Nick Tustison

A computational framework is described that was developed for quantitative analysis of hyperpolarized helium-3 MR lung ventilation image data. This computational framework was applied to a study consisting of 55 subjects (47 asthmatic and eight normal). Each subject was imaged before and after respiratory challenge and also underwent spirometry. Approximately 1600 image features were calculated from the lungs in each image. Both the image and 27 spirometric features were ranked based on their ability to characterize clinical diagnosis using a mutual information-based feature subset selection algorithm. It was found that the top image features perform much better compared with the current clinical gold-standard spirometric values when considered individually. Interestingly, it was also found that spirometric values are relatively orthogonal to these image feature values in terms of informational content.

Abstract A 4-D B-spline model has been created to accurately capture the dynamic motion of the heart from tagged magnetic resonance imaging (MRI). The knot planes of the 4-D B-spline model surfaces deform to fit three orthogonal sets of tag surfaces for all frames within the constraints of the model's spatio-temporal internal energy. As an important byproduct, the points defined by the three-dimensional (3-D) intersections of the three sets of orthogonal tag planes, which we refer to as myocardial beads, can easily be determined ...

abstract Research involving the quantification of left ventricular myocardial strain from cardiac tagged magnetic resonance imaging (MRI) is extensive. Two different imaging geometries are commonly employed by these methodologies to extract longitudinal deformation. We refer to these imaging geometries as either parallel or spiral. In the spiral configuration, four long-axis tagged image slices which intersect along the long-axis of the left ventricle are collected and in the parallel configuration, contiguous tagged long-axis ...

abstract In this paper, an improved framework for estimation of 3-D left-ventricular deformations from tagged MRI is presented. Contiguous short-and long-axis tagged MR images are collected and are used within a 4-D B-Spline based deformable model to determine 4-D displacements and strains. An initial 4-D B-spline model fitted to sparse tag line data is first constructed by minimizing a 4-D Chamfer distance potential-based energy function for aligning isoparametric planes of the model with tag line locations; ...

We present research in which both left and right ventricular deformation is estimated from tagged cardiac magnetic resonance imaging using volumetric deformable models constructed from nonuniform rational B-splines (NURBS). The four model types considered and compared for the left ventricle include two Cartesian NURBS models--one with a cylindrical parameter assignment and one with a prolate spheroidal parameter assignment. The remaining two are non-Cartesian, i.e., prolate spheroidal and cylindrical each with their respective prolate spheroidal and cylindrical parameter assignment regimes. These choices were made based on the typical shape of the left ventricle. For each frame starting with end-diastole, a NURBS model is constructed by fitting two surfaces with the same parameterization to the corresponding set of epicardial and endocardial contours from which a volumetric model is created. Using normal displacements of the three sets of orthogonal tag planes as well as displacements of contour/tag line intersection points and tag plane intersection points, one can solve for the optimal homogeneous coordinates, in a weighted least squares sense, of the control points of the deformed NURBS model at end-diastole using quadratic programming. This allows for subsequent nonrigid registration of the biventricular model at end-diastole to all later time frames. After registration of the model to all later time points, the registered NURBS models are temporally lofted in order to create a comprehensive four-dimensional NURBS model. From the lofted model, we can extract three-dimensional myocardial deformation fields and corresponding Lagrangian and Eulerian strain maps which are local measures of nonrigid deformation. The results show that, in the case of simulated data, the quadratic Cartesian NURBS models with the cylindrical and prolate spheroidal parameter assignments outperform their counterparts in predicting normal strain. The decreased complexity associated with the Cartesian model with the cylindrical parameter assignment prompted its use for subsequent calculations. Lagrangian strains in three canine data, a normal human, and a patient with history of myocardial infarction are presented. Eulerian strains for the normal human data are also included.

ABSTRACT Recent innovations in hyperpolarized helium-3 magnetic resonance imaging (MRI) include the employment of MR tagging techniques for assessment of pulmonary deformation. Historically, such tagging methods have been successfully applied to cardiac research inspiring the development of computational techniques for the quantitative analysis of myocardial deformation. We present related research which concerns the calculation of kinematic quantities, such as displacement and strain, in the lung. Utilizing the high contrast tag lines as landmarks in the images and a recently developed fast n-D B-spline approximation algorithm, we fit a parametric object to the sparse tag line data to smoothly interpolate the underlying deformation field and, subsequently, extract lung kinematic information. We present results from a single human volunteer

Introduction: Advanced Normalization Tools (ANTs) is an image registration program that has been validated for analysis of several tissues, but not previously applied to the disc. The objectives of this study were to optimize ANTs for disc image registration and to validate ANTs strain measurements by comparison with Vic2D, a commercially available software previously applied to quantify disc strain, before

ABSTRACT Although 3He MRI permits compelling visualization of the pulmonary air spaces, quantitation of absolute ventilation is difficult due to confounds such as field inhomogeneity and relative intensity differences between image acquisition; the latter complicating longitudinal investigations of ventilation variation with respiratory alterations. To address these potential difficulties, we present a 4-D segmentation and normalization approach for intra-subject quantitative analysis of lung hyperpolarized 3He MRI. After normalization, which combines bias correction and relative intensity scaling between longitudinal data, partitioning of the lung volume time series is performed by iterating between modeling of the combined intensity histogram as a Gaussian mixture model and modulating the spatial heterogeneity tissue class assignments through Markov random field modeling. Evaluation of the algorithm was retrospectively applied to a cohort of 10 asthmatics between 19-25 years old in which spirometry and 3He MR ventilation images were acquired both before and after respiratory exacerbation by a bronchoconstricting agent (methacholine). Acquisition was repeated under the same conditions from 7 to 467 days (mean +/- standard deviation: 185 +/- 37.2) later. Several techniques were evaluated for matching intensities between the pre and post-methacholine images with the 95th percentile value histogram matching demonstrating superior correlations with spirometry measures. Subsequent analysis evaluated segmentation parameters for assessing ventilation change in this cohort. Current findings also support previous research that areas of poor ventilation in response to bronchoconstriction are relatively consistent over time.

ABSTRACT The utitlity of pulmonary functional imaging techniques, such as hyperpolarized 3He MRI, has encouraged their inclusion in research studies for longitudinal assessment of disease progression and the study of treatment effects. We present methodology for performing voxelwise statistical analysis of ventilation maps derived from hyper­ polarized 3He MRI which incorporates multivariate template construction using simultaneous acquisition of IH and 3He images. Additional processing steps include intensity normalization, bias correction, 4-D longitudinal segmentation, and generation of expected ventilation maps prior to voxelwise regression analysis. Analysis is demonstrated on a cohort of eight individuals with diagnosed cystic fibrosis (CF) undergoing treatment imaged five times every two weeks with a prescribed treatment schedule.

Precise focusing is essential for transcranial MRI-guided focused ultrasound (TcMRgFUS) to minimize collateral damage to non-diseased tissues and to achieve temperatures capable of inducing coagulative necrosis at acceptable power deposition levels. CT is usually used for this refocusing but requires a separate study (CT) ahead of the TcMRgFUS procedure. The goal of this study was to determine whether MRI using an appropriate sequence would be a viable alternative to CT for planning ultrasound refocusing in TcMRgFUS. We tested three MRI pulse sequences (3D T1 weighted 3D volume interpolated breath hold examination (VIBE), proton density weighted 3D sampling perfection with applications optimized contrasts using different flip angle evolution and 3D true fast imaging with steady state precision T2-weighted imaging) on patients who have already had a CT scan performed. We made detailed measurements of the calvarial structure based on the MRI data and compared those so-called 'virtual CT' to detailed measurements of the calvarial structure based on the CT data, used as a reference standard. We then loaded both standard and virtual CT in a TcMRgFUS device and compared the calculated phase correction values, as well as the temperature elevation in a phantom. A series of Bland-Altman measurement agreement analyses showed T1 3D VIBE as the optimal MRI sequence, with respect to minimizing the measurement discrepancy between the MRI derived total skull thickness measurement and the CT derived total skull thickness measurement (mean measurement discrepancy: 0.025; 95% CL (-0.22-0.27); p = 0.825). The T1-weighted sequence was also optimal in estimating skull CT density and skull layer thickness. The mean difference between the phase shifts calculated with the standard CT and the virtual CT reconstructed from the T1 dataset was 0.08 ± 1.2 rad on patients and 0.1 ± 0.9 rad on phantom. Compared to the real CT, the MR-based correction showed a 1 °C drop on the maximum temperature elevation in the phantom (7% relative drop). Without any correction, the maximum temperature was down 6 °C (43% relative drop). We have developed an approach that allows for a reconstruction of a virtual CT dataset from MRI to perform phase correction in TcMRgFUS.

The purpose of this study was to investigate whether diffusion tensor imaging (DTI) of the corticospinal tract (CST) is a reliable surrogate for intraoperative macrostimulation through the deep brain stimulation (DBS) leads. The authors hypothesized that the distance on MRI from the DBS lead to the CST as determined by DTI would correlate with intraoperative motor thresholds from macrostimulations through the same DBS lead. The authors retrospectively reviewed pre- and postoperative MRI studies and intraoperative macrostimulation recordings in 17 patients with Parkinson disease (PD) treated by DBS stimulation. Preoperative DTI tractography of the CST was coregistered with postoperative MRI studies showing the position of the DBS leads. The shortest distance and the angle from each contact of each DBS lead to the CST was automatically calculated using software-based analysis. The distance measurements calculated for each contact were evaluated with respect to the intraoperative voltage thresholds that elicited a motor response at each contact. There was a nonsignificant trend for voltage thresholds to increase when the distances between the DBS leads and the CST increased. There was a significant correlation between the angle and the voltage, but the correlation was weak (coefficient of correlation [R] = 0.36). Caution needs to be exercised when using DTI tractography information to guide DBS lead placement in patients with PD. Further studies are needed to compare DTI tractography measurements with other approaches such as microelectrode recordings and conventional intraoperative MRI-guided placement of DBS leads.

To develop and validate a method for acquiring helium-3 ((3) He) and proton ((1) H) three-dimensional (3D) image sets of the human lung with isotropic spatial resolution within a 10-s breath-hold by using compressed sensing (CS) acceleration, and to assess the fidelity of undersampled images compared with fully sampled images. The undersampling scheme for CS acceleration was optimized and tested using (3) He ventilation data. Rapid 3D acquisition of both (3) He and (1) H data during one breath-hold was then implemented, based on a balanced steady-state free-precession pulse sequence, by random undersampling of k-space with reconstruction by means of minimizing the L1 norm and total variance. CS-reconstruction fidelity was evaluated quantitatively by comparing fully sampled and retrospectively undersampled image sets. Helium-3 and (1) H 3D image sets of the lung with isotropic 3.9-mm resolution were acquired during a single breath-hold in 12 s and 8 s using acceleration factors of 2 and 3, respectively. Comparison of fully sampled and retrospectively undersampled (3) He and (1) H images yielded mean absolute errors <10% and structural similarity indices >0.9. By randomly undersampling k-space and using CS reconstruction, high-quality (3) He and (1) H 3D image sets with isotropic 3.9-mm resolution can be acquired within an 8-s breath-hold. Magn Reson Med, 2014. © 2014 Wiley Periodicals, Inc.

Segmenting and quantifying gliomas from MRI is an important task for diagnosis, planning intervention, and for tracking tumor changes over time. However, this task is complicated by the lack of prior knowledge concerning tumor location, spatial extent, shape, possible displacement of normal tissue, and intensity signature. To accommodate such complications, we introduce a framework for supervised segmentation based on multiple modality intensity, geometry, and asymmetry feature sets. These features drive a supervised whole-brain and tumor segmentation approach based on random forest-derived probabilities. The asymmetry-related features (based on optimal symmetric multimodal templates) demonstrate excellent discriminative properties within this framework. We also gain performance by generating probability maps from random forest models and using these maps for a refining Markov random field regularized probabilistic segmentation. This strategy allows us to interface the supervised learning capabilities of the random forest model with regularized probabilistic segmentation using the recently developed ANTsR package-a comprehensive statistical and visualization interface between the popular Advanced Normalization Tools (ANTs) and the R statistical project. The reported algorithmic framework was the top-performing entry in the MICCAI 2013 Multimodal Brain Tumor Segmentation challenge. The challenge data were widely varying consisting of both high-grade and low-grade glioma tumor four-modality MRI from five different institutions. Average Dice overlap measures for the final algorithmic assessment were 0.87, 0.78, and 0.74 for "complete", "core", and "enhanced" tumor components, respectively.

Rapid regional fluctuations in GABA may result in inhomogeneous concentrations throughout the brain parenchyma. The goal of this study is to provide further insight into the natural distribution of GABA throughout the brain and thus determine if a surrogate site may be used for spectroscopy when evaluating motor diseases, neurological disorders, or psychiatric dysfunction. In this prospective study, eight healthy volunteers underwent spectroscopic evaluation of the frontal lobe, occipital lobe, lateral temporal lobe, basal ganglia, and both hippocampi using a spin echo variant of a J-difference editing method. Knowledge of the relative peak intensities of the macromolecule peaks at 3 ppm and 0.9 ppm was used to correct the contribution of co-edited macromolecules to the GABA peak at 3 ppm. The GABA values were internally referenced to NAA. Linear regression was used to normalize the effect of regional tissue-fraction variation on the GABA/NAA values. A one-way ANOVA was performed with Tukey's multiple comparison test to compare the normalized GABA/NAA values in each pair of locations. After accounting for the macromolecule contribution to the GABA signal and correction for tissue fraction variation, the normalized GABA/NAA ratios differ significantly between the six brain locations (p<0.001). Pairwise comparisons of the corrected normalized GABA/NAA ratios show statistically significant variation between the frontal lobe and the basal ganglia, frontal and lateral temporal lobes, and frontal lobe and right hippocampus. Variations in the normalized GABA/NAA ratios trend toward significance between the frontal lobe and left hippocampus, occipital lobe and the frontal lobe, occipital lobe and basal ganglia, and occipital lobe and right hippocampus.. Our study suggests that GABA concentration is inhomogeneous throughout the parenchyma. Studies evaluating the role of GABA must carefully consider voxel placement when incorporating spectroscopy.

To use diffusion-tensor (DT) magnetic resonance (MR) imaging in patients with essential tremor who were treated with transcranial MR imaging-guided focused ultrasound lesion inducement to identify the structural connectivity of the ventralis intermedius nucleus of the thalamus and determine how DT imaging changes correlated with tremor changes after lesion inducement. With institutional review board approval, and with prospective informed consent, 15 patients with medication-refractory essential tremor were enrolled in a HIPAA-compliant pilot study and were treated with transcranial MR imaging-guided focused ultrasound surgery targeting the ventralis intermedius nucleus of the thalamus contralateral to their dominant hand. Fourteen patients were ultimately included. DT MR imaging studies at 3.0 T were performed preoperatively and 24 hours, 1 week, 1 month, and 3 months after the procedure. Fractional anisotropy (FA) maps were calculated from the DT imaging data sets for all time points in all patients. Voxels where FA consistently decreased over time were identified, and FA change in these voxels was correlated with clinical changes in tremor over the same period by using Pearson correlation. Ipsilateral brain structures that showed prespecified negative correlation values of FA over time of -0.5 or less included the pre- and postcentral subcortical white matter in the hand knob area; the region of the corticospinal tract in the centrum semiovale, in the posterior limb of the internal capsule, and in the cerebral peduncle; the thalamus; the region of the red nucleus; the location of the central tegmental tract; and the region of the inferior olive. The contralateral middle cerebellar peduncle and bilateral portions of the superior vermis also showed persistent decrease in FA over time. There was strong correlation between decrease in FA and clinical improvement in hand tremor 3 months after lesion inducement (P < .001). DT MR imaging after MR imaging-guided focused ultrasound thalamotomy depicts changes in specific brain structures. The magnitude of the DT imaging changes after thalamic lesion inducement correlates with the degree of clinical improvement in essential tremor.

Non-uniform distribution of inspired gas within the lung, termed ventilation heterogeneity, is present in patients with even mild asthma. Current evidence strongly supports ventilation heterogeneity as a fundamental derangement of lung function in asthma that contributes per se to hypoxemia and airway hyper-responsiveness. An extreme example of ventilation heterogeneity is the identification by hyperpolarized gas MRI of lung regions with no ventilation, termed filling defects. Lung filling defects in patients with asthma can persist over time, increase in size with methacholine-induced bronchospasm and more likely are caused by obstruction of the peripheral and not the proximal airways. Ventilation heterogeneity can be quantified in the conducting and acinar lung zones with the multiple gas washout method, and in the acinar zone does not fully resolve following bronchodilator treatment in patients with asthma. In prospective studies, the degree of ventilation heterogeneity at baseline predicts airway hyper-responsiveness and response to corticosteroid dose titration. An important unanswered question is the relationship of airways inflammation to ventilation heterogeneity. In consideration of the importance of ventilation heterogeneity in its pathobiology, asthma is more a focal disorder with regional pathology akin to regional ileitis and not the generalized disorder of the airways as it has been viewed in the past.

ABSTRACT Motivated by previous work [16] and recent diffeomorphic image registration developments in which the characteristic velocity field is represented by spatiotemporal B-splines [3], we present a diffeomorphic B-spline-based image registration algorithm combining and extending these techniques. The advancements of the proposed framework over previous work include a preconditioned gradient descent algorithm and potential weighting of the metric gradient permitting, among other things, enforcement of stationary boundary conditions. In addition to theoretical and practical discussions of our contribution, we also describe its parallelized implementation as open source in the Insight Toolkit and perform an evaluation on publicly available brain data.

ABSTRACT Publicly available scientific resources help establish evaluation standards, provide a platform for teaching and may improve reproducibility. Version 4 of the Insight ToolKit ( ITK $^{\text{4}}$) seeks to establish new standards in publicly available image registration methodology. In this work, we provide an overview and preliminary evaluation of the revised toolkit against registration based on the previous major ITK version (3.20). Furthermore, we propose a nomenclature that may be used to discuss registration frameworks via schematic representations. In total, the ITK$^{\text{4}}$ contribution is intended as a structure to support reproducible research practices, will provide a more extensive foundation against which to evaluate new work in image registration and also enable application level programmers a broad suite of tools on which to build.

Publicly available scientific resources help establish evaluation standards, provide a platform for teaching and improve reproducibility. Version 4 of the Insight ToolKit (ITK(4)) seeks to establish new standards in publicly available image registration methodology. ITK(4) makes several advances in comparison to previous versions of ITK. ITK(4) supports both multivariate images and objective functions; it also unifies high-dimensional (deformation field) and low-dimensional (affine) transformations with metrics that are reusable across transform types and with composite transforms that allow arbitrary series of geometric mappings to be chained together seamlessly. Metrics and optimizers take advantage of multi-core resources, when available. Furthermore, ITK(4) reduces the parameter optimization burden via principled heuristics that automatically set scaling across disparate parameter types (rotations vs. translations). A related approach also constrains steps sizes for gradient-based optimizers. The result is that tuning for different metrics and/or image pairs is rarely necessary allowing the researcher to more easily focus on design/comparison of registration strategies. In total, the ITK(4) contribution is intended as a structure to support reproducible research practices, will provide a more extensive foundation against which to evaluate new work in image registration and also enable application level programmers a broad suite of tools on which to build. Finally, we contextualize this work with a reference registration evaluation study with application to pediatric brain labeling.

Diffeomorphic mappings are central to image registration due largely to their topological properties and success in providing biologically plausible solutions to deformation and morphological estimation problems. Popular diffeomorphic image registration algorithms include those characterized by time-varying and constant velocity fields, and symmetrical considerations. Prior information in the form of regularization is used to enforce transform plausibility taking the form of physics-based constraints or through some approximation thereof, e.g., Gaussian smoothing of the vector fields [a la Thirion's Demons (Thirion, 1998)]. In the context of the original Demons' framework, the so-called directly manipulated free-form deformation (DMFFD) (Tustison et al., 2009) can be viewed as a smoothing alternative in which explicit regularization is achieved through fast B-spline approximation. This characterization can be used to provide B-spline "flavored" diffeomorphic image registration solutions with several advantages. Implementation is open source and available through the Insight Toolkit and our Advanced Normalization Tools (ANTs) repository. A thorough comparative evaluation with the well-known SyN algorithm (Avants et al., 2008), implemented within the same framework, and its B-spline analog is performed using open labeled brain data and open source evaluation tools.

The ability to reconstruct multivariate approximating or interpolating functions from sampled data finds many practical applications in medical image analysis. Parameterized reconstruction methods employing B-splines have typically utilized least-squares methodology for data fitting. For large sample sets, solving the resulting linear system is computationally demanding as well as susceptible to ill-conditioning. We present a generalization of a previously proposed fast

ABSTRACT The recent discovery of methodological flaws in experimental design and analysis in neuroscience research has raised concerns over the validity of certain techniques used in routine analyses and their corresponding findings. Such concerns have centered around selection bias whereby data is inadvertently manipulated such that the resulting analysis produces falsely increased statistical significance, i.e. type I errors. This has been illustrated recently in flv1RI studies, with excessive flexibility in data collection, and general experimental design issues. Current work from our group has shown how this problem extends to generic voxel-based analysis (and certain technique derivatives such as tract- based spatial statistics) using fractional anisotropy images derived from diffusion tensor imaging. In this work, we demonstrate how this circularity principle can potentially extend to the well-known optimized voxel-based morphometry technique for assessing cortical density differences whereby the principal cause of experimental corruption is due to normalization strategy. Specifically, the popular sum­ of-squared-differences (SSD) metric explicitly optimizes statistical findings potentially inflating type I errors. Additional experimentation demonstrates that this problem is not restricted to the SSD metric but extends to other commonly used metrics such as mutual information, neighborhood cross correlation, and Demons.

ABSTRACT Numerous studies have explored the relationship between cortical structure and brain development, cognitive function, and functional connectivity. The highly convoluted cortical topography makes manual measurements arduous and often impractical given the population sizes necessary for sufficient statistical power. Computational techniques have permitted large-scale studies as they provide robust and reliable localized measurements characterizing the cortex with little or no human intervention. Particularly useful to the neuroscience community are publicly available tools, such as the popular surface-based Freesurfer, which facilitate the testing and refinement of hypotheses. In this paper, we introduce the volume-based Advanced Normalization Tools (ANTs) cortical thickness automated pipeline comprising well-vetted components such as SyGN (multivariate template construction), SyN (image registration), N4 (bias correction), Atropos (n-tissue segmentation), and DiReCT (cortical thickness) all developed as part of the ANTs open science effort. Complementing the open source aspect of ANTs we demonstrate its utility using the publicly available IXI data set.

Study objectives were to develop, validate, and apply a method to measure three-dimensional (3D) internal strains in intact human discs under axial compression. A custom-built loading device applied compression and permitted load-relaxation outside of the magnet while also maintaining compression and hydration during imaging. Strain was measured through registration of 300 μm isotropic resolution images. Excellent registration accuracy was achieved, with 94% and 65% overlap of disc volume and lamellae compared to manual segmentation, and an average Hausdorff, a measure of distance error, of 0.03 and 0.12 mm for disc volume and lamellae boundaries, respectively. Strain maps enabled qualitative visualization and quantitative regional annulus fibrosus (AF) strain analysis. Axial and circumferential strains were highest in the lateral AF and lowest in the anterior and posterior AF. Radial strains were lowest in the lateral AF, but highly variable. Overall, this study provided new methods that will be valuable in the design and evaluation surgical procedures and therapeutic interventions.

... distinct methods are available. However, few comparison studies are reported [see Boesen HBM 2003 and Fennema-Notestine 2006]. This study may be the first to use both ... reported in the abstract. Brian Avants1, Arno Klein2, Nicholas Tustison1, Jue Woo1 and James C. Gee1 ...

Hyperpolarized gas MR virtual colonography was performed in plastic phantoms and in the dog colon. (3)He was laser polarized in a prototype commercial system. 2D and 3D gradient echo sequences were used to image the noble gas-filled structures. The hyperpolarized (3)He within the plastic tube and colon lumen produced high signal, providing excellent contrast from the surrounding structures. The virtual colonoscopic analysis of the canine dataset allowed visualization of the colonic features and the colonic wall from inside the colon. (3)He colonoscopy is a novel technique to visualize the colon with MRI with the application of an inert gaseous endoluminal contrast agent.

PurposeTo propose and test the feasibility of a novel method for quantifying 3-D regional pulmonary kinematics from hyperpolarized helium-3 tagged MRI in human subjects using a tailored image processing pipeline and a recently developed nonrigid registration framework.Materials and MethodsFollowing image acquisition, inspiratory and expiratory tagged helium-3 MR images were preprocessed using various image filtering techniques to enhance the tag surfaces. Segmentation of the three orthogonal sets of tag planes in each lung produced distinct point-set representations of the tag surfaces. Using these labeled point-sets, deformation fields and corresponding strain maps were obtained via nonrigid point-set registration. Kinematic analysis was performed on three volunteers.ResultsTag lines in inspiratory and expiratory images were co-registered producing a continuous 3-D correspondence mapping. Average displacement and directional strains were calculated in three subjects in the inferior, mid, and superior portions of the right and left lungs. As expected, the predominant direction of displacements with expiration is from inferior to superior.ConclusionKinematic quantitation of pulmonary motion using tagged helium-3 MRI is feasible using the applied image preprocessing filtering techniques and nonrigid point-set registration. Potential benefits from regional pulmonary kinematic quantitation include the facilitation of diagnosis and local assessment of disease progression.