Electrocardiographic prediction of left ventricular hypertrophy in women and men with left bundle branch block - Comparison of QRS duration, amplitude and voltage-time-integral

A novel ECG based tool for diagnosing pericardial effusion: A case-control study

BACKGROUND

Current electrocardiographic (ECG) criteria for diagnosing pericardial effusion are limited by low sensitivity. This study aimed to evaluate traditional ECG criteria within a contemporary patient cohort and to compare the diagnostic accuracy of a novel ECG-based score, the ARENA score, with conventional low-voltage criteria for the detection of pericardial effusion.

METHODS

A retrospective case-control study was conducted at a university hospital, including consecutive patients who underwent both echocardiography and ECG, regardless of admission diagnosis. Patients were divided into derivation and validation cohorts, each comprising individuals with and without pericardial effusion (≥1.0 cm). ECGs were analyzed using traditional low-voltage criteria and the ARENA score, with sensitivity, specificity, accuracy and likelihood ratios calculated for both.

RESULTS

A total of 244 patients were included, with 104 presenting with pericardial effusion and 140 without. These patients were divided into a derivation cohort (n = 100) and a validation cohort (n = 144). In the validation cohort, sensitivity was 5.6 % (95 % CI: 0.0 %-11.7 %) with traditional criteria and 51.9 % (95 % CI: 38.5 %-65.3 %) with the ARENA score (p < 0.001). Specificity was 92.2 % (95 % CI: 86.7 %-97.8 %) for traditional criteria and 82.2 % (95 % CI: 74.3 %-90.1 %) for the ARENA score (p = 0.047). Accuracy in the validation cohort was 59.7 % (95 % CI: 51.7 %-67.7 %) for traditional criteria and 70.8 % (95 % CI: 63.4 %-78.2 %) for the ARENA score (p = 0.048).

CONCLUSIONS

The ARENA score demonstrates higher sensitivity and accuracy in detecting pericardial effusion compared to traditional low-voltage ECG criteria, though with a modest reduction in specificity.

QRS 3D Voltage-Time Integral in Narrow QRS Complex - Establishing the Normal Reference Range

Background

Vectorcardiographic 3D QRS voltage-time integral (VTI QRS-3D ) is a novel marker of ventricular dyssynchrony pertinent for cardiac resynchronization therapy. It may have additional clinical utility but its normal reference ranges have not been established. We sought to define reference ranges for VTI QRS-3D in healthy individuals.

Methods

We retrospectively analyzed 12-lead ECGs of healthy adults (2010-2014) and compared them to patients with cardiomyopathy with reduced ejection fraction (EF) <50%. Using the Kors matrix, 12-lead ECGs with QRS duration ≤120 ms were converted to vectorcardiographic X, Y, and Z leads. VTI QRS-3D was calculated as the instantaneous root-mean-square (3D) voltage integrated over the QRS duration. Reference range limits were defined as the 2.5th to 97.5th percentiles respectively for healthy females and males in age groups 18-34, 35-54 and ≥55 years.

Results

The study included 468 healthy adults (age 44.6 ± 17.0 years; 63.9% female) and 314 patients with cardiomyopathy (age 62.1 ± 14.0 years; 34.4% female). VTI QRS-3D was significantly larger in the cardiomyopathy patients compared to the healthy population (48.2±21.4 vs. 38.1±9.3 µVs, p<0.0001). Increased age and female sex were significant predictors of lower VTI QRS-3D in the healthy population (both p<0.0001). VTI QRS-3D reference ranges for respective age groups for healthy females were 23.2-55.0, 23.9-56.4 and 19.6-50.9 µVs, and for healthy males were 29.9-57.2, 28.2-56.7 and 21.4-55.9 µVs.

Conclusion

VTI QRS-3D is higher at younger age in healthy population, male sex and in patients having cardiomyopathy with reduced EF. Age and sex need to be accounted for using VTI QRS-3D as a marker for structural heart disease.

Evaluation of electrocardiographic criteria for predicting left ventricular hypertrophy and dilation in presence of left bundle branch block

BACKGROUND

The utility of standard published electrocardiographic (ECG) criteria for left ventricular hypertrophy (LVH) in patients with left bundle branch block (LBBB) is not established. We have previously shown that in ECGs demonstrating LBBB, QRS duration outperforms vectorcardiographic X, Y, Z lead and root-mean-squared (3D) amplitudes and voltage-time-integrals in diagnosing LVH and dilation. We sought to evaluate diagnostic yields of published LVH criteria versus QRS duration for ECG based diagnosis of LVH and dilation in presence of LBBB.

METHODS

We included adult patients with typical LBBB having ECG and transthoracic echocardiogram performed within 3 months of each other in 2010-2020. We obtained area under receiver-operator characteristic curve (AUC) for QRS duration and each of the published ECG LVH criteria to predict increased LV mass indexed (↑LVMi, women >95 g/m2, men >115 g/m2) and LV end diastolic volume indexed (↑LVEDVi, women >61 mL/m2, men >74 mL/m2).

RESULTS

Among 413 adults (53 % women, age 73 ± 12 yr) with LBBB, the traditional LVH criteria performed poorly to detect ↑LVMi or ↑LVEDVi. Cornell voltage-duration product had the highest AUCs (↑LVMi 0.634, ↑LVEDVi 0.580). QRS duration had a higher AUC for diagnosing ↑LVMi (women 0.657, men 0.703) and ↑LVEDVi (women 0.668, men 0.699) compared to any other criteria.

CONCLUSIONS

In patients with LBBB, prolonged QRS duration (women ≥150 ms, men ≥160 ms) is a superior predictor of LVH and dilation than traditional ECG-based LVH criteria.

Machine learning to classify left ventricular hypertrophy using ECG feature extraction by variational autoencoder

Background

Traditional ECG criteria for left ventricular hypertrophy (LVH) have low diagnostic yield. Machine learning (ML) can improve ECG classification.

Methods

ECG summary features (rate, intervals, axis), R-wave, S-wave and overall-QRS amplitudes, and QRS/QRST voltage-time integrals (VTIs) were extracted from 12-lead, vectorcardiographic X-Y-Z-lead, and root-mean-square (3D) representative-beat ECGs. Latent features were extracted by variational autoencoder from X-Y-Z and 3D representative-beat ECGs. Logistic regression, random forest, light gradient boosted machine (LGBM), residual network (ResNet) and multilayer perceptron network (MLP) models using ECG features and sex, and a convolutional neural network (CNN) using ECG signals, were trained to predict LVH (left ventricular mass indexed in women >95 g/m², men >115 g/m²) on 225,333 adult ECG-echocardiogram (within 45 days) pairs. AUROCs for LVH classification were obtained in a separate test set for individual ECG variables, traditional criteria and ML models.

Results

In the test set (n=25,263), AUROC for LVH classification was higher for ML models using ECG features (LGBM 0.790, MLP 0.789, ResNet 0.788) as compared to the best individual variable (VTIQRS-3D 0.677), the best traditional criterion (Cornell voltage-duration product 0.647) and CNN using ECG signal (0.767). Among patients without LVH who had a follow-up echocardiogram >1 (closest to 5) years later, LGBM false positives, compared to true negatives, had a 2.63 (95% CI 2.01, 3.45)-fold higher risk for developing LVH (p<0.0001).

Conclusions

ML models are superior to traditional ECG criteria to classify-and predict future-LVH. Models trained on extracted ECG features, including variational autoencoder latent variables, outperformed CNN directly trained on ECG signal.

Different methods of 3D QRS area calculation from vectorcardiographic X, Y, and Z Leads

3DQRSarea is a strong marker for cardiac resynchronization therapy and can be obtained by taking the (i) summation or the (ii) difference of the areas subtended by positive and negative deflections in X, Y, Z vectorcardiographic electrocardiogram (ECG) leads. We correlated both methods with the instantaneous-absolute-3D-voltage-time-integral (VTIQRS-3D). 3DQRSarea consistently underestimated the VTIQRS -3D, but the summation method was a closer and more reliable approximation. The dissimilarity was less apparent in left bundle branch block (r2 summation .996 vs. difference .972) and biventricular paced ECGs (r2 .996 vs. .957) but was more apparent in normal ECGs (r2 .988 vs. .653).

Artificial Intelligence Interpretation of the Electrocardiogram: A State-of-the-Art Review

PURPOSE OF REVIEW

Artificial intelligence (AI) is transforming electrocardiography (ECG) interpretation. AI diagnostics can reach beyond human capabilities, facilitate automated access to nuanced ECG interpretation, and expand the scope of cardiovascular screening in the population. AI can be applied to the standard 12-lead resting ECG and single-lead ECGs in external monitors, implantable devices, and direct-to-consumer smart devices. We summarize the current state of the literature on AI-ECG.

RECENT FINDINGS

Rhythm classification was the first application of AI-ECG. Subsequently, AI-ECG models have been developed for screening structural heart disease including hypertrophic cardiomyopathy, cardiac amyloidosis, aortic stenosis, pulmonary hypertension, and left ventricular systolic dysfunction. Further, AI models can predict future events like development of systolic heart failure and atrial fibrillation. AI-ECG exhibits potential in acute cardiac events and non-cardiac applications, including acute pulmonary embolism, electrolyte abnormalities, monitoring drugs therapy, sleep apnea, and predicting all-cause mortality. Many AI models in the domain of cardiac monitors and smart watches have received Food and Drug Administration (FDA) clearance for rhythm classification, while others for identification of cardiac amyloidosis, pulmonary hypertension and left ventricular dysfunction have received breakthrough device designation. As AI-ECG models continue to be developed, in addition to regulatory oversight and monetization challenges, thoughtful clinical implementation to streamline workflows, avoiding information overload and overwhelming of healthcare systems with false positive results is necessary. Research to demonstrate and validate improvement in healthcare efficiency and improved patient outcomes would be required before widespread adoption of any AI-ECG model.

Electrocardiographic Z-axis QRS-T voltage-time-integral in patients with typical right bundle branch block - Correlation with echocardiographic right ventricular size and function

BACKGROUND

Right bundle branch block (RBBB) can be benign or associated with right ventricular (RV) functional and structural abnormalities. Our aim was to evaluate QRS-T voltage-time-integral (VTI) compared to QRS duration and lead V1 R' as markers for RV abnormalities.

METHODS

We included adults with an ECG demonstrating RBBB and echocardiogram obtained within 3 months of each other, between 2010 and 2020. VTIQRS and VTIQRST were obtained for 12 standard ECG leads, reconstructed vectorcardiographic X, Y, Z leads and root-mean-squared (3D) ECG. Age, sex and BSA-adjusted linear regressions were used to assess associations of QRS duration, amplitudes, VTIs and lead V1 R' duration/VTI with echocardiographic tricuspid annular plane systolic excursion (TAPSE), RV tissue Doppler imaging S', basal and mid diameter, and systolic pressure (RVSP).

RESULTS

Among 782 patients (33% women, age 71 ± 14 years) with RBBB, R' duration in lead V1 was modestly associated with RV S', RV diameters and RVSP (all p ≤ 0.03). QRS duration was more strongly associated with RV diameters (both p < 0.0001). AmplitudeQRS-Z was modestly correlated with all 5 RV echocardiographic variables (all p ≤ 0.02). VTIR'-V1 was more strongly associated with TAPSE, RV S' and RVSP (all p ≤ 0.0003). VTIQRS-Z and VTIQRST-Z were among the strongest correlates of the 5 RV variables (all p < 0.0001). VTIQRST-Z.√BSA cutoff of ≥62 μVsm had sensitivity 62.7% and specificity 65.7% for predicting ≥3 of 5 abnormal RV variables (AUC 0.66; men 0.71, women 0.60).

CONCLUSION

In patients with RBBB, VTIQRST-Z is a stronger predictor of RV dysfunction and adverse remodeling than QRS duration and lead V1 R'.