# Risk evaluation of ventricular tachycardia using wavelet transform irregularity of the high-resolution elect rocard iog ram P. Lewandowski I O. Meste 2 R. Maniewski I T. Mroczka 1"3 K. Steinbach 3 H. Rix 2 1Institute of Biocybernetics & Biomedical Engineering PAS, Department of Biomeasurements & Biocontrol, Warsaw, Poland 2Universit6 de Nice, Sophia Antipolis, France 3Ludwig Boltzmann Arrhythmia Research Institute, Vienna, Austria Abstract--A new method for analysis of high-resolution ECG signals using a wavelet transform based on a modified Morlet function is presented. A polynomial filter is used to reduce low-frequency, high-amplitude noise components in the analysed signals. The method is tested on test ECG signals with simulated late potentials and finally verified on two post-infarction patient (PP) groups: 62 PPs with ventricular tachycardia and 44 PPs without arrhythmia. A new quantitative parameter, the irregularity factor, is proposed for discrimination between the study groups. The results show a significant difference in the parameter values for tachycardia patients compared with those for patients without arrhythmia. The sensitivity of the proposed method is 85 %, and the specificity is 93 %. Keywords--High-resolution ECG, Late potentials, Spectro-temporal analysis, Wavelet transform, Ventricular tachycardia Med. Biol. Eng. Comput., 2000, 38, 666-673 J 1 Introduction HIGH-RESOLUTIONELECTROCARDIOGRAPHY(HR ECG) is an approved and widely used non-invasive method for detecting ventricular late potentials. These low-amplitude signals of higher frequency in the terminal portion of the QRS complex are proven to be markers of risk for ventricular tachycardia, particularly in patients after myocardial infarction (BREITHARDT et al., 1983; ZIMMERMANN et al., 1985; KUCHAR et al., 1986; GOMES et al., 1989). The HR ECG signals, usually recorded in the X, Y and Z leads, are first pre-processed by filtering, averaged and then analysed using various time-amplitude, frequency-domain, or spectro- temporal methods. Finally, the results are evaluated by the calculation of quantitative parameter(s) for arrhythmia risk evaluation. Various methods of analysis have been developed for the detection of micropotentials. Time-domain analysis, proposed by SIMSON (1981), is based on the bidirectional high-pass filtering of averaged ECG signals recorded in three X, Y, Z orthogonal leads. From these filtered signals, the so-called modulus of the Simson vector is calculated as the root mean square of all the leads x/X 2 + y2 + Z 2. Despite some limita- tions, this method is widely used and is recognised as a standard Correspondence should be addressed to Prof. R. Maniewski; emaih roman@hrabia.ibib.waw.pl First received 28 January 2000 and in final form 16August 2000 MBEC online number: 20003518 © IFMBE:2000 in HR ECG analysis (BREITHARDT et al., 1991). Time-ampli- tude parameters obtained from the Simson vector are, in many cases, effective criteria for late-potential (LP) detection. Besides time-domain methods, classical frequency analysis using a fast Fourier transform (CAIN et al., 1984, 1985; PIERCE et al., 1989; HABERL et al., 1989), as well as methods based on autoregressive modelling (SCHELS et al., 1991; LEWANDOWSKI et al., 1994) has been proposed. Analysis is usually carried out on the terminal part of the QRS complex and ST segment of the ECG signal. For the evaluation of obtained spectra, various quantification parameters, based mainly on the spectral area ratio, have been proposed. These parameters allow for detection of high-frequency components representing late potentials at the end of the QRS complex and ST segment. However, time-domain methods, as well as frequency- domain analysis, do not allow for the effective detection of LPs because of the presence of excessive noise and the time- varying frequency spectrum of the ECG signal. Combined time-frequency analysis, e.g. the analysis of short data segments shifted with a constant time step, often called short time Fourier transform or spectro-temporal mapping (LANDER et al., 1992), is more reliable for analysis ofHR ECG. in this case, the results of analysis are presented in the form of a three- dimensional map of time, frequency and power spectral density (PSD). To evaluate the results, various methods of quantifica- tion of spectro-temporal distributions have been proposed, such as: the normality factor introduced by HABERL et al. (1989), the spectral factor proposed by LEWANDOWSKI et al. (1994) and spectral turbulence analysis proposed by KELEN et al. (1991). 666 Medical & Biological Engineering & Computing 2000, Vol. 38