Blood flow hamper in the heart's interior detected with wave propagation techniques

Amsterdam 13 September 2001Three-dimensional visualisation and quantification in cardiovascular images was the topic of a presentation held by ir. Joris Schaap from the University Medical Center in Leiden (LUMC) at the Dutch National Visualisation Days. The audience were shown some examples of high rendering techniques and algorithms, developed at the LUMC laboratory, to study the quality of vascular ventricles, based on Magnetic Resonance (MR) and Computed Tomography (CT) image scanner material. The results of this lab research are used in the hospital to diagnose heart patients.

Advertisement

A set of MR images have been taken from a volunteer to provide a basis for studying the quality of vascular ventricles, explained ir. Schaap. In order to do this, these 2D images are utilised to create a 3D representation of the ventricles. The technique applied at the LUMC consists in a combination of both segmentation and wave propagation. With segmentation, the parts belonging to the ventricle can be identified. An imaginary wave is then sent through these parts.

In turn, CT images have the great advantage that their grey values represent a true correspondence with the real density of tissue and bone in the patient's body. As such, the opacity curve allows the researcher to make certain types of connective tissue visible or not by applying thresholds, as ir. Schaap briefly noted. At the LUMC lab, radial slices are created along the radial axis of the heart and various techniques are applied to cut details out of the images which cover underlying structures.

The results are used to determine fairly good the structure and flow properties of the ventricle, displaying also possible obstacles in the vessel which could hamper the blood flow through it. Maximum intensity projections are carried out and markers placed and projected to find the vessel path. From the moment the end is reached, the algorithm stops. Light vessel parts are characterised by a high intensity through which the wave can travel fast, whereas dark parts tend to obstruct the wave, thus proving that there might be an obstacle.

For the researchers, it is now possible to analyse the backtrack along the wave path by perpendicularly following the steepest descent, according to ir. Schaap. This is surely the fastest way across the vessel but not the shortest one. The path is often not straight because the lightest vessel parts are not always to be found in the middle. Therefore, the path has to be corrected by means of an algorithm in order to detect the vessel afterwards in the data set. Once the path has been straightened out, it is possible to generate a fly-through across the vessel to study its interior for an accurate diagnosis.


Leslie Versweyveld

[Medical IT News][Calendar][Virtual Medical Worlds Community][News on Advanced IT]