Arianne Van Muiswinkel

Soap-bubble visualization and quantitative analysis of 3D coronary magnetic resonance angiograms

In order to compare coronary magnetic resonance angiography (MRA) data obtained with different scanning methodologies, adequate visualization and presentation of the coronary MRA data need to be ensured. Furthermore, an objective quantitative comparison between images acquired with different scanning methods is desirable. To address this need, a software tool (“Soap‐Bubble”) that facilitates visualization and quantitative comparison of 3D volume targeted coronary MRA data was developed. In the present implementation, the user interactively specifies a curved subvolume (enclosed in the 3D coronary MRA data set) that closely encompasses the coronary arterial segments. With a 3D Delaunay triangulation and a parallel projection, this enables the simultaneous display of multiple coronary segments in one 2D representation. For objective quantitative analysis, frequently explored quantitative parameters such as signal‐to‐noise ratio (SNR); contrast‐to‐noise ratio (CNR); and vessel length, sharpness, and diameter can be assessed. The present tool supports visualization and objective, quantitative comparisons of coronary MRA data obtained with different scanning methods. The first results obtained in healthy adults and in patients with coronary artery disease are presented. Magn Reson Med 48:658–666, 2002.

Magnetic resonance spectroscopy

Magnetic resonance spectroscopy (MRS) is noninvasive and may be readily combined with magnetic resonance imaging (MRI). Attention has focussed on proton (1H) and phosphorus (31P) MRS, and studies have been undertaken by using single voxels or many voxels simultaneously (chemical‐shift imaging, magnetic resonance spectroscopic imaging). The latter is more difficult and prone to artefact but potentially yields significantly more information. 1H MRS has principally yielded data on concentrations of N‐acetyl aspartate (NAA), choline, creatine, and phosphocreatine. NAA is located primarily within neurons, and reduction of the ratio of NAA to choline, creatine, and phosphocreatine is a marker of neu‐ronal loss and dysfunction. This technique may be useful as a noninvasive tool for localizing epileptogenic foci, but its role requires further evaluation. As with all functional imaging methods, coregistration with high‐quality MRI is essential for interpreting data. 1H MRS can be used also to estimate cerebral concentrations of several neurotrans‐mitters: glutamate, glutamine, and γ‐aminobutyric acid (GABA). This may prove useful for characterizing the neurometabolic profiles of patients with different epilepsy syndromes and for evaluating the effects of medical and surgical treatments. 31P MRS can detect adenosine tri‐phosphate, phosphodiesters, phosphomonoesters, phosphocreatine, and inorganic phosphate, and estimate intra‐cerebral pH. Abnormalities that have been associated with epileptogenic brain areas include increased inorganic phosphate, reduced phosphomonoesters, and increased pH. Only small numbers of patients have been studied, however, so that conclusions are not definitive, and the clinical role of this technique is not yet established.

Grey-value-based 3D registration of functional MRI time-series: comparison of interpolation order and similarity measure

The analysis of functional MR images of the brain such as FMRI and neuro perfusion is significantly limited by movement of the head during image acquisition. Already small motions introduce artifacts in voxel-based statistical analysis and restrict the assessment of functional information. The retrospective compensation of head motion is usually addressed by image registration techniques which spatially align the images of the time-series. In this paper we investigate the relevance of intermediate interpolation during the registration process, similarity measure and optimization scheme by means of statistical consistency of the registration results. Experiments show that cubic and quartic interpolation remarkably improve the consistency when compared to linear methods. The use of larger interpolation kernels, however, does not result in further improvements. Measures based on the mean squared error are successfully applied to FMRI time- series which provide constant tissue-to-image transfer. However, they are not suitable for neuro perfusion imaging since the change of image intensity during the inflow of the contrast agent affords measures typically applied in multi- modality registration. Our results indicate that a recently proposed measure based on local correlation is preferable to mutual information in the case of neuro perfusion.

Image Registration for Distortion Correction in Diffusion Tensor Imaging

In diffusion tensor imaging the calculation of functional information is limited by head movement and eddy current-induced image distortion. In this paper the application of image registration for distortion correction is investigated. In particular, a 3D affine and a dedicated transformation which is adapted to the type of distortion and the similarity measures mutual information and local correlation are compared to each other. The registration results are quantitatively evaluated by analyzing their consistency properties. Visual inspection shows that registration generally improves the quality of the functional information. The consistency tests reveal that both transformations provide similar registration results which is remarkable since the dedicated transformation does not take advantage of modeling of the underlying imaging physics. Furthermore, it is shown that local correlation similarity is an interesting alternative to mutual information. The registration of a DTI series with local correlation is more consistent and takes only about one minute for calculation.

Automatic knee cartilage delineation using inheritable segmentation

We present a fully automatic method for segmentation of knee joint cartilage from fat suppressed MRI. The method first applies 3-D model-based segmentation technology, which allows to reliably segment the femur, patella, and tibia by iterative adaptation of the model according to image gradients. Thin plate spline interpolation is used in the next step to position deformable cartilage models for each of the three bones with reference to the segmented bone models. After initialization, the cartilage models are fine adjusted by automatic iterative adaptation to image data based on gray value gradients. The method has been validated on a collection of 8 (3 left, 5 right) fat suppressed datasets and demonstrated the sensitivity of 83±6% compared to manual segmentation on a per voxel basis as primary endpoint. Gross cartilage volume measurement yielded an average error of 9±7% as secondary endpoint. For cartilage being a thin structure, already small deviations in distance result in large errors on a per voxel basis, rendering the primary endpoint a hard criterion.

Model-based vessel segmentation using an elliptic cylinder

For viewing vascular structures in 3D MRA with bloodpool contrast agent via maximum intensity projection, it is necessary to suppress other occluding vessels which are also contrasted. These can be located very close to the vessels of interest. In this paper, an automated selection method based on the use of an elliptically cross-sectioned cylinder model is proposed. This model of local shape is used to define the speed function for a front propagation algorithm, which extracts a vessel axis between two selected points. The boundary is then reconstructed using local shape parameters from the fitted cylinder model. A comparison of automated and manual segmentations showed a mean selection accuracy of 94% using elliptical model, and a mean surface deviation of 1.1mm.