Creatig a 3-D Thorax Model with Reconstruction of Contours of Body Organs Used in
Determining Epicardial Maps

Marek Doros

Institute of Biocybernetics and Biomedical Engineering PAS, Warsaw, Poland

Correspondence: M Doros, Institute of Biocybernetics and Biomedical Engineering PAS, Ks.Trojdena 4, 02-109 Warsaw, Poland.
E-mail: Marek.Doros@ibib.waw.pl, phone +48 22 6599143 ex 411, fax +48 22 6597030

---

---

1.  Introduction

The purpose of this work was to present and test the method of creating the 3D thorax model with reconstruction of the contours of body organs for non-invasive determination of the epicardial maps on a basis of body surface potential maps (BSPMs). For that purpose two kinds of data are necessary: i) distribution of BSPMs (measured by HPM-7100 system of Fucuda Denshi [Czerwińska et al., 2000]), and ii) specific conductivity and sizes of organs of a chest.

In elaborated method, the specific conductivity data are taken from Rush tables [Deok, 1988]. Sizes of respective organs of a chest (the image data) used to construction of the 3D model are obtained from the radiographic chest images. It is less expensive (but less accurate) than using the CT images [Doros, 2002]. At first the radiographic chest image is being enhanced [Jain, 1990]. Then, there are manually made contours of respective organs. This implementation is based on the fact that the time of the manual contour creation, when applied to the already processed image where the edges are enhanced, is shorter than the corresponding time needed to create contour on the original image.

2.  Necessary Image Data, Creating 3D Models

The process of creating the contours of selected chest organs was made using the radiographic enhanced images. Laplacian operator [Jain, 1990] has been used for edge enhancement.

As mentioned in a previous chapter, the time of the manual contour creation on the image with edges enhanced is shorter that the time needed to create contour on the original image.

Accurate model

The accurate reference model is constructed on a basis of the CT thorax cross-sections at the considered thorax region (Fig.1a, b, rows 1¸7).

Approximate model

In the creation of the approximate model either the image data of the average model constructed on a basis of anatomical CT atlases, or the approximate image data based on the contours of the enhanced radiographic images (Fig.1c) combined with the external chest sizes are applied. The average model is being corrected using the approximate image data and correction coefficients [Czerwińska et al., 2000].

                                a)                                                        b)                                                            c)

Figure 1. a) location of rows of electrodes on the chest surface, b) locations of chest cross-sections, c) scheme of the enhanced radiographic image with contours of respective organs: l’i - width of a thorax, l’Li  - length of the projection of a left lung, l’Ri - length of the projection of a right lung, l’Hi - length of the projection of the heart.

3.  Determining the Epicardial Maps, Discussion of Results

Figure 2. a) BSPM and  epicardial maps calculated for: b) accurate model, c) approximated model.

The method was tested on 3 healthy subjects. For each case for given instants of ECG there have been determined the epicardial maps on a basis of BSPM’s using either the approximate model or the model based on CT image data (accurate model). There has been made the comparison of the respective epicardial maps. As the example, Fig.2 displays the BSPM (Fig.2a) and epicardial maps calculated for the models: accurate (constructed on a basis of CT data - Fig.2b), and approximate created with reconstruction of the contours of the body organs (Fig.2c). It is seen, that for both maps the distribution of local extremes is similar, only their values differ about 15%.

On a basis of the carried out examinations it has been confirmed, that the thorax models created with reconstruction of the contours of organs taken from radiographic chest images do not introduce the considerable distortions to the distribution of epicardial potentials. Finally, the method allows determining the epicardial potentials a less expensive way than the method based on the CT image data.

Acknowledgements

This work was supported by National CSR Research Project No 4 T11E 013 24

References

Czerwińska A., Doros M., Kolebska K. Non-invasive cardiac diagnosis method based on the simulation of electrical field generated by myocardium. Biocybernetics and Biomedical Engineering, vol.20, No 1, pp. 47-58, PWN – Warszawa 2000.

Deok W. Origins of the impedance change in impedance cardiolography by a three dimensional finite element model. IEEE Trans. of Biomed. Eng. vol 35, no 12, pp.65-75, 1988.

Doros M. Using the computer simulation to the reconstruction of contours of the thorax organs for creating the 3d model. Proc. XXIV Int. Workshop ASIS 2002, Knov (Czech Republic), pp. 103-108, September 2002.

Jain A.K. Fundamentals of digital image processing. Prentice Hall International, 1990.