Skip to main content

Über dieses Buch

This book constitutes the thoroughly refereed post-conference proceedings of the Second International Workshop on Statistical Atlases and Computational Models of the Heart: Imaging and Modelling Challegenges, STACOM 2011, held in conjunction with MICCAI 2011, in Toronto, Canada, in September 2011.

The 28 revised full papers were carefully reviewed and selected from numerous submissions. The papers are organized in topical sections on EP simulation challenge, motion tracking challenge, segmentation challenge, and regular papers.



EP Simulation Challenge

EP Challenge - STACOM’11: Forward Approaches to Computational Electrophysiology Using MRI-Based Models and In-Vivo CARTO Mapping in Swine Hearts

Our broad aim is to integrate experimental measurements (electro-cardiographic and MR) and cardiac computer models, for a better understanding of transmural wave propagation in individual hearts. In this paper, we first describe the acquisition and processing of the data provided to the EP simulation challenge organized at STACOM’11. The measurements were obtained in two swine hearts (i.e., one healthy and one with chronic infarction) and comprise in-vivo electro-anatomical CARTO maps (e.g., surfacic endo-/epicardial depolarization maps and bipolar voltage maps recorded in sinus rhythm), and high-resolution ex-vivo diffusion-weighted DW-MR images (voxel size < 1mm3). We briefly detail how we built anisotropic 3D MRI-based models for these two hearts, with fiber directions obtained using DW-MRI methods (which also allowed for infarct identification). We then focus on applications in cardiac modelling concerning propagation of depolarization wave, by employing forward mathematical approaches. Specifically, we present simulation results for the depolarization wave using a fast, macroscopic monodomain formalism (i.e., the two-variable Aliev-Panfilov model) and comparisons with measured depolarization times. We also include simulations obtained using the healthy heart and a simple Eikonal model, as well as a complex bidomain model. The results demonstrate small differences between computed isochrones using these computer models; specifically, we calculated a mean error ± S.D. of 2.8 ± 1.67 ms between Aliev-Panfilov and Eikonal models, and 6.1 ± 3.9 ms between Alie-Panfilov and bidomain models, respectively.
Mihaela Pop, Maxime Sermesant, Tommaso Mansi, Eugene Crystal, Sudip Ghate, Jatin Relan, Charles Pierre, Yves Coudiere, Jennifer Barry, Ilan Lashevsky, Beiping Qiang, Elliot R. McVeigh, Nicholas Ayache, Graham A. Wright

Personalisation of a 3D Ventricular Electrophysiological Model, Using Endocardial and Epicardial Contact Mapping and MRI

Personalisation, i.e. parameter estimation of a cardiac ElectroPhysiology (EP) model is needed to build patient-specific models, which could then be used to understand and predict the complex dynamics involved in patient’s pathology. In this paper, we present an EP model personalisation approach applied to an infarcted porcine heart, using contact mapping data and Diffusion Tensor MRI. The contact mapping data was gathered during normal sinus rhythm, on the ventricles in-vivo, endocardially as well as epicardially, using a CARTO mapping system. The Diffusion Tensor MRI was then obtained ex-vivo, in order to have the true cardiac fibre orientations, for the infarcted heart. Both datasets were then used to build and personalise the 3D ventricular electrophysiological model, with the proposed personalisation approach. Secondly, the effect of using only endocardial mapping or epicardial mapping measurements, on the personalised EP model was also tested.
Jatin Relan, Maxime Sermesant, Hervé Delingette, Nicholas Ayache

Transmural Electrophysiologic and Scar Imaging on Porcine Heart with Chronic Infarction

Myocardial scar is the most common substrate for malignant arrhythmia and cardiac arrest. Radiofrequency ablation, as one of the emerging mainstream therapies, is subject to limited success rate because of the inadequate assessment of scar substrates that currently relies on electrophysiologic (EP) map acquired on endocardial and occasionally epicardial surfaces. As myocardial scar is often complex with shapes varying with the depth of the myocardium, endocardial and epicardial maps may differ substantially, and may fail to identify mid-wall fibrosis that exist in ~30% of patients with nonischemic cardiomyopathy. Alternatively, noninvasive and transmural scar delineation by current imaging techniques does not always show electrically altered functional substrates. Participating in CESC’11, we presented a new application of the previously developed method of transmural EP imaging, where epicardial unipolar electrograms acquired by CARTO together with MRI-derived ventricular anatomical data of a porcine heart with chronicle myocardial infarction were used for computing the transmural EP dynamics and subsequently classifying conduction blocks of the porcine heart. Validation was performed versus CARTO electroanatomic maps on the epicardium and endocardium, as well as DW-MRI enhanced anatomical scars. This allowed detailed examinations of the reported method in computing transmural EP anomalies using only surface data and without any condition-specific knowledge a priori, which could not be achieved with either current EP mapping or medical imaging techniques alone.
Linwei Wang, Fady Dawoud, Ken C. L. Wong, Heye Zhang, Huafeng Liu, Albert C. Lardo, Pengcheng Shi

Motion Tracking Challenge

A Multimodal Database for the 1 st Cardiac Motion Analysis Challenge

This paper describes the acquisition of the multimodal database used in the 1 st Cardiac Motion Analysis Challenge. The database includes magnetic resonance (MR) and 3D ultrasound (3DUS) datasets from a dynamic phantom and 15 datasets from healthy volunteers. The MR acquisition included cine steady state free precession (SSFP), whole-heart turbo field echo (TFE), and 4D tagged MR (tMR) sequences. From the SSFP images, the end diastolic anatomy was extracted using a deformable model of the left ventricle (LV). The LV model was mapped to the tMR coordinates using DICOM information. From the LV model, 12 landmarks were generated (4 walls at 3 ventricular levels). These landmarks were manually tracked in the tMR data over the whole cardiac cycle by two observes using an in-house application with 4D visualization capabilities. Finally, the LV model was registered to the 3DUS coordinates using a point based similarity transform. Four institutions responded to the challenge by providing motion estimates for the data. Preliminary results are presented for one of the volunteer data sets.
Catalina Tobon-Gomez, Mathieu De Craene, Annette Dahl, Stam Kapetanakis, Gerry Carr-White, Anja Lutz, Volker Rasche, Patrick Etyngier, Sebastian Kozerke, Tobias Schaeffter, Chiara Riccobene, Yves Martelli, Oscar Camara, Alejandro F. Frangi, Kawal S. Rhode

Automatic Cardiac Motion Tracking Using Both Untagged and 3D Tagged MR Images

We present a fully automatic framework for cardiac motion tracking based on non-rigid image registration for the analysis of myocardial motion using both untagged and 3D tagged MR images. We detect and track anatomical landmarks in the heart and combine this with intensity-based motion tracking to allow accurately model cardiac motion while significantly reduce the computational complexity. A collaborative similarity measure simultaneously computed in three LA views is employed to register a sequence of images taken during the cardiac cycle to a reference image taken at end-diastole. We then integrate a valve plane tracker into the framework which uses short-axis and long-axis untagged MR images as well as 3D tagged images to estimate a fully four-dimensional motion field of the left ventricle.
Haiyan Wang, Wenzhe Shi, Xiahai Zhuang, Simon Duckett, KaiPin Tung, Philip Edwards, Reza Razavi, Sebastien Ourselin, Daniel Rueckert

An Incompressible Log-Domain Demons Algorithm for Tracking Heart Tissue

We describe an application of the previously proposed iLogDemons algorithm to the STACOM motion-tracking challenge data. The iLogDemons algorithm is a consistent and efficient framework for tracking left-ventricle heart tissue using an elastic incompressible non-linear registration algorithm based on the LogDemons algorithm. This method has shown promising results when applied to previous data-sets. Along with having the advantages of the LogDemons algorithm such as computing deformations that are invertible with smooth inverse, the method has the added advantage of allowing physiological constraints to be added to the deformation model. The registration is entirely performed in the log-domain with the incompressibility constraint strongly ensured and applied directly in the demons minimisation space. Strong incompressibility is ensured by constraining the stationary velocity fields that parameterise the transformations to be divergence-free in the myocardium. The method is applied to a data-set of 15 volunteers and one phantom, each with echocardiography, cine-MR and tagged-MR images. We are able to obtain reasonable results for each modality and good results for echocardiography images with respect to quality of the registration and computed strain curves.
Kristin McLeod, Adityo Prakosa, Tommaso Mansi, Maxime Sermesant, Xavier Pennec

Temporal Diffeomorphic Free Form Deformation (TDFFD) Applied to Motion and Deformation Quantification of Tagged MRI Sequences

This paper presents strain quantification results obtained from the Tagged Magnetic Resonance Imaging (TMRI) sequences acquired for the 1 st cardiac Motion Analysis Challenge (cMAC). We applied the Temporal Diffeomorphic Free Form Deformation (TDFFD) algorithm to the phantom and the 15 healthy volunteers of the cMAC database. The TDFFD was modified in two ways. First, we modified the similarity metric to incorporate frame to frame intensity differences. Second, on volunteer sequences, we performed the tracking backward in time since the first frames did not show the contrast between blood and myocardium, making these frames poor choices of reference.
On the phantom, we propagated a grid adjusted to tag lines to all frames for visually assessing the influence of the different algorithmic parameters. The weight between the two metric terms appeared to be a critical parameter for making a compromise between good tag tracking while preventing drifts and avoiding tag jumps. For each volunteer, a volumetric mesh was defined in the Steady-State Free Precession (SSFP) image, at the closest cardiac time from the last frame of the tagging sequence. Uniform strain patterns were observed over all myocardial segments, as physiologically expected.
Mathieu De Craene, Catalina Tobon-Gomez, Constantine Butakoff, Nicolas Duchateau, Gemma Piella, Kawal S. Rhode, Alejandro F. Frangi

Motion Analysis with Quadrature Filter Based Registration of Tagged MRI Sequences

Analysis of tagged MRI is a valuable tool for assessing regional myocardial function. One major obstacle for existing methods based on feature extraction and registration is the desaturation of the tagging grid over time. We propose a method based on quadrature filters that is invariant to changes in intensity, robust with respect to the grid geometry and provides a dense motion field that allows for the analysis of both global and local movements. A multi-scale and multi-resolution scheme is used to cover different scales of motion and to speed up registration. The described method has been integrated into a prototypical application and applied to a phantom data set and 15 volunteer data sets provided by the STACOM’11. The automatic detection of the 4D motion field took about 130 minutes per MRI data set and about 90 minutes per US data set and resulted in plausible motion fields, which will be quantitatively assessed within the motion tracking challenge at MICCAI 2011.
Lennart Tautz, Anja Hennemuth, Heinz-Otto Peitgen

Segmentation Challenge

Left Ventricular Segmentation Challenge from Cardiac MRI: A Collation Study

This paper presents collated results from the left ventricular (LV) cardiac MRI segmentation challenge as part of STACOM’11. Clinical cases from patients with myocardial infarction (100 test and 100 validation cases) were randomly selected from the Cardiac Atlas Project (CAP) database. Two independent sets of expert (manual) segmentation from different sources that are available from the CAP database were included in this study. Automated segmentations from five groups were contributed in the challenge. The total number of cases with segmentations from all seven raters was 18. For these cases, a ground truth “consensus” segmentation was estimated based on all raters using an Expectation-Maximization (EM) method (the STAPLE algorithm).
Avan Suinesiaputra, Brett R. Cowan, J. Paul Finn, Carissa G. Fonseca, Alan H. Kadish, Daniel C. Lee, Pau Medrano-Gracia, Simon K. Warfield, Wenchao Tao, Alistair A. Young

Automatic Segmentation of the Myocardium in Cine MR Images Using Deformable Registration

This paper proposes a system to automatically segment the left ventricle in cardiac MR cine images. Individual frames are segmented using a shortest path algorithm and temporal consistency is enforced through the backward and forward deformation fields of an inverse consistent deformable registration. In addition, a segmentation of the mitral valve plane is obtained from long axis images. This algorithm was applied to 95 datasets as part of the STACOM’11 4D LV Segmentation Challenge. We analyze the results and evaluate the strengths and weaknesses of our system.
Marie-Pierre Jolly, Christoph Guetter, Xiaoguang Lu, Hui Xue, Jens Guehring

Layered Spatio-temporal Forests for Left Ventricle Segmentation from 4D Cardiac MRI Data

In this paper we present a new method for fully automatic left ventricle segmentation from 4D cardiac MR datasets. To deal with the diverse dataset, we propose a machine learning approach using two layers of spatio-temporal decision forests with almost no assumptions on the data nor explicitly specifying the segmentation rules. We introduce 4D spatio-temporal features to classification with decision forests and propose a method for context aware MR intensity standardization and image alignment. The second layer is then used for the final image segmentation. We present our first results on the STACOM LV Segmentation Challenge 2011 validation datasets.
Ján Margeta, Ezequiel Geremia, Antonio Criminisi, Nicholas Ayache

Myocardial Segmentation Using Contour-Constrained Optical Flow Tracking

Despite the important role of object tracking using the Optical Flow (OF) in computer graphics applications, it has a limited role in segmenting speckle-free medical images such as magnetic resonance images of the heart. In this work, we propose a novel solution of the OF equation that allows incorporating additional constraints of the shape of the segmented object. We formulate a cost function that include the OF constraint in addition to myocardial contour properties such as smoothness and elasticity. The method is totally different from the common naïve combination of OF estimation within the active contour model framework. The technique is applied to dataset of 20 patients and comparison with manual segmentation shows sensitivity and specificity levels of 93% and 99% respectively is obtained through the challenge validation system.
Ahmed S. Fahmy, Ahmed O. Al-Agamy, Ayman Khalifa

Regular Papers

Optimization for Multi-Region Segmentation of Cardiac MRI

We introduce a new multi-region model for simultaneous segmentation of the left and right ventricles, myocardium and the left ventricular papillary muscles in MRI. The model enforces geometric constraints such as inclusion and exclusion between the regions, which makes it possible to correctly segment different regions even though the intensity distributions are identical. We efficiently optimize the model using Lagrangian duality which is faster and more memory efficient than current state of the art. As the optimization is based on global techniques, the resulting segmentations are independent of initialization. We evaluate our approach on two benchmarks with competitive results.
Johannes Ulén, Petter Strandmark, Fredrik Kahl

Analysis of Catheter-Based Registration with Vessel-Radius Weighting of 3D CT Data to 2D X-ray for Cardiac Catheterisation Procedures in a Phantom Study

X-ray fluoroscopy is routinely used to guide cardiac catheterisations due to its real-time imaging capability and high device visibility, but lacks depth information and poorly visualizes the heart itself. A novel 2D-3D image registration method was developed that can augment 2D fluoroscopy by overlaying 3D CT cardiac images that have excellent soft-tissue information. The method relies on the catheterisation of two vessels during the procedure and globally minimizing a vessel-radius-weighted distance error between the vessel centrelines, segmented from the 3D data, and corresponding catheters reconstructed from biplane X-ray fluoroscopy. Validation of the algorithm was carried out using a glass heart phantom with catheters inserted into combinations of six of its vessels. Results show that registration with the coronary sinus resulted in an average 3D-TRE between 0.55 and 9.1 mm, with the best tested pair being the coronary sinus and descending aorta. The algorithm will be useful for guiding cardiac cauterization procedures and also for co-registration of data for the purposes of biophysical cardiac modelling.
Michael Truong, Thomas Gordon, Reza Razavi, Graeme Penney, Kawal S. Rhode

Myocardial Contractility and Regional Work throughout the Cardiac Cycle Using FEM and MRI

The role of myocardial contractile force in the progression of cardiovascular diseases such as heart failure (HF) has been the focus of many studies. In order to better understand the mechanisms underlying compromised contractility, finite element (FE) modelling of ventricular mechanics is a useful tool. Distributions of active fibre stress during systole were estimated using left ventricular (LV) FE models that incorporated in vivo MRI tagging data and concurrent LV endocardial pressure recordings to parameterise a time-varying model of myocardial contraction. For five canine hearts, the calcium dependent contractile stress increased to peaks ranging from 33 kPa to 57 kPa during systole. Regional distributions of fibre stretch, stress, and myocardial work were examined in each case. Using this type of integrative biophysical modelling to compare normal and pathological cases will elucidate the underlying physiological mechanisms of cardiac mechanical dysfunction.
Vicky Y. Wang, Daniel B. Ennis, Brett R. Cowan, Alistair A. Young, Martyn P. Nash

Variability of the Human Cardiac Laminar Structure

The cardiac fiber architecture has an important role in electrophysiology, in mechanical functions of the heart, and in remodeling processes. The variability of the fibers is the focus of various studies in different species. However, the variability of the laminar sheets is still not well known especially in humans. In this paper, we present preliminary results on a quantitative study on the variability of the human cardiac laminar structure. We show that the laminar structure has a complex variability and we show the possible presence of two populations of laminar sheets. Bimodal distributions of the intersection angle of the third eigenvector of the diffusion tensor have been observed in 10 ex vivo healthy human hearts. Additional hearts will complete the study and further characterize the different populations of cardiac laminar sheets.
Hervé Lombaert, Jean-Marc Peyrat, Laurent Fanton, Farida Cheriet, Hervé Delingette, Nicholas Ayache, Patrick Clarysse, Isabelle Magnin, Pierre Croisille

Polynomial Regression Based Edge Filtering for Left Ventricle Tracking in 3D Echocardiography

Automated detection of endocardial borders in 3D echocardiography is a challenging task. Part of the reason for this is the endocardial boundary leads to alternating edge characteristics that vary over a cardiac cycle. The maximum gradient (MG), step criterion (STEP) and max flow/min cut (MFMC) edge detectors have been previously applied for the endocardial edge detection problem. In this paper, a local polynomial regression based method (LPR) is introduced for filtering the STEP results. For each endocardial model point, (1) the surface is parametrized locally around the point, (2) a polynomial regression is applied on the STEP edges in the parametric domain, and (3) the fitted polynomial is evaluated at the origin of the parametric domain to determine the endocardial edge position. The effectiveness of the introduced method is validated via comparative analyses among the MFMC, STEP, and first & second degree LPR methods.
Engin Dikici, Fredrik Orderud

A Multi-image Graph Cut Approach for Cardiac Image Segmentation and Uncertainty Estimation

Registration and segmentation uncertainty may be important information to convey to a user when automatic image analysis is performed. Uncertainty information may be used to provide additional diagnostic information to traditional analysis of cardiac function. In this paper, we develop a framework for the automatic segmentation of the cardiac anatomy from multiple MR images. We also define the registration and segmentation uncertainty and explore its use for diagnostic purposes. Our framework uses cardiac MR image sequences that are widely available in clinical practice. We improve the performance of the cardiac segmentation algorithms by combining information from multiple MR images using a graph-cut based segmentation. We evaluate this framework on images from 32 subjects: 13 patients with ischemic cardiomyopathy, 14 patients with dilated cardiomyopathy and 5 normal volunteers. Our results indicate that the proposed method is capable of producing segmentation results with very high robustness and high accuracy with minimal user interaction across all subject groups. We also show that registration and segmentation uncertainties are good indicators for segmentation failures as well as good predictors for the functional abnormality of the subject.
Wenzhe Shi, Xiahai Zhuang, Robin Wolz, Duckett Simon, KaiPin Tung, Haiyan Wang, Sebastien Ourselin, Philip Edwards, Reza Razavi, Daniel Rueckert

Toward Clinically-Feasible Noninvasive Electrophysiological Imaging: Investigating the Impact of Local Anatomical Details

Noninvasive Cardiac electrophysiological (EP) imaging aims to compute cardiac electrical dynamics from body surface potential. Anatomical data acquisition and processing computations, to reconstruct detailed geometry of heart and torso, are complex and time consuming tasks that are incompatible with clinical requirements. Our ultimate goal is to improve noninvasive EP imaging techniques toward clinical feasibility by investigating the minimum anatomical information. As the first step toward this goal, in this study we investigate the impact of local geometrical details on cardiac EP imaging. It is known that, global geometrical factors such as size, position and orientation of heart are important in noninvasive electrocardiography problem; but the effect of local geometrical details is unknown and it is difficult to accurately capture. We hypothesize that, as long as global geometrical parameters are captured, local details of realistic cardiac geometry do not significantly impact diagnostic effectiveness of cardiac EP imaging. We verify this hypothesis by developing simple geometrical model instead of realistic heart that enables us to measure local anatomical error, and applying it in EP imaging for detection of myocardial infarction. The results computed based on simple geometrical model are comparable to that of the realistic heart geometry. Thus, it confirms our hypothesis that discarding local geometrical details does not affect diagnostic cardiac EP imaging. The findings of this study pave the road for further studies on tomographic input data processing toward clinical feasibility.
Azar Rahimi, Hongda Mao, Pengcheng Shi, Linwei Wang

A 3D+Time Spatio-temporal Model for Joint Segmentation and Registration of Sparse Cardiac Cine MR Image Stacks

We previously developed a hybrid spatio-temporal method for the segmentation of the left ventricle in 2D+time magnetic resonance (MR) image sequences and here extend this model-based approach towards 3D+time sparse stacks of cine MR images with random orientation. The presented method combines an explicit landmark based statistical geometric model of the inter-subject variability at the end-diastolic and end-systolic time frames with an implicit geometric model that constraints the intra-subject frame-to-frame temporal deformations through deterministic non-rigid image registration of adjacent frames. This hybrid model is driven by both local and global intensity similarity, resulting in a combined spatio-temporal segmentation and registration approach. The advantage of our hybrid model is that the segmentation of all image slices and of the whole sequence can be performed at once, guided by shape and intensity information of all time frames. In addition, prior shape and intensity knowledge are incorporated in order to cope with ambiguity in the images, while keeping training requirements limited.
An Elen, Jeroen Hermans, Hadewich Hermans, Frederik Maes, Paul Suetens

Statistical Atlas of Human Cardiac Fibers: Comparison with Abnormal Hearts

Criteria of normality of the cardiac fibers are important in cardiomyopathies. In this paper, we investigate the differences in the cardiac fiber structures between 10 hearts classified as healthy and 6 hearts classified as abnormal, and determine if properties of the cardiac fiber structures can be discriminants for abnormality. We compare the variability of the fiber directions from abnormal hearts to an atlas of healthy hearts. The human atlas of the cardiac fiber structures is built with an automated framework based on symmetric Log-domain diffeomorphic demons. We study the angular variability of the different fiber structures. Our preliminary results might suggest that a higher variability of the fiber structure directions could possibly characterize abnormality of a heart.
Hervé Lombaert, Jean-Marc Peyrat, Laurent Fanton, Farida Cheriet, Hervé Delingette, Nicholas Ayache, Patrick Clarysse, Isabelle Magnin, Pierre Croisille

Maximum Likelihood Correction of Shape Bias Arising from Imaging Protocol: Application to Cardiac MRI

To establish a fair comparison between shape models derived from different imaging protocols, a mapping correcting local bias must be applied. In this paper, a multi-dimensional statistical model has been investigated to correct the systematic differences between Steady-State Free Precession (SSFP) and Gradient Recalled Echo (GRE) cardiac MRI protocols. This statistical model makes use of the Maximum Likelihood (ML) approach to estimate the local parameters of the respective GRE and SSFP distributions. Once those parameters are known, a local mapping can be applied. We applied this method to 46 normal volunteers who were imaged with both protocols. The SSFP model was estimated from the corresponding GRE model and validation was performed with leave-one-out experiments. The error was examined in both the local model parameters and the clinically important global mass and volume estimates. Results showed that the systematic bias around the apex and papillary muscles could be locally corrected and that the mapping also provided a global correction in cavity volume (average error of 0.4 ±12.4 ml) and myocardial mass (− 1.2 ±11.1 g).
Pau Medrano-Gracia, David A. Bluemke, Brett R. Cowan, J. Paul Finn, Carissa G. Fonseca, João A. C. Lima, Avan Suinesiaputra, Alistair A. Young

Volumetric Modeling Electromechanics of the Heart

Heart is an electromechanical coupled organ, thus it is important to integrate electrical and mechanical functions when building a computational model of the heart. The existing models either treat electrical and mechanical functions separately, or follow a so-called ”one-way” electromechanical coupling. However, electrical and mechanical functions of the heart are depended on each other, and realistic simulation results can only be achieved when such coupled relationship is considered. In this paper, we propose a generic model to simulate electromechanics of the heart that takes both electromechanical coupling and mechanoelectrical feedback into account. The model contains four components: cardiac electrophysiological model, electromechanical coupling, cardiac mechanics model and mechanoelectrical feedback. We report numerical simulations of a cube to provide an insight of the electromechanical coupled behavior of our model. Experiments have also been performed on a biventricular heart which present physiological plausible values, such as transmembrane potential (TMP) maps and strain maps.
Hongda Mao, Linwei Wang, Ken C. L. Wong, Huafeng Liu, Pengcheng Shi

Matching Sparse Sets of Cardiac Image Cross-Sections Using Large Deformation Diffeomorphic Metric Mapping Algorithm

The purpose of this study is to illustrate the application of large deformation diffeomorphic metric mapping to perform registration among sparsely sampled cardiac magnetic resonance imaging (MRI) data. To evaluate the performance of this method, we use two sets of data: 1) contours that are generated from sparsely sampled left ventricular sections and extracted from short axis cardiac MRI of patients with hypertrophic cardiomyopathy and 2) left ventricular surface mesh that is generated from higher resolution cardiac computed tomography image. We present two different discrepancy criteria, one based on a measure that is embedded in the dual of a reproducing kernel Hilbert space of functions for curves and the other is based on a geometric soft matching distance between a surface and a curve.
Siamak Ardekani, Aastha Jain, Saurabh Jain, Theodore P. Abraham, Maria R. Abraham, Stefan Zimmerman, Raimond L. Winslow, Michael I. Miller, Laurent Younes

VURTIGO: Visualization Platform for Real-Time, MRI-Guided Cardiac Electroanatomic Mapping

Guidance of electrophysiological (EP) procedures by magnetic resonance imaging (MRI) has significant advantages over x-ray fluoroscopy. Display of electroanatomic mapping (EAM) during an intervention fused with a prior MR volume and DE-MRI derived tissue classification should improve the accuracy of cardiac resynchronization therapy (CRT) for ventricular arrhythmias. Improved accuracy in the spatial localization of recorded EP points will produce an EAM to constrain and customize patient-specific cardiac electroanatomic models being developed for understanding the patterns of arrhythmogenic slow conduction zones causing reentry circuits and treatment planning. The Vurtigo software presented here is a four dimensional (3D+time) real-time visualization application for guiding interventions capable of displaying prior volumes, real-time MRI scan planes, EAM (voltage or activation times), segmented models, and tracked catheters. This paper will describe the architecture and features of Vurtigo followed by the application example of guiding percutaneous cardiac electroanatomic mapping in porcine models.
Perry E. Radau, Stefan Pintilie, Roey Flor, Labonny Biswas, Samuel O. Oduneye, Venkat Ramanan, Kevan A. Anderson, Graham A. Wright

Validation of a Novel Method for the Automatic Segmentation of Left Atrial Scar from Delayed-Enhancement Magnetic Resonance

Delayed-enhancement magnetic resonance imaging is an effective technique for imaging left atrial (LA) scars both pre- and post- radio-frequency ablation for the treatment of atrial fibrillation. Existing techniques for LA scar segmentation require expert manual interaction, making them tedious and prone to high observer variability. In this paper, a novel automatic segmentation algorithm for segmenting LA scar was validated using digital phantoms and clinical data from 11 patients. The performance of the approach was compared to the two leading semi-automatic techniques and the ground truth of manual segmentations by 2 expert observers. The novel approach was shown to be accurate in terms of Dice coefficient, robust to typical image intensity variability, and much faster in terms of execution time.
Rashed Karim, Aruna Arujuna, Alex Brazier, Jaswinder Gill, C. Aldo Rinaldi, Michael Cooklin, Mark O’Neill, Reza Razavi, Tobias Schaeffter, Daniel Rueckert, Kawal S. Rhode

Cardiac Motion Estimation Using Covariant Derivatives and Helmholtz Decomposition

Quantification of cardiac function is important for the assessment of abnormalities and response to therapy. We present a method to reconstruct dense cardiac motion from sparse features in tagging MRI, decomposed into solenoidal and irrotational parts using multi-scale Helmholtz decomposition. Reconstruction is based on energy minimization using covariant derivatives exploiting prior knowledge about the motion field. The method is tested on cardiac motion images. Experiments on phantom data show that both covariant derivatives and multi-scale Helmholtz decomposition improve motion field reconstruction.
Alessandro Becciu, Remco Duits, Bart J. Janssen, Luc M. J. Florack, Hans C. van Assen

Temporal Diffeomorphic Motion Analysis from Echocardiographic Sequences by Using Intensity Transitivity Consistency

Quantitative motion analysis from echocardiography is an important yet challenging problem. We develop a motion estimation algorithm for echocardiographic sequences based on diffeomorphic image registration in which the velocity field is spatiotemporally smooth. The novelty of this work is that we propose a functional of the velocity field which minimizes the intensity consistency error of the local unwarped frames. The consistency error is measured as the sum of squared difference of the four frames evolving to any time point between the two inner frames of them. The estimated spatiotemporal transformation has maximum local transitivity consistency. We validate our method by using simulated images with known ground truth and real ultrasound datasets, experiment results indicate that our motion estimation method is more accurate than other methods.
Zhijun Zhang, David J. Sahn, Xubo Song


Weitere Informationen

Premium Partner