T-wave changes in patients with hemodynamic evidence of systolic or diastolic overload of the left ventricle: a retrospective study on 168 patients with isolated chronic aortic valve disease

An Artificial Intelligence Analysis of Electrocardiograms for the Clinical Diagnosis of Cardiovascular Diseases: A Narrative Review

Artificial intelligence (AI) applied to cardiovascular disease (CVD) is enjoying great success in the field of scientific research. Electrocardiograms (ECGs) are the cornerstone form of examination in cardiology and are the most widely used diagnostic tool because they are widely available, inexpensive, and fast. Applications of AI to ECGs, especially deep learning (DL) methods using convolutional neural networks (CNNs), have been developed in many fields of cardiology in recent years. Deep learning methods provide valuable support for rapid ECG interpretation, demonstrating a diagnostic capability overlapping with specialists in the diagnosis of CVD by a classical analysis of macroscopic changes in the ECG trace. Through photoplethysmography, wearable devices can obtain single-derivative ECGs for the recognition of AI-diagnosed arrhythmias. In addition, CNNs have been developed that recognize no macroscopic electrocardiographic changes and can predict, from a 12-lead ECG, atrial fibrillation, even from sinus rhythm; left and right ventricular function; hypertrophic cardiomyopathy; acute coronary syndromes; or aortic stenosis. The fields of application are many, but numerous are the limitations, mainly associated with the reliability of the acquired data, an inability to verify black box processes, and medico-legal and ethical problems. The challenge of modern medicine is to recognize the limitations of AI and overcome them.

Electrocardiogram Changes in the Spectrum of TTNtv Dilated Cardiomyopathy: Accuracy and Predictive Value of a New Index for LV-Changes Identification

BACKGROUND

Truncating TTN variants (TTNtv) are the main cause of dilated cardiomyopathy (DCM). The dynamic nature of this entity has previously been described. Based on own empirical observations and previous evidences, this study assessed repolarisation patterns and the possible association with morphological and functional status of TTNtv-DCM patients.

METHODS

Electrocardiograms (ECGs) of index patients with TTNtv-DCM and their relatives were included and matched in time with an echocardiogram. All individuals were classified into five phenotype groups: 1) Reduced left ventricular ejection fraction (LVEF <50%); 2) Recovered LVEF: at least 10% increase and LVEF >30% after optimal medical treatment; 3) Borderline phenotype (mildly enlarged ventricle and/or hyper-trabeculation); 4) Genotype positive, phenotype negative; and 5) Non-carriers. All electrocardiograms were evaluated by two blinded observers in qualitative and quantitative terms [T index (mm)=Σ T-wave amplitude (V5, V6, II, aVF)] and these data were compared with demographic and clinical information. The Δ T-index was calculated in those individuals with more than one electrocardiogram.

RESULTS

Seventy-eight (78) electrocardiograms were included (46% female, mean age 50 years). T-index and prevalence of an abnormal T-wave had significantly different results among the groups (p<0.0001). Age and haemodynamic factors were shown to be ECG-modifiers, especially in phenotype-negative patients. T-index enabled individuals with reduced LVEF (<2.5) to be identified and to differentiate patients with favourable and unfavourable responses to treatment (Δ T index >3.5 and ≤2, respectively).

CONCLUSION

Repolarisation changes enabled characterisation of the spectrum of TTNtv-DCM. The T-index identified potential carriers and patients with the worst profiles of the spectrum of the disease.

Deep Learning-Based Algorithm for Detecting Aortic Stenosis Using Electrocardiography

Background

Severe, symptomatic aortic stenosis (AS) is associated with poor prognoses. However, early detection of AS is difficult because of the long asymptomatic period experienced by many patients, during which screening tools are ineffective. The aim of this study was to develop and validate a deep learning–based algorithm, combining a multilayer perceptron and convolutional neural network, for detecting significant AS using ECGs.

Methods and Results

This retrospective cohort study included adult patients who had undergone both ECG and echocardiography. A deep learning–based algorithm was developed using 39 371 ECGs. Internal validation of the algorithm was performed with 6453 ECGs from one hospital, and external validation was performed with 10 865 ECGs from another hospital. The end point was significant AS (beyond moderate). We used demographic information, features, and 500‐Hz, 12‐lead ECG raw data as predictive variables. In addition, we identified which region had the most significant effect on the decision‐making of the algorithm using a sensitivity map. During internal and external validation, the areas under the receiver operating characteristic curve of the deep learning–based algorithm using 12‐lead ECG for detecting significant AS were 0.884 (95% CI, 0.880–0.887) and 0.861 (95% CI, 0.858–0.863), respectively; those using a single‐lead ECG signal were 0.845 (95% CI, 0.841–0.848) and 0.821 (95% CI, 0.816–0.825), respectively. The sensitivity map showed the algorithm focused on the T wave of the precordial lead to determine the presence of significant AS.

Conclusions

The deep learning–based algorithm demonstrated high accuracy for significant AS detection using both 12‐lead and single‐lead ECGs.

Electrocardiographic characteristics indicating a risk of irreversibly impaired myocardial function in chronic aortic regurgitation

In order to define which of selected ECG variables could indicate irreversibly impaired myocardial function in chronic aortic regurgitation 54 patients were stratified according to normal (> or = 50%; Group A, n = 41) or subnormal radionuclide left ventricular ejection fraction (LVEF < 50%; Group B, n = 13) late after aortic valve replacement. Preoperatively, Group B patients had a significantly greater QRS duration, greater R-peak time (RPT) prolongation in I, V5 or V6, greater RPT relative to the S-peak time of the maximum S in V1, V2 or V3 (R-peak delay) and a greater negative T-wave in I or V6, as compared with Group A. These ECG variables together with preoperative angiocardiographic LVEF and end-systolic volume index were subjected to stepwise linear discriminant analysis. The maximum RPT, angio-LVEF and the maximum RPT relative to the S-peak time of the maximum S in V1, V2 or V3 emerged as the most promising variables. Of of Group A patients 82.9% and 84.6% of Group B patients were correctly classified by the three variables. If applied separately, APT prolongation or the presence of the R-peak delay in the left-sided leads, although less sensitive, have reasonably high specificity as risk indicators of irreversibly impaired chamber function, their positive predictive value being 60 and 62.5%, respectively. In conjunction with preoperative LVEF the diagnostic contribution of the two ECG variables amounts to the greatest overall separation of postoperatively preserved from irreversibly impaired systolic function.

Repetitively negative changing T waves at 24-h electrocardiographic monitors in patients with the chronic fatigue syndrome. Left ventricular dysfunction in a cohort

This study surveys the occurrence of repetitively negative to flat T waves, alternating with normal upright T waves in 24-h electrocardiographic recordings from a subspecialty infectious diseases outpatient practice during the years 1982 to 1990. Patients with normal resting electrocardiogram in the assayed leads, but with repetitively inverted to isoelectric abnormal T waves at Holter monitors, were considered to have abnormal readings. A total of 300 patients had undergone a 24-h Holter monitor. This group included 24 individuals with chronic fatigue syndrome (CFS). This population was restricted to individuals 50 years old or younger, and the patients with CFS are compared with the patients without CFS. One of the more striking differences between the two groups was the difference in abnormal Holter readings. The patients with CFS all had abnormal Holter readings, while 22.4 percent patients without CFS had abnormal readings (p < 0.01). We further report the occurrence of mild left ventricular dysfunction in 8 of 60 patients in continuing studies of this population with CFS, younger than 50 years old, and with no risk factors for coronary artery disease. All 60 patients with CFS showed repetitively flat to inverted T waves alternating with normal T waves. Stress multiple gated acquisitions (MUGAs) (labeled erythrocytes with stannous pyrophosphate) were abnormal in eight patients with CFS. Although resting ejection fractions (EFs) were normal (mean, 60 percent), with increasing work loads (Kilopon meters [Kpms]), gross left ventricular dysfunction occurred. The fatigue of patients with CFS may be related to subtle cardiac dysfunction occurring at work loads common to ordinary living.