Principles of Magnetocardiographic Maps —
Classification and CAD Detection

Illya Chaikovsky^a, Daniela Katz^b, Mark Katz^b

^aUniversity Witten/Herdecke, Essen, Germany,
^bEssen University Hospital, Cardiology department,Essen,Germany

Correspondence: I Chaikovsky, University Witten/Herdecke, 45134 Waldsaum 1 Essen,Germany.
E-mail: Illya_6@hotmail.com, phone +49 201 1059132, fax +49 201 1059140

1.  Introduction

Magnetocardiography (MCG) is supposed to be a sensitive and useful method for detection of a various heart diseases, first of all coronary artery disease (CAD). A reasonable and physiologically related evaluation of MCG examination has to be developed making this method acceptable for clinical routine. A variety of methods and indicators for medical analysis are already available in the current stage of MCG development, so that a discussion of each merit and pitfalls appears in order. These indicators describe different aspects of magnetocardiographic images, such as homogeneity of maps, direction of magnetic moments or current vectors and so on. Development of integral parameter, taking into consideration all physiologically important characteristics of maps will allow classifying every image in terms of normal or pathological current distribution, which is "end point" of magnetocardiogram analysis.

Aim of this study was to develop an integral classification score for evaluation of magnetocardiographic maps within ST-T interval.

2.  Material and Methods

57 healthy volunteers and 131 patients with proved CAD (stenosis more then 50% of at least one coronary vessel) were examined.

MCG recordings were taken at 9 pre-thoracic sites using a four-channel SQUID-magnetometer (SQUID AG, Essen) in an unshielded setting within a 20 by 20 rectangular grid with a 4 cm pitch over the precordial area. Data were recorded at each registration point for 30 seconds with simultaneous registration of lead II of the surface ECG.

For further analysis the current density vectors (CDV) were reconstructed from the stored data. The calculation of these maps is based on an "inverse problem solution" implemented in the software package MAGwin. CDV maps were generated every 10 ms within the ST-T interval starting with the J-point. A classification system with a scale from 0 to 4 was employed.

This classification system is based on the following consideration: The electrical generator during repolarization can be formulated as an extended current source located in the border zone separating excited and non-excited zones of myocardium. This excitation wavefront, integrated in a medium with homogeneous conductivity, should be directed left-downwards, the normal direction for ventricular repolarisation. Two types of current are represented by this model: so-called "the impressed currents", due to the transmembraneous potential gradient and passive volume currents, generated by "impressed" currents. These volume currents constitute two vortices, which are symmetrical and equal. The resulting map has an "ideal" dipole structure.

Additional excitation wavefronts may appear because of some pathophysiological processes in the myocardium, such as ischemia. These pathological wavefronts will lead to the non-dipole structure of the maps. Another aspect, which has to be taken into account that inhomogeneity of conductivity results in asymmetry and deformation of the vortexes.

3.  Results

A flow-chart for visual map classification within the ST-T interval was developed (see Fig.1).

Based on this classification system each CDV map in course of the ST-T interval was analyzed visually by two independent observers with a scale from 0 ("ideal") to 4 (grossly abnormal) depending on the extent of abnormalities.

Mean class value in the control group resulted in 1.35 ± 0.54 and in 2.91 ± 0.67 for the CAD patient group (p< 0.01). Sensitivity of MCG examination with Class 2 as a threshold was 84 % and 73% respectively.

Figure 1.   Flow-chart for classification of current dencity vector maps within ST-T interval.

4.  Discussion

Proposed score system depends to some degree on the experience of the observers, with the need for a training procedure. The implementation of an automatic process following the described criteria for each category should improve this aspect.

5.  Conclusions

MCG measurement seems to be a sensitive method for CAD detection. The classification of maps makes it possible to determine a score value, which can be used for evaluation of MCG examination results in a reproducible way.