Electrocardiographic Machine Learning to Predict Left Ventricular Diastolic Dysfunction in Asian Young Male Adults

Association between secondhand smoke exposure and incident heart failure: The Multi-Ethnic Study of Atherosclerosis (MESA)

AIMS

There are no studies on the association between secondhand smoke (SHS) exposure and incident heart failure (HF). This cohort study aimed to examine the associations of self-reported and urinary cotinine-assessed SHS exposure with incident HF.

METHODS AND RESULTS

This study included 5548 non-active smoking participants aged 45-84 years and free of known cardiovascular diseases and HF at baseline who self-reported SHS exposure time in the Multi-Ethnic Study of Atherosclerosis (MESA) at baseline (2000-2002). A cohort subset of 3376 non-active smoking participants underwent urinary cotinine measurements. HF events were verified by medical records or death certificates and ascertained from baseline through 2019. Multivariable Cox proportional hazards regression analysis was used with adjustment for demographic variables, traditional cardiovascular risk factors, physical activity, tobacco pack-years and medications. During a median follow-up of 17.7 years, 353 and 196 HF events were identified in the self-report cohort and cohort subset, respectively. In the self-report cohort, compared with the SHS unexposed group (0 h/week), the highest tertile of the SHS exposed group (7-168 h/week) was not associated with incident HF (hazard ratio [HR] 0.70, 95% confidence interval [CI] 0.49-1.00; p = 0.052). In contrast, in the cohort subset, participants with detectable urinary cotinine >7.07 ng/ml had a higher risk of incident HF than those with undetectable urinary cotinine ≤7.07 ng/ml (HR 1.45, 95% CI 1.03-2.06; p = 0.034). There were no significant heterogeneities in HF risk by age, sex, race/ethnicity, or past smoking status.

CONCLUSION

Secondhand smoke exposure reflected by modestly increased urinary cotinine (>7.07 ng/ml) rather than self-report in non-active smokers was associated with a 40-50% higher risk of any HF event.

Phonocardiogram transfer learning-based CatBoost model for diastolic dysfunction identification using multiple domain-specific deep feature fusion

Left ventricular diastolic dyfunction detection is particularly important in cardiac function screening. This paper proposed a phonocardiogram (PCG) transfer learning-based CatBoost model to detect diastolic dysfunction noninvasively. The Short-Time Fourier Transform (STFT), Mel Frequency Cepstral Coefficients (MFCCs), S-transform and gammatonegram were utilized to perform four different representations of spectrograms for learning the representative patterns of PCG signals in two-dimensional image modality. Then, four pre-trained convolutional neural networks (CNNs) such as VGG16, Xception, ResNet50 and InceptionResNetv2 were employed to extract multiple domain-specific deep features from PCG spectrograms using transfer learning, respectively. Further, principal component analysis and linear discriminant analysis (LDA) were applied to different feature subsets, respectively, and then these different selected features are fused and fed into CatBoost for classification and performance comparison. Finally, three typical machine learning classifiers such as multilayer perceptron, support vector machine and random forest were employed to compared with CatBoost. The hyperparameter optimization of the investigated models was determined through grid search. The visualized result of the global feature importance showed that deep features extracted from gammatonegram by ResNet50 contributed most to classification. Overall, the proposed multiple domain-specific feature fusion based CatBoost model with LDA achieved the best performance with an area under the curve of 0.911, accuracy of 0.882, sensitivity of 0.821, specificity of 0.927, F1-score of 0.892 on the testing set. The PCG transfer learning-based model developed in this study could aid in diastolic dysfunction detection and could contribute to non-invasive evaluation of diastolic function.

Electrocardiographic and cardiometabolic risk markers of left ventricular diastolic dysfunction in physically active adults: CHIEF heart study

Aim

This study was aimed to investigate the association of cardiometabolic and ECG markers with left ventricular diastolic dysfunction (LVDD) in physically active Asian young adults, which has not been clarified in prior studies.

Methods and results

A total of 2,019 men aged 18–43 years were included from the military in Taiwan. All the subjects underwent anthropometric, hemodynamic, and blood metabolic marker measurements. Physical fitness was investigated by time for a 3,000-m run. LVDD was defined by presence of either one of the three echocardiographic criteria: (1) mitral inflow E/A ratio < 0.8 with a peak E velocity of > 50 cm/s, (2) tissue Doppler lateral mitral annulus e′ <10 cm/s, and (3) E/e′ ratio > 14. Multiple logistic regressions with adjustments for age, physical fitness, and pulse rate were conducted to determine the association of cardiometabolic and ECG markers with LVDD. The prevalence of LVDD was estimated to be 4.16% (N = 84). Of the cardiometabolic markers, central obesity, defined as waist circumference ≥ 90 cm, was the only independent marker of LVDD [odds ratio (OR) and 95% confidence interval: 2.97 (1.63–5.41)]. There were no association for hypertension, prediabetes, and dyslipidemia. Of the ECG markers, left atrial enlargement and incomplete right bundle branch block/intraventricular conduction delay were the independent ECG markers of LVDD [OR: 2.98 (1.28–6.94) and 1.94 (1.09–3.47), respectively]. There was borderline association for Cornell-based left ventricular hypertrophy and inferior T wave inversion [OR: 1.94 (0.97–3.63) and 2.44 (0.98–6.08), respectively].

Conclusion

In the physically active Asian young male adults, central obesity and some ECG markers for left heart abnormalities were useful to identify LVDD.

Machine Learning for Electrocardiographic Features to Identify Left Atrial Enlargement in Young Adults: CHIEF Heart Study

Background

Left atrial enlargement (LAE) is associated with cardiovascular events. Machine learning for ECG parameters to predict LAE has been performed in middle- and old-aged individuals but has not been performed in young adults.

Methods

In a sample of 2,206 male adults aged 17–43 years, three machine learning classifiers, multilayer perceptron (MLP), logistic regression (LR), and support vector machine (SVM) for 26 ECG features with or without 6 biological features (age, body height, body weight, waist circumference, and systolic and diastolic blood pressure) were compared with the P wave duration of lead II, the traditional ECG criterion for LAE. The definition of LAE is based on an echocardiographic left atrial dimension > 4 cm in the parasternal long axis window.

Results

The greatest area under the receiver operating characteristic curve is present in machine learning of the SVM for ECG only (77.87%) and of the MLP for all biological and ECG features (81.01%), both of which are superior to the P wave duration (62.19%). If the sensitivity is fixed to 70–75%, the specificity of the SVM for ECG only is up to 72.4%, and that of the MLP for all biological and ECG features is increased to 81.1%, both of which are higher than 48.8% by the P wave duration.

Conclusions

This study suggests that machine learning is a reliable method for ECG and biological features to predict LAE in young adults. The proposed MLP, LR, and SVM methods provide early detection of LAE in young adults and are helpful to take preventive action on cardiovascular diseases.