Anisotropic Voxel Size Research Articles

Lesion delineation framework for vestibular schwannoma, meningioma and brain metastasis for gamma knife radiosurgery using stereotactic magnetic resonance images

Background and objectiveGKRS is an effective treatment for smaller intracranial tumors with a high control rate and low risk of complications. Target delineation in medical MR images is essential in the planning of GKRS and follow-up. A deep learning–based algorithm can effectively segment the targets from medical images and has been widely explored. However, state-of-the-art deep learning–based target delineation uses fixed sizes, and the isotropic voxel size may not be suitable for stereotactic MR images which use different anisotropic voxel sizes and numbers of slices according to the lesion size and location for clinical GKRS planning. This study developed an automatic deep learning–based segmentation scheme for stereotactic MR images. MethodsWe retrospectively collected stereotactic MR images from 506 patients with VS, 1,069 patients with meningioma and 574 patients with BM who had been treated using GKRS; the lesion contours and individual T1W+C and T2W MR images were extracted from the GammaPlan system. The three-dimensional patching-based training strategy and dual-pathway architecture were used to manage inconsistent FOVs and anisotropic voxel size. Furthermore, we used two-parametric MR image as training input to segment the regions with different image characteristics (e.g., cystic lesions) effectively. ResultsOur results for VS and BM demonstrated that the model trained using two-parametric MR images significantly outperformed the model trained using single-parametric images with median Dice coefficients (0.91, 0.05 versus 0.90, 0.06, and 0.82, 0.23 versus 0.78, 0.34, respectively), whereas predicted delineations in meningiomas using the dual-pathway model were dominated by single-parametric images (median Dice coefficients 0.83, 0.17 versus 0.84, 0.22). Finally, we combined three data sets to train the models, achieving the comparable or even higher testing median Dice (VS: 0.91, 0.07; meningioma: 0.83, 0.22; BM: 0.84, 0.23) in three diseases while using two-parametric as input. ConclusionsOur proposed deep learning–based tumor segmentation scheme was successfully applied to multiple types of intracranial tumor (VS, meningioma and BM) undergoing GKRS and for segmenting the tumor effectively from stereotactic MR image volumes for use in GKRS planning.

The ImSURE phantoms: a digital dataset for radiomic software benchmarking and investigation

In radiology and oncology, radiomic models are increasingly employed to predict clinical outcomes, but their clinical deployment has been hampered by lack of standardisation. This hindrance has driven the international Image Biomarker Standardisation Initiative (IBSI) to define guidelines for image pre-processing, standardise the formulation and nomenclature of 169 radiomic features and share two benchmark digital phantoms for software calibration. However, to better assess the concordance of radiomic tools, more heterogeneous phantoms are needed. We created two digital phantoms, called ImSURE phantoms, having isotropic and anisotropic voxel size, respectively, and 90 regions of interest (ROIs) each. To use these phantoms, we designed a systematic feature extraction workflow including 919 different feature values (obtained from the 169 IBSI-standardised features considering all possible combinations of feature aggregation and intensity discretisation methods). The ImSURE phantoms will allow to assess the concordance of radiomic software depending on interpolation, discretisation and aggregation methods, as well as on ROI volume and shape. Eventually, we provide the feature values extracted from these phantoms using five open-source IBSI-compliant software.

MRI-based assessment of proximal femur strength compared to mechanical testing.

Half of the women who sustain a hip fracture would not qualify for osteoporosis treatment based on current DXA-estimated bone mineral density criteria. Therefore, a better approach is needed to determine if an individual is at risk of hip fracture from a fall. The objective of this study was to determine the association between radiation-free MRI-derived bone strength and strain simulations compared to results from direct mechanical testing of cadaveric femora. Imaging was conducted on a 3-Tesla MRI scanner using two sequences: one balanced steady-state free precession sequence with 300μm isotropic voxel size and one spoiled gradient echo with anisotropic voxel size of 234×234×1500μm. Femora were dissected free of soft-tissue and 4350-ohm strain-gauges were securely applied to surfaces at the femoral shaft, inferior neck, greater trochanter, and superior neck. Cadavers were mechanically tested with a hydraulic universal test frame to simulate loading in a sideways fall orientation. Sideways fall forces were simulated on MRI-based finite element meshes and bone stiffness, failure force, and force for plastic deformation were computed. Simulated bone strength metrics from the 300μm isotropic sequence showed strong agreement with experimentally obtained values of bone strength, with stiffness (r=0.88, p=0.0002), plastic deformation point (r=0.89, p<0.0001), and failure force (r=0.92, p<0.0001). The anisotropic sequence showed similar trends for stiffness, plastic deformation point, and failure force (r=0.68, 0.70, 0.84; p=0.02, 0.01, 0.0006, respectively). Surface strain-gauge measurements showed moderate to strong agreement with simulated magnitude strain values at the greater trochanter, superior neck, and inferior neck (r=-0.97, -0.86, 0.80; p≤0.0001, 0.003, 0.03, respectively). The findings from this study support the use of MRI-based FE analysis of the hip to reliably predict the mechanical competence of the human femur in clinical settings.

Local up-sampling and morphological analysis of low-resolution magnetic resonance images

Limitations in the resolution of acquired images, which are due to sensor manufacturing and acquisition conditions, are reduced with the help of algorithms that enhance the spatial resolution by assigning pixel values that are interpolated or approximated from known pixels. We propose a variant of the moving least-squares approximation for image up-sampling, with a specific focus on biomedical MR images. For each evaluation point, we locally compute the best approximation by minimizing a weighted least-squares error between the input data and their approximation with an implicit function. The proposed approach provides a continuous approximation, an accuracy and extrapolation capabilities higher than previous work, and a lower computational cost. As main application, we consider the up-sampling of low field MR images, where the volumetric and meshless properties of the approximation allow us to easily process images with anisotropic voxel size by rescaling the image and inter-slices resolution. Finally, we include the resolution rescaling into a pipeline that performs a morphological characterization of 3D anatomical districts, which has been developed with a focus on rheumatoid arthritis evolution and provides a more accurate segmentation as an input to quantitative analysis.

Abstract WP205: Pipeline for Automated White Matter Hyperintensity Segmentation in Patients With Acute Ischemic Stroke: The MRI-GENIE Study

Introduction: White matter hyperintensity volume (WMHv) is an important and highly heritable cerebrovascular phenotype; however, manual or semi-automated approaches to clinically acquired MRI analysis hinder large-scale studies in acute ischemic stroke (AIS). In this work, we develop a high-throughput, fully automated WMHv analysis pipeline for clinical fluid-attenuated inversion recovery (FLAIR) images to facilitate rapid genetic discovery in AIS. Methods: Automated WMHv extraction from multiple subjects relies on significant pre-processing of medical scans, including co-registration of the images. To reduce the effects of anisotropic voxel sizes, each FLAIR image is upsampled using bi-cubic interpolation. Brain extraction is performed using RObust Brain EXtraction (ROBEX). Images are then registered to an in-house FLAIR template using Advanced Normalization Tools (ANTs). The spatial covariation of WMH is learned through principal component analysis (PCA) of manual outlines from 100 subjects. Areas of leukoaraiosis are identified and separated from other lesions using the PCA modes. Volumes are then computed using non-interpolated slices for each subject. Standard deviation (SD) in WMHv (9 subjects; 6 raters each) is calculated as a measure of variability. Good agreement between automated and manual outlines is assessed in 358 subjects (automated WMHv within 3SD of manual WMHv). Results: As part of the MRI - Gen etics I nterface E xploration (MRI-GENIE) study, WMHv were calculated on a set of 2703 FLAIR images of patients from 12 independent AIS cohorts (sites). Results are shown in Figure 1. Comparing manual and automated WMHv shows that 88% of the automated WMHv fall within 3 SD from the manual WMHv, suggesting good agreement. Conclusion: WMHv segmentation using a fully-automated pipeline for analysis of clinical MRIs is both feasible and accurate. Ongoing analysis of the extracted WMHv is expected to advance current knowledge of risks and outcomes in AIS.

A New Variational Method for Bias Correction and Its Applications to Rodent Brain Extraction.

Brain extraction is an important preprocessing step for further analysis of brain MR images. Significant intensity inhomogeneity can be observed in rodent brain images due to the high-field MRI technique. Unlike most existing brain extraction methods that require bias corrected MRI, we present a high-order and L0 regularized variational model for bias correction and brain extraction. The model is composed of a data fitting term, a piecewise constant regularization and a smooth regularization, which is constructed on a 3-D formulation for medical images with anisotropic voxel sizes. We propose an efficient multi-resolution algorithm for fast computation. At each resolution layer, we solve an alternating direction scheme, all subproblems of which have the closed-form solutions. The method is tested on three T2 weighted acquisition configurations comprising a total of 50 rodent brain volumes, which are with the acquisition field strengths of 4.7 Tesla, 9.4 Tesla and 17.6 Tesla, respectively. On one hand, we compare the results of bias correction with N3 and N4 in terms of the coefficient of variations on 20 different tissues of rodent brain. On the other hand, the results of brain extraction are compared against manually segmented gold standards, BET, BSE and 3-D PCNN based on a number of metrics. With the high accuracy and efficiency, our proposed method can facilitate automatic processing of large-scale brain studies.

Diffusion Tensor Imaging and Fiber Tractography in Children with Craniosynostosis Syndromes.

Patients with craniosynostosis syndromes caused by mutations in FGFR-2, FGFR-3, and TWIST1 genes are characterized by having prematurely fused skull sutures and skull base synchondroses, which result in a skull deformity and are accompanied by brain anomalies, including altered white matter microarchitecture. In this study, the reliability and reproducibility of DTI fiber tractography was investigated in these patients. The outcomes were compared with those of controls. DTI datasets were acquired with a 1.5T MR imaging system with 25 diffusion gradient orientations (voxel size = 1.8 × 1.8 × 3.0 mm(3), b-value = 1000 s/mm(2)). White matter tracts studied included the following: corpus callosum, cingulate gyrus, fornix, corticospinal tracts, and medial cerebellar peduncle. Tract pathways were reconstructed with ExploreDTI in 58 surgically treated patients with craniosynostosis syndromes and 7 controls (age range, 6-18 years). Because of the brain deformity and abnormal ventricular shape and size, DTI fiber tractography was challenging to perform in patients with craniosynostosis syndromes. To provide reliable tracts, we adapted standard tracking protocols. Fractional anisotropy was equal to that in controls (0.44 versus 0.45 ± 0.02, P = .536), whereas mean, axial, and radial diffusivity parameters of the mean white matter were increased in patients with craniosynostosis syndromes (P < .001). No craniosynostosis syndrome-specific difference in DTI properties was seen for any of the fiber tracts studied in this work. Performing DTI fiber tractography in patients with craniosynostosis syndromes was difficult due to partial volume effects caused by an anisotropic voxel size and deformed brain structures. Although these patients have a normal fiber organization, increased diffusivity parameters suggest abnormal microstructural tissue properties of the investigated white matter tracts.

MRI upsampling using feature-based nonlocal means approach.

In magnetic resonance imaging (MRI), spatial resolution is limited by several factors such as acquisition time, short physiological phenomena, and organ motion. The acquired image usually has higher resolution in two dimensions (the acquisition plane) in comparison with the third dimension, resulting in highly anisotropic voxel size. Interpolation of these low resolution (LR) images using standard techniques, such as linear or spline interpolation, results in distorted edges in the planes perpendicular to the acquisition plane. This poses limitation on conducting quantitative analyses of LR images, particularly on their voxel-wise analysis and registration. We have proposed a new non-local means feature-based technique that uses structural information of a high resolution (HR) image with a different contrast and interpolates the LR image. In this approach, the similarity between voxels is estimated using a feature vector that characterizes the laminar pattern of the brain structures, resulting in a more accurate similarity measure in comparison with conventional patch-based approach. This technique can be applied to LR images with both anisotropic and isotropic voxel sizes. Experimental results conducted on brain MRI scans of patients with brain tumors, multiple sclerosis, epilepsy, as well as schizophrenic patients and normal controls show that the proposed method is more accurate, requires fewer computations, and thus is significantly faster than a previous state-of-the-art patch-based technique. We also show how the proposed method may be used to upsample regions of interest drawn on LR images.

Structural and mechanical parameters of trabecular bone estimated from in vivo high‐resolution magnetic resonance images at 3 tesla field strength

To evaluate the performance of a new 3 Tesla (T) high-resolution trabecular bone (TB) imaging technique at two resolution regimens in terms of serial reproducibility and sensitivity. The left distal tibial metaphysis of seven healthy volunteers was imaged at three time-points using a FLASE (fast large-angle spin-echo) pulse sequence at 137 x 137 x 410 mum(3) and (160 mum)(3) voxel sizes. Image artifacts, motion degradation, and serial image volume misalignments were controlled to maximize reproducibility of image-derived measures of scale, topology, orientation in terms of structural anisotropy, and finite-element derived Young's and shear moduli. Coefficients of variation (CV) and intraclass correlation coefficients (ICC) for structural and mechanical parameters were evaluated as measures of reproducibility and reliability. The ability of structural and mechanical parameters to distinguish between subjects was tested by analysis of variance. Reproducibility was generally higher in the anisotropic data (CVs 1-5% versus 1-9% for isotropic images). Anisotropic voxel size yielded greater measurement reliability (ICCs 0.75-0.99, mean = 0.92 versus 0.62-0.99, mean = 0.86) and better discrimination of the seven subjects (75% versus 50% of the possible comparisons were significantly different [P < 0.05]) except for measures of structural anisotropy and topology. Isotropic resolution improved detection of structural orientation and permitted visualization of small perforations in longitudinal trabecular plates not detected at anisotropic resolution. Improved image acquisition and processing techniques enhance reproducibility of structural and mechanical parameters derived from high-resolution MRI of metaphyseal bone in the distal tibia.

Fast prospective registration of in vivo MR images of trabecular bone microstructure in longitudinal studies

In micro-MRI studies of trabecular bone designed to evaluate structural changes in response to intervention, follow-up scan volumes do not typically align exactly with the baseline scan volumes due to the orientation and placement of the anatomic location, here the distal tibia, relative to the scanner coordinates. Failure of accurate registration of the follow-up to the baseline images introduces errors due to the inherent anisotropy in the trabecular network and anisotropic voxel size. In this work it is shown that these limitations can be overcome by incorporating on-line prospective registration into the data acquisition protocol. The technique is based on a short 3D localizer scan of 1 mm isotropic resolution prior to acquiring the high-resolution images. During the follow-up exam, localizer images are registered on-site with an algorithm relying on a fast Fourier transform for maximizing the correlation between baseline and follow-up localizers. Transformation parameters obtained in this manner are then fed into the scanner software so that the imaging slab for the high-resolution follow-up images is automatically positioned consistent with that of the baseline scan. Based on phantom and human subject studies it is shown that prospective registration yields very close matching between baseline and follow-up imaging volumes.

Visualizing human brain surface from T 1-weighted MR images using texture-mapped triangle meshes

We describe a novel method for visualizing brain surface from anatomical magnetic resonance images (MRIs). The method utilizes standard 2D texture mapping capabilities of OpenGL graphics language. It combines the benefits of volume rendering and triangle-mesh rendering, allowing fast and realistic-looking brain surface visualizations. Consequently, relatively low-resolution triangle meshes can be used while the texture images provide the necessary details. The mapping is optimized to provide good texture-image resolution for the triangles with respect to their original sizes in the 3D MRI volume. The actual 2D texture images are generated by depth integration from the original MRI data. Our method adapts to anisotropic voxel sizes without any need to interpolate the volume data into cubic voxels, and it is very well suited for visualizing brain anatomy from standard T 1-weighted MR images. Furthermore, other OpenGL objects and techniques can be easily combined, for example, to use cut planes, to show other surfaces and objects, and to visualize functional data in addition to the anatomical information.