location of coronary arterY stenosis using bspm
and new approach to ecg signal decomposition

M. Sobieszczańska, J. Jagielski, L. Rusiecki
Department of Pathophysiology, Wroclaw Medical University
Marcinkowskiego 1, 50-368 Wrocław, POLAND

Abstract: The study was undertaken to assess the features of QRST isointegral maps and the electric myocardium activity estimated by the ECG signal decomposition in the patients with symptomatic myocardial ischemic disease, presenting no changes in resting 12-lead ECG, and with documented one-vessel or multi-vessel coronary artery lesions. The potential minimum extremes of the QRST isointegral maps from the patient groups revealed no significant differences as compared with the controls, excluding the cases with three-vessel CAD. Contrarily, the QRST maps displayed the unique patterns of the integral negativity distribution in relation to the coronary artery narrowing sites. The ECG signal decomposition showed the disabled electric activity of the particular portions of myocardium in some of the CAD patients, especially in those with multi-vessel lesions, however it was of less sensitivity than QRST isointegral mapping.

INTRODUCTION

It is commonly known that up to 50% patients with considerable stenosis of coronary arteries do not manifest, when pain free, the relevant abnormalities in resting ECG [1]. Body surface potential mapping (BSPM) is regarded to be of higher sensitivity to the regional ischemic lesions within myocardium [2]. The previous studies considered a variety of body surface maps recorded under specific clinical arrangements, such as resting, exercise, post-exercise states or during PTCA procedure to evaluate the BSPM accuracy in identifying the site of coronary lesion in patients with CAD. These papers rendered conflicting appraisals of the BSPM clinical usefulness for the CAD location detecting [3,4,5,6].

A goal of the presented study was to investigate the features of the isointegral QRST maps in the patients with single-vessel or multi-vessel CAD having no alternations in the resting 12-lead ECG. The additional verification of myocardial electric activity in the examined patients was performed by using the ECG signal decomposition process.

METHODS

Technique

The BSPM procedure was performed using HPM-7100 Fukuda Denshi system, enabling simultaneous recording of the ECG signals from 87 electrodes arranged in a matrix covering entire anterior (59 leads) and posterior (28 leads) thoracic surfaces. The ECG data were averaged and sampled at the rate of 1000 Hz, with the Wilson’s central terminal as reference, in order to create isointegral maps for the whole QRST interval. The QRST maps displayed the thorax as a cylinder unrolled from the right axillary line with the front designated by A-I, and the back by J-M electrode rows.

The second method applied was a new approach to the ECG signal decomposition based on the Fourier transform. The algorithm enhanced the Electron Spin Resonance (ESR) spectrum resolution. The computer program named RKU [7] converted the experimental spectrum of ECG signals into a spectrum with the higher resolution. This method enables of correlating the individual peaks of the newly created ECG curve with the particular portions of the myocardium. Using this method the electric activity of the anterior (A), lateral (L), inferior (I), and posterior (P) walls was analyzed.

Material

A group of 50 patients (43 males and 7 females) with symptomatic coronary artery disease (CAD), with informed consent given, were enrolled to the study. Their age ranged from 37 to 78 years (mean, 54,7±10,2 years). The recruited patients demonstrated the clinical evidence of angina pectoris episodes documented angiographically by critical luminal stenosis (>70%) confined to a single coronary artery, i.e., RCA, LAD, or Cx. According to this criteria, the patients were divided to the groups R (n=10), L (n=18), and C (n=7), respectively. In addition, the group M (n=15), presenting the multi-vessel CAD with significant narrowings (>50%) in the all of the three coronary arteries, was constituted. The resting 12-lead ECG registered during painless periods revealed no abnormalities in the all of the examined patients. None of the patients had a proven history of previous myocardial infarction, either clinical evidence of congestive heart failure, or conduction disturbances. The all patients underwent the BSPM recordings being in the resting, angina-free state.

The group-mean values of potential minimum extremes, likewise their location on the QRST maps for each of the patient groups were compared with the parameters from the 30 age-matched control subjects with no cardiac involvement.

RESULTS

The group-mean values of the potential minimum extremes for the QRST isointegral maps were calculated for the particular patient groups: R, -42,3 (± 17,6)mVs; L, -55,9 (±18,5)mVs; C, -21,9 (±5,29)mVs; and M, -106,1 (±61,1)mVs. In turn, the mean potential minimum value for the control subjects was -48,5 (±12,7)mVs. It was found that only in the group with multi-vessel CAD the QRST integral minimum extremes were significantly (with p<0,0002) more negative comparing with the group-mean minimums for the controls, likewise with the all remain patients groups (R, L, and C).

The minimum extremes appeared on the QRST integral maps in the R group in the area extending from A3-7 to F3-5, as well as from L2-6 to M2-6, in the L group from D4-5 to H2-3, in the C group from D1-3 to H1-3, and in the M group heterogenously from A1-7 to H1-7. The negative potentials in the R group were situated over the right anterior thorax and around the back. The L group showed the integral negativity over the anteroinferior thorax, leftward. In the C group, the potential minimums covered the entire inferior thorax.

The QRST negativity distributions in the CAD groups contrasted with the normal QRST integral distribution, where the negative potentials occupied the right anterosuperior torso (A-D3-7) and the back region (L-M). See Fig. 1.

The ECG signal decomposition in the group R revealed the significantly decreased, as compared with the standard, electric activity of the A wall in 22% of the cases, of the P wall in 78%, of the I wall in 33%, and no case of the L wall disability. In the LAD group, the attenuated electric activity was found for the A wall in 29% of the patients, for the L wall in 17%, for the P wall in 64%, and for the I wall in 50% of the cases. The C group estimated in this aspect showed, as follows: the A wall was disabled in 25% of the patients, the L wall in 15%, the P wall in 75%, and the I wall in 50% of the cases. Finally, the most remarkable findings were noted in the M group, where the considerably diminished electric activity was shown by the A wall in 70% of the cases, by the L wall in 60%, by the P wall also in 60%, and by the I wall in the 90% of the patients. Furthermore, in 70% of the M group patients the concomitant electric disability of these all four portions of the myocardium was observed.

Figure 1. Schemes of the QRST integral distribution patterns for the normal (N), RCA (R), LAD (L) and Cx (C) groups

DISCUSSION

Because of shortcomings caused by a scarcity of the examined patients’ number, the results of the presented investigations should be considered as preliminary ones. However, it is noteworthy that the negative potential distribution on the QRST integral maps revealed an unique characteristics in reference to the affected vessel.

The central positivity seen on the QRST isointegral maps in the normals over anterior torso was replaced in the each of the patient groups by the negativity, which was significantly intensified for the three-vessel CAD patients.

The specific distribution for the group with RCA lesions comprised the negativity located on the right anterior torso coupled with the minimum on the back. The pattern for the patients with LAD stenosis showed a negativity extended over left anterior thorax and downward, and the characteristic distribution for the group with Cx artery presented the wide negativity spread over the entire inferior zone of the thorax.

The observed surface distributions are in certain keeping with the patterns for the QRST integrals reported by Green et al. [1] and by Montague et al. [4]. However, making a direct comparison between the findings noted by the various authors are difficult because of the divergence of the techniques used.

The presented study proved that QRST integral negativity distribution is of more specificity with regard to the CAD location in comparison with the electric myocardial activity estimation by the ECG signal decomposition.

Finally, one should bear in mind the fact that searching for a precise correlation between the affected vessel and the damaged portion of the myocardium can be confusing., which results from the anatomic variations, steal syndrome, presence of anastomoses, or thorax size and resistance.

REFERENCES

[1] L.S. Green, R.L. Lux, C.W. Haws, “Detection and localization of coronary artery disease with body surface mapping in patients with normal electrocardiograms,” Circulation, vol. 76, pp. 1290-1297, 1987.

[2] N.C. Flowers, L. Horan, “Body surface potential mapping,” in Cardiac Electrophysiology. From Cell to Bedside. D.P. Zipes, J. Jalife (eds.), 3^rd ed. Philadelphia: W.B. Sanders, 2000, pp. 737-746.

[3] M.R. Sridharan, L.G. Horan, C. Hand, et al., “Use of Body Surface Maps to Identify Vessel Site of Coronary Occlusion,” J. Electrocardiology, vol. 22 (Suppl.), pp. 72-81, 1989.

[4] T.J. Montague, F.X. Witkowski, R.M. Miller, et al., “Exercise Body Surface Potential Mapping in Single and Multiple Coronary Artery Disease,” Chest, vol. 97, pp. 1333-1342, 1990.

[5]  M. Shenasa, D. Hamel, J. Nasmith, et al., “Body Surface Potential Mapping of ST-segment Shift in Patients Undergoing Percutaneous Transluminal Coronary Angioplasty,” J. Electrocardiology, vol., 26, pp. 25-41, 1993.

[6] H. Miyakoda, T. Kinugawa, K. Ogino, et al., “QRST integral analysis of body surface electrocardiographic mapping for assessing exercise induced changes in the spatial distribution of local repolarization properties in patients with coronary artery disease and in patients with previous anterior infarction,” J. Electrocardiology, vol. 32, pp. 123-136, 1999.

[7] R. Krzyminiewski, “Computer enhancement of complex spectroscopic spectra resolution,” Molecular Physics Reports, vol. 6, pp. 174-179, 1994.