1. Introduction

Magnetocardiography (MCG) is a novel noninvasive multichannel mapping technique to record cardiac electromagnetic signals, generated by the same ionic currents underlying the ECG [Siltanen, 1989]. The MCG has similar morphological features as the ECG, such as QRS complex, T-, P-, and U-waves, but there are some fundamental differences. The MCG is more sensitive to tangential currents in the heart than the ECG, and it is also sensitive to vortex currents, which cannot be detected by the ECG [Siltanen, 1989]. In normal heart, the main direction of the activation wavefront is radial, from endocardium to epicardium. For these reasons, MCG may show ischemia-induced deviations from the normal direction of depolarization and repolarization with better accuracy than the ECG. MCG is affected less by conductivity variations in the body (lungs, muscles, skin) than ECG [Nenonen et al., 1994]. In addition, because MCG is a fully noncontact method, the problems in the skin-electrode contact encountered in ECG are avoided [Siltanen et al., 1989]. To improve signal to noise ratio, sensitive MCG recordings are typically performed in a magnetically shielded room [Nenonen et al., 1997].

Body surface potential mapping (BSPM) covers an extensive area of the thorax offering better spatial scope and resolution compared to the standard ECG due to sampling in areas of the thoracic surface not covered by the conventional six precordial leads. It has been reported to be more sensitive in the detection of myocardial infarction than the 12-lead ECG [Kornreich et al., 1993]. In acute myocardial infarction the optimal electrode positions for recognition of different types of myocardial infarction have been found to be located in positions other than the conventional ones of the 12-lead ECG. The time consuming placement of electrodes is the major disadvantage of body surface potential mapping.

Combining the body surface potential mapping and multichannel magnetocardiography allows a comprehensive study of electromagnetic fields generated by the heart in different clinical settings. These multichannel mapping techniques have made it possible also to look at the ischemia-induced spatial changes both in the magnetic and electric fields over the thorax.

2. Electromagnetic Changes Caused by Ischemia

Myocardial ischemia causes changes in the electrophysiological properties of the myocardium. The resting membrane potential decreases and the upstroke velocity of the action potential and its velocity is lowered [Janse et al., 1981]. The conduction velocity decreases markedly and unhomogenously in the ischemic areas resulting in the dispersion of the activation wavefront. Acute transient ischemia and on the other hand infarcted necrotic regions therefore have different electromagnetic properties. Furthermore, in acute myocardial infarction after experimental occlusion of a coronary artery, opposite ST-segment and baseline shifts on direct current MCG have been found, suggesting MCG to have an advantage over ECG in the detection of injury currents [Cohen et al., 1975].

3. Exercise MCG in Healthy Controls

In healthy subjects, ST-T segment changes have been reported in MCG mapping after pharmacological stress testing, not visible in 32-lead ECG [Brockmeier et al., 1997]. These changes included ST segment depressions, as well as T-wave inversions with monopolar negative MCG maps after stress. The changes in MCG were concluded to be of non-pathological origin, and the authors suggested that they could have been induced by circular, so-called vortex currents not detectable by ECG. Significant amplitude changes in the MCG, not detectable in the BSPM have also been reported in a another study [Takala et al., 1998]. At the T-wave apex, a bending of the spatial zero-field line from rest to the cessation of stress was found to take place, and monopolar group-mean difference maps at the T-wave apex as well as during the ST segment were detected.

Due to these physiological changes reported the ischemia detection in stress MCG cannot be based solely in similar amplitude parameters used in standard ECG testing. Instead, novel multichannel mapping techniques have made it possible also to look at the ischemia-induced spatial changes in the cardiac magnetic field.

4. Exercise Induced Ischemia in MCG

Cohen et al. reported the first stress MCG study, in which during exercise the ST segment in the MCG became depressed and the baseline TQ segment elevated in direct current MCG [Cohen et al., 1983]. The baseline elevation disappeared after termination of stress more rapidly than the ST depression. The ischemic ST depression in the MCG was mostly due to a steady injury current in the TQ segment, only detectable in the direct current MCG but not in the ECG. The first non-direct current stress MCG recording was performed to a coronary artery disease patient [Saarinen et al., 1974, 1989]. The 12-lead ECG and single channel MCG from 3 different locations were recorded after physical stress. The ratio of the ischemic ST segment depression to R-wave amplitude after stress was even greater in the MCG than in the ECG.
In another study MCG was recorded in rest 5 patients with acute myocardial infarction and after physical stress testing in 11 patients with coronary artery disease [Seese et al., 1995]. Minimum norm method was used for the construction of the current density during the ST-segment. The results were compared with coronary angiography and myocardial scintigraphy. Injury currents during transient exercise-induced ischemia were in most cases directed from the ischemic area to the non-ischemic area, whereas the injury currents in acute infarction flew in the opposite direction.

In a very recent study 27 single vessel coronary artery disease patients and 17 healthy controls were studied with stress MCG [Hänninen et al., 2000]. Physical stress was induced using a nonmagnetic stress ergometer. The ischemia was found to induce a change in the magnetic field orientation in the ST segment and at the T-wave apex, quantified by using a novel surface gradient method (Figs 1 and 2). This change in the magnetic field orientation after stress compared to rest measurement separate the coronary artery disease patients from controls. In the ST segment this rotation of the magnetic field was most prominent in the subgroup of patients with stenosis in left anterior descending coronary artery. In the subgroup of patients with stenosis in the left circumflex coronary artery the map became monopolar after stress. At the T-wave apex, the changes in the field orientation were most prominent in the subgroup of patients with stenosis in the right coronary artery.

Figure 1. The group mean ST segment magnetic field distribution and interpolated spatial field gradient of 17 healthy controls and 27 coronary artery disease patients (CAD) at rest and at cessation of exercise. Spatial distribution of the magnetic field component Bz at rest (A) and at cessation of exercise (C). The line connecting the magnetic field maximum and minimum is marked by an arrow. The change in the magnetic field orientation can be seen as a rotation of the arrow. Interpolated gradient arrows at rest (B) and at cessation of exercise (D). The peak gradient is marked by a circled arrow, which is enlargened. The map direction in A and C parallels the direction of the maximal field gradient in B and D, respectively. Continuous line: positive field distribution, dotted line: negative field distribution, the step between two consecutive lines is 0.1 pT.

Figure 2. The group mean T-wave apex magnetic field distribution and interpolated spatial field gradient of 17 healthy controls and 8 patients with stenosis in right coronary artery (RCA) at rest and 4 minutes postexercise. Spatial distribution of the magnetic field component Bz at rest (A) and 4 minutes postexercise (C). Interpolated gradient arrows at rest (B) and 4 minutes postexercise (D). Technical description is the same as for Figure 1 for the ST-segment. The step between two consecutive lines is 0.5 pT.

5. Myocardial Infarction in MCG

In a study on healthy subjects and post-infarction patients with abnormal ECG, the cardiac magnetic field during QT-interval was characterized by the field strength, field width, and field orientation defined as the location of the center of gravity [van Leeuwen et al., 1999]. The MCG field orientation could separate the postinfarction patients from healthy controls at rest, and these MCG changes were more prominent in patients with multivessel disease.
Lant et al. reported the MCG and the body surface potential mapping to be complementary [Lant et al., 1990]. The most profound abnormalities in the MCG maps in anterior myocardial infarction were found in the QRS-complex, while abnormalities in inferior infarction were found during the whole repolarization period. On the contrary, the potential mapping showed significant differences between anterior and inferior infarctions during QRS-complex only. The repolarization abnormalities detected in MCG maps could have been caused by prolonged ischemic injury increasing tangential current flow through the spared subendocardial tissue.

6. Chronic Ischemia and Myocardial Viability

The magnetic field orientation, width and strenght during QT-interval could separate CAD patients with normal ECGs at rest from healthy subjects [van Leeuwen et al., 1999]. The authors suggested that MCG at rest might contain relevant additional information to ECG, possibly reflecting silent ischemia.
Leder et al. investigated two infarction patients (one anterior and one inferior) and two healthy controls with multichannel magnetocardiography, magnetic resonance imaging (MRI), coronary angiography and Tl-SPECT [Leder et al., 1997]. Current densities were calculated for the endocardial surface of the left ventricle using minimum norm least square criterion during ventricular depolarization, and applied to predefined myocardial geometry acquired from MR images. The main finding was markedly decreased regional current densities in the areas corresponding to the infarcted segments of the heart defined by the clinical reference methods. The authors concluded, however, that further studies are needed in order to define the spatial resolution and the diagnostic performance of this method.

Current density estimation (CDE) was also tested in ischemia localization in three different settings: in simulation studies, in patients with single vessel coronary artery disease (CAD), and in patients three-vessel coronary artery disease and regional or global left ventricular dysfunction [Pesola et al., 1999]. First, in simulation studies zero- and second-order regularizations were tested by placing distrubuted sources on a triangulated epicardial surface on the left ventricle. The source locations were simulated to represent the myocardial regions vascularized by the three main coronary arteries, and the CDE was found to localize extended sources regions. In the second part, CDEs were calculated from the ST segment difference (recovery-rest) signals of four single vessel disease patients. The injury currents were found to match to myocardial regions supplied by the coronary artery in question: to the anterior and septal area for the left anterior descending coronary artery, and to the posterior wall for the right coronary artery stenosis patients. In the third part of the study CDE signals were compared to 18F-2-deoxyglucose (FDG) positron emission tomography (PET) images to differentiate between scarred and viable myocardium in dyskinetic ventricular areas (Fig. 3). The areas of low CDE amplitude were found to match with the scar regions, whereas high CDE amplitude was found to correlate with areas of viable, ischemic cardiac tissue.

Figure 3. Left: The FDG-PET result of a three-vessel disease CAD patient showing a scar in the lateral region. Right: The current density estimation (CDE) of the same patient corresponding to the PET result.

7. Advantages and Disadvantages of Magnetocardiography

Multichannel MCG enables simultaneous recording of the signal over a large area without time consuming electrode attachment. Therefore, in the MCG signal analysis, restriction to one or few recording locations is not needed. Instead, spatial features of MCG maps can be utilized to detect coronary heart disease.

MCG still has of course also some disadvantages compared to electrical mapping. With present available technology the measurements still have to be performed in a shielded room, which excludes the use of method as a quick bed-side test. Furthermore, the MCG recording system is sensitive to metal objects, which is exclusion criteria for some patients.

8. Conclusions

When developing magnetocardiographic methods for the detection of myocardial ischemia, the precordial negative component during repolarization detected in the studies on healthy volunteers should be considered as a normal response to physical exercise. In addition to ST segment amplitude changes similar to those used in the standard 12-lead ECG analysis, the novel multichannel mapping techniques enable to also look at the spatial changes of cardiac magnetic field.

Transient acute myocardial ischemia causes significant changes in the magnetocardiogram of the coronary artery disease patients. Exercise-induced ischemia changes the polarity of the magnetic isofield map in the ST segment and at the T-wave. This change can be quantified with the surface gradient method.

Localization of chronic myocardial ischemia and viable but dyskinetic myocardial tissue using the current density estimation (CDE) method is a promising, clinically important application for magnetocardiography.

Further studies of magnetocardiography in the detection of acute and chronic myocardial ischemia are still needed. Furthermore, dynamic analysis of cardiac magnetic fields during provoked ischemia may improve the ischemia detection and localization in MCG.

References