Alexandra Groth

Automatic whole heart segmentation in static magnetic resonance image volumes

We present a fully automatic segmentation algorithm for the whole heart (four chambers, left ventricular myocardium and trunks of the aorta, the pulmonary artery and the pulmonary veins) in cardiac MR image volumes with nearly isotropic voxel resolution, based on shape-constrained deformable models. After automatic model initialization and reorientation to the cardiac axes, we apply a multi-stage adaptation scheme with progressively increasing degrees of freedom. Particular attention is paid to the calibration of the MR image intensities. Detailed evaluation results for the various anatomical heart regions are presented on a database of 42 patients. On calibrated images, we obtain an average segmentation error of 0.76mm.

Comprehensive validation of computational fluid dynamics simulations of in‐vivo blood flow in patient‐specific cerebral aneurysms

PURPOSE Recently, image-based computational fluid dynamic (CFD) simulations have been proposed to investigate the local hemodynamics inside human cerebral aneurysms. It was suggested that the knowledge of the computed three-dimensional flow fields can be used to assist clinical risk assessment and treatment decision making. Therefore, it was desired to know the reliability of CFD for cerebral blood flow simulation, and be able to provide clinical feedback. However, the validations were not yet comprehensive as they lack either patient-specific boundary conditions (BCs) required for CFD simulations or quantitative comparison methods. METHODS In this study, based on a recently proposed in-vitro quantitative CFD evaluation approach via virtual angiography, the CFD evaluation was extended from phantom to patient studies. In contrast to previous work, patient-specific blood flow rates obtained by transcranial color coded Doppler ultrasound measurements were used to impose CFD BCs. Virtual angiograms (VAs) were constructed which resemble clinically acquired angiograms (AAs). Quantitative measures were defined to thoroughly evaluate the correspondence of the detailed flow features between the AAs and the VAs, and thus, the reliability of CFD simulations. RESULTS The proposed simulation pipeline provided a comprehensive validation method of CFD simulation for reproducing cerebral blood flow, with a focus on the aneurysm region. Six patient cases were tested and close similarities were found in terms of spatial and temporal variations of contrast agent (CA) distribution between AAs and VAs. For patient #1 to #5, discrepancies of less than 11% were found for the relative root mean square errors in time intensity curve comparisons from characteristic vasculature positions. For patient #6, where the CA concentration curve at vessel inlet cannot be directly extracted from the AAs and given as a BC, deviations about 20% were found. CONCLUSIONS As a conclusion, the reliability of the CFD simulations was well confirmed. Besides, it was shown that the accuracy of CFD simulations was closely related to the input BCs.

Phantom-based experimental validation of computational fluid dynamics simulations on cerebral aneurysms.

PURPOSE Recently, image-based computational fluid dynamics (CFD) simulation has been applied to investigate the hemodynamics inside human cerebral aneurysms. The knowledge of the computed three-dimensional flow fields is used for clinical risk assessment and treatment decision making. However, the reliability of the application specific CFD results has not been thoroughly validated yet. METHODS In this work, by exploiting a phantom aneurysm model, the authors therefore aim to prove the reliability of the CFD results obtained from simulations with sufficiently accurate input boundary conditions. To confirm the correlation between the CFD results and the reality, virtual angiograms are generated by the simulation pipeline and are quantitatively compared to the experimentally acquired angiograms. In addition, a parametric study has been carried out to systematically investigate the influence of the input parameters associated with the current measuring techniques on the flow patterns. RESULTS Qualitative and quantitative evaluations demonstrate good agreement between the simulated and the real flow dynamics. Discrepancies of less than 15% are found for the relative root mean square errors of time intensity curve comparisons from each selected characteristic position. The investigated input parameters show different influences on the simulation results, indicating the desired accuracy in the measurements. CONCLUSIONS This study provides a comprehensive validation method of CFD simulation for reproducing the real flow field in the cerebral aneurysm phantom under well controlled conditions. The reliability of the CFD is well confirmed. Through the parametric study, it is possible to assess the degree of validity of the associated CFD model based on the parameter values and their estimated accuracy range.

Automatic multi-model-based segmentation of the left atrium in cardiac MRI scans

Model-based segmentation approaches have been proven to produce very accurate segmentation results while simultaneously providing an anatomic labeling for the segmented structures. However, variations of the anatomy, as they are often encountered e.g. on the drainage pattern of the pulmonary veins to the left atrium, cannot be represented by a single model. Automatic model selection extends the model-based segmentation approach to handling significant variational anatomies without user interaction. Using models for the three most common anatomical variations of the left atrium, we propose a method that uses an estimation of the local fit of different models to select the best fitting model automatically. Our approach employs the support vector machine for the automatic model selection. The method was evaluated on 42 very accurate segmentations of MRI scans using three different models. The correct model was chosen in 88.1% of the cases. In a second experiment, reflecting average segmentation results, the model corresponding to the clinical classification was automatically found in 78.0% of the cases.

Quantitative evaluation of virtual angiography for interventional X-ray acquisitions

Virtual angiography has been recently proposed [1] as a method to indirectly validate the accuracy of computational fluid dynamic (CFD) simulations. In doing so, X-ray angiograms are created from CFD results and compared with the acquired data. In this paper, we apply the extended virtual angiography method [2] which has been only tested on an imaged phantom to clinical data for the first time. To obtain realistic CFD simulations, the contrast agent injection (CA) curve and patient specific blood flow measured by transcranial colour coded Doppler ultrasound (TCCD) are used as boundary conditions. In a quantitative comparison, the spatial and temporal variation of CA concentration of both clinical and virtual angiograms showed good correspondence.

Robust left ventricular myocardium segmentation for multi-protocol MR

For a number of cardiac procedures like the treatments of ventricular tachycardia (VT), coronary artery disease (CAD) and heart failure (HF) both anatomical as well as vitality information about the left ventricular myocardium are required. To this end, two images for the anatomical and functional information, respectively, must be acquired and analyzed, e.g. using two different 3D MR protocols. To enable automatic analysis, a workflow has been proposed1 which allows to integrate the vitality information extracted from the functional image data into a patient-specific anatomical model generated from the anatomical image. However, in the proposed workflow the extraction of accurate vitality information from the functional image depends to a large extend on the accuracy of both the anatomical model and the mapping of the model to the functional image. In this paper we propose and evaluate methods for improving these two aspects. More specifically, on one hand we aim to improve the segmentation of the often low-contrast left ventricular epicardium in the anatomical 3D MR images by introducing a patient-specific shape-bias. On the other hand, we introduce a registration approach that facilitates the mapping of the anatomical model to images acquired by different protocols and modalities, such as functional 3D MR. The new methods are evaluated on clinical MR data, for which considerable improvements can be achieved.

Clinical study of model-based blood flow quantification on cerebrovascular data

Diagnosis and treatment decisions of cerebrovascular diseases are currently based on structural information like the endovascular lumen. In future, clinical diagnosis will increasingly be based on functional information which gives direct information about the physiological parameters and, hence, is a direct measure for the severity of the pathology. In this context, an important functional quantity is the volumetric blood flow over time. The proposed flow quantification method uses contrasted X-ray images from cerebrovascular interventions and a model of contrast agent dispersion to estimate the flow parameters from the spatial and temporal development of the contrast agent concentration through the vascular system. To evaluate the model-based blood flow quantification under realistic circumstances, dedicated cerebrovascular data has been acquired during clinical interventions. To this aim, a clinical protocol for this novel procedure has been defined and optimized. For the verification of the measured flow results ultrasound Doppler measurements have been performed acting as reference measurements. The clinical data available so far indicates the ability of the proposed flow model to explain the in-vivo transport of contrast agent in blood. The flow quantification results show good correspondence of flow waveform and mean volumetric flow rate with the accomplished ultrasound measurements before or after angiography.

Experimental validation and sensitivity analysis for CFD simulations of cerebral aneurysms

In this work, by exploiting a phantom aneurysm model, we illustrate the correlation between experimental data and computational fluid dynamics (CFD) simulation results under well controlled conditions. This is difficult to achieve with clinical patient cases where several uncertainties are present. Quantitative measures are defined for CFD validation by virtual angiography. In addition, a parametric study has been carried out to systematically investigate the sensitivity of current measuring technique on the flow pattern.

Real-time implementation of a multiresolution motion-compensating temporal filter on general-purpose hardware

A data-driven algorithmic structure on a standard PC was developed for a block-based motion compensated temporal filtering in real time. The major time limiting factor of the algorithm was identified as the irregular memory access mainly caused by the layered multi-resolution representation of the input frames. As a result, data is transferred from main memory to cache multiple times leading to memory-dominated critical paths in execution. In order to improve the cache utilization, the computations have been rearranged to process the complete signal on the cached subset of data. The input frames are now divided into super-lines, which are subsets of data containing the relevant information to calculate one line of motion vectors and to filter the corresponding image lines. Only when a set of data is no longer used nor for motion vector analysis nor for filtering the images themselves it is replaced by data of different layers or lines. Due to these data-driven techniques the cache capacity miss rate is reduced to less than 0.8%. As a result, images are processed at a rate of more than 44 fps on a standard PC (Intel dual-processor Xeon, 1.8 GHz), as opposed to 1 fps in the standard implementation.

Deep Learning-Based Boundary Detection for Model-Based Segmentation with Application to MR Prostate Segmentation

Model-based segmentation (MBS) has been successfully used for the fully automatic segmentation of anatomical structures in medical images with well defined gray values due to its ability to incorporate prior knowledge about the organ shape. However, the robust and accurate detection of boundary points required for the MBS is still a challenge for organs with inhomogeneous appearance such as the prostate and magnetic resonance (MR) images, where the image contrast can vary greatly due to the use of different acquisition protocols and scanners at different clinical sites. In this paper, we propose a novel boundary detection approach and apply it to the segmentation of the whole prostate in MR images. We formulate boundary detection as a regression task, where a convolutional neural network is trained to predict the distances between a surface mesh and the corresponding boundary points. We have evaluated our method on the Prostate MR Image Segmentation 2012 challenge data set with the results showing that the new boundary detection approach can detect boundaries more robustly with respect to contrast and appearance variations and more accurately than previously used features. With an average boundary distance of 1.71 mm and a Dice similarity coefficient of 90.5%, our method was able to segment the prostate more accurately on average than a second human observer and placed first out of 40 entries submitted to the challenge at the writing of this paper.