The signal-averaged ECG: time-domain analysis

An artificial intelligence-enabled ECG algorithm for identifying ventricular premature contraction during sinus rhythm

Background

Ventricular premature complex (VPC) is a common arrhythmia in clinical practice. VPC could trigger ventricular tachycardia/fibrillation or VPC-induced cardiomyopathy in susceptible patients. Existing screening methods require prolonged monitoring and are limited by cost and low yield when the frequency of VPC is low. Twelve-lead electrocardiogram (ECG) is low cost and widely used. We aimed to identify patients with VPC during normal sinus rhythm (NSR) using artificial intelligence (AI) and machine learning-based ECG reading.

Methods

We developed AI-enabled ECG algorithm using a convolutional neural network (CNN) to detect the ECG signature of VPC presented during NSR using standard 12-lead ECGs. A total of 2515 ECG records from 398 patients with VPC were collected. Among them, only ECG records of NSR without VPC (1617 ECG records) were parsed.

Results

A total of 753 normal ECG records from 387 patients under NSR were used for comparison. Both image and time-series datasets were parsed for the training process by the CNN models. The computer architectures were optimized to select the best model for the training process. Both the single-input image model (InceptionV3, accuracy: 0.895, 95% confidence interval [CI] 0.683–0.937) and multi-input time-series model (ResNet50V2, accuracy: 0.880, 95% CI 0.646–0.943) yielded satisfactory results for VPC prediction, both of which were better than the single-input time-series model (ResNet50V2, accuracy: 0.840, 95% CI 0.629–0.952).

Conclusions

AI-enabled ECG acquired during NSR permits rapid identification at point of care of individuals with VPC and has the potential to predict VPC episodes automatically rather than traditional long-time monitoring.

Evidence of altered cardiac electrophysiology following prolonged androgenic anabolic steroid use

The non-therapeutic use of androgenic anabolic steroids (AAS) is associated with sudden cardiac death. Despite this, there is no proposed mechanism by which this may occur. Signal-averaged ECG (SAECG) allows the assessment of cardiac electrical stability, reductions of which are a known risk factor for cardiac arrhythmias. The aim of the present study was to examine cardiac electrical stability using SAECG in a group (n = 15) of long-term AAS users (AAS use 21.3 ± 3.1 years) compared with a group (n = 15) of age-matched weight lifters (WL) and age-matched sedentary controls [C (n = 15)]. AS, WL and C underwent SAECG analysis at rest and following an acute bout of exercise to volitional exhaustion. SAECGs were analyzed using a 40 Hz filter and were averaged over 200 beats. Results indicate a non-significant trend for increased incidence of abnormal SAECG measures at rest in AS (P = 0.55). However, AS demonstrated a significantly higher incidence of abnormalities of SAECG following exercise than C or WL (P < 0.05). In conclusion, the higher incidence of abnormal SAECG measurements immediately post-exercise in the AAS group places them at a greater risk of sudden death. The present study provides a strong contraindication to the use of AAS.

Late potentials and ventricular arrhythmias in acromegaly

BACKGROUND

Sudden death and increased prevalence of ventricular arrhythmias have already been described in acromegaly. Although late potentials (LPs) have been proved to be a new technique in detecting patients at risk for ventricular tachyarrhythmias its use in acromegaly is still unknown.

METHODS

We studied 70 acromegalic patients [32 males, 38 females; age 49+/-12 years (mean+/-S.D.)] and 70 control subjects age- and sex-matched [(35 males and 35 females; 46+/-12 years (mean+/-S.D.)]. Besides hormonal tests, we performed the following cardiovascular investigations: ECG, 24-h ECG Holter monitoring, echocardiography, and signal-averaged ECG (SAECG) time-domain analysis.

RESULTS

LPs occurrence was significantly higher in acromegalic patients as compared to the control group (22.9% vs. 2.9%; p=0.001). A greater duration of disease in patients with positive LPs compared to negative ones was pointed out (18 vs. 12 years; p=0.024). In the group of acromegalic patients with positive LPs we observed a significant association with premature ventricular complexes (PVCs) detected by means of 24-h Holter ECG recording (13 out of 15 patients: 86.7%; p=0.024). The positivity or negativity of LPs proved to be significantly associated with Lown scale PVC trends recorded by 24-h Holter ECG (p=0.014). In the group of patients with left ventricular hypertrophy a significant and pathological worsening of SAECG signals (QRS, LAS, RMS) was documented.

CONCLUSIONS

We observed a higher prevalence of LPs in acromegaly which significantly correlated with Lown scale of PVCs.

Use of signal-averaged electrocardiography in the evaluation of arrhythmogenic right ventricular cardiomyopathy in Boxers

OBJECTIVE

To assess signal-averaged electrocardiography (SAECG) for evaluation of Boxers with arrhythmogenic right ventricular cardiomyopathy (ARVC) and identify dogs at risk for sudden death (SD) or death related to congestive heart failure (CHF).

DESIGN

Prospective study.

ANIMALS

94 Boxers with ARVC and 49 clinically normal non-Boxers (controls).

PROCEDURE

Boxers were screened for ARVC, and severity was estimated by use of echocardiography, 24-hour ambulatory ECG, and SAECG. Statistical evaluation was performed to identify significant differences in SAECG variables relative to clinical outcome, frequency of ventricular arrhythmias, and systolic function. Sensitivity, specificity, and positive and negative predictive values were evaluated for each SAECG variable for occurrence of SD or death related to CHF. Late potentials were also evaluated as a predictor of cardiac-related death.

RESULTS

Differences were detected in SAECG variables on the basis of clinical outcome, systolic function, and frequency of ventricular arrhythmias. More severely affected dogs had significantly more abnormal SAECG findings. The presence of late potentials, defined as 2 abnormal root mean square values (of 4), was associated with high sensitivity, specificity, and negative predictive value for cardiac-related SD or death secondary to CHF CONCLUSIONS AND CLINICAL RELEVANCE: Results suggest that SAECG is a useful noninvasive diagnostic test to evaluate dogs affected with ARVC and identify individuals at risk for cardiac-related death.

Multiresolution decomposition of the signal-averaged ECG using the mallat approach for prediction of arrhythmic events after myocardial infarction

The aim of this study was to analyze the ability of the multiresolution decomposition of the signal-averaged electrocardiogram (ECG) to discriminate between patients who develop life-threatening ventricular arrhythmias after myocardial infarction and those who do not and to compare the predictive values of this approach with those obtained from the analysis of ventricular late potentials in the time domain. Signal-averaged ECGs of 769 prospectively included patients were analyzed. A total of 42 arrhythmic events occurred during the follow-up period. For numerical calculations of wavelet analysis, the total and relative energies of the QRS complex were obtained in seven frequency bands. The combination of the relative energy in the frequency bands 7.8-15.6 Hz and 62.5-125 Hz enhanced statistical performance as compared with the time-domain parameters (positive predictive accuracy, 11.3 vs 8.2%). Combining wavelet transform and time-domain parameters enhanced the predictive values even more (positive predictive accuracy, 14.3%) compared with applying each method alone.