Computer-Interpreted Electrocardiograms: Benefits and Limitations

Artificial neural networks for ECG interpretation in acute coronary syndrome: A scoping review

INTRODUCTION

The electrocardiogram (ECG) is a crucial diagnostic tool in the Emergency Department (ED) for assessing patients with Acute Coronary Syndrome (ACS). Despite its widespread use, the ECG has limitations, including low sensitivity of the STEMI criteria to detect Acute Coronary Occlusion (ACO) and poor inter-rater reliability. Emerging ECG features beyond the traditional STEMI criteria show promise in improving early ACO diagnosis, but complexity hinders widespread adoption. The potential integration of Artificial Neural Networks (ANN) holds promise for enhancing diagnostic accuracy and addressing reliability issues in ECG interpretation for ACO symptoms.

METHODS

Ovid MEDLINE, CINAHL, EMBASE, Cochrane, PubMed and Scopus were searched from inception through to 8th of December 2023. A thorough search of the grey literature and reference lists of relevant articles was also performed to identify additional studies. Articles were included if they reported the use of ANN for ECG interpretation of Acute Coronary Syndrome in the Emergency Department patients.

RESULTS

The search yielded a total of 244 articles. After removing duplicates and excluding non-relevant articles, 14 remained for analysis. There was significant heterogeneity in the types of ANN models used and the outcomes assessed, making direct comparisons challenging. Nevertheless, ANN appeared to demonstrate higher accuracy than physician interpreters for the evaluated outcomes and this proved independent of both specialty and years of experience.

CONCLUSIONS

The interpretation of ECGs in patients with suspected ACS using ANN appears to be accurate and potentially superior when compared to human interpreters and computerised algorithms. This appears consistent across various ANN models and outcome variables. Future investigations should emphasise ANN interpretation of ECGs in patients with ACO, where rapid and accurate diagnosis can significantly benefit patients through timely access to reperfusion therapies.

Prospects for artificial intelligence-enhanced electrocardiogram as a unified screening tool for cardiac and non-cardiac conditions: an explorative study in emergency care

Aims

Current deep learning algorithms for automatic ECG analysis have shown notable accuracy but are typically narrowly focused on singular diagnostic conditions. This exploratory study aims to investigate the capability of a single deep learning model to predict a diverse range of both cardiac and non-cardiac discharge diagnoses based on a single ECG collected in the emergency department.

Methods and results

In this study, we assess the performance of a model trained to predict a broad spectrum of diagnoses. We find that the model can reliably predict 253 ICD codes (81 cardiac and 172 non-cardiac) in the sense of exceeding an AUROC score of 0.8 in a statistically significant manner.

Conclusion

The model demonstrates proficiency in handling a wide array of cardiac and non-cardiac diagnostic scenarios, indicating its potential as a comprehensive screening tool for diverse medical encounters.

Artificial Intelligence for Cardiovascular Care-Part 1: Advances: JACC Review Topic of the Week

Recent artificial intelligence (AI) advancements in cardiovascular care offer potential enhancements in diagnosis, treatment, and outcomes. Innovations to date focus on automating measurements, enhancing image quality, and detecting diseases using novel methods. Applications span wearables, electrocardiograms, echocardiography, angiography, genetics, and more. AI models detect diseases from electrocardiograms at accuracy not previously achieved by technology or human experts, including reduced ejection fraction, valvular heart disease, and other cardiomyopathies. However, AI's unique characteristics necessitate rigorous validation by addressing training methods, real-world efficacy, equity concerns, and long-term reliability. Despite an exponentially growing number of studies in cardiovascular AI, trials showing improvement in outcomes remain lacking. A number are currently underway. Embracing this rapidly evolving technology while setting a high evaluation benchmark will be crucial for cardiology to leverage AI to enhance patient care and the provider experience.

Comparison of emergency medicine specialist, cardiologist, and chat-GPT in electrocardiography assessment

INTRODUCTION

ChatGPT, developed by OpenAI, represents the cutting-edge in its field with its latest model, GPT-4. Extensive research is currently being conducted in various domains, including cardiovascular diseases, using ChatGPT. Nevertheless, there is a lack of studies addressing the proficiency of GPT-4 in diagnosing conditions based on Electrocardiography (ECG) data. The goal of this study is to evaluate the diagnostic accuracy of GPT-4 when provided with ECG data, and to compare its performance with that of emergency medicine specialists and cardiologists.

METHODS

This study has received approval from the Clinical Research Ethics Committee of Hitit University Medical Faculty on August 21, 2023 (decision no: 2023-91). Drawing on cases from the "150 ECG Cases" book, a total of 40 ECG cases were crafted into multiple-choice questions (comprising 20 everyday and 20 more challenging ECG questions). The participant pool included 12 emergency medicine specialists and 12 cardiology specialists. GPT-4 was administered the questions in a total of 12 separate sessions. The responses from the cardiology physicians, emergency medicine physicians, and GPT-4 were evaluated separately for each of the three groups.

RESULTS

In the everyday ECG questions, GPT-4 demonstrated superior performance compared to both the emergency medicine specialists and the cardiology specialists (p < 0.001, p = 0.001). In the more challenging ECG questions, while Chat-GPT outperformed the emergency medicine specialists (p < 0.001), no significant statistical difference was found between Chat-GPT and the cardiology specialists (p = 0.190). Upon examining the accuracy of the total ECG questions, Chat-GPT was found to be more successful compared to both the Emergency Medicine Specialists and the cardiologists (p < 0.001, p = 0.001).

CONCLUSION

Our study has shown that GPT-4 is more successful than emergency medicine specialists in evaluating both everyday and more challenging ECG questions. It performed better compared to cardiologists on everyday questions, but its performance aligned closely with that of the cardiologists as the difficulty of the questions increased.

Explaining deep learning for ECG analysis: Building blocks for auditing and knowledge discovery

Deep neural networks have become increasingly popular for analyzing ECG data because of their ability to accurately identify cardiac conditions and hidden clinical factors. However, the lack of transparency due to the black box nature of these models is a common concern. To address this issue, explainable AI (XAI) methods can be employed. In this study, we present a comprehensive analysis of post-hoc XAI methods, investigating the glocal (aggregated local attributions over multiple samples) and global (concept based XAI) perspectives. We have established a set of sanity checks to identify saliency as the most sensible attribution method. We provide a dataset-wide analysis across entire patient subgroups, which goes beyond anecdotal evidence, to establish the first quantitative evidence for the alignment of model behavior with cardiologists' decision rules. Furthermore, we demonstrate how these XAI techniques can be utilized for knowledge discovery, such as identifying subtypes of myocardial infarction. We believe that these proposed methods can serve as building blocks for a complementary assessment of the internal validity during a certification process, as well as for knowledge discovery in the field of ECG analysis.

ECG-based cardiac arrhythmias detection through ensemble learning and fusion of deep spatial-temporal and long-range dependency features

Cardiac arrhythmia is one of the prime reasons for death globally. Early diagnosis of heart arrhythmia is crucial to provide timely medical treatment. Heart arrhythmias are diagnosed by analyzing the electrocardiogram (ECG) of patients. Manual analysis of ECG is time-consuming and challenging. Hence, effective automated detection of heart arrhythmias is important to produce reliable results. Different deep-learning techniques to detect heart arrhythmias such as Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Transformer, and Hybrid CNN-LSTM were proposed. However, these techniques, when used individually, are not sufficient to effectively learn multiple features from the ECG signal. The fusion of CNN and LSTM overcomes the limitations of CNN in the existing studies as CNN-LSTM hybrids can extract spatiotemporal features. However, LSTMs suffer from long-range dependency issues due to which certain features may be ignored. Hence, to compensate for the drawbacks of the existing models, this paper proposes a more comprehensive feature fusion technique by merging CNN, LSTM, and Transformer models. The fusion of these models facilitates learning spatial, temporal, and long-range dependency features, hence, helping to capture different attributes of the ECG signal. These features are subsequently passed to a majority voting classifier equipped with three traditional base learners. The traditional learners are enriched with deep features instead of handcrafted features. Experiments are performed on the MIT-BIH arrhythmias database and the model performance is compared with that of the state-of-art models. Results reveal that the proposed model performs better than the existing models yielding an accuracy of 99.56%.

Fundamental and Practical Feasibility of Electrocardiogram Reconstruction from Photoplethysmogram

Electrocardiogram (ECG) reconstruction from contact photoplethysmogram (PPG) would be transformative for cardiac monitoring. We investigated the fundamental and practical feasibility of such reconstruction by first replicating pioneering work in the field, with the aim of assessing the methods and evaluation metrics used. We then expanded existing research by investigating different cycle segmentation methods and different evaluation scenarios to robustly verify both fundamental feasibility, as well as practical potential. We found that reconstruction using the discrete cosine transform (DCT) and a linear ridge regression model shows good results when PPG and ECG cycles are semantically aligned-the ECG R peak and PPG systolic peak are aligned-before training the model. Such reconstruction can be useful from a morphological perspective, but loses important physiological information (precise R peak location) due to cycle alignment. We also found better performance when personalization was used in training, while a general model in a leave-one-subject-out evaluation performed poorly, showing that a general mapping between PPG and ECG is difficult to derive. While such reconstruction is valuable, as the ECG contains more fine-grained information about the cardiac activity as well as offers a different modality (electrical signal) compared to the PPG (optical signal), our findings show that the usefulness of such reconstruction depends on the application, with a trade-off between morphological quality of QRS complexes and precise temporal placement of the R peak. Finally, we highlight future directions that may resolve existing problems and allow for reliable and robust cross-modal physiological monitoring using just PPG.

Automatic triage of twelve-lead electrocardiograms using deep convolutional neural networks: a first implementation study

Aims

Expert knowledge to correctly interpret electrocardiograms (ECGs) is not always readily available. An artificial intelligence (AI)-based triage algorithm (DELTAnet), able to support physicians in ECG prioritization, could help reduce current logistic burden of overreading ECGs and improve time to treatment for acute and life-threatening disorders. However, the effect of clinical implementation of such AI algorithms is rarely investigated.

Methods and results

Adult patients at non-cardiology departments who underwent ECG testing as a part of routine clinical care were included in this prospective cohort study. DELTAnet was used to classify 12-lead ECGs into one of the following triage classes: normal, abnormal not acute, subacute, and acute. Performance was compared with triage classes based on the final clinical diagnosis. Moreover, the associations between predicted classes and clinical outcomes were investigated. A total of 1061 patients and ECGs were included. Performance was good with a mean concordance statistic of 0.96 (95% confidence interval 0.95-0.97) when comparing DELTAnet with the clinical triage classes. Moreover, zero ECGs that required a change in policy or referral to the cardiologist were missed and there was a limited number of cases predicted as acute that did not require follow-up (2.6%).

Conclusion

This study is the first to prospectively investigate the impact of clinical implementation of an ECG-based AI triage algorithm. It shows that DELTAnet is efficacious and safe to be used in clinical practice for triage of 12-lead ECGs in non-cardiology hospital departments.

BRAVEHEART: Open-source software for automated electrocardiographic and vectorcardiographic analysis

BACKGROUND AND OBJECTIVES

Electrocardiographic (ECG) and vectorcardiographic (VCG) analyses are used to diagnose current cardiovascular disease and for risk stratification for future adverse cardiovascular events. With increasing use of digital ECGs, research into novel ECG/VCG parameters has increased, but widespread computer-based ECG/VCG analysis is limited because there are no currently available, open-source, and easily customizable software packages designed for automated and reproducible analysis.

METHODS AND RESULTS

We present BRAVEHEART, an open-source, modular, customizable, and easy to use software package implemented in the MATLAB programming language, for scientific analysis of standard 12-lead ECGs acquired in a digital format. BRAVEHEART accepts a wide variety of digital ECG formats and provides complete and automatic ECG/VCG processing with signal denoising to remove high- and low-frequency artifact, non-dominant beat identification and removal, accurate fiducial point annotation, VCG construction, median beat construction, customizable measurements on median beats, and output of measurements and results in numeric and graphical formats.

CONCLUSIONS

The BRAVEHEART software package provides easily customizable scientific analysis of ECGs and VCGs. We hope that making BRAVEHART available will allow other researchers to further the field of ECG/VCG analysis without having to spend significant time and resources developing their own ECG/VCG analysis software and will improve the reproducibility of future studies. Source code, compiled executables, and a detailed user guide can be found at http://github.com/BIVectors/BRAVEHEART. The source code is distributed under the GNU General Public License version 3.

The HEARTS ECG workshop: a novel approach to resident and student ECG education

OBJECTIVES

ECG interpretation is a life-saving skill in emergency medicine (EM), and a core competency in undergraduate medical curricula; however, confidence for residents/students is low. We developed a novel educational intervention-the HEARTS ECG workshop-that provides a systematic approach to ECG interpretation, teaches EM residents through the process of teaching medical students and highlights emergency management.

METHODS

We used the Kern Approach to Curriculum Development. A review of ECG education literature and a targeted needs assessment of local students/residents led to goals and objectives including systematic ECG interpretation with clinical relevance. ECGs were selected based on a national consensus of EM program directors and categorized into 5 common emergency presentations. The educational strategy included content based on HEARTS approach (Heart rate/rhythm, Electrical conduction, Axis, R-wave progression, Tall/small voltages, and ST/T changes), and methods including flipped classroom and near-peer teaching. Evaluation and feedback were based on the Kirkpatrick program evaluation. The workshop was piloted with 6 junior EM residents and 58 medical students, and repeated with nine residents and 68 students from four medical schools.

RESULTS

Residents and students agreed or strongly agreed that the workshop improved their perceived ability (100% and 95%, respectively) and confidence (77% and 88%, respectively) in interpreting ECGs. Reports of ECG interpretation causing anxiety declined from pre-workshop (61% and 83% respectively) to post-workshop (38% and 37% respectively). Residents reported behavior change: 3 months after the workshop, 92.3% reported ongoing use of the HEARTS approach clinically and through teaching medical students on shifts. Reported workshop strengths included the pre-workshop material, the clinical application, facilitator-to-learner ratio, interactivity, the ease of remembering and applying the HEARTS mnemonic, and the iterative application of the approach. Suggested changes included longitudinal sessions with graded difficulty, and allocating more time for introductory material for ease of understanding.

CONCLUSION

The HEARTS ECG workshop is an innovative pedagogical method that can be adapted for all levels of training. Future directions include integration in undergraduate medical and EM residency curricula, and workshops for physicians to update ECG interpretation skills.

Impact of Computer-Interpreted ECGs on the Accuracy of Healthcare Professionals

The interpretation of electrocardiograms (ECGs) involves a dynamic interplay between computerized ECG interpretation (CEI) software and human overread. However, the impact of computer ECG interpretation on the performance of healthcare professionals remains largely unexplored. The aim of this study was to evaluate the interpretation proficiency of various medical professional groups, with and without access to the CEI report. Healthcare professionals from diverse disciplines, training levels, and countries sequentially interpreted 60 standard 12-lead ECGs, demonstrating both urgent and nonurgent findings. The interpretation process consisted of 2 phases. In the first phase, participants interpreted 30 ECGs with clinical statements. In the second phase, the same 30 ECGs and clinical statements were randomized and accompanied by a CEI report. Diagnostic performance was evaluated based on interpretation accuracy, time per ECG (in seconds [s]), and self-reported confidence (rated 0 [not confident], 1 [somewhat confident], or 2 [confident]). A total of 892 participants from various medical professional groups participated in the study. This cohort included 44 (4.9%) primary care physicians, 123 (13.8%) cardiology fellows-in-training, 259 (29.0%) resident physicians, 137 (15.4%) medical students, 56 (6.3%) advanced practice providers, 82 (9.2%) nurses, and 191 (21.4%) allied health professionals. The inclusion of the CEI was associated with a significant improvement in interpretation accuracy by 15.1% (95% confidence interval, 14.3-16.0; P < 0.001), decrease in interpretation time by 52 s (-56 to -48; P < 0.001), and increase in confidence by 0.06 (0.03-0.09; P = 0.003). Improvement in interpretation accuracy was seen across all professional subgroups, including primary care physicians by 12.9% (9.4-16.3; P = 0.003), cardiology fellows-in-training by 10.9% (9.1-12.7; P < 0.001), resident physicians by 14.4% (13.0-15.8; P < 0.001), medical students by 19.9% (16.8-23.0; P < 0.001), advanced practice providers by 17.1% (13.3-21.0; P < 0.001), nurses by 16.2% (13.4-18.9; P < 0.001), allied health professionals by 15% (13.4-16.6; P < 0.001), physicians by 13.2% (12.2-14.3; P < 0.001), and nonphysicians by 15.6% (14.3-17.0; P < 0.001).CEI integration improves ECG interpretation accuracy, efficiency, and confidence among healthcare professionals.

Personalized, intuitive & visual QT-prolongation monitoring using patient-specific QTc threshold with pseudo-coloring and explainable AI

BACKGROUND

Drug-induced QT-prolongation increases the risk of TdP arrhythmia attacks and sudden cardiac death. However, measuring the QT-interval and determining a precise cut-off QT/QTc value that could put a patient at risk of TdP is challenging and influenced by many factors including female sex, drug-free baseline, age, genetic predisposition, and bradycardia.

OBJECTIVES

This paper presents a novel approach for intuitively and visually monitoring QT-prolongation showing a potential risk of TdP, which can be adjusted according to patient-specific risk factors, using a pseudo-coloring technique and explainable artificial intelligence (AI).

METHODS

We extended the development and evaluation of an explainable AI-based technique- visualized using pseudo-color on the ECG signal, thus intuitively 'explaining' how its decision was made -to detect QT-prolongation showing a potential risk of TdP according to a cut-off personalized QTc value (using Bazett's ∆QTc > 60 ms relative to drug-free baseline and Bazett's QTc > 500 ms as examples), and validated its performance using a large number of ECGs (n = 5050), acquired from a clinical trial assessing the effects of four known QT-prolonging drugs versus placebo on healthy subjects. We compared this new personalized approach to our previous study that used a more general approach using the QT-nomogram.

RESULTS AND CONCLUSIONS

The explainable AI-based algorithm can accurately detect QT-prolongation when adjusted to a personalized patient-specific cut-off QTc value showing a potential risk of TdP. Using ∆QTc > 60 ms relative to drug-free baseline and QTc > 500 ms as examples, the algorithm yielded a sensitivity of 0.95 and 0.79, and a specificity of 0.95 and 0.98, respectively. We found that adjusting pseudo-coloring according to Bazett's ∆QTc > 60 ms relative to a drug-free baseline personalized to each patient provides better sensitivity than using Bazett's QTc > 500 ms, which could underestimate a potentially clinically significant QT-prolongation with bradycardia.

Compared with physician overread, computer is less accurate but helpful in interpretation of electrocardiography for ST-segment elevation myocardial infarction

INTRODUCTION

Previous studies have demonstrated varying sensitivity and specificity of computer-interpreted electrocardiography (CIE) in identifying ST-segment elevation myocardial infarction (STEMI). This study aims to evaluate the accuracy of contemporary computer software in recognizing electrocardiography (ECG) signs characteristic of STEMI compared to emergency physician overread in clinical practice.

MATERIAL AND METHODS

In this retrospective observational single-center study, we reviewed the records of patients in the emergency department (ED) who underwent ECGs and troponin tests. Both the Philips DXL 16-Lead ECG. Algorithm and on-duty emergency physicians interpreted each standard 12‑lead ECG. The sensitivity and specificity of computer interpretation and physician overread ECGs for the definite diagnosis of STEMI were calculated and compared.

RESULTS

Among the 9340 patients included in the final analysis, 133 were definitively diagnosed with STEMI. When "computer-reported infarct or injury" was used as the indicator, the sensitivity was 87.2% (95% CI 80.3% to 92.4%) and the specificity was 86.2% (95% CI 85.5% to 86.9%). When "physician-overread STEMI" was used as the indicator, the sensitivity was 88.0% (95% CI 81.2% to 93.0%) and the specificity was 99.9% (95% CI 99.8% to 99.9%). The area under the receiver operating characteristic curve for physician-overread STEMI and computer-reported infarct or injury were 0.939 (95% CI 0.907 to 0.972) and 0.867 (95% CI 0.834 to 0.900), respectively.

CONCLUSIONS

This study reveals that while the sensitivity of the computer in recognizing ECG signs of STEMI is similar to that of physicians, physician overread of ECGs is more specific and, therefore, more accurate than CIE.

Missing occlusions: Quality gaps for ED patients with occlusion MI

BACKGROUND

ST-elevation Myocardial Infarction (STEMI) guidelines encourage monitoring of false positives (Code STEMI without culprit) but ignore false negatives (non-STEMI with occlusion myocardial infarction [OMI]). We evaluated the hospital course of emergency department (ED) patients with acute coronary syndrome (ACS) using STEMI vs OMI paradigms.

METHODS

This retrospective chart review examined all ACS patients admitted through two academic EDs, from June 2021 to May 2022, categorized as 1) OMI (acute culprit lesion with TIMI 0-2 flow, or acute culprit lesion with TIMI 3 flow and peak troponin I >10,000 ng/L; or, if no angiogram, peak troponin >10,000 ng/L with new regional wall motion abnormality), 2) NOMI (Non-OMI, i.e. MI without OMI) or 3) MIRO (MI ruled out: no troponin elevation). Patients were stratified by admission for STEMI. Initial ECGs were reviewed for automated interpretation of "STEMI", and admission/discharge diagnoses were compared.

RESULTS

Among 382 patients, there were 141 OMIs, 181 NOMIs, and 60 MIROs. Only 40.4% of OMIs were admitted as STEMI: 60.0% had "STEMI" on ECG, and median door-to-cath time was 103 min (IQR 71-149). But 59.6% of OMIs were not admitted as STEMI: 1.3% had "STEMI" on ECG (p < 0.001) and median door-to-cath time was 1712 min (IQR 1043-3960; p < 0.001). While 13.9% of STEMIs were false positive and had a different discharge diagnosis, 32.0% of Non-STEMIs had OMI but were still discharged as "Non-STEMI."

CONCLUSIONS

STEMI criteria miss a majority of OMI, and discharge diagnoses highlight false positive STEMI but never false negative STEMI. The OMI paradigm reveals quality gaps and opportunities for improvement.

Machine learning for the diagnosis of acute coronary syndrome using a 12-lead ECG: a systematic review

OBJECTIVES

Prompt diagnosis of acute coronary syndrome (ACS) using a 12-lead electrocardiogram (ECG) is a critical task for emergency physicians. While computerized algorithms for ECG interpretation are limited in their accuracy, machine learning (ML) models have shown promise in several areas of clinical medicine. We performed a systematic review to compare the performance of ML-based ECG analysis to clinician or non-ML computerized ECG interpretation in the diagnosis of ACS for emergency department (ED) or prehospital patients.

METHODS

We searched Medline, Embase, Cochrane Central, and CINAHL databases from inception to May 18, 2022. We included studies that compared ML algorithms to either clinicians or non-ML based software in their ability to diagnose ACS using only a 12-lead ECG, in adult patients experiencing chest pain or symptoms concerning for ACS in the ED or prehospital setting. We used QUADAS-2 for risk of bias assessment. Prospero registration CRD42021264765.

RESULTS

Our search yielded 1062 abstracts. 10 studies met inclusion criteria. Five model types were tested, including neural networks, random forest, and gradient boosting. In five studies with complete performance data, ML models were more sensitive but less specific (sensitivity range 0.59-0.98, specificity range 0.44-0.95) than clinicians (sensitivity range 0.22-0.93, specificity range 0.63-0.98) in diagnosing ACS. In four studies that reported it, ML models had better discrimination (area under ROC curve range 0.79-0.98) than clinicians (area under ROC curve 0.67-0.78). Heterogeneity in both methodology and reporting methods precluded a meta-analysis. Several studies had high risk of bias due to patient selection, lack of external validation, and unreliable reference standards for ACS diagnosis.

CONCLUSIONS

ML models have overall higher discrimination and sensitivity but lower specificity than clinicians and non-ML software in ECG interpretation for the diagnosis of ACS. ML-based ECG interpretation could potentially serve a role as a "safety net", alerting emergency care providers to a missed acute MI when it has not been diagnosed. More rigorous primary research is needed to definitively demonstrate the ability of ML to outperform clinicians at ECG interpretation.

A cardiologist-like computer-aided interpretation framework to improve arrhythmia diagnosis from imbalanced training datasets

Arrhythmias can pose a significant threat to cardiac health, potentially leading to serious consequences such as stroke, heart failure, cardiac arrest, shock, and sudden death. In computer-aided electrocardiogram interpretation systems, the inclusion of certain classes of arrhythmias, which we term "aggressive" or "bullying," can lead to the underdiagnosis of other "vulnerable" classes. To address this issue, a method for arrhythmia diagnosis is proposed in this study. This method combines morphological-characteristic-based waveform clustering with Bayesian theory, drawing inspiration from the diagnostic reasoning of experienced cardiologists. The proposed method achieved optimal performance in macro-recall and macro-precision through hyperparameter optimization, including spliced heartbeats and clusters. In addition, with increasing bullying by aggressive arrhythmias, our model obtained the highest average recall and the lowest average drop in recall on the nine vulnerable arrhythmias. Furthermore, the maximum cluster characteristics were found to be consistent with established arrhythmia diagnostic criteria, lending interpretability to the proposed method.

Fourier space approach for convolutional neural network (CNN) electrocardiogram (ECG) classification: A proof-of-concept study

There has been a proliferation of machine learning (ML) electrocardiogram (ECG) classification algorithms reaching > 85% accuracy for various cardiac pathologies. Although the accuracy within institutions might be high, models trained at one institution might not be generalizable enough for accurate detection when deployed in other institutions due to differences in type of signal acquisition, sampling frequency, time of acquisition, device noise characteristics and number of leads. In this proof-of-concept study, we leverage the publicly available PTB-XL dataset to investigate the use of time-domain (TD) and frequency-domain (FD) convolutional neural networks (CNN) to detect myocardial infarction (MI), ST/T-wave changes (STTC), atrial fibrillation (AFIB) and sinus arrhythmia (SARRH). To simulate interinstitutional deployment, the TD and FD implementations were also compared on adapted test sets using different sampling frequencies 50 Hz, 100 Hz and 250 Hz, and acquisition times of 5 s and 10s at 100 Hz sampling frequency from the training dataset. When tested on the original sampling frequency and duration, the FD approach showed comparable results to TD for MI (0.92 FD - 0.93 TD AUROC) and STTC (0.94 FD - 0.95 TD AUROC), and better performance for AFIB (0.99 FD - 0.86 TD AUROC) and SARRH (0.91 FD - 0.65 TD AUROC). Although both methods were robust to changes in sampling frequency, changes in acquisition time were detrimental to the TD MI and STTC AUROCs, at 0.72 and 0.58 respectively. Alternatively, the FD approach was able to maintain the same level of performance, and, therefore, showed better potential for interinstitutional deployment.

A Comprehensive Comparison of Six Publicly Available Algorithms for Localization of QRS Complex on Electrocardiograph

The QRS complex is the most prominent feature of the electrocardiogram (ECG) that is used as a marker to identify the cardiac cycles. Identification of QRS complex locations enables arrhythmia detection and heart rate variability estimation. Therefore, accurate and consistent localization of the QRS complex is an important component of automated ECG analysis which is necessary for the early detection of cardiovascular diseases. This study evaluates the performance of six popular publicly available QRS complex detection methods on a large dataset of over half a million ECGs in a diverse population of patients. We found that a deep-learning method that won first place in the 2019 Chinese physiological challenge (CPSC-1) outperforms the remaining five QRS complex detection methods with an F1 score of 98.8% and an absolute sdRR error of 5.5 ms. We also examined the stratified performance of the studied methods on various cardiac conditions. All six methods had a lower performance in the detection of QRS complexes in ECG signals of patients with pacemakers, complete atrioventricular block, or indeterminate cardiac axis. We also concluded that, in the presence of different cardiac conditions, CPSC-1 is more robust than Pan-Tompkins which is the most popular model for QRS complex detection. We expect that this study can potentially serve as a guide for researchers on the appropriate QRS detection method for their target applications.Clinical Relevance-This study highlights the overall performance of publicly available QRS detection algorithms in a large dataset of diverse patients. We showed that there are specific cardiac conditions that are associated with the poor performance of QRS detection algorithms and may adversely influence the performance of algorithms that rely on accurate and reliable QRS detection.

Electrocardiographic QTc as a Surrogate Measure of Cardiac Risk in Children, Adolescents, and Young Adults With Eating Disorders

The study goal was to investigate electrocardiographic findings, including corrected QT interval (QTc), in patients aged 8 to 23 with eating disorders (EDs) at presentation, compared with an age-and sex-matched control population. We retrospectively reviewed 200 ED patients, and 200 controls. Blinded electrocardiograms (ECGs) were interpreted by an expert reader, and QT intervals corrected using the Bazett formula. Eating disorder patients were 89.5% female, with mean age 16.4 years and median percent median body mass index (BMI)-for-age (%mBMI)a of 91.1%. In ED patients, QTc was significantly shorter than controls (399.6 vs 415.0msec, P < .001). After adjusting for height, %mBMI, sex, magnesium level, and bradycardia, mean QTc duration in patients with anorexia nervosa-restricting subtype (AN-R) was significantly shorter than other ED patients (P = .010). Higher %mBMI was associated with shorter QTc duration (P = .041) after adjusting for height, magnesium, bradycardia, and Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) diagnosis. Within the ED group, no significant association was identified between QTc and medications, electrolytes, or inpatient status.

An Arrhythmia Classification Model Based on Vision Transformer with Deformable Attention

The electrocardiogram (ECG) is a highly effective non-invasive tool for monitoring heart activity and diagnosing cardiovascular diseases (CVDs). Automatic detection of arrhythmia based on ECG plays a critical role in the early prevention and diagnosis of CVDs. In recent years, numerous studies have focused on using deep learning methods to address arrhythmia classification problems. However, the transformer-based neural network in current research still has a limited performance in detecting arrhythmias for the multi-lead ECG. In this study, we propose an end-to-end multi-label arrhythmia classification model for the 12-lead ECG with varied-length recordings. Our model, called CNN-DVIT, is based on a combination of convolutional neural networks (CNNs) with depthwise separable convolution, and a vision transformer structure with deformable attention. Specifically, we introduce the spatial pyramid pooling layer to accept varied-length ECG signals. Experimental results show that our model achieved an F1 score of 82.9% in CPSC-2018. Notably, our CNN-DVIT outperforms the latest transformer-based ECG classification algorithms. Furthermore, ablation experiments reveal that the deformable multi-head attention and depthwise separable convolution are both efficient in extracting features from multi-lead ECG signals for diagnosis. The CNN-DVIT achieved good performance for the automatic arrhythmia detection of ECG signals. This indicates that our research can assist doctors in clinical ECG analysis, providing important support for the diagnosis of arrhythmia and contributing to the development of computer-aided diagnosis technology.

CatBoost-based improved detection of P-wave changes in sinus rhythm and tachycardia conditions: a lead selection study

Examining P-wave morphological changes in Electrocardiogram (ECG) is essential for characterizing atrial arrhythmias. However, standard 12-lead ECGsuffer from diagnostic redundancy due to low signal-to-noise ratio of P-waves. To address this issue, various optimal leads have been proposed for improved atrial activity recording, but the right selection among these leads is crucial for enhancing diagnostic efficacy. This study proposes an automated lead selection technique using the CatBoost machine learning (ML) model to improve the detection of P-wave changes among optimal bipolar leads under different heart rates. ECGs were obtained from healthy participants with a mean age of 25 ± 3.81 years (34% women), including 114 in sinus rhythm (SR) and 38 in sinus tachycardia (ST). The recordings were made using a newly designed atrial lead system (ALS), standard limb lead (SLL), modified limb lead (MLL), modified Lewis lead (LLM) and P-lead. P-wave features and Atrioventricular (AV) ratio were extracted for statistical analysis and ML classification. The optimum ML model was chosen to identify the best-performing optimal lead, which was selected based on the SLL metrics among different ML classifiers. CatBoost was found to outperform the other ML models in SLL-II with the highest accuracy and sensitivity of 0.82 and 0.90, respectively. The CatBoost model, amid other optimal leads, gave the best results for AL-I and AL-II (0.86 and 0.83 in accuracy and 0.91 and 0.93 in sensitivity). The developed CatBoost model selected AL-I and AL-II as the top two best-performing optimal leads for the enhanced acquisition of P-wave changes, which may be useful for diagnosing atrial arrhythmias.

Machine Learning and Artificial Intelligence in Surgical Research

Machine learning, a subtype of artificial intelligence, is an emerging field of surgical research dedicated to predictive modeling. From its inception, machine learning has been of interest in medical and surgical research. Built on traditional research metrics for optimal success, avenues of research include diagnostics, prognosis, operative timing, and surgical education, in a variety of surgical subspecialties. Machine learning represents an exciting and developing future in the world of surgical research that will not only allow for more personalized and comprehensive medical care.

Does the Electrocardiogram Machine Interpretation Affect the Ability to Accurately Diagnose ST-Elevation Myocardial Infarction by Emergency Physicians?

INTRODUCTION

An ST-elevation myocardial infarction (STEMI) can portend significant morbidity and mortality to the patient and therefore must be rapidly diagnosed by an emergency medicine (EM) physician. The primary aim of this study is to determine whether EM physicians are more or less likely to accurately diagnose STEMI on an electrocardiogram (ECG) if they are blinded to the ECG machine interpretation as opposed to if they are provided the ECG machine interpretation.

METHODS

We performed a retrospective chart review of adult patients over 18 years of age admitted to our large, urban tertiary care center with a diagnosis of STEMI from January 1, 2016, to December 31, 2017. From these patients' charts, we selected 31 ECGs to create a quiz that was presented twice to a group of emergency physicians. The first quiz contained the 31 ECGs without the computer interpretations revealed. The second quiz, presented to the same physicians 2 weeks later, contained the same set of ECGs with the computer interpretations revealed. Physicians were asked "Based on the ECG above, is there a blocked coronary artery present causing a STEMI?"

RESULTS

Twenty-five EM physicians completed two 31-question ECG quizzes for a total of 1550 ECG interpretations. On the first quiz with computer interpretations blinded, the overall sensitivity in identifying a "true STEMI" was 67.2% with an overall accuracy of 65.6%. On the second quiz in which the ECG machine interpretation was revealed, the overall sensitivity was 66.4% with an accuracy of 65.8 % in correctly identifying a STEMI. The differences in sensitivity and accuracy were not statistically significant.

CONCLUSION

This study demonstrated no significant difference in physicians blinded versus those unblinded to computer interpretations of possible STEMI.

Electronic monitoring of potential adverse drug events related to lopinavir/ritonavir and hydroxychloroquine during the first wave of COVID-19

During Switzerland's first wave of COVID-19, clinical pharmacy activities during medical rounds in Geneva University Hospitals were replaced by targeted remote interventions. We describe using the electronic PharmaCheck system to screen high-risk situations of adverse drug events (ADEs), particularly targeting prescriptions of lopinavir/ritonavir (LPVr) and hydroxychloroquine (HCQ) in the presence of contraindications or prescriptions outside institutional guidelines. Of 416 patients receiving LPVr and/or HCQ, 182 alerts were triggered for 164 (39.4%) patients. The main associated risk factors of ADEs were drug-drug interactions, QTc interval prolongation, electrolyte disorder and inadequate LPVr dosage. Therapeutic optimisation recommended by a pharmacist or proposals for additional monitoring were accepted in 80% (n=36) of cases. Combined with pharmacist contextualisation to the clinical context, PharmaCheck made it possible to successfully adapt clinical pharmacist activities by switching from a global to a targeted analysis mode in an emergency context.

Hierarchical deep learning with Generative Adversarial Network for automatic cardiac diagnosis from ECG signals

Cardiac disease is the leading cause of death in the US. Accurate heart disease detection is critical to timely medical treatment to save patients' lives. Routine use of the electrocardiogram (ECG) is the most common method for physicians to assess the cardiac electrical activities and detect possible abnormal conditions. Fully utilizing the ECG data for reliable heart disease detection depends on developing effective analytical models. In this paper, we propose a two-level hierarchical deep learning framework with Generative Adversarial Network (GAN) for ECG signal analysis. The first-level model is composed of a Memory-Augmented Deep AutoEncoder with GAN (MadeGAN), which aims to differentiate abnormal signals from normal ECGs for anomaly detection. The second-level learning aims at robust multi-class classification for different arrhythmia identification, which is achieved by integrating the transfer learning technique to transfer knowledge from the first-level learning with the multi-branching architecture to handle the data-lacking and imbalanced data issues. We evaluate the performance of the proposed framework using real-world ECG data from the MIT-BIH arrhythmia database. Experimental results show that our proposed model outperforms existing methods that are commonly used in current practice.

Comparison of two artificial intelligence-augmented ECG approaches: Machine learning and deep learning

BACKGROUND

Artificial intelligence-augmented ECG (AI-ECG) refers to the application of novel AI solutions for complex ECG interpretation tasks. A broad variety of AI-ECG approaches exist, each having differing advantages and limitations relating to their creation and application.

PURPOSE

To provide illustrative comparison of two general AI-ECG modeling approaches: machine learning (ML) and deep learning (DL).

METHOD COMPARISON

Two AI-ECG algorithms were developed to carry out two separate tasks using ML and DL, respectively. ML modeling techniques were used to create algorithms designed for automatic wide QRS complex tachycardia differentiation into ventricular tachycardia and supraventricular tachycardia. A DL algorithm was formulated for the task of comprehensive 12‑lead ECG interpretation. First, we describe the ML models for WCT differentiation, which rely upon expert domain knowledge to identify and formulate ECG features (e.g., percent monophasic time-voltage area [PMonoTVA]) that enable strong diagnostic performance. Second, we describe the DL method for comprehensive 12‑lead ECG interpretation, which relies upon the independent recognition and analysis of a virtually incalculable number of ECG features from a vast collection of standard 12‑lead ECGs.

CONCLUSION

We have showcased two different AI-ECG methods, namely ML and DL respectively. In doing so, we highlighted the strengths and weaknesses of each approach. It is essential for investigators to understand these differences when attempting to create and apply novel AI-ECG solutions.

ECG Classification Using an Optimal Temporal Convolutional Network for Remote Health Monitoring

Increased life expectancy in most countries is a result of continuous improvements at all levels, starting from medicine and public health services, environmental and personal hygiene to the use of the most advanced technologies by healthcare providers. Despite these significant improvements, especially at the technological level in the last few decades, the overall access to healthcare services and medical facilities worldwide is not equally distributed. Indeed, the end beneficiary of these most advanced healthcare services and technologies on a daily basis are mostly residents of big cities, whereas the residents of rural areas, even in developed countries, have major difficulties accessing even basic medical services. This may lead to huge deficiencies in timely medical advice and assistance and may even cause death in some cases. Remote healthcare is considered a serious candidate for facilitating access to health services for all; thus, by using the most advanced technologies, providing at the same time high quality diagnosis and ease of implementation and use. ECG analysis and related cardiac diagnosis techniques are the basic healthcare methods providing rapid insights in potential health issues through simple visualization and interpretation by clinicians or by automatic detection of potential cardiac anomalies. In this paper, we propose a novel machine learning (ML) architecture for the ECG classification regarding five heart diseases based on temporal convolution networks (TCN). The proposed design, which implements a dilated causal one-dimensional convolution on the input heartbeat signals, seems to be outperforming all existing ML methods with an accuracy of 96.12% and an F1 score of 84.13%, using a reduced number of parameters (10.2 K). Such results make the proposed TCN architecture a good candidate for low power consumption hardware platforms, and thus its potential use in low cost embedded devices for remote health monitoring.

Image based deep learning in 12-lead ECG diagnosis

Background

The electrocardiogram is an integral tool in the diagnosis of cardiovascular disease. Most studies on machine learning classification of electrocardiogram (ECG) diagnoses focus on processing raw signal data rather than ECG images. This presents a challenge for models in many areas of clinical practice where ECGs are printed on paper or only digital images are accessible, especially in remote and regional settings. This study aims to evaluate the accuracy of image based deep learning algorithms on 12-lead ECG diagnosis.

Methods

Deep learning models using VGG architecture were trained on various 12-lead ECG datasets and evaluated for accuracy by testing on holdout test data as well as data from datasets not seen in training. Grad-CAM was utilized to depict heatmaps of diagnosis.

Results

The results demonstrated excellent AUROC, AUPRC, sensitivity and specificity on holdout test data from datasets used in training comparable to the best signal and image-based models. Detection of hidden characteristics such as gender were achieved at a high rate while Grad-CAM successfully highlight pertinent features on ECGs traditionally used by human interpreters.

Discussion

This study demonstrates feasibility of image based deep learning algorithms in ECG diagnosis and identifies directions for future research in order to develop clinically applicable image based deep-learning models in ECG diagnosis.

Normal limits of electrocardiogram in Africans and their consequences on the prevalence of left ventricular hypertrophy in hypertensive individuals: Insights from the TAHES study

AIMS

To determine normal limits for major ECG variables, and the electrocardiographic impact of hypertension, in a rural sub-Saharan African setting.

METHODS

This cross-sectional study included adults aged ≥25 years from Tanvè Health Study (TAHES) cohort. ECG were recorded at rest at 25 mm/s using a standard 12‑lead device. Wave amplitudes and durations were measured. Corrected QT interval (QTc) was calculated using Bazett's formula. Sokolow-Lyon, Cornell and Peguero-Lo Presti criteria were determined to assess left ventricular hypertrophy (LVH).

RESULTS

ECG was recorded among 997 out of 1407 TAHES participants. After exclusion of subjects with hypertension or diabetes, normal limits, defined as the 2nd and 98th percentiles, were evaluated in 622 healthy participants (median: 37 years; 60.1% women). The following limits were established in men (women): heart rate: 50 to 100 (55 to 102) beats/min, P wave duration: 80 to 120 (80 to 120) ms, PR interval: 120 to 200 (120 to 200) ms, QTc: 315 to 470 (323 to 465) ms, QRS duration: 50 to 120 (50 to 110) ms. Upper limits (in millimeter) for the Sokolow-Lyon, Cornell and Peguero-Lo Presti for men (women) were 47 (38), 30 (22) and 39 (30), respectively, all above current reference limits. The prevalence of LVH in hypertensive subjects according to these criteria were lower than those estimated according to current LVH criteria.

CONCLUSION

The normal limits of ECG variables determined in this African population differ from those in Caucasians, indicating that ethnicity must be considered in ECG interpretation.

Interpretable Machine Learning Techniques in ECG-Based Heart Disease Classification: A Systematic Review

Heart disease is one of the leading causes of mortality throughout the world. Among the different heart diagnosis techniques, an electrocardiogram (ECG) is the least expensive non-invasive procedure. However, the following are challenges: the scarcity of medical experts, the complexity of ECG interpretations, the manifestation similarities of heart disease in ECG signals, and heart disease comorbidity. Machine learning algorithms are viable alternatives to the traditional diagnoses of heart disease from ECG signals. However, the black box nature of complex machine learning algorithms and the difficulty in explaining a model's outcomes are obstacles for medical practitioners in having confidence in machine learning models. This observation paves the way for interpretable machine learning (IML) models as diagnostic tools that can build a physician's trust and provide evidence-based diagnoses. Therefore, in this systematic literature review, we studied and analyzed the research landscape in interpretable machine learning techniques by focusing on heart disease diagnosis from an ECG signal. In this regard, the contribution of our work is manifold; first, we present an elaborate discussion on interpretable machine learning techniques. In addition, we identify and characterize ECG signal recording datasets that are readily available for machine learning-based tasks. Furthermore, we identify the progress that has been achieved in ECG signal interpretation using IML techniques. Finally, we discuss the limitations and challenges of IML techniques in interpreting ECG signals.

Risk stratification of patients who present with chest pain and have normal troponins using a machine learning model

BACKGROUND

Risk stratification tools exist for patients presenting with chest pain to the emergency room and have achieved the recommended negative predictive value (NPV) of 99%. However, due to low positive predictive value (PPV), current stratification tools result in unwarranted investigations such as serial laboratory tests and cardiac stress tests (CSTs).

AIM

To create a machine learning model (MLM) for risk stratification of chest pain with a better PPV.

METHODS

This retrospective cohort study used de-identified hospital data from January 2016 until November 2021. Inclusion criteria were patients aged > 21 years who presented to the ER, had at least two serum troponins measured, were subsequently admitted to the hospital, and had a CST within 4 d of presentation. Exclusion criteria were elevated troponin value (> 0.05 ng/mL) and missing values for body mass index. The primary outcome was abnormal CST. Demographics, coronary artery disease (CAD) history, hypertension, hyperlipidemia, diabetes mellitus, chronic kidney disease, obesity, and smoking were evaluated as potential risk factors for abnormal CST. Patients were also categorized into a high-risk group (CAD history or more than two risk factors) and a low-risk group (all other patients) for comparison. Bivariate analysis was performed using a χ² test or Fisher’s exact test. Age was compared by t test. Binomial regression (BR), random forest, and XGBoost MLMs were used for prediction. Bootstrapping was used for the internal validation of prediction models. BR was also used for inference. Alpha criterion was set at 0.05 for all statistical tests. R software was used for statistical analysis.

RESULTS

The final cohort of the study included 2328 patients, of which 245 (10.52%) patients had abnormal CST. When adjusted for covariates in the BR model, male sex [risk ratio (RR) = 1.52, 95% confidence interval (CI): 1.2-1.94, P < 0.001)], CAD history (RR = 4.46, 95%CI: 3.08-6.72, P < 0.001), and hyperlipidemia (RR = 3.87, 95%CI: 2.12-8.12, P < 0.001) remained statistically significant. Incidence of abnormal CST was 12.2% in the high-risk group and 2.3% in the low-risk group (RR = 5.31, 95%CI: 2.75-10.24, P < 0.001). The XGBoost model had the best PPV of 24.33%, with an NPV of 91.34% for abnormal CST.

CONCLUSION

The XGBoost MLM achieved a PPV of 24.33% for an abnormal CST, which is better than current stratification tools (13.00%-17.50%). This highlights the beneficial potential of MLMs in clinical decision-making.

Computerized electrocardiogram data transformation enables effective algorithmic differentiation of wide QRS complex tachycardias

BACKGROUND

Accurate automated wide QRS complex tachycardia (WCT) differentiation into ventricular tachycardia (VT) and supraventricular wide complex tachycardia (SWCT) can be accomplished using calculations derived from computerized electrocardiogram (ECG) data of paired WCT and baseline ECGs.

OBJECTIVE

Develop and trial novel WCT differentiation approaches for patients with and without a corresponding baseline ECG.

METHODS

We developed and trialed WCT differentiation models comprised of novel and previously described parameters derived from WCT and baseline ECG data. In Part 1, a derivation cohort was used to evaluate five different classification models: logistic regression (LR), artificial neural network (ANN), Random Forests [RF], support vector machine (SVM), and ensemble learning (EL). In Part 2, a separate validation cohort was used to prospectively evaluate the performance of two LR models using parameters generated from the WCT ECG alone (Solo Model) and paired WCT and baseline ECGs (Paired Model).

RESULTS

Of the 421 patients of the derivation cohort (Part 1), a favorable area under the receiver operating characteristic curve (AUC) by all modeling subtypes: LR (0.96), ANN (0.96), RF (0.96), SVM (0.96), and EL (0.97). Of the 235 patients of the validation cohort (Part 2), the Solo Model and Paired Model achieved a favorable AUC for 103 patients with (Solo Model 0.87; Paired Model 0.95) and 132 patients without (Solo Model 0.84; Paired Model 0.95) a corroborating electrophysiology procedure or intracardiac device recording.

CONCLUSION

Accurate WCT differentiation may be accomplished using computerized data of (i) the WCT ECG alone and (ii) paired WCT and baseline ECGs.

Detection of arrhythmia in 12-lead varied-length ECG using multi-branch signal fusion network

Objective.Automatic detection of arrhythmia based on electrocardiogram (ECG) plays a critical role in early prevention and diagnosis of cardiovascular diseases. With the increase in widely available digital ECG data and the development of deep learning, multi-class arrhythmia classification based on automatic feature extraction of ECG has become increasingly attractive. However, the majority of studies cannot accept varied-length ECG signals and have limited performance in detecting multi-class arrhythmias.Approach.In this study, we propose a multi-branch signal fusion network (MBSF-Net) for multi-label classification of arrhythmia in 12-lead varied-length ECG. Our model utilizes the complementary power between different structures, which include Inception with depthwise separable convolution (DWS-Inception), spatial pyramid pooling (SPP) Layer, and multi-scale fusion Resnet (MSF-Resnet). The proposed method can extract features from each lead of 12-lead ECG recordings separately and then effectively fuse the features of each lead by integrating multiple convolution kernels with different receptive fields, which can achieve the information of complementation between different angles of the ECG signal. In particular, our model can accept 12-lead ECG signals of arbitrary length.Main results.The experimental results show that our model achieved an overall classification F1 score of 83.8% in the 12-lead ECG data of CPSC-2018. In addition, the F1 score of the MBSF-Net performed best among the MBF-Nets which are removed the SPP layer from MBSF-Net. In comparison with the latest ECG classification algorithms, the proposed model can be applied in varied-length signals and has an excellent performance, which not only can fully retain the integrity of the original signals, but also eliminates the cropping/padding signal beforehand when dealing with varied-length signal database.Significance.MBSF-Net provides an end-to-end multi-label classification model with outperfom performance, which allows detection of disease in varied-length signals without any additional cropping/padding. Moreover, our research is beneficial to the development of computer-aided diagnosis.

Artificial intelligence in cardiology: Hope for the future and power for the present

Cardiovascular disease (CVD) is the principal cause of mortality and morbidity globally. With the pressures for improved care and translation of the latest medical advances and knowledge to an actionable plan, clinical decision-making for cardiologists is challenging. Artificial Intelligence (AI) is a field in computer science that studies the design of intelligent agents which take the best feasible action in a situation. It incorporates the use of computational algorithms which simulate and perform tasks that traditionally require human intelligence such as problem solving and learning. Whilst medicine is arguably the last to apply AI in its everyday routine, cardiology is at the forefront of AI revolution in the medical field. The development of AI methods for accurate prediction of CVD outcomes, non-invasive diagnosis of coronary artery disease (CAD), detection of malignant arrythmias through wearables, and diagnosis, treatment strategies and prediction of outcomes for heart failure (HF) patients, demonstrates the potential of AI in future cardiology. With the advancements of AI, Internet of Things (IoT) and the promotion of precision medicine, the future of cardiology will be heavily based on these innovative digital technologies. Despite this, ethical dilemmas regarding the implementation of AI technologies in real-world are still unaddressed.

Device agnostic AI-based analysis of ambulatory ECG recordings

Deep Convolutional Neural Networks (DCNNs) have been shown to provide improved performance over traditional heuristic algorithms for the detection of arrhythmias from ambulatory ECG recordings. However, these DCNNs have primarily been trained and tested on device-specific databases with standardized electrode positions and uniform sampling frequencies. This work explores the possibility of training a DCNN for Atrial Fibrillation (AF) detection on a database of single‑lead ECG rhythm strips extracted from resting 12‑lead ECGs. We then test the performance of the DCNN on recordings from ambulatory ECG devices with different recording leads and sampling frequencies. We developed an extensive proprietary resting 12‑lead ECG dataset of 549,211 patients. This dataset was randomly split into a training set of 494,289 patients and a testing set of the remaining 54,922 patients. We trained a 34-layer convolutional DCNN to detect AF and other arrhythmias on this dataset. The DCNN was then validated on two Physionet databases commonly used to benchmark automated ECG algorithms (1) MIT-BIH Arrhythmia Database and (2) MIT-BIH Atrial Fibrillation Database. Validation was performed following the EC57 guidelines, with performance assessed by gross episode and duration sensitivity and positive predictive value (PPV). Finally, validation was also performed on a selection of rhythm strips from an ambulatory ECG patch that a committee of board-certified cardiologists annotated. On MIT-BIH, The DCNN achieved a sensitivity of 100% and 84% PPV in detecting episodes of AF. and 100% sensitivity and 94% PPV in quantifying AF episode duration. On AFDB, The DCNN achieved a sensitivity of 94% and PPV of 98% in detecting episodes of AF, and 98% sensitivity and 100% PPV in quantifying AF episode duration. On the patch database, the DCNN demonstrated performance that was closely comparable to that of a cardiologist. The results indicate that DCNN models can learn features that generalize between resting 12‑lead and ambulatory ECG recordings, allowing DCNNs to be device agnostic for detecting arrhythmias from single‑lead ECG recordings and enabling a range of clinical applications.

A Deep Learning Algorithm for Detecting Acute Pericarditis by Electrocardiogram

(1) Background: Acute pericarditis is often confused with ST-segment elevation myocardial infarction (STEMI) among patients presenting with acute chest pain in the emergency department (ED). Since a deep learning model (DLM) has been validated to accurately identify STEMI cases via 12-lead electrocardiogram (ECG), this study aimed to develop another DLM for the detection of acute pericarditis in the ED. (2) Methods: This study included 128 ECGs from patients with acute pericarditis and 66,633 ECGs from patients visiting the ED between 1 January 2010 and 31 December 2020. The ECGs were randomly allocated based on patients to the training, tuning, and validation sets, at a 3:1:1 ratio. We used raw ECG signals to train a pericarditis-DLM and used traditional ECG features to train a machine learning model. A human–machine competition was conducted using a subset of the validation set, and the performance of the Philips automatic algorithm was also compared. STEMI cases in the validation set were extracted to analyze the DLM ability of differential diagnosis between acute pericarditis and STEMI using ECG. We also followed the hospitalization events in non-pericarditis cases to explore the meaning of false-positive predictions. (3) Results: The pericarditis-DLM exceeded the performance of all participating human experts and algorithms based on traditional ECG features in the human–machine competition. In the validation set, the pericarditis-DLM could detect acute pericarditis with an area under the receiver operating characteristic curve (AUC) of 0.954, a sensitivity of 78.9%, and a specificity of 97.7%. However, our pericarditis-DLM also misinterpreted 10.2% of STEMI ECGs as pericarditis cases. Therefore, we generated an integrating strategy combining pericarditis-DLM and a previously developed STEMI-DLM, which provided a sensitivity of 73.7% and specificity of 99.4%, to identify acute pericarditis in patients with chest pains. Compared to the true-negative cases, patients with false-positive results using this strategy were associated with higher risk of hospitalization within 3 days due to cardiac disorders (hazard ratio (HR): 8.09; 95% confidence interval (CI): 3.99 to 16.39). (4) Conclusions: The AI-enhanced algorithm may be a powerful tool to assist clinicians in the early detection of acute pericarditis and differentiate it from STEMI using 12-lead ECGs.

The West Jutland Tele-Electrocardiogram Registry (WEJU-tECG): content, data quality, and research potential

AIM

To present the content, data quality, and research potential of the West Jutland Tele-Electrocardiogram Registry (WEJU-tECG).

METHODS

Danish patients reporting symptoms indicating heart disease in the prehospital setting are subjected to a 12-lead tele-electrocardiogram (ECG) in the ambulance, which is digitally sent to a local tele-centre. WEJU-tECG is a newly established Danish registry containing information from the individual tele-ECGs received at the Regional Hospital West Jutland tele-centre.

RESULTS

WEJU-tECG holds extracted information from all tele-ECGs with a valid Civil Personal Register number between 2011 and 2020. WEJU-tECG contains information on patient characteristics, tele-ECG data (including a computerised tele-ECG interpretation), vital signs, and time information. A unique Civil Personal Register number allows individual-level linkage between WEJU-tECG and other Danish registries and enables complete follow-up. WEJU-tECG contains 43,696 tele-ECGs from 29,489 different patient contacts among 20,280 different patients. WEJU-tECG contains 5566 patients with ST-segment deviations. The median age is 67 years and 45% are women. Completeness is highest for time information (100% for all variables), tele-ECG data (99% for heart rate, the specific intervals and axes, and QRS duration, and 86% for J-point deviation), and patient characteristics (100% for all variables). Completeness is lowest for vital signs (13% for systolic, diastolic, and mean arterial blood pressure, and 12% for blood oxygen saturation). The computerised tele-ECG interpretation had a negative predictive value of 80% for ST-segment elevation myocardial infarction and 94% for non-ST-segment elevation myocardial infarction and a positive predictive value of 45% for ST-segment elevation myocardial infarction and 32% for non-ST-segment elevation myocardial infarction.

CONCLUSIONS

WEJU-tECG is a novel population-based tele-ECG registry with high research potential.

[Artificial intelligence-based ECG analysis: current status and future perspectives-Part 1 : Basic principles]

Even though electrocardiography is a diagnostic procedure that is now more than 100 years old, medicine cannot do without it. On the contrary, interest in the procedure and its clinical significance is even increasing again. Reports on the evaluation of electrocardiograms (ECGs) with the aid of artificial intelligence (AI) are also responsible for this. Using machine learning and in particular deep learning, both AI subfields, completely new perspectives of ECG evaluation and interpretation arise. The weaknesses inherent in classical computer-assisted ECG evaluation appear to be overcome. This two-part overview deals with AI-based ECG analysis. Part 1 introduces basic aspects of the procedure. Part 2, which is published separately, is devoted to the current state of research and discusses the available studies. In addition, possible scenarios of future application of AI in ECG analysis are discussed.

[Artificial intelligence-based ECG analysis: current status and future perspectives : Part 2: Recent studies and future]

Während grundlegende Aspekte der Anwendung von künstlicher Intelligenz (KI) zur Elektrokardiogramm(EKG)-Analyse in Teil 1 dieser Übersicht behandelt wurden, beschäftigt sich die vorliegende Arbeit (Teil 2) mit einer Besprechung von aktuellen Studien zum praktischen Einsatz dieser neuen Technologien und Aspekte ihrer aktuellen und möglichen zukünftigen Anwendung. Die Anzahl der zum Thema KI-basierte EKG-Analyse publizierten Studien steigt seit 2017 rasant an. Dies gilt vor allem für Untersuchungen, die Deep Learning (DL) mit künstlichen neuronalen Netzen (KNN) einsetzen. Inhaltlich geht es nicht nur darum, die Schwächen der klassischen EKG-Diagnostik mit Hilfe von KI zu überwinden und die diagnostische Güte des Verfahrens zu verbessern, sondern auch die Funktionalität des EKGs zu erweitern. Angestrebt wird die Erkennung spezieller kardiologischer und nichtkardiologischer Krankheitsbilder sowie die Vorhersage zukünftiger Krankheitszustände, z. B. die zukünftige Entwicklung einer linksventrikulären Dysfunktion oder das zukünftige Auftreten von Vorhofflimmern. Möglich wird dies, indem KI mittels DL in riesigen EKG-Datensätzen subklinische Muster findet und für die Algorithmen-Entwicklung nutzt. Die KI-unterstützte EKG-Analyse wird somit zu einem Screening-Instrument und geht weit darüber hinaus, nur besser als ein Kardiologe zu sein. Die erzielten Fortschritte sind bemerkenswert und sorgen in Fachwelt und Öffentlichkeit für Aufmerksamkeit und Euphorie. Bei den meisten Studien handelt es sich allerdings um Proof-of-Concept-Studien. Häufig werden private (institutionseigene) Daten verwendet, deren Qualität unklar ist. Bislang ist nur selten eine klinische Validierung der entwickelten Algorithmen in anderen Kollektiven und Szenarien erfolgt. Besonders problematisch ist, dass der Weg, wie KI eine Lösung findet, bislang meistens verborgen bleibt (Blackbox-Charakter). Damit steckt die KI-basierte Elektrokardiographie noch in den Kinderschuhen. Unbestritten ist aber schon absehbar, dass das EKG als einfach anzuwendendes und beliebig oft wiederholbares diagnostisches Verfahren auch in Zukunft nicht nur weiterhin unverzichtbar sein wird, sondern durch KI an klinischer Bedeutung gewinnen wird.

Athlete ECG T-wave abnormality interpretation patterns by non-experts

Background

The presence of T-wave abnormalities (TWA) on an athlete's electrocardiogram (ECG) presents as a diagnostic challenge for physicians. Types of TWA patterns classified as abnormal by inexperienced readers have not been systematically analyzed.

Methods

ECGs from the 2011-2015 National Football League Scouting Combine (initially interpreted by general cardiologists) were retrospectively reviewed by expert sports cardiologists with strict application of the 2017 International Criteria. Patterns of TWA that were altered from the original interpretation were analyzed.

Results

The study included 1643 athletes (mean age 22 years). There was a 67 % reduction in the number of athletes with any TWA (p < 0.001) with 111 ECGs changed to normal. Inferior TWA was the most common interpreted initial ECG abnormality altered followed by anterior and lateral.

Discussion

This analysis revealed an initial high rate of TWA by non-expert readers. Tailored education programs to physicians who interpret athlete ECGs should highlight these specific T-wave patterns. We see this as an opportunity to make more clinicians aware of ECG interpretation guidelines as sports trained cardiologists are mostly self-taught.

Automatic Multichannel Electrocardiogram Record Classification Using XGBoost Fusion Model

There is an increasing demand for automatic classification of standard 12-lead electrocardiogram signals in the medical field. Considering that different channels and temporal segments of a feature map extracted from the 12-lead electrocardiogram record contribute differently to cardiac arrhythmia detection, and to the classification performance, we propose a 12-lead electrocardiogram signal automatic classification model based on model fusion (CBi-DF-XGBoost) to focus on representative features along both the spatial and temporal axes. The algorithm extracts local features through a convolutional neural network and then extracts temporal features through bi-directional long short-term memory. Finally, eXtreme Gradient Boosting (XGBoost) is used to fuse the 12-lead models and domain-specific features to obtain the classification results. The 5-fold cross-validation results show that in classifying nine categories of electrocardiogram signals, the macro-average accuracy of the fusion model is 0.968, the macro-average recall rate is 0.814, the macro-average precision is 0.857, the macro-average F1 score is 0.825, and the micro-average area under the curve is 0.919. Similar experiments with some common network structures and other advanced electrocardiogram classification algorithms show that the proposed model performs favourably against other counterparts in F1 score. We also conducted ablation studies to verify the effect of the complementary information from the 12 leads and the auxiliary information of domain-specific features on the classification performance of the model. We demonstrated the feasibility and effectiveness of the XGBoost-based fusion model to classify 12-lead electrocardiogram records into nine common heart rhythms. These findings may have clinical importance for the early diagnosis of arrhythmia and incite further research. In addition, the proposed multichannel feature fusion algorithm can be applied to other similar physiological signal analyses and processing.