Standard 12-lead Electrocardiogram Research Articles

Deep neural networks detect regional wall motion abnormalities and preclinical cardiovascular disease from 12-lead ECGs.

Identifying regional wall motion abnormalities (RWMAs) is critical for diagnosing and risk stratifying patients with cardiovascular disease, particularly ischemic heart disease. We hypothesized that a deep neural network could accurately identify patients with regional wall motion abnormalities from a readily available standard 12-lead electrocardiogram (ECG). This observational, retrospective study included patients who were treated at Beth Israel Deaconess Medical Center and had an ECG and echocardiogram performed within 14 days of each other between 2008 and 2019. We trained a convolutional neural network to detect the presence of RWMAs, qualitative global right ventricular (RV) hypokinesis, and varying degrees of left ventricular dysfunction (left ventricular ejection fraction [LVEF] ≤50%, LVEF ≤40%, and LVEF ≤35%) identified by echocardiography, using ECG data alone. Patients were randomly split into development (80%) and test sets (20%). Model performance was assessed using area under the receiver operating characteristic curve (AUC). Cox proportional hazard models adjusted for age and sex were performed to estimate the risk of future acute coronary events. The development set consisted of 19,837 patients (mean age 66.7±16.4; 46.7% female) and the test set comprised of 4,953 patients (mean age 67.5±15.8 years; 46.5% female). On the test dataset, the model accurately identified the presence of RWMA, RV hypokinesis, LVEF ≤50%, LVEF ≤40%, and LVEF ≤35% with AUCs of 0.87 (95% CI 0.858-0.882), 0.888 (95% CI 0.878-0.899), 0.923 (95% CI 0.914-0.933), 0.93 (95% CI 0.921-0.939), and 0.876 (95% CI 0.858-0.896), respectively. Among patients with normal biventricular function at the time of the index ECG, those classified as having RMWA by the model were 3 times the risk (age- and sex-adjusted hazard ratio, 2.8; 95% CI 1.9-3.9) for future acute coronary events compared to those classified as negative. We demonstrate that a deep neural network can help identify regional wall motion abnormalities and reduced LV function from a 12-lead ECG and could potentially be used as a screening tool for triaging patients who need either initial or repeat echocardiographic imaging.

An antisense lncRNA mediated mechanism controls desmoplakin gene expression

Abstract Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): Department of Innovation, Research and University of the Autonomous Province of Bolzano-South Tyrol (Italy). Background Desmosomal genes are established as definitive genes for Arrhythmogenic Cardiomyopathy (ACM). The desmoplakin encoding gene, DSP, is amongst the most frequently mutated genes in ACM patients. While rare genetic variants within desmosomal genes can cause ACM, common variants might also affect cardiac conduction traits in the general population. Purpose 1) To explore the association between common genetic variants in ACM desmosomal genes (DSC2, DSG2, PKP2, JUP, DSP) and cardiac conduction traits. 2) To investigate the effect of a lncRNA, identified from downstream analyses, on DSP gene expression. Methods P-wave, PR, QRS and QT intervals were measured using standard 12-lead electrocardiograms (ECG) in 4295 participants to the Cooperative Health Research in South Tyrol (CHRIS) study (data release 2). Samples were genotyped using the Illumina HumanOmniExpressExome array and imputed on the 1000 Genome Phase 1 panel. We conducted a genetic association study on 3458 variants within the linkage disequilibrium blocks that encompass desmosomal genes, assuming an additive genetic model, at 1.9x10-4 statistical significance. Mendelian Randomization (MR) analyses integrating ECG traits and Genotype-Tissue Expression (GTEx) database transcription levels were performed. A specific GapmeR was used to downregulate the antisense RP3-512B11.3 lncRNA expression in human induced pluripotent stem cell derived cardiomyocytes (hiPSC-CMs). hiPSC-CMs lysates were analyzed by digital droplet PCR and Wes™ immunoassay system. Results The rs2744389 variant in DSP gene is associated with a shorter QRS interval (-1.10 ms per effect allele copy; p=3.5x10-6). In the GTEx database, rs2744389 is identified as an expression quantitative trait locus for the antisense lncRNA. MR results support a causal effect of the lncRNA on QRS. To assess the hypothesis that the lncRNA regulates DSP gene expression, we identified a Gapmer able to downregulate the lncRNA expression by approximately 60% in hiPSC-CMs. The lncRNA downregulation led to a significant increase of DSP expression, both at mRNA and protein level, compared to hiPSC-CMs treated with a negative control Gapmer (DSP/RPP30: 16.39±3.32 vs 11.84±1.79; p=0.0313), (Desmoplakin/total protein: 2.44±0.29 vs 1.66±0.11; p=0.0156). Conclusions Starting with the identification of an association between a common genetic variant in DSP gene and a significantly shorter QRS interval in a general population sample, we further unrevealed a new mechanism regulating DSP expression through an antisense lncRNA. While requiring further evidence, these findings might pave the way to a potentially novel therapeutic approach for DSP mutations carriers characterized by desmoplakin protein deficit, as for example patients affected by ACM, but also by dilated cardiomyopathy, cardiocutaneous diseases such as the Carvajal syndrome, and some cancer conditions.

Mobile ECG devices in pediatric cardiology: efficacy and accuracy of smart watch and hand-held device in healthy children

Abstract Background and Aim Pediatric arrhythmias require accurate electrocardiogram (ECG) assessment for diagnosis and management. Mobile ECG devices, such as smart watch based and hand-held devices, offer potential advantages in terms of portability and accessibility. This study aimed to compare the efficancy and diagnostic accuracy of ECG tracings obtained using a smart watch and a hand-held device with standard 12-lead ECG tracings in healthy children. Methods ECG data were acquired from healthy pediatric subjects (n=30) aged 4 to 18 years using the smart watch, hand-held device, and standard 12-lead ECG. Single leads were acquired sequentially. The time required for ECG acquisition was recorded for each device. Subsequently, ECG leads were analyzed for voltage, duration, and morphology of waves and intervals, as well as ST segments. Qualitative assessment of leads included morphology comparison of waves and complexes to standard 12-lead ECG and using a physician based quality rating scale. Assessment was performed by two blinded pediatric electrophysiologists. Results We found significant differences in the time required to acquire ECG tracings between the smart watch, standard 12-lead ECG, and hand-held device, with the smart watch being notably slower (median 15 min, 4 min, 1 min, respectively). ECG leads obtained with the smart watch and hand-held device demonstrated comparability to standard 12-lead ECG tracings in terms of quantitative key parameters, including voltage (mV) and duration (ms) of waves and intervals. Qualitative assessment revealed good correlation of the hand-held device with standard 12-lead ECG, while the smart watch performed less well. Conclusions This study suggests that mobile ECG devices offer a diagnostic accuracy comparable to that of standard 12-lead ECGs in healthy children. While the smart watch may be slower in ECG acquisition, acquisition of all 12 standard ECG leads is feasible in this population. The potential of both included mobile devices for pediatric usage in specific clinical scenarios, such as remote monitoring or emergencies, is promising.Graphical Abstract

Time-dependent activation and repolarization changes following left bundle branch area pacing versus conventional biventricular pacing in patients with dyssynchronous heart failure

Abstract Background Cardiac Resynchronization Therapy (CRT) is the standard treatment for patients with dyssynchronous heart failure. Despite the effectiveness of conventional biventricular pacing (BiVP), 6-16% of patients develop ventricular tachyarrhythmias (VTAs) within a year of implantation. The relationship between BiVP and the development of VTAs is still under debate. During BiVP, the left-sided epicardium-to-endocardium activation creates a non-physiological activation and hence a non-physiological repolarization, which might trigger VTAs. A novel alternative treatment option to BiVP that on the contrary induces a left ventricular endocardium-to-epicardium activation, is left bundle branch area pacing (LBBAP). We hypothesized that LBBAP leads to a more homogenous repolarization than conventional BiVP. Purpose To compare time-dependent activation and repolarization changes in patients with dyssynchronous heart failure receiving LBBAP versus BiVP. Methods Patients eligible for CRT who underwent LBBAP (n=21) or BiVP (n=94) were retrospectively included. Standard 12-lead electrocardiograms obtained prior to implantation [baseline], shortly following implantation (0-2 days, [acute phase]), and after prolonged pacing during follow-up (31-180 days, [chronic phase]) were analyzed. QRS duration and QRS area were calculated through reconstruction of the vectorcardiogram and further analyzed together with corrected QT (QTc) and corrected Tpeak-Tend (Tp-e,c) intervals using a semi-automatic approach in MATLAB. Differences between LBBAP and BiVP patients were determined using the Mann-Whitney U-test. Time-dependent changes within groups were determined using the Wilcoxon signed-rank tests. Significance was defined as a p-value ≤0.05. Results (see figure): Compared to baseline, QRS duration and QRS area decreased significantly during the acute phase (both p < 0.01) and stabilized during chronic pacing for both BiVP and LBBAP patients. Differences in QRS duration or QRS area between LBBAP and BiVP were not significant at any of the time points. In the acute phase, BiVP resulted in a temporary and significant prolongation of both QTc and Tp-e,c compared to baseline (both p < 0.01). However, there was a subsequent shortening observed between the acute and chronic phases for both QTc and Tp-e,c (both p < 0.01). Compared to baseline, LBBAP acutely shortened Tp-e,c (p = 0.02) and tended to shorten QTc. During chronic pacing, LBBAP shortened QTc and Tp-e,c compared to baseline (p = 0.01 and p < 0.01). The decreases in QTc and Tp-e,c were larger in LBBAP than in BiVP patients in the acute and chronic pacing phase (all p <0.01). Conclusion CRT induces activation changes in both BiVP and LBBAP patients to a comparable extent. While BiVP leads to an acute increase in repolarization dispersions, LBBAP reduces markers of repolarization dispersion compared to baseline. The diminished dispersion of repolarization might contribute to an antiarrhythmic effect with LBBAP.

Comparison of electrocardiogram (ECG) waveforms and centralized ECG measurements between a wearable mobile ECG device and a standard 12-lead ECG

Comparison of electrocardiogram (ECG) waveforms and centralized ECG measurements between a wearable mobile ECG device and a standard 12-lead ECG

The Danish Nationwide Electrocardiogram (ECG) Cohort

The electrocardiogram (ECG) is a non-invasive diagnostic tool holding significant clinical importance in the diagnosis and risk stratification of cardiac disease. However, access to large-scale, population-based digital ECG data for research purposes remains limited and challenging. Consequently, we established the Danish Nationwide ECG Cohort to provide data from standard 12-lead digital ECGs in both pre- and in-hospital settings, which can be linked to comprehensive Danish nationwide administrative registers on health and social data with long-term follow-up. The Danish Nationwide ECG Cohort is an open real-world cohort including all patients with at least one digital pre- or in-hospital ECG in Denmark from January 01, 2000, to December 31, 2021. The cohort includes data on standardized and uniform ECG diagnostic statements and ECG measurements including global parameters as well as lead-specific measures of waveform amplitudes, durations, and intervals. Currently, the cohort comprises 2,485,987 unique patients with a median age at the first ECG of 57 years (25th–75th percentiles, 40–71 years; males, 48%), resulting in a total of 11,952,430 ECGs. In conclusion, the Danish Nationwide ECG Cohort represents a novel and extensive population-based digital ECG dataset for cardiovascular research, encompassing both pre- and in-hospital settings. The cohort contains ECG diagnostic statements and ECG measurements that can be linked to various nationwide health and social registers without loss to follow-up.

Convolutional neural network (CNN)-enabled electrocardiogram (ECG) analysis: a comparison between standard twelve-lead and single-lead setups.

Artificial intelligence (AI) has shown promise in the early detection of various cardiac conditions from a standard 12-lead electrocardiogram (ECG). However, the ability of AI to identify abnormalities from single-lead recordings across a range of pathological conditions remains to be systematically investigated. This study aims to assess the performance of a convolutional neural network (CNN) using a single-lead (D1) rather than a standard 12-lead setup for accurate identification of ECG abnormalities. We designed and trained a lightweight CNN to identify 20 different cardiac abnormalities on ECGs, using data from the PTB-XL dataset. With a relatively simple architecture, the network was designed to accommodate different combinations of leads as input (<100,000 learnable parameters). We compared various lead setups such as the standard 12-lead, D1 alone, and D1 paired with an additional lead. This has been corrected to “The CNN based on single-lead ECG (D1) achieved satisfactory performance compared to the standard 12-lead framework (average percentage AUC difference: −8.7%). Notably, for certain diagnostic classes, there was no difference in the diagnostic AUC between the single-lead and the standard 12-lead setups. When a second lead was detected in the CNN in addition to D1, the AUC gap was further reduced to an average percentage difference of -2.8% compared with that of the standard 12-lead setup. A relatively lightweight CNN can predict different classes of cardiac abnormalities from D1 alone and the standard 12-lead ECG. Considering the growing availability of wearable devices capable of recording a D1-like single-lead ECG, we discuss how our findings contribute to the foundation of a large-scale screening of cardiac abnormalities.

Prevalence of electrocardiographic abnormalities among Iranian children and adolescents and associations with blood pressure and obesity: findings from the SHED LIGHT study.

There are few studies for detecting rhythm abnormalities among healthy children and adolescents. The aim of the study was to investigate the prevalence of abnormal electrocardiographic findings in the young Iranian population and its association with blood pressure and obesity. A total of 15084 children and adolescents were examined in a randomly selected population of Tehran city, Iran, between October 2017 and December 2018. Anthropometric values and blood pressure measurements were also assessed. A standard 12-lead electrocardiogram was recorded by a unique recorder, and those were examined by electrophysiologists. All students mean age was 12.3 ± 3.1 years (6-18 years), and 52% were boys. A total of 2900 students (192.2/1000 persons; 95% confidence interval 186-198.6) had electrocardiographic abnormalities. The rate of electrocardiographic abnormalities was higher in boys than girls (p < 0.001). Electrocardiographic abnormalities were significantly higher in thin than obese students (p < 0.001), and there was a trend towards hypertensive individuals to have more electrocardiographic abnormalities compared to normotensive individuals (p = 0.063). Based on the multivariable analysis, individuals with electrocardiographic abnormalities were less likely to be girls (odds ratio 0.745, 95% confidence interval 0.682-0.814) and had a lower body mass index (odds ratio 0.961, 95% confidence interval 0.944-0.979). In this large-scale study, there was a high prevalence of electrocardiographic abnormalities among young population. In addition, electrocardiographic findings were significantly influenced by increasing age, sex, obesity, and blood pressure levels. This community-based study revealed the implications of electrocardiographic screening to improve the care delivery by early detection.

Downsloping TP Segment in the Precordial Leads on a Standard 12-lead Electrocardiogram: Suspected Cardiac Impulse-tapping Artifact.

We herein report an 80-year-old man showing a downsloping TP segment together with an increase in the height of the T wave in the precordial leads on a standard 12-lead electrocardiogram (ECG). Separately, an 87-year-old woman showed only a downsloping TP segment in the precordial leads on a standard 12-lead ECG. Neither patient reported chest pain or dyspnea when ECGs was obtained. This downsloping TP segment in the precordial leads on the standard 12-lead ECG is thought to be due to a cardiac impulse-tapping artifact. Differential diagnoses are also discussed.

Correlation between mitral valve area and left atrial function in rheumatic mitral valve stenosis patients.

Rheumatic heart disease (RHD) continues to be one of the leading causes of cardiovascular morbidity and mortality. The mitral valve frequently develops mitral stenosis (MS), and it is the most prevalent valve lesion in patients with chronic RHD. Left atrial (LA) functional impairment is associated with rheumatic MS. The aim of this study was to evaluate the association between LA function and mitral valve area (MVA) in rheumatic MS patients, and to assess the echocardiographic parameters in sinus rhythm and atrial fibrillation (AF) patients. This was a cross-sectional, descriptive study that involved patients with rheumatic MS. Patients underwent a standard 12-lead electrocardiogram and echocardiographic examination. MVA was assessed and correlated with LA function. Comparison was made between sinus rhythm and AF patients. Eighty-one patients with rheumatic MS were included in this study, with 71.6% of them having associated MR. MVA showed a statistically highly significant positive correlation with LA and right ventricular (RV) function, and a statistically significant/highly significant negative correlation with their dimensions. A higher percentage of patients with severe MS was in AF (58.1%). There was a positive correlation between LA function and MVA in rheumatic MS patients. AF was related to the severity of MS.

The Reliability of the Apple Watch's Electrocardiogram.

Background An electrocardiogram (ECG) is a standard tool used to detectvarious cardiovascular abnormalities. Detection sensitivity for atrial fibrillation (AF) was recently shown to be greatly increased by using short, intermittent ECG recordings.Modern mobile ECG recording devices that can monitor patients' heart activities around the clock have made this a reality.The Apple Watch is one of these portable ECG devices that can detect heart rhythms and is approved by the American FDA for screening and detecting AF. Objectives We compared the results of the Apple Watch I lead ECG with conventional ECG results to assess the sensitivity and specificity of the Apple Watch I lead ECG. We then determined the abnormalities that can be detected by the Apple Watch I lead ECG. Methods This study was conducted on outpatient cardiac clinics at King Abdullah bin Abdulaziz University Hospital (KAAUH), and Prince Sultan Cardiac Center (PSCC),from May to October 2021. A standard 12-lead ECG was recorded and compared with the Apple Watch I lead ECG. A total of 469 ECG comparisons were included in this study and evaluated by two investigators. The data on patient demographics, medical and medication history, and ECG data were reviewed and analyzed using IBMSPSSStatistics for Windows, Version 23 (Released 2015; IBM Corp., Armonk, New York, United States). Results No significant differences were seen between the Apple Watch and the 12-lead ECG in terms of the studied ECG characteristics. A significant and strong positive correlation between the heart rate measurements in the 12-lead ECG and Apple Watch ECG was documented. The most commonly found abnormalities in the Apple Watch ECG wereAF in 37 (7.9%), followed by first-degree atrioventricular (AV) block in 32 (6.8%). The sensitivity of Apple Watch's automated interpretation to detect an AF was 99.54%, while the manual interpretation yielded a sensitivity of 100%. Conclusion The results of this study demonstrated a robust relationship between the 12-lead ECG and Apple Watch ECG in the diagnosis of arrhythmias. Consequently, cardiac patients may consider the Apple Watch ECG a trustworthy remote monitoring technique.

Diagnostic Accuracy of Single-Lead Electrocardiograms Using the Kardia Mobile App and the Apple Watch 4: Validation Study.

To date, the 12-lead electrocardiogram (ECG) is the gold standard for cardiological diagnosis in clinical settings. With the advancements in technology, a growing number of smartphone apps and gadgets for recording, visualizing, and evaluating physical performance as well as health data is available. Although this new smart technology is innovative and time- and cost-efficient, less is known about its diagnostic accuracy and reliability. This study aimed to examine the agreement between the mobile single-lead ECG measurements of the Kardia Mobile App and the Apple Watch 4 compared to the 12-lead gold standard ECG in healthy adults under laboratory conditions. Furthermore, it assessed whether the measurement error of the devices increases with an increasing heart rate. This study was designed as a prospective quasi-experimental 1-sample measurement, in which no randomization of the sampling was carried out. In total, ECGs at rest from 81 participants (average age 24.89, SD8.58 years; n=58, 72% male) were recorded and statistically analyzed. Bland-Altman plots were created to graphically illustrate measurement differences. To analyze the agreement between the single-lead ECGs and the 12-lead ECG, Pearson correlation coefficient (r) and Lin concordance correlation coefficient (CCCLin) were calculated. The results showed a higher agreement for the Apple Watch (mean deviation QT: 6.85%; QT interval corrected for heart rate using Fridericia formula [QTcF]: 7.43%) than Kardia Mobile (mean deviation QT: 9.53%; QTcF: 9.78%) even if both tend to underestimate QT and QTcF intervals. For Kardia Mobile, the QT and QTcF intervals correlated significantly with the gold standard (rQT=0.857 and rQTcF=0.727; P<.001). CCCLin corresponded to an almost complete heuristic agreement for the QT interval (0.835), whereas the QTcF interval was in the range of strong agreement (0.682). Further, for the Apple Watch, Pearson correlations were highly significant and in the range of a large effect (rQT=0.793 and rQTcF=0.649; P<.001). CCCLin corresponded to a strong heuristic agreement for both the QT (0.779) and QTcF (0.615) intervals. A small negative correlation between the measurement error and increasing heart rate could be found of each the devices and the reference. Smart technology seems to be a promising and reliable approach for nonclinical health monitoring. Further research is needed to broaden the evidence regarding its validity and usability in different target groups.

Validation of an artificial intelligence model for 12-lead ECG interpretation

Abstract Background The electrocardiogram (ECG) is one of the most accessible and comprehensive diagnostic tools to assess cardiac abnormalities. However, automated ECG interpretation remains inferior to physician interpretation in terms of accuracy and reliability. Purpose This study evaluated the accuracy of an AI-powered ECG model in providing a precise diagnosis of 12-lead ECGs and compared its diagnostic performance to primary care physicians and cardiologists through extensive benchmarking. Methods A deep neural network (DNN) was trained on standard 12-lead ECGs to detect 38 diagnoses (grouped into 6 categories: rhythm, conduction abnormalities, chamber enlargement, infarction, ectopy, and axis), denoting the most common types of electrocardiographic abnormalities. Performance of AI-powered ECG diagnosis was evaluated on an independent test set annotated by consensus of two expert cardiologists. Benchmarking was performed against three individual primary care physicians and six individual cardiologists who independently annotated the same ECG test set. The key metrics used to compare performances were positive predictive value (PPV), negative predictive value (NPV), Sensitivity, Specificity, and F1 score. Results A total of 931,344 standard 12-lead ECGs from 172,750 patients were used to train a DNN. The independent test set had 11,932 annotated ECG labels. The model attained an overall mean F1 score of 0.921, sensitivity 0.910 (0.889–0.931), specificity 0.968 (0.954–0.981), PPV 0.939 (0.919–0.958), and NPV 0.965 (0.951–0.979) [Figure 1]. In all 6 diagnostic categories, the DNN achieved higher mean F1 scores than the mean cardiologist and primary care physician (Rhythm 0.951 vs. 0.892 vs. 0.734; Conduction abnormalities 0.883 vs. 0.824 vs. 0.693; Chamber enlargement 0.970 vs. 0.761 vs. 0.562; Infarction 0.918 vs. 0.853 vs. 0.781; Ectopy 0.966 vs. 0.951 vs. 0.897; Axis 0.909 vs. 0.644 vs. 0.528, respectively). The ability of the DNN to identify atrial fibrillation achieved nearly perfect performance (PPV of 0.989 and NPV of 0.990). Diagnostic performance surpassed primary care physicians and was non-inferior to cardiologists based on the F1 scores for all individual diagnoses. Conclusions Our results demonstrate the AI-powered ECG model’s ability to accurately identify electrocardiographic abnormalities from the 12-lead ECG, showcasing its utility as clinical tool for healthcare professionals.

ECG-based deep learning for detecting epicardial coronary occlusion in acute myocardial infarction

Abstract Background One third of Non-ST-elevation myocardial infarction (NSTEMI) patients have occlusion myocardial infarction (OMI) associated with poor short- and long-term outcomes due to delayed invasive management. Purpose We sought to develop and externally validate a versatile artificial intelligence (AI)-model detecting OMI on single standard 12-lead electrocardiograms (ECGs) and compare its performance to existing state-of-the-art criteria. Methods An AI model was developed using 18,616 ECGs from 10,692 unique contacts (22.9% OMI) of 10,543 patients (age 66±14 years, 65.9% males) with acute coronary syndrome (ACS) originating from an international database and a tertiary care center. This AI model was tested on a holdout set of 3,254 ECGs from 2,263 unique contacts (20% OMI) of 2,222 patients (age 62±14 years, 67% males) and compared with STEMI criteria and annotations of ECG experts in detecting OMI on 12-lead ECGs using sensitivity, specificity, and predictive values. Results The OMI AI model achieved an AUROC of 0.941 (95% CI: 0.926, 0.954) in identifying the primary outcome of OMI on 12-lead ECGs in the holdout set [Figure 1A], yielding robust performance across genders and age subgroups (ranging from 0.907 to 0.951 AUROC) [Figure 1B]. Sensitivity in detecting OMI was significantly higher for OMI AI model compared to STEMI criteria (82.6% [95% CI: 78.9%, 86.1%] vs. 34.4% [95% CI: 30.0%, 38.8%]) and statistically equal compared to ECG experts 75.9% [95% CI: 71.9%, 80%] in identifying OMI [Table 1]. The OMI AI Model showed statistically superior performance compared to STEMI criteria and equal (non-inferior) performance to ECG experts when evaluated using industry-standard metrics. Conclusions This sizeable validation of an AI ECG model demonstrates the ability of AI to detect acute OMI and suggests a potential in improving ACS patient triage and need for immediate revascularization.AI Model Performance on Holdout SetBenchmark Comparison

Abstract 14207: Mobile ECG Atrial Fibrillation Prediction Enhancement Powered by Standard ECG Algorithm With Artificial Intelligence

Introduction: Artificial intelligence (AI) using electrocardiogram (ECG) enabled to predict atrial fibrillation (AF) in patients without documented AF. It needs 12-lead ECG with tremendous data learning. Mobile 1-lead ECG is more convenient to surveil cardiac rhythm with simple measurement. However, AI-enabled arrhythmia predictability by mobile ECG is limited due to single channel utilization and longer duration for arrhythmia diagnosis. Hypothesis: We aimed to enhance mobile ECG AF prediction AI algorithm integrated with 12-lead ECG deep learning model. Methods: Based on 552,372 12-lead ECG data of 318,321 patients, a statistical AF prediction model employing a deep-learning approach was constituted. The raw data of 6,792 1-lead mobile ECGs were acquired from 6,792 patients for about 1 minute at 250Hz. A statistical AF prediction model with mobile ECG employing a deep-learning approach was constituted. Resnet structure was utilized to distinguish subtle changes of the vicinity of P-wave. Both 12-lead ECG and mobile ECG were allocated to training, validation, testing datasets in a 6:2:2 ratio. Then, we conducted transfer learning using the standard 12-lead ECG’s deep learning model to improve performance of 1-lead mobile ECG deep learning model. Results: AF was annotated in 26,541 (4.8%) with 12-lead ECG whereas 1,443 (21.2%) with 1-lead mobile ECG. The area under the curve (AUC) value for predicting AF was 0.910 with 12-lead ECG, and 0.721 with mobile ECG . The predictive performance of mobile ECG was 79.7% in accuracy, 47.5% in sensitivity and 50.3% in F1-score. The AUC value of mobile ECG after applying transfer learning based on 12-lead ECG for AF prediction was increased to 0.905 with accuracy of 89.1%, sensitivity of 71.3% and F1-score of 57.7%. Conclusions: Integration with deep learning algorithm of standard 12-lead ECG significantly improved the model performance of mobile ECG AF prediction model compared to 1-lead mobile ECG only based model. Easy application of mobile ECG with enhanced AF predictability might serve a more convenient method as a pre-emptive assistive tool to provide probabilistic prediction for PAF screening rather than 12-lead ECG.

Abstract 17037: Validation of Deep Learning System for Comprehensive 12-Lead ECG Interpretation

Introduction: The electrocardiogram (ECG) is a widely available diagnostic tool for evaluating cardiac patients. Although automated ECG interpretation has made significant progress, it has yet to match the accuracy demonstrated by physicians. Hypothesis: In this study, we hypothesized that an artificial intelligence based ECG system can achieve comparable performance to physicians in accurately identifying 20 essential ECG patterns. Methods: An AI-powered system comprising six deep neural networks (DNNs) was trained to identify 20 diagnostic patterns from 12-lead ECGs categorized into six groups: rhythm, infarction, conduction abnormalities, ectopy, chamber enlargement, and axis. An independent test set with the consensus of two expert cardiologists was used as a reference standard. We compared the system's performance to that of three General Practitioners (GPs) and six individual cardiologists, using F1 scores as the evaluation metric. Results: The AI system was trained on 932,711 standard 12-lead ECGs from 173,949 patients. The independent test set comprised 11,932 annotated ECG labels. Figure 1 shows the respective F1 scores of the DNNs, average GP and average cardiologist as follows: Rhythm: 0.957 vs. 0.771 vs. 0.905; Infarction: 0.925 vs. 0.780 vs. 0.852; Conduction abnormalities: 0.893 vs. 0.714 vs. 0.851; Ectopy: 0.966 vs. 0.896 vs. 0.951; Chamber enlargement: 0.972 vs. 0.562 vs. 0.773; Axis: 0.897 vs. 0.601 vs. 0.685. The AI system's diagnostic performance exceeded that of GPs and was on par with cardiologists for all individual diagnostic patterns. Conclusions: The AI-powered ECG system is able to accurately identify electrocardiographic abnormalities from the 12-lead ECG, highlighting its potential as a clinical tool for healthcare professionals.

Abstract 17298: Deep Learning Electrocardiogram Detecting Acute Coronary Occlusion in Myocardial Infarction Presenting Without ST-Elevation

Introduction: One third of Non-ST-elevation myocardial infarction (NSTEMI) patients present with an acutely occluded culprit coronary artery (OMI), associated with poor short and long-term outcomes due to delayed identification and consequent delayed invasive management. Hypothesis: We sought to develop and validate an artificial intelligence (AI) model detecting OMI on single standard 12-lead electrocardiograms (ECGs) and compare its performance to existing standard criteria. Methods: An AI model was developed using 18,616 ECGs from 10,692 unique contacts (22.9% OMI) of 10,543 patients (age 66±14 years, 65.9% males) with acute coronary syndrome (ACS) from an international cohort. This AI model was tested on an international test set of 3,254 ECGs from 2,263 unique contacts (20% OMI) of 2,222 patients (age 62±14 years, 67% males) and compared with STEMI criteria and annotations of ECG experts in detecting OMI on 12-lead ECGs using sensitivity, specificity, predictive values and time to OMI diagnosis. OMI was based on a combination of angiographic and biomarker outcomes. Results: The AI model achieved an area under the curve (AUC) of 0.941 (95% CI: 0.926-0.954) in identifying the primary outcome of OMI, with superior performance (accuracy 90.7% [95% CI: 89.5-91.9], sensitivity 82.6% [95% CI: 78.9-86.1], specificity 92.8 [95% CI: 91.5-93.9]) compared to STEMI criteria (accuracy 84.9% [95% CI: 83.5-86.3], sensitivity 34.4% [95% CI: 30.0-38.8], specificity 97.6% [95% CI: 96.8-98.2]) and similar performance compared to ECG experts (accuracy 91.2% [95% CI: 90.0-92.4], sensitivity 75.9% [95% CI: 71.9-80.0], specificity 95.0 [95% CI: 94.0-96.0]) ( Figure 1 ). Conclusions: The present novel ECG AI model demonstrates superior accuracy in detecting acute OMI when compared to the STEMI criteria. Its external and international validation suggests its potential to improve ACS patient triage with timely referral for immediate revascularization.

Cross-Sectional Relationship Between Atrial Conduction Delay and Arterial Stiffness in Patients with Obstructive Sleep Apnea.

Prolonged P-wave duration (PWD), which indicates atrial conduction delay, is a potent precursor of atrial fibrillation (AF) that may be induced by obstructive sleep apnea (OSA). The cardio-ankle vascular index (CAVI), which is an arterial stiffness parameter, is elevated in patients with OSA; moreover, an increased CAVI is associated with atrial conduction delay through left atrium enlargement in association with left ventricular diastolic dysfunction. We aimed to examine the relationship between the CAVI and PWD in patients with OSA. We included patients with a sinus rhythm who underwent overnight polysomnography. We measured the PWD and CAVI on standard 12-lead electrocardiograms; further, we analyzed the relationship between PWD and CAVI. We analyzed data from 300 participants (men, 89.0%; mean age, 52.3 ± 13.1 years; and body mass index, 26.2 ± 3.9 kg/m2). The mean PWD was 104.4 ± 10.4 ms while the mean CAVI was 7.5 ± 1.5. PWD was significantly correlated with CAVI (r = 0.478, p < 0.001); additionally, PWD and CAVI were directly associated with OSA severity (p = 0.002 and p = 0.002, respectively). Multivariate regression analysis revealed an independent significant correlation of PWD and CAVI with OSA severity. In patients with OSA, an increase in arterial stiffness is associated with atrial conduction delay.

Six-Lead Electrocardiography Enables Identification of Rhythm and Conduction Anomalies of Patients in the Telemedicine-Based, Hospital-at-Home Setting: A Prospective Validation Study.

The hospital-at-home (HAH) model is a viable alternative for conventional in-hospital stays worldwide. Serum electrolyte abnormalities are common in acute patients, especially in those with many comorbidities. Pathologic changes in cardiac electrophysiology pose a potential risk during HAH stays. Periodical electrocardiogram (ECG) tracing is therefore advised, but few studies have evaluated the accuracy and efficiency of compact, self-activated ECG devices in HAH settings. This study aimed to evaluate the reliability of such a device in comparison with a standard 12-lead ECG. We prospectively recruited consecutive patients admitted to the Sheba Beyond Virtual Hospital, in the HAH department, during a 3-month duration. Each patient underwent a 12-lead ECG recording using the legacy device and a consecutive recording by a compact six-lead device. Baseline patient characteristics during hospitalization were collected. The level of agreement between devices was measured by Cohen's kappa coefficient for inter-rater reliability (Ϗ). Fifty patients were included in the study (median age 80 years, IQR 14). In total, 26 (52%) had electrolyte disturbances. Abnormal D-dimer values were observed in 33 (66%) patients, and 12 (24%) patients had elevated troponin values. We found a level of 94.5% raw agreement between devices with regards to nine of the options included in the automatic read-out of the legacy device. The calculated Ϗ was 0.72, classified as a substantial consensus. The rate of raw consensus regarding the ECG intervals' measurement (PR, RR, and QT) was 78.5%, and the calculated Ϗ was 0.42, corresponding to a moderate level of agreement. This is the first report to our knowledge regarding the feasibility of using a compact, six-lead ECG device in the setting of an HAH to be safe and bearing satisfying agreement level with a legacy, 12-lead ECG device, enabling quick, accessible arrythmia detection in this setting. Our findings bear a promise to the future development of telemedicine-based hospital-at-home methodology.

Agreement of Computerized QT and QTc Interval Measurements Between Both Bedside and Expert Nurses Using Electronic Calipers.

In hospitalized patients, QT/QTc (heart rate corrected) prolongation on the electrocardiogram (ECG) increases the risk of torsade de pointes. Manual measurements are time-consuming and often inaccurate. Some bedside monitors automatically and continuously measure QT/QTc; however, the agreement between computerized versus nurse-measured values has not been evaluated. The aim of this study was to examine the agreement between computerized QT/QTc and bedside and expert nurses who used electronic calipers. This was a prospective observational study in 3 intensive care units. Up to 2 QT/QTc measurements (milliseconds) per patient were collected. Bland-Altman test was used to analyze measurement agreement. A total of 54 QT/QTc measurements from 34 patients admitted to the ICU were included. The mean difference (bias) for QT comparisons was as follows: computerized versus expert nurses, -11.04 ± 4.45 milliseconds (95% confidence interval [CI], -2.3 to -19.8; P = .016), and computerized versus bedside nurses, -13.72 ± 6.70 (95% CI, -0.70 to -26.8; P = .044). The mean bias for QTc comparisons was as follows: computerized versus expert nurses, -12.46 ± 5.80 (95% CI, -1.1 to -23.8; P = .035), and computerized versus bedside nurses, -18.49 ± 7.90 (95% CI, -3.0 to -33.9; P = .022). Computerized QT/QTc measurements calculated by bedside monitor software and measurements performed by nurses were in close agreement; statistically significant differences were found, but differences were less than 20 milliseconds (on-half of a small box), indicating no clinical significance. Computerized measurements may be a suitable alternative to nurse-measured QT/QTc. This could reduce inaccuracies and nurse burden while increasing adherence to practice recommendations. Further research comparing computerized QT/QTc from bedside monitoring to standard 12-lead electrocardiogram in a larger sample, including non-ICU patients, is needed.