imaging Data Research Articles

TACA-RNet: Tri-Axis Based Context-Aware Reverse Network for Multimodal Brain Tumor Segmentation

Brain tumor segmentation using Magnetic Resonance Imaging (MRI) is vital for clinical decision making. Traditional deep learning-based studies using convolutional neural networks have predominantly processed MRI data as two-dimensional slices, leading to the loss of contextual information. While three-dimensional (3D) convolutional layers represent an advancement, they have not fully exploited pathological information according to the three-axis nature of 3D MRI data—axial, coronal, and sagittal. Recognizing these limitations, we introduce a Tri-Axis based Context-Aware Reverse Network (TACA-RNet). This innovative approach leverages the unique 3D spatial orientations of MRI, learning crucial information on brain anatomy and pathology. We incorporated three specialized modules: a Tri-Axis Channel Reduction module for optimizing feature dimensions, a MultiScale Contextual Fusion module for aggregating multi-scale features and enhancing spatial discernment, and a 3D Axis Reverse Attention module for the precise delineation of tumor boundaries. The TACA-RNet leverages three specialized modules to enhance the understanding of tumor characteristics and spatial relationships within MRI data by fully utilizing its tri-axial structure. Validated on the Brain Tumor Segmentation Challenge 2018 and 2020 datasets, the TACA-RNet demonstrated superior performances over contemporary methodologies. This underscores the critical role of leveraging the three-axis structure of MRI to enhance segmentation accuracy.

MRI super-resolution using similarity distance and multi-scale receptive field based feature fusion GAN and pre-trained slice interpolation network

Challenges arise in achieving high-resolution Magnetic Resonance Imaging (MRI) to improve disease diagnosis accuracy due to limitations in hardware, patient discomfort, long acquisition times, and high costs. While Convolutional Neural Networks (CNNs) have shown promising results in MRI super-resolution, they often don't look into the structural similarity and prior information available in consecutive MRI slices. By leveraging information from sequential slices, more robust features can be obtained, potentially leading to higher-quality MRI slices. We propose a multi-slice two-dimensional (2D) MRI super-resolution network that combines a Generative Adversarial Network (GAN) with feature fusion and a pre-trained slice interpolation network to achieve three-dimensional (3D) super-resolution. The proposed model requires consecutively acquired three low-resolution (LR) MRI slices along a specific axis, and achieves the reconstruction of the MRI slices in the remaining two axes. The network effectively enhances both in-plane and out-of-plane resolution along the sagittal axis while addressing computational and memory constraints in 3D super-resolution. The proposed generator has a in-plane and out-of-plane Attention (IOA) network that fuses both in-plane and out-plane features of MRI dynamically. In terms of out-of-plane attention, the network merges features by considering the similarity distance between features and for in-plane attention, the network employs a two-level pyramid structure with varying receptive fields to extract features at different scales, ensuring the inclusion of both global and local features. Subsequently, to achieve 3D MRI super-resolution, a pre-trained slice interpolation network is used that takes two consecutive super-resolved MRI slices to generate a new intermediate slice. To further enhance the network performance and perceptual quality, we introduce a feature up-sampling layer and a feature extraction block with Scaled Exponential Linear Unit (SeLU). Moreover, our super-resolution network incorporates VGG loss from a fine-tuned VGG-19 network to provide additional enhancement. Through experimental evaluations on the IXI dataset and BRATS dataset, using the peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM) and the number of training parameters, we demonstrate the superior performance of our method compared to the existing techniques. Also, the proposed model can be adapted or modified to achieve super-resolution for both 2D and 3D MRI data.

Value of aortic volumes assessed by automated segmentation of 3D MRI data in patients with thoracic aortic dilatation: A case-control study

PurposeThe purpose of this study was to investigate the benefit of aortic volumes compared to diameters or cross-sectional areas on three-dimensional (3D) magnetic resonance imaging (MRI) in discriminating between patients with dilated aorta and matched controls. Materials and methodsSixty-two patients (47 men and 15 women; median age, 66 years; age range: 33–86 years) with tricuspid aortic valve and ascending thoracic aorta aneurysm (TAV-ATAA) and 43 patients (35 men and 8 women; median age, 51 years; age range: 17–76 years) with bicuspid aortic valve and dilated ascending aorta (BAV) were studied. One group of 54 controls matched for age and sex to patients with TAV-ATAA (39 men and 15 women; median age, 68 years; age range: 33–81 years) and one group of 42 controls matched for age and sex to patients with BAV (34 men and 8 women; median age, 50 years; age range: 17–77 years) were identified. All participants underwent 3D MRI, used for 3D-segmentation for measuring aortic length, maximal diameter, maximal cross-sectional area (CSA) and volume for the ascending aorta. ResultsAn increase in ascending aorta volume (TAV-ATAA: +107%; BAV: +171% vs. controls; P < 0.001) was found, which was three times greater than the increase in diameter (TAV-ATAA: +29%; BAV: +40% vs. controls; P < 0.001). In differentiating patients with TAV-ATAA from their controls, the indexed ascending aorta volume showed better performances (AUC, 0.935 [95% confidence interval (CI): 0.882–0.989]; accuracy, 88.7% [95% CI: 82.9–94.5]) than indexed ascending aorta length (P < 0.001), indexed ascending aorta maximal diameter (P = 0.003) and indexed ascending aorta maximal CSA (P = 0.03). In differentiating patients with BAV from matched controls, indexed ascending aorta volume showed significantly better performances performance (AUC, 0.908 [95% CI: 0.829–0.987]; accuracy, 88.0% [95% CI: 80.9–95.0]) than indexed ascending aorta length (P = 0.02) and not different from indexed ascending aorta maximal diameter (P = 0.07) or from indexed ascending aorta maximal CSA (P = 0.27) ConclusionAortic volume measured by 3D-MRI integrates both elongation and luminal dilatation, resulting in greater classification performance than maximal diameter and length in differentiating patients with dilated ascending aorta or aneurysm from controls.

Virtual Reality as Tool for Bioprinting Quality Inspection: A Proof of Principle.

As virtual reality (VR) has drastically evolved over the past few years, the field of applications of VR flourished way beyond the gaming industry. While commercial VR solutions might be available, there is a need to develop a workflow for specific applications. Bioprinting represents such an example. Here, complex 3D data is generated and needs to be visualized in the context of quality control. We demonstrate that the transfer to a commercially available VR software is possible by introducing an optimized workflow. In the present work, we developed a workflow for the visualization of the critical quality attribute (cQA) cell distribution in bioprinted (extrusion-based) samples in VR. The cQA cell distribution is directly influenced by the pre-processing step mixing of cell material in the bioink. Magnetic Resonance Imaging (MRI) was used as an analytical tool to generate spatially resolved 2.5 and 3D data of the bioprinted objects. A sample with poor quality in respect of the cQA cell distribution was identified as its inhomogeneous cell distribution could be displayed spatially resolved in VR. The described workflow facilitates the usage of VR as a tool for quality inspection in the field of bioprinting and represents a powerful tool for visualization of complex 3D MRI data.

Unsupervised Image Registration towards Enhancing Performance and Explainability in Cardiac and Brain Image Analysis.

Magnetic Resonance Imaging (MRI) typically recruits multiple sequences (defined here as “modalities”). As each modality is designed to offer different anatomical and functional clinical information, there are evident disparities in the imaging content across modalities. Inter- and intra-modality affine and non-rigid image registration is an essential medical image analysis process in clinical imaging, as for example before imaging biomarkers need to be derived and clinically evaluated across different MRI modalities, time phases and slices. Although commonly needed in real clinical scenarios, affine and non-rigid image registration is not extensively investigated using a single unsupervised model architecture. In our work, we present an unsupervised deep learning registration methodology that can accurately model affine and non-rigid transformations, simultaneously. Moreover, inverse-consistency is a fundamental inter-modality registration property that is not considered in deep learning registration algorithms. To address inverse consistency, our methodology performs bi-directional cross-modality image synthesis to learn modality-invariant latent representations, and involves two factorised transformation networks (one per each encoder-decoder channel) and an inverse-consistency loss to learn topology-preserving anatomical transformations. Overall, our model (named “FIRE”) shows improved performances against the reference standard baseline method (i.e., Symmetric Normalization implemented using the ANTs toolbox) on multi-modality brain 2D and 3D MRI and intra-modality cardiac 4D MRI data experiments. We focus on explaining model-data components to enhance model explainability in medical image registration. On computational time experiments, we show that the FIRE model performs on a memory-saving mode, as it can inherently learn topology-preserving image registration directly in the training phase. We therefore demonstrate an efficient and versatile registration technique that can have merit in multi-modal image registrations in the clinical setting.

3D U-Net Improves Automatic Brain Extraction for Isotropic Rat Brain Magnetic Resonance Imaging Data.

Brain extraction is a critical pre-processing step in brain magnetic resonance imaging (MRI) analytical pipelines. In rodents, this is often achieved by manually editing brain masks slice-by-slice, a time-consuming task where workloads increase with higher spatial resolution datasets. We recently demonstrated successful automatic brain extraction via a deep-learning-based framework, U-Net, using 2D convolutions. However, such an approach cannot make use of the rich 3D spatial-context information from volumetric MRI data. In this study, we advanced our previously proposed U-Net architecture by replacing all 2D operations with their 3D counterparts and created a 3D U-Net framework. We trained and validated our model using a recently released CAMRI rat brain database acquired at isotropic spatial resolution, including T2-weighted turbo-spin-echo structural MRI and T2*-weighted echo-planar-imaging functional MRI. The performance of our 3D U-Net model was compared with existing rodent brain extraction tools, including Rapid Automatic Tissue Segmentation, Pulse-Coupled Neural Network, SHape descriptor selected External Regions after Morphologically filtering, and our previously proposed 2D U-Net model. 3D U-Net demonstrated superior performance in Dice, Jaccard, center-of-mass distance, Hausdorff distance, and sensitivity. Additionally, we demonstrated the reliability of 3D U-Net under various noise levels, evaluated the optimal training sample sizes, and disseminated all source codes publicly, with a hope that this approach will benefit rodent MRI research community.Significant Methodological Contribution: We proposed a deep-learning-based framework to automatically identify the rodent brain boundaries in MRI. With a fully 3D convolutional network model, 3D U-Net, our proposed method demonstrated improved performance compared to current automatic brain extraction methods, as shown in several qualitative metrics (Dice, Jaccard, PPV, SEN, and Hausdorff). We trust that this tool will avoid human bias and streamline pre-processing steps during 3D high resolution rodent brain MRI data analysis. The software developed herein has been disseminated freely to the community.

Effect of data leakage in brain MRI classification using 2D convolutional neural networks

In recent years, 2D convolutional neural networks (CNNs) have been extensively used to diagnose neurological diseases from magnetic resonance imaging (MRI) data due to their potential to discern subtle and intricate patterns. Despite the high performances reported in numerous studies, developing CNN models with good generalization abilities is still a challenging task due to possible data leakage introduced during cross-validation (CV). In this study, we quantitatively assessed the effect of a data leakage caused by 3D MRI data splitting based on a 2D slice-level using three 2D CNN models to classify patients with Alzheimer’s disease (AD) and Parkinson’s disease (PD). Our experiments showed that slice-level CV erroneously boosted the average slice level accuracy on the test set by 30% on Open Access Series of Imaging Studies (OASIS), 29% on Alzheimer’s Disease Neuroimaging Initiative (ADNI), 48% on Parkinson’s Progression Markers Initiative (PPMI) and 55% on a local de-novo PD Versilia dataset. Further tests on a randomly labeled OASIS-derived dataset produced about 96% of (erroneous) accuracy (slice-level split) and 50% accuracy (subject-level split), as expected from a randomized experiment. Overall, the extent of the effect of an erroneous slice-based CV is severe, especially for small datasets.

Alzheimer's Disease Classification Using 2D Convolutional Neural Networks.

Alzheimer's disease (AD) is a non-treatable and non-reversible disease that affects about 6% of people who are 65 and older. Brain magnetic resonance imaging (MRI) is a pseudo-3D imaging technology that is widely used for AD diagnosis. Convolutional neural networks with 3D kernels (3D CNNs) are often the default choice for deep learning based MRI analysis. However, 3D CNNs are usually computationally costly and data-hungry. Such disadvantages post a barrier of using modern deep learning techniques in the medical imaging domain, in which the number of data that can be used for training is usually limited. In this work, we propose three approaches that leverage 2D CNNs on 3D MRI data. We test the proposed methods on the Alzheimer's Disease Neuroimaging Initiative dataset across two popular 2D CNN architectures. The evaluation results show that the proposed method improves the model performance on AD diagnosis by 8.33% accuracy or 10.11% auROC compared with the ResNet-based 3D CNN model, while significantly reducing the training time by over 89%. We also discuss the potential causes for performance improvement and the limitations. We believe this work can serve as a strong baseline for future researchers.

A Holographic Augmented Reality Interface for Visualizing of MRI Data and Planning of Neurosurgical Procedures.

The recent introduction of wireless head-mounted displays (HMD) promises to enhance 3D image visualization by immersing the user into 3D morphology. This work introduces a prototype holographic augmented reality (HAR) interface for the 3D visualization of magnetic resonance imaging (MRI) data for the purpose of planning neurosurgical procedures. The computational platform generates a HAR scene that fuses pre-operative MRI sets, segmented anatomical structures, and a tubular tool for planning an access path to the targeted pathology. The operator can manipulate the presented images and segmented structures and perform path-planning using voice and gestures. On-the-fly, the software uses defined forbidden-regions to prevent the operator from harming vital structures. In silico studies using the platform with a HoloLens HMD assessed its functionality and the computational load and memory for different tasks. A preliminary qualitative evaluation revealed that holographic visualization of high-resolution 3D MRI data offers an intuitive and interactive perspective of the complex brain vasculature and anatomical structures. This initial work suggests that immersive experiences may be an unparalleled tool for planning neurosurgical procedures.

Computer-Aided Diagnosis of Alzheimer's Disease through Weak Supervision Deep Learning Framework with Attention Mechanism.

Alzheimer’s disease (AD) is the most prevalent neurodegenerative disease causing dementia and poses significant health risks to middle-aged and elderly people. Brain magnetic resonance imaging (MRI) is the most widely used diagnostic method for AD. However, it is challenging to collect sufficient brain imaging data with high-quality annotations. Weakly supervised learning (WSL) is a machine learning technique aimed at learning effective feature representation from limited or low-quality annotations. In this paper, we propose a WSL-based deep learning (DL) framework (ADGNET) consisting of a backbone network with an attention mechanism and a task network for simultaneous image classification and image reconstruction to identify and classify AD using limited annotations. The ADGNET achieves excellent performance based on six evaluation metrics (Kappa, sensitivity, specificity, precision, accuracy, F1-score) on two brain MRI datasets (2D MRI and 3D MRI data) using fine-tuning with only 20% of the labels from both datasets. The ADGNET has an F1-score of 99.61% and sensitivity is 99.69%, outperforming two state-of-the-art models (ResNext WSL and SimCLR). The proposed method represents a potential WSL-based computer-aided diagnosis method for AD in clinical practice.

Printable 3D vocal tract shapes from MRI data and their acoustic and aerodynamic properties

A detailed understanding of how the acoustic patterns of speech sounds are generated by the complex 3D shapes of the vocal tract is a major goal in speech research. The Dresden Vocal Tract Dataset (DVTD) presented here contains geometric and (aero)acoustic data of the vocal tract of 22 German speech sounds (16 vowels, 5 fricatives, 1 lateral), each from one male and one female speaker. The data include the 3D Magnetic Resonance Imaging data of the vocal tracts, the corresponding 3D-printable and finite-element models, and their simulated and measured acoustic and aerodynamic properties. The dataset was evaluated in terms of the plausibility and the similarity of the resonance frequencies determined by the acoustic simulations and measurements, and in terms of the human identification rate of the vowels and fricatives synthesized by the artificially excited 3D-printed vocal tract models. According to both the acoustic and perceptual metrics, most models are accurate representations of the intended speech sounds and can be readily used for research and education.

Hepatic tumor classification using texture and topology analysis of non-contrast-enhanced three-dimensional T1-weighted MR images with a radiomics approach

The purpose of this study is to evaluate the accuracy for classification of hepatic tumors by characterization of T1-weighted magnetic resonance (MR) images using two radiomics approaches with machine learning models: texture analysis and topological data analysis using persistent homology. This study assessed non-contrast-enhanced fat-suppressed three-dimensional (3D) T1-weighted images of 150 hepatic tumors. The lesions included 50 hepatocellular carcinomas (HCCs), 50 metastatic tumors (MTs), and 50 hepatic hemangiomas (HHs) found respectively in 37, 23, and 33 patients. For classification, texture features were calculated, and also persistence images of three types (degree 0, degree 1 and degree 2) were obtained for each lesion from the 3D MR imaging data. We used three classification models. In the classification of HCC and MT (resp. HCC and HH, HH and MT), we obtained accuracy of 92% (resp. 90%, 73%) by texture analysis, and the highest accuracy of 85% (resp. 84%, 74%) when degree 1 (resp. degree 1, degree 2) persistence images were used. Our methods using texture analysis or topological data analysis allow for classification of the three hepatic tumors with considerable accuracy, and thus might be useful when applied for computer-aided diagnosis with MR images.

Retrospective distortion correction of diffusion tensor imaging data by semi-elastic image fusion – Evaluation by means of anatomical landmarks

ObjectiveDiffusion tensor imaging (DTI) based on echo-planar imaging (EPI) can suffer from geometric image distortions in comparison to conventional anatomical magnetic resonance imaging (MRI). Therefore, DTI-derived information, such as fiber tractography (FT) used for treatment planning of brain tumors, might be associated with spatial inaccuracies when linearly projected on anatomical MRI. Hence, a non-linear, semi-elastic image fusion shall be evaluated in this study that aims at correcting for image distortions in DTI. Patients and methodsIn a sample of 27 patient datasets, 614 anatomical landmark pairs were retrospectively defined in DTI and T1- or T2-weighted three-dimensional (3D) MRI data. The datasets were processed by a commercial software package (Elements Image Fusion .0; Brainlab AG, Munich, Germany) providing rigid and semi-elastic fusion functionalities, such as DTI distortion correction. To quantify the displacement prior to and after semi-elastic fusion, the Euclidian distances of rigidly and elastically fused landmarks were evaluated by means of descriptive statistics and Bland-Altman plot. ResultsFor rigid and semi-elastic fusion mean target registration errors of 3.03 ± 2.29 mm and 2.04 ± 1.95 mm were found, respectively, with 91% of the evaluated landmarks moving closer to their position determined in T1- or T2-weighted 3D MRI data after distortion correction. Most efficient correction was achieved for non-superficial landmarks showing distortions up to 1 cm. ConclusionThis study indicates that semi-elastic image fusion can be used for retrospective distortion correction of DTI data acquired for image guidance, such as DTI FT as used for a broad range of clinical indications.

A multilinear tongue model derived from speech related MRI data of the human vocal tract

We present a multilinear statistical model of the human tongue that captures anatomical and tongue pose related shape variations separately. The model is derived from 3D magnetic resonance imaging data of 11 speakers sustaining speech related vocal tract configurations. To extract model parameters, we use a minimally supervised method based on an image segmentation approach and a template fitting technique. Furthermore, we use image denoising to deal with possibly corrupt data, palate surface information reconstruction to handle palatal tongue contacts, and a bootstrap strategy to refine the obtained shapes. Our evaluation shows that, by limiting the degrees of freedom for the anatomical and speech related variations, to 5 and 4, respectively, we obtain a model that can reliably register unknown data while avoiding overfitting effects. Furthermore, we show that it can be used to generate plausible tongue animation by tracking sparse motion capture data.

Rapid Interactive and Intuitive Segmentation of 3D Medical Images Using Radial Basis Function Interpolation

Segmentation is one of the most important parts of medical image analysis. Manual segmentation is very cumbersome, time-consuming, and prone to inter-observer variability. Fully automatic segmentation approaches require a large amount of labeled training data and may fail in difficult or abnormal cases. In this work, we propose a new method for 2D segmentation of individual slices and 3D interpolation of the segmented slices. The Smart Brush functionality quickly segments the region of interest in a few 2D slices. Given these annotated slices, our adapted formulation of Hermite radial basis functions reconstructs the 3D surface. Effective interactions with less number of equations accelerate the performance and, therefore, a real-time and an intuitive, interactive segmentation of 3D objects can be supported effectively. The proposed method is evaluated on 12 clinical 3D magnetic resonance imaging data sets and are compared to gold standard annotations of the left ventricle from a clinical expert. The automatic evaluation of the 2D Smart Brush resulted in an average Dice coefficient of 0.88 ± 0.09 for the individual slices. For the 3D interpolation using Hermite radial basis functions, an average Dice coefficient of 0.94 ± 0.02 is achieved.

Enabling Real-Time Volume Rendering of Functional Magnetic Resonance Imaging on an iOS Device.

Powerful non-invasive imaging technologies like computed tomography (CT), ultrasound, and magnetic resonance imaging (MRI) are used daily by medical professionals to diagnose and treat patients. While 2D slice viewers have long been the standard, many tools allowing 3D representations of digital medical data are now available. The newest imaging advancement, functional MRI (fMRI) technology, has changed medical imaging from viewing static to dynamic physiology (4D) over time, particularly to study brain activity. Add this to the rapid adoption of mobile devices for everyday work and the need to visualize fMRI data on tablets or smartphones arises. However, there are few mobile tools available to visualize 3D MRI data, let alone 4D fMRI data. Building volume rendering tools on mobile devices to visualize 3D and 4D medical data is challenging given the limited computational power of the devices. This paper describes research that explored the feasibility of performing real-time 3D and 4D volume raycasting on a tablet device. The prototype application was tested on a 9.7" iPad Pro using two different fMRI datasets of brain activity. The results show that mobile raycasting is able to achieve between 20 and 40 frames per second for traditional 3D datasets, depending on the sampling interval, and up to 9 frames per second for 4D data. While the prototype application did not always achieve true real-time interaction, these results clearly demonstrated that visualizing 3D and 4D digital medical data is feasible with a properly constructed software framework.

Quantification of total fetal brain volume using 3D MR imaging data acquired in utero.

Interpretation of magnetic resonance (MR) imaging of the fetal brain in utero is primarily undertaken using 2D images to provide anatomical information about structural abnormalities. It is now possible to obtain 3D image acquisitions that allow measurement of fetal brain volumes that are potentially useful clinically. The aim of our current work is to provide reference values of total brain volumes obtained from a cohort of low risk fetuses with no abnormalities on ante-natal ultrasonography and in utero MR imaging. Images from volume MR acquisitions of 132 fetuses were used to extract brain volumes by manual segmentation. Reproducibility and reliability were assessed by analysis of the results of two subgroups who had repeated measurements made by the primary and a secondary observer. Intra-observer and inter-observer agreement was high with no statistically significant differences between and within observers (p = 0.476 and p = 0.427, respectively). The results of the brain volume assessments are presented graphically with mean and 95% prediction limits alongside estimates of normal growth rates. We have shown that fetal brain volumes can be reliably extracted from in utero MR (iuMR) imaging 3D datasets with a high degree of reproducibility. The resultant data could potentially be used as a reference tool in the clinical setting. © 2016 John Wiley & Sons, Ltd.

Six-Fold Acceleration of High-Spatial Resolution 3D SPACE MRI of the Knee Through Incoherent k-Space Undersampling and Iterative Reconstruction-First Experience.

The aim of this study was to prospectively test the hypothesis that 6-fold acceleration of a 3-dimensional (3D) turbo spin echo (TSE) magnetic resonance imaging (MRI) pulse sequence with k-space undersampling and iterative reconstruction is feasible for fast high spatial resolution MRI of the knee, while yielding similar image quality and diagnostic performance when compared with a conventional 2-dimensional (2D) TSE MRI standard. The study was approved by the institutional review board. A 10-minute isotropic 3D TSE knee protocol consisting of accelerated intermediate-weighted (repetition time, 900 milliseconds; echo time, 29 milliseconds; voxel size, 0.5 × 0.5 × 0.5 mm; acquisition time, 4:45 minutes) and fat-saturated T2-weighted (repetition time, 900 milliseconds; echo time, 92 milliseconds; voxel size, 0.5 × 0.5 × 0.5 mm; acquisition time, 5:10 minutes) SPACE (sampling perfection with application optimized contrast using different flip angle evolutions) sequence prototypes was compared against a 20-minute 2D TSE standard protocol. The accelerated SPACE sequences were equipped with an optional variable-density poisson-disc pattern as an undersampling mask. An undersampling factor of 0.17 was chosen (6-fold acceleration compared with an acquisition with full sampling). An iterative, sensitivity encoding-type reconstruction with L1 norm-based regularization term was used. The study was performed on a 3 T MRI system using a 15-channel transmit/receive knee coil. The study groups included 15 asymptomatic volunteers and 15 patient volunteers. Quantitative and qualitative assessments were performed by 2 observers. Outcome variables included signal and contrast-to-noise ratio, image quality, and diagnostic accuracy. Qualitative and quantitative measurements were statistically analyzed using nonparametric tests. P values of less than 0.01 were considered significant. The signal-to-noise ratios of 2D and 3D MRI were similar with the exception of fluid, which was brighter on 2D MRI. Relevant contrast-to-noise ratios of 2D MRI were higher than 3D MRI; however, observer ratings for satisfaction, image quality, and visibility of anatomic structures were similar for 2D and 3D MRI. There was moderate to excellent interobserver (κ = 0.54-1.00) and intermethod (κ = 0.54-1.00) agreement for assessing menisci, cartilage, ligaments, cartilage, and bone. Two-dimensional and 3D MRI had similar sensitivity (100%/100%, respectively) and specificity (87%/75%, respectively) for detecting 9 meniscal tears (P = 1.00). We demonstrate the successful clinical implementation of 3D TSE MRI with incoherent k-space undersampling and iterative reconstruction for 6-fold accelerated high spatial resolution isotropic 3D MRI data acquisition. Our preliminary assessments suggest similar image quality and diagnostic performance of a comprehensive 10-minute 3D TSE MRI prototype protocol and 20-minute TSE MRI standard protocol.

The Generalized Log-Ratio Transformation: Learning Shape and Adjacency Priors for Simultaneous Thigh Muscle Segmentation.

We present a novel probabilistic shape representation that implicitly includes prior anatomical volume and adjacency information, termed the generalized log-ratio (GLR) representation. We demonstrate the usefulness of this representation in the task of thigh muscle segmentation. Analysis of the shapes and sizes of thigh muscles can lead to a better understanding of the effects of chronic obstructive pulmonary disease (COPD), which often results in skeletal muscle weakness in lower limbs. However, segmenting these muscles from one another is difficult due to a lack of distinctive features and inter-muscular boundaries that are difficult to detect. We overcome these difficulties by building a shape model in the space of GLR representations. We remove pose variability from the model by employing a presegmentation-based alignment scheme. We also design a rotationally invariant random forest boundary detector that learns common appearances of the interface between muscles from training data. We combine the shape model and the boundary detector into a fully automatic globally optimal segmentation technique. Our segmentation technique produces a probabilistic segmentation that can be used to generate uncertainty information, which can be used to aid subsequent analysis. Our experiments on challenging 3D magnetic resonance imaging data sets show that the use of the GLR representation improves the segmentation accuracy, and yields an average Dice similarity coefficient of 0.808 ±0.074, comparable to other state-of-the-art thigh segmentation techniques.

A Comparative Study of Automated Segmentation Methods for Use in a Microwave Tomography System for Imaging Intracerebral Hemorrhage in Stroke Patients

Microwave technology offers the possibility for pre-hospital stroke detection as we have previously demonstrated using non-imaging diagnostics. The focus in this paper is on image-based diagnostics wherein the technical and computational complexities of image reconstruction are a challenge for clinical realization. Herein we investigate whether information about a patient’s brain anatomy obtained prior to a stroke event can be used to facilitate image-based stroke diagnostics. A priori information can be obtained by segmenting the patient’s head tissues from magnetic resonance images. Expert manual segmentation is presently the gold standard, but it is laborious and subjective. A fully automatic method is thus desirable. This paper presents an evaluation of several such methods using both synthetic magnetic resonance imaging (MRI) data and real data from four healthy subjects. The segmentation was performed on the full 3D MRI data, whereas the electromagnetic evaluation was performed using a 2D slice. The methods were evaluated in terms of: i) tissue classification accuracy over all tissues with respect to ground truth, ii) the accuracy of the simulated electromagnetic wave propagation through the head, and iii) the accuracy of the image reconstruction of the hemorrhage. The segmentation accuracy was measured in terms of the degree of overlap (Dice score) with the ground truth. The electromagnetic simulation accuracy was measured in terms of signal deviation relative to the simulation based on the ground truth. Finally, the image reconstruction accuracy was measured in terms of the Dice score, relative error of dielectric properties, and visual comparison between the true and reconstructed intracerebral hemorrhage. The results show that accurate segmentation of tissues (Dice score = 0.97) from the MRI data can lead to accurate image reconstruction (relative error = 0.24) for the intracerebral hemorrhage in the subject’s brain. They also suggest that accurate automated segmentation can be used as a surrogate for manual segmentation and can facilitate the rapid diagnosis of intracerebral hemorrhage in stroke patients using a microwave imaging system.