The evolution of ambulatory ECG monitoring

Artificial Intelligence and Its Role in Diagnosing Heart Failure: A Narrative Review

Heart failure (HF) is prevalent globally. It is a dynamic disease with varying definitions and classifications due to multiple pathophysiologies and etiologies. The diagnosis, clinical staging, and treatment of HF become complex and subjective, impacting patient prognosis and mortality. Technological advancements, like artificial intelligence (AI), have been significant roleplays in medicine and are increasingly used in cardiovascular medicine to transform drug discovery, clinical care, risk prediction, diagnosis, and treatment. Medical and surgical interventions specific to HF patients rely significantly on early identification of HF. Hospitalization and treatment costs for HF are high, with readmissions increasing the burden. AI can help improve diagnostic accuracy by recognizing patterns and using them in multiple areas of HF management. AI has shown promise in offering early detection and precise diagnoses with the help of ECG analysis, advanced cardiac imaging, leveraging biomarkers, and cardiopulmonary stress testing. However, its challenges include data access, model interpretability, ethical concerns, and generalizability across diverse populations. Despite these ongoing efforts to refine AI models, it suggests a promising future for HF diagnosis. After applying exclusion and inclusion criteria, we searched for data available on PubMed, Google Scholar, and the Cochrane Library and found 150 relevant papers. This review focuses on AI's significant contribution to HF diagnosis in recent years, drastically altering HF treatment and outcomes.

Low-Power Long-Term Ambulatory Electrocardiography Monitor of Three Leads with Beat-to-Beat Heart Rate Measurement in Real Time

A low-power long-term ambulatory ECG monitor was developed for the acquisition, storage and processing of three simultaneous leads DI, aVF and V2 with a beat-to-beat heart rate measurement in real time. It provides long-term continuous ECG recordings until 84 h. The monitor uses a QRS complex detection algorithm based on the continuous wavelet transform with splines, which automatically selects the scale for the analysis of ECG records with different sampling frequencies. It includes a lead-off detection to continuously monitor the electrode connections and a real-time system of visual and acoustic alarms to alert users of abnormal conditions in its operation. The monitor presented is based in an ADS1294 analogue front end with four channels, 24-bit analog-to-digital converters and programmable gain amplifiers, a low-power dual-core ESP32 microcontroller, a microSD memory for data storage in a range of 4 GB to 32 GB and a 1.4 in thin-film transistor liquid crystal display (LCD) variant with a resolution of 128 × 128 pixels. It has programmable sampling rates of 250, 500 and 1000 Hz; a bandwidth of 0 Hz to 50% of the selected sampling rate; a CMRR of -105 dB; an input margin of ±2.4 V; a resolution of 286 nV; and a current consumption of 50 mA for an average battery life of 84 h. The ambulatory ECG monitor was evaluated with the commercial data-acquisition system BIOPAC MP36 and its module for ECG LABEL SS2LB, simultaneously comparing the morphologies of two ECG records and obtaining a correlation of 91.78%. For the QRS detection in real time, the implemented algorithm had an error less than 5%. The developed ambulatory ECG monitor can be used for the analysis of the dynamics of the heart rate variability in long-term ECG records and for the development of one's own databases of ECG recordings of normal subjects and patients with cardiovascular and noncardiovascular diseases.

Ventricular tachycardia risk prediction with an abbreviated duration mobile cardiac telemetry

Background

Ventricular tachycardia (VT) occurs intermittently, unpredictably, and has potentially lethal consequences.

Objective

Our aim was to derive a risk prediction model for VT episodes ≥10 beats detected on 30-day mobile cardiac telemetry based on the first 24 hours of the recording.

Methods

We included patients who were monitored for 2 to 30 days in the United States using full-disclosure mobile cardiac telemetry, without any VT episode ≥10 beats on the first full recording day. An elastic net prediction model was derived for the outcome of VT ≥10 beats on monitoring days 2 to 30. Potential predictors included age, sex, and electrocardiographic data from the first 24 hours: heart rate; premature atrial and ventricular complexes occurring as singlets, couplets, triplets, and runs; and the fastest rate for each event. The population was randomly split into training (70%) and testing (30%) samples.

Results

In a population of 19,781 patients (mean age 65.3 ± 17.1 years, 43.5% men), with a median recording time of 18.6 ± 9.6 days, 1510 patients had at least 1 VT ≥10 beats. The prediction model had good discrimination in the testing sample (area under the receiver-operating characteristic curve 0.7584, 95% confidence interval 0.7340-0.7829). A model excluding age and sex had an equally good discrimination (area under the receiver-operating characteristic curve 0.7579, 95% confidence interval 0.7332-0.7825). In the top quintile of the score, more than 1 in 5 patients had a VT ≥10 beats, while the bottom quintile had a 98.2% negative predictive value.

Conclusion

Our model can predict risk of VT ≥10 beats in the near term using variables derived from 24-hour electrocardiography, and could be used to triage patients to extended monitoring.

Analysis and postprocessing of ECG or heart rate data from wearable devices beyond the proprietary cloud and app infrastructure of the vendors

Background

The impact of medical-grade wearable electrocardiographic (ECG) recording technology is increasing rapidly. A wide range of different portable smartphone-connected ECG and heart rate trackers is available on the market. Smart ECG devices are especially valuable to monitor either supraventricular arrhythmias or prolonged QT intervals to avoid drug-induced life-threatening arrhythmias. However, frequent false alarms or false-positive arrhythmia results from wearable devices are unwanted. Therefore, for clinical evaluation, it should be possible to measure and evaluate the biosignals of the wearables independent of the manufacturer.

Objective

Unlike radiological devices that do support the universal digital imaging and communications in medicine standard, these medical-grade devices do not yet support a secure standardized exchange pathway between sensors, smartphones/smartwatches, and end services such as cloud storage or universal Web-based application programming interface (API) access. Consequently, postprocessing of recorded ECGs or heart rate interval data requires a whole toolbox of customized software technologies.

Methods/Results

Various methods for measuring and analyzing nonstandardized ECG and heart rate data are proposed, including online measurement of ECG waveforms within a PDF, access to data using manufacturer-specific software development kits, and access to biosignals using modern Web APIs.

Conclusion

With the appropriate workaround, modern software technologies such as JavaScript and PHP allow health care providers and researchers to easily and instantly access necessary and important signal measurements on demand.

The role of non-invasive devices for the telemonitoring of heart failure patients

Heart failure (HF) patients represent one of the most prevalent as well as one of the most fragile population encountered in the cardiology and internal medicine departments nowadays. Estimated to account for around 26 million people worldwide, diagnosed patients present a poor prognosis and quality of life with a clinical history accompanied by repeated hospital admissions caused by an exacerbation of their chronic condition. The frequent hospitalizations and the extended hospital stays mean an extremely high economic burden for healthcare institutions. Meanwhile, the number of chronically diseased and elderly patients is continuously rising, and a lack of specialized physicians is evident. To cope with this health emergency, more efficient strategies for patient management, more accurate diagnostic tools, and more efficient preventive plans are needed. In recent years, telemonitoring has been introduced as the potential answer to solve such needs. Different methodologies and devices have been progressively investigated for effective home monitoring of cardiologic patients. Invasive hemodynamic devices, such as CardioMEMS™, have been demonstrated to be reducing hospitalizations and mortality, but their use is however restricted to limited cases. The role of external non-invasive devices for remote patient monitoring, instead, is yet to be clarified. In this review, we summarized the most relevant studies and devices that, by utilizing non-invasive telemonitoring, demonstrated whether beneficial effects in the management of HF patients were effective.

Cross-instrument feasibility, validity, and reproducibility of wireless heart rate monitors: Novel opportunities for extended daily life monitoring

Wired ambulatory monitoring of the electrocardiogram (ECG) is an established method used by researchers and clinicians. Recently, a new generation of wireless, compact, and relatively inexpensive heart rate monitors have become available. However, before these monitors can be used in scientific research and clinical practice, their feasibility, validity, and reproducibility characteristics have to be investigated. Therefore, we tested how two wireless heart rate monitors (i.e., the Ithlete photoplethysmography (PPG) finger sensor and the Cortrium C3 ECG monitor perform against an established wired reference method (the VU-AMS ambulatory ECG monitor). Monitors were tested on cross-instrument and test-retest reproducibility in a controlled laboratory setting, while feasibility was evaluated in protocolled ambulatory settings at home. We found that the Cortrium and the Ithlete monitors showed acceptable agreement with the VU-AMS reference in laboratory setting. In ambulatory settings, assessments were feasible with both wireless devices although more valid data were obtained with the Cortrium than with the Ithlete. We conclude that both monitors have their merits under controlled laboratory settings where motion artefacts are minimized and stationarity of the ECG signal is optimized by design. These findings are promising for long-term ambulatory ECG measurements, although more research is needed to test whether the wireless devices' feasibility, validity, and reproducibility characteristics also hold in unprotocolled daily life settings with natural variations in posture and activities.

Ambulatory External Electrocardiography Monitoring: Holter, Extended Holter, Mobile Cardiac Telemetry Monitoring

Ambulatory external electrocardiography (AECG) monitoring is effective as an evidence-based diagnostic tool when suspicion for cardiac arrhythmia is high. Multiple modalities of AECG monitoring exist, with unique advantages and limitations that predict effectiveness in a variety of clinical settings. Knowledge of these characteristics allows appropriate use of AECG, maximizing patient adherence, diagnostic yield, and cost-effectiveness. In addition, new technology has allowed the development of a modern generation of devices that offer increased efficacy and functionality compared with Holter monitors.

Validation of the Polar V800 heart rate monitor and comparison of artifact correction methods among adults with hypertension

Heart rate variability (HRV) measurements via ambulatory monitors have become common. We examined the validity of recording R-R intervals using the Polar V800^™ compared to 12-lead electrocardiograms (ECG) among middle-aged (44.7±10.1years); overweight to obese (29.8±4.3 kg.m^-2) adults (n = 25) with hypertension (132.3±12.2/ 84.3±10.2 mmHg). After resting for 5-min in the supine position, R-R intervals were simultaneously recorded using the Polar V800^™ and the 12-lead ECG. Artifacts present in uncorrected (UN) R-R intervals were corrected with the Kubios HRV Premium (ver. 3.2.) automatic (AC) and threshold-based (TBC) correction, and manual correction (MC) methods. Intra-class correlation coefficients (ICC), Bland-Altman limits of agreement (LoA), and effect sizes (ES) were calculated. We detected 71 errors with the Polar V800^™ for an error rate of 0.85%. The bias (LoAs), ES, and ICC between UN and ECG R-R intervals were 0.69ms (-215.80 to +214.42ms), 0.004, and 0.79, respectively. Correction of artifacts improved the agreeability between the Polar V800^™ and ECG HRV measures. The biases (LoAs) between the AC, TBC, and MC and ECG R-R intervals were 3.79ms (-130.32 to +137.90ms), 1.16ms (-92.67 to +94.98ms), and 0.37ms (-41.20 to +41.94ms), respectively. The ESs of AC, TBC, and MC were 0.024, 0.008, and 0.002, and ICCs were 0.91, 0.95, and 1.00, respectively. R-R intervals measured using the Polar V800^™ compared to 12-lead ECG were comparable in adults with hypertension, especially after the artifacts corrected by MC. However, TBC correction also yielded acceptable results.

Application of Modern Multi-Sensor Holter in Diagnosis and Treatment

Modern Holter devices are very trendy tools used in medicine, research, or sport. They monitor a variety of human physiological or pathophysiological signals. Nowadays, Holter devices have been developing very fast. New innovative products come to the market every day. They have become smaller, smarter, cheaper, have ultra-low power consumption, do not limit everyday life, and allow comfortable measurements of humans to be accomplished in a familiar and natural environment, without extreme fear from doctors. People can be informed about their health and 24/7 monitoring can sometimes easily detect specific diseases, which are normally passed during routine ambulance operation. However, there is a problem with the reliability, quality, and quantity of the collected data. In normal life, there may be a loss of signal recording, abnormal growth of artifacts, etc. At this point, there is a need for multiple sensors capturing single variables in parallel by different sensing methods to complement these methods and diminish the level of artifacts. We can also sense multiple different signals that are complementary and give us a coherent picture. In this article, we describe actual interesting multi-sensor principles on the grounds of our own long-year experiences and many experiments.

Deep learning models for electrocardiograms are susceptible to adversarial attack

Electrocardiogram (ECG) acquisition is increasingly widespread in medical and commercial devices, necessitating the development of automated interpretation strategies. Recently, deep neural networks have been used to automatically analyze ECG tracings and outperform physicians in detecting certain rhythm irregularities¹. However, deep learning classifiers are susceptible to adversarial examples, which are created from raw data to fool the classifier such that it assigns the example to the wrong class, but which are undetectable to the human eye^2,3. Adversarial examples have also been created for medical-related tasks^4,5. However, traditional attack methods to create adversarial examples do not extend directly to ECG signals, as such methods introduce square-wave artefacts that are not physiologically plausible. Here we develop a method to construct smoothed adversarial examples for ECG tracings that are invisible to human expert evaluation and show that a deep learning model for arrhythmia detection from single-lead ECG⁶ is vulnerable to this type of attack. Moreover, we provide a general technique for collating and perturbing known adversarial examples to create multiple new ones. The susceptibility of deep learning ECG algorithms to adversarial misclassification implies that care should be taken when evaluating these models on ECGs that may have been altered, particularly when incentives for causing misclassification exist.

[Prinzmetal angina. Questions of pathogenesis, clinic, diagnosis and treatment]

Current problems of Prinzmetal angina (vasospastic angina, variant angina) considers in this review. Attention is drawn to early diagnosis, which should be comprehensive, taking into account possible atypical courses and the development of complications. The important role of electrocardiographic monitoring (including using implantable recorders) is highlighted. It is emphasized that patients with cardiac arrhythmias, syncope are at high risk of developing sudden cardiac death. In this category of patients, it is recommended to timely determine the indications for implantation of a cardioverter - defibrillator. Authors consider the prospects of using new methods of treatment of angina pectoris.

Ambulatory cardiac bio-signals: From mirage to clinical reality through a decade of progress

BACKGROUND

Health monitoring is shifting towards continuous, ambulatory and clinically comparable wearable devices. Telemedicine and remote diagnosis could harness the capability of mobile cardiac health information, as the technology on bio-physical signal monitoring has improved significantly.

OBJECTIVES

The purpose of this review article is (1) to systematically assess the viability of ambulatory electrocardiography (ECG), (2) to provide a systems level understanding of a broad spectrum of wearable heart signal monitoring approaches and (3) to identify areas of improvement in the existing technology needed to attain clinical grade diagnosis.

RESULTS

Based on the included literature, we have identified (1) that the developments in ECG monitoring through wearable devices are reaching feasibility, and are capable of delivering diagnostic and prognostic information, (2) that reliable sensing is the major bottleneck in the entire process of ambulatory monitoring, (3) that there is a strong need for artificial intelligence and machine learning techniques to parse and infer the biosignals and (4) that aspects of wearer comfort has largely been ignored in the prevailing developments, which can become a key factor for consumer acceptance.

CONCLUSIONS

Cardiac health information is crucial for diagnosis and prevention of several disease onsets. Mobile and continuous monitoring can aid avoiding risks involved with acute symptoms. The health information obtained through continuous monitoring can serve as the BigData of heart signals, and can facilitate new treatment methods and devise effective health policies.

Mobile Personal Health Monitoring for Automated Classification of Electrocardiogram Signals in Elderly

Mobile electrocardiogram (ECG) monitoring is an emerging area that has received increasing attention in recent years, but still real-life validation for elderly residing in low and middle-income countries is scarce. We developed a wearable ECG monitor that is integrated with a self-designed wireless sensor for ECG signal acquisition. It is used with a native purposely designed smartphone application, based on machine learning techniques, for automated classification of captured ECG beats from aged people. When tested on 100 older adults, the monitoring system discriminated normal and abnormal ECG signals with a high degree of accuracy (97%), sensitivity (100%), and specificity (96.6%). With further verification, the system could be useful for detecting cardiac abnormalities in the home environment and contribute to prevention, early diagnosis, and effective treatment of cardiovascular diseases, while keeping costs down and increasing access to healthcare services for older persons.

Ambulatory monitoring of myocardial ischemia in the 21st century-an opportunity for high frequency QRS analysis

Ambulatory monitoring represents an effective tool for the assessment of silent and transient myocardial ischemia during routine daily activities. Incidence of silent ischemia can provide important prognostic information about patients with coronary artery disease or acute coronary syndrome, as well as about post-myocardial infarction patients. The current technological progress enables development of powerful and miniaturized wearable devices for Holter monitoring. Higher sampling rates, dynamic range, and extended computational and storage capacity allow for considering of more complex methodological solutions such as high-frequency QRS analysis for diagnosing myocardial ischemia. Implementation of suitable methodologies for advanced detection of myocardial ischemia into modern ambulatory monitoring devices creates the potential of making the ambulatory myocardial ischemia monitoring a valuable diagnostic tool in clinical practice.

Diagnostic accuracy of a smartphone electrocardiograph in dogs: Comparison with standard 6-lead electrocardiography

The diagnostic accuracy of a smartphone electrocardiograph (ECG) in evaluating heart rhythm and ECG measurements was evaluated in 166 dogs. A standard 6-lead ECG was acquired for 1 min in each dog. A smartphone ECG tracing was simultaneously recorded using a single-lead bipolar ECG recorder. All ECGs were reviewed by one blinded operator, who judged if tracings were acceptable for interpretation and assigned an electrocardiographic diagnosis. Agreement between smartphone and standard ECG in the interpretation of tracings was evaluated. Sensitivity and specificity for the detection of arrhythmia were calculated for the smartphone ECG. Smartphone ECG tracings were interpretable in 162/166 (97.6%) tracings. A perfect agreement between the smartphone and standard ECG was found in detecting bradycardia, tachycardia, ectopic beats and atrioventricular blocks. A very good agreement was found in detecting sinus rhythm versus non-sinus rhythm (100% sensitivity and 97.9% specificity). The smartphone ECG provided tracings that were adequate for analysis in most dogs, with an accurate assessment of heart rate, rhythm and common arrhythmias. The smartphone ECG represents an additional tool in the diagnosis of arrhythmias in dogs, but is not a substitute for a 6-lead ECG. Arrhythmias identified by the smartphone ECG should be followed up with a standard ECG before making clinical decisions.

A Smartphone Application to Diagnose the Mechanism of Pediatric Supraventricular Tachycardia

Smartphone applications that record a single-lead ECG are increasingly available. We sought to determine the utility of a smartphone application (AliveCor) to record supraventricular tachycardia (SVT) and to distinguish atrioventricular reentrant tachycardia (AVRT) from atrioventricular nodal reentrant tachycardia (AVNRT) in pediatric patients. A prior study demonstrated that interpretation of standard event and Holter monitors accurately identifies the tachycardia mechanism in only 45 % of recordings. We performed an IRB-approved prospective study in pediatric patients undergoing an ablation for SVT. Tracings were obtained by placing the smartphone in three different positions on the chest (PI-horizontal, PII-rotated 60° clockwise, and PIII-rotated 120° clockwise). Two blinded pediatric electrophysiologists jointly analyzed a pair of sinus and tachycardia tracings in each position. Tracings with visible retrograde P waves were classified as AVRT. The three positions were compared by Chi-square test. Thirty-seven patients (age 13.7 ± 2.8 years) were enrolled in the study. Twenty-four had AVRT, and 13 had AVNRT. One hundred and eight pairs of tracings were obtained. The correct diagnosis was made in 27/37 (73 %) with position PI, 28/37 (76 %) with PII, and 20/34 (59 %) with PIII (p = 0.04 for PII vs. PIII and p = NS for other comparisons). A single-lead ECG obtained with a smartphone monitor can successfully record SVT in pediatric patients and can predict the SVT mechanism at least as well as previously published reports of Holter monitors, along with the added convenience of not requiring patients to carry a dedicated monitor.

Cost Analysis and Clinical Outcomes of Ambulatory Care Monitoring in Medicare Patients: Describing the Diagnostic Odyssey

Objectives: The diagnostic sequence and costs for arrhythmia detection utilizing Holter ambulatory ECG monitoring have not been well studied. The objective of the current study was to characterize the number of patients and associated costs incurred in the diagnosis, additional monitoring, clinical events and sequelae after an initial Holter monitor in Medicare patients with arrhythmia-the diagnostic odyssey. Methods: We performed a retrospective, longitudinal claims analysis using a 5% random sample of Medicare beneficiaries' claims from the Fee-for-Service (FFS) Standard Analytic Files. The analysis was limited to patients with full benefits for 1 year prior and 2 years post the index 24- or 48-hour Holter event, no prior arrhythmia or Holter. Results: The group of greatest interest was the "No results" category, since these 1,976 patients (11.1% of the total 17,887 patients evaluated) reflected the failure of repeat Holter monitoring to either detect clinical events or diagnose disease. In spite of this failure, there was a total allowed charge of more than $45 million or slightly more than $23,000 per involved patient. When extrapolated over the entire Medicare FFS population, this category was estimated to cost more than $900 million over the 2-year study period. Conclusions: Additional diagnostic paradigms need to be explored to improve upon these patient and system outcomes, where repeat monitoring frequently did not yield a diagnosis and patients continued to experience clinical events.