Frequency Response Characteristics Required for Detection of T-Wave Alternans During Ambulatory ECG Monitoring

Machine learning based detection of T-wave alternans in real ambulatory conditions

BACKGROUND AND OBJECTIVE

T-wave alternans (TWA) is a fluctuation in the repolarization morphology of the ECG. It is associated with cardiac instability and sudden cardiac death risk. Diverse methods have been proposed for TWA analysis. However, TWA detection in ambulatory settings remains a challenge due to the absence of standardized evaluation metrics and detection thresholds.

METHODS

In this work we use traditional TWA analysis signal processing-based methods for feature extraction, and two machine learning (ML) methods, namely, K-nearest-neighbor (KNN) and random forest (RF), for TWA detection, addressing hyper-parameter tuning and feature selection. The final goal is the detection in ambulatory recordings of short, non-sustained and sparse TWA events.

RESULTS

We train ML methods to detect a wide variety of alternant voltage from 20 to 100 μV, i.e., ranging from non-visible micro-alternans to TWA of higher amplitudes, to recognize a wide range in concordance to risk stratification. In classification, RF outperforms significantly the recall in comparison with the signal processing methods, at the expense of a small lost in precision. Despite ambulatory detection stands for an imbalanced category context, the trained ML systems always outperform signal processing methods.

CONCLUSIONS

We propose a comprehensive integration of multiple variables inspired by TWA signal processing methods to fed learning-based methods. ML models consistently outperform the best signal processing methods, yielding superior recall scores.

A multichannel decision-level fusion method for T wave alternans detection

Sudden cardiac death (SCD) is one of the most prominent causes of death among patients with cardiac diseases. Since ventricular arrhythmia is the main cause of SCD and it can be predicted by T wave alternans (TWA), the detection of TWA in the body-surface electrocardiograph (ECG) plays an important role in the prevention of SCD. But due to the multi-source nature of TWA, the nonlinear propagation through thorax, and the effects of the strong noises, the information from different channels is uncertain and competitive with each other. As a result, the single-channel decision is one-sided while the multichannel decision is difficult to reach a consensus on. In this paper, a novel multichannel decision-level fusion method based on the Dezert-Smarandache Theory is proposed to address this issue. Due to the redistribution mechanism for highly competitive information, higher detection accuracy and robustness are achieved. It also shows promise to low-cost instruments and portable applications by reducing demands for the synchronous sampling. Experiments on the real records from the Physikalisch-Technische Bundesanstalt diagnostic ECG database indicate that the performance of the proposed method improves by 12%-20% compared with the one-dimensional decision method based on the periodic component analysis.

Nonparametric signal processing validation in T-wave alternans detection and estimation

Although a number of methods have been proposed for T-Wave Alternans (TWA) detection and estimation, their performance strongly depends on their signal processing stages and on their free parameters tuning. The dependence of the system quality with respect to the main signal processing stages in TWA algorithms has not yet been studied. This study seeks to optimize the final performance of the system by successive comparisons of pairs of TWA analysis systems, with one single processing difference between them. For this purpose, a set of decision statistics are proposed to evaluate the performance, and a nonparametric hypothesis test (from Bootstrap resampling) is used to make systematic decisions. Both the temporal method (TM) and the spectral method (SM) are analyzed in this study. The experiments were carried out in two datasets: first, in semisynthetic signals with artificial alternant waves and added noise; second, in two public Holter databases with different documented risk of sudden cardiac death. For semisynthetic signals (SNR = 15 dB), after the optimization procedure, a reduction of 34.0% (TM) and 5.2% (SM) of the power of TWA amplitude estimation errors was achieved, and the power of error probability was reduced by 74.7% (SM). For Holter databases, appropriate tuning of several processing blocks, led to a larger intergroup separation between the two populations for TWA amplitude estimation. Our proposal can be used as a systematic procedure for signal processing block optimization in TWA algorithmic implementations.

A multilead scheme based on periodic component analysis for T-wave alternans analysis in the ECG

T-wave alternans (TWA) is a cardiac phenomenon that appears in the electrocardiogram (ECG) and is associated with the mechanisms leading to sudden cardiac death (SCD). In this study, we propose the use of a multilead TWA analysis scheme that combines the Laplacian likelihood ratio (LLR) method and periodic component analysis (piCA), an eigenvalue decomposition technique whose aim is to extract the most periodic sources of the signal. The proposed scheme is evaluated in different scenarios--from synthetic signals to stress test ECGs--and is compared to other reported schemes based on the LLR method. Results demonstrate that the piCA-based scheme provides a superior ability to detect TWA than previously reported schemes, and has the potential to improve the prognostic value of testing for TWA.

Methodological principles of T wave alternans analysis: a unified framework

Visible T wave alternans (TWA) in the electrocardiogram (ECG) had been regarded as an infrequent phenomenon during the first 80 years of electrocardiography. Nevertheless, computerized analysis changed this perception. In the last two decades, a variety of techniques for automatic TWA analysis have been proposed. These techniques have allowed researchers to detect nonvisible TWA in a wide variety of clinical and experimental conditions. Such studies have recently shown that TWA is related to cardiac instability and increased arrhythmogenicity. Comparison of TWA analysis methods is a difficult task due to the diversity of approaches. In this paper, we propose a unified framework which holds the existing methods. In the light of this framework, the methodological principles of the published TWA analysis schemes are compared and discussed. This framework may have an important role to develop new approaches to this problem.

Noninvasive sudden death risk stratification by ambulatory ECG-based T-wave alternans analysis: evidence and methodological guidelines

Extensive experimental and clinical evidence supports the utility of T-wave alternans (TWA) as a marker of risk for ventricular fibrillation. This entity appears to reflect the fundamental arrhythmogenic property of enhanced dispersion of repolarization. This relationship probably accounts for its relative ubiquity in patients with diverse types of cardiac disease, as has been recognized with the development of analytical tools. A basic premise of this review is that ambulatory ECG monitoring of TWA as patients experience the provocative stimuli of daily activities can expose latent electrical instability in individuals at heightened risk for arrhythmias. We will discuss the literature that supports this concept and summarize the current state of knowledge regarding the use of routine ambulatory ECGs to evaluate TWA for arrhythmia risk stratification. The dynamic, nonspectral modified moving average analysis method for assessing TWA, which is compatible with ambulatory ECG monitoring, is described along with methodological guidelines for its implementation. Finally, the rationale for combined monitoring of autonomic markers along with TWA will be presented.

Ambulatory ECG monitoring of T-Wave alternans for arrhythmia risk assessment

Experimental and clinical studies indicate a basic linkage between T-wave alternans (TWA) and susceptibility to malignant arrhythmias. In a variety of clinical populations with elevated risk of ventricular tachyarrhythmias, Fast Fourier Transform (FFT)-based assessment of TWA during fixed-rate atrial pacing or bicycle ergometry has shown predictive ability for arrhythmic events. However, after more than a decade since the introduction of TWA testing in human subjects, few studies have explored its utility in ambulatory ECG (AECG) recordings. This gap probably relates to major technical obstacles associated with monitoring of ambulatory subjects, including motion artifact and the requirement of data stationarity, which mandates fixing heart rate. To circumvent these difficulties, we devised a time-domain method, "Modified Moving Average Beat Analysis" (MMA) to determine TWA level accurately in freely moving subjects. Recently, MMA analysis was employed to analyze ambulatory ECG (AECG) records of post-myocardial infarction patients who were were at low risk of arrhythmic death. An increased risk of arrhythmic death was predicted by TWA level above the 75th percentile of controls (p<.05). Thus, the predictive power of TWA obtained with MMA analysis from AECG records obtained appears promising.

Ambulatory electrocardiogram-based tracking of T wave alternans in postmyocardial infarction patients to assess risk of cardiac arrest or arrhythmic death

INTRODUCTION

This is the first study to assess T wave alternans (TWA) analyzed from routine ambulatory electrocardiograms (AECGs) to identify postmyocardial infarction (post-MI) patients at increased risk for arrhythmic events.

METHODS AND RESULTS

The new method of modified moving average (MMA) analysis was used to measure TWA magnitude in 24-hour AECGs from ATRAMI, a prospective study of 1,284 post-MI patients. Using a nested case-control approach, we defined cases as patients who experienced cardiac arrest due to documented ventricular fibrillation or arrhythmic death during the follow-up period of 21 +/- 8 months. We analyzed 15 cases and 29 controls matched for sex, age, site of MI, left ventricular ejection fraction, thrombolysis, and beta-blockade therapy. TWA was reported as the maximum 15-second value at three predetermined times associated with cardiovascular stress: maximum heart rate, 8:00 A.M., and maximum ST segment deviation. TWA increased significantly from baseline in both leads at each time point (P <<0.01) in cases and controls. TWA in V5 increased more in cases than controls during peak heart rate (P = 0.005) and at 8:00 A.M. (P = 0.02). A 4- to 7-fold higher odds of life-threatening arrhythmias was predicted by TWA level above the 75th percentile during maximum heart rate in leads V1 (odds ratio [OR] 4.2, 95% confidence interval [CI]: 1.1-16.3, P = 0.04) and V5 (OR 7.9, 95% CI: 1.9-33.1, P = 0.005). TWA at 8:00 A.M. also predicted risk in leads V1 (OR = 5.0, 95% CI: 1.2-20.5, P = 0.02) and V5 (OR = 4.2, 95% CI: 1.1-16.3, P = 0.04).

CONCLUSION

TWA measurement from routine 24-hour AECGs is a promising approach for risk stratification for cardiac arrest and arrhythmic death in relatively low-risk post-MI patients.

Modified moving average analysis of T-wave alternans to predict ventricular fibrillation with high accuracy

T-wave alternans is a marker of cardiac electrical instability with the potential for arrhythmia risk stratification. The modified moving average method was developed to measure alternans in settings with artifacts, noise, and nonstationary data. Algorithms were developed and performance characteristics were validated with simulated electrocardiograms (ECGs). Experimental laboratory ECGs with dynamically changing alternans values were analyzed. Alternans values estimated by modified moving average analysis correlated strongly with input alternans values (r(2) = 0.9999). Rapidly changing alternans levels and phase reversals did not perturb the measurement. When heart rate was increased from 60 to 180 beats/min, with T-wave alternans apex moving from 237 to 103 ms after the R wave, the measured alternans peak varied <5% from input value. Simulated 50- to 1,000-microV motion artifact spikes typical of treadmill ECGs produced inaccuracies <2%. Alternans values in experimental laboratory study using standard electrodes tracked vulnerability to myocardial ischemia-induced ventricular fibrillation with 100% sensitivity and specificity at a cut point of 0.75 mV. Modified moving average analysis is a robust method that precisely measures T-wave alternans in settings with artifacts, noise, and nonstationary data typical of clinical ECGs and yields an accurate estimate of risk for ventricular fibrillation.

Electrical behavior of T-wave polarity alternans in patients with congenital long QT syndrome

OBJECTIVES

This study was designed to evaluate the incidence and characteristics of onset of T-wave polarity alternans (TWPA) in patients with long QT syndrome.

BACKGROUND

The T-wave alternans is a phenomenon that consists of beat-to-beat variability in the amplitude, morphology, and sometimes polarity of the T-wave, and it may trigger life-threatening arrhythmias.

METHODS

The 24-h Holter recordings of 11 patients with congenital long QT syndrome were studied. Episodes of TWPA with 10 or more consecutive cycles were selected and analyzed as follows: 1) mean cycle length (MCL) and QTc interval duration (QTcI) of the episodes of TWPA and the 10 cycles preceding and succeeding the TWPA; 2) MCL and QTcI of the third, second, and first minute before onset (Mn_3, Mn_2, Mn_1); 3) MCL and QTcI from the tenth to the first cycle immediately preceding the onset of TWPA (R_10 to R_1); 4) MCL and QTcI from the first to the fourteenth cycle during alternans (R0 to R14); 5) MCL and QTcI from the first to the tenth cycle immediately succeeding TWPA (R+1 to R+10); 6) linear correlation (Lnc) between QT interval and cycle length (CL) (LncQT/CL) during alternans and for the 10 preceding cycles; 7) Lnc between the first three alternans cycles and episode duration (Lnc 3CL/EpD); and 8) difference between the longest and shortest QTc interval. We also selected episodes consisting of four or more consecutive cycles in order to analyze daily rhythms of the phenomenon.

RESULTS

The TWPA was observed in 5 (45%) out of the 11 patients studied. The alternans process is initiated by a sudden shortening of the first alternans cycle without previous heart rate changes and ends at the moment when prolongation of the cycle tends to occur. LncQT/ CL-alternans: r = 0.38 +/- 0.2 (p = 0.20); without alternans: r = 0.81 +/- 0.06 (p = 0.01). Lnc 3CL/EpD: r = 0.002 (p = 0.992). The QTc difference during alternans: 312.0 +/- 52.1 ms; without alternans: 86.0 +/- 36.4 ms (p = 0.001). Daily rhythm: 71% of the episodes occurred between 8 AM and 8 PM, with higher incidence during the morning.

CONCLUSIONS

The TWPA was dependent on the cardiac CL; there was loss of the LncQT/CL and an increase in the QT interval variability. Like other biological variables, T-wave polarity alternans has a higher density during the morning.