Atrial Fibrillation, Atrial Premature Complexes, Sinus Rhythm, Ventricular Premature Complexes, Ventricular Tachycardia, Nonsustained
Conditions
Keywords
Automated ECG diagnosis
Brief summary
Single-channel electrocardiograms (lead I of 12-lead surface ECG; 30 seconds) will be collected from subjects/patients at 11 clinical centers in Germany to train an Artificial Intelligence in the automatic diagnosis of regular and irregular heart rhythms. Heart rhythms of interest are normal sinus rhythm (SR), atrial fibrillation (AF), atrial premature beats (APBs), ventricular premature beats (VPBs), and nonsustained ventricular tachycardia (VT). Per diagnosis, 20,000 ECGs are required, for a total of 100,000 ECGs to be obtained from approximately 10,000 subjects/patients.
Detailed description
In phase 1 of a research project titled 'Prevention of stroke and sudden cardiac death by Recording of 1-Channel Electrocardiograms' (PRICE), a total of 100,000 30-sec single-channel ECGs (lead I of 12-lead surface ECG) will be collected from approximately 10,000 subjects/patients at 11 participating clinical centers in Germany. Relevant baseline clinical patient characteristics will also be recorded. The ECGs, diagnosed by an experienced electrophysiologist (diagnostic gold standard), will be fed into an Artificial Intelligence (AI) for the automatic detection of normal sinus rhythm (SR), atrial fibrillation (AF), atrial premature beats (APBs), ventricular premature beats (VPBs), and nonsustained ventricular tachycardia (VT). It is expected that the overall diagnostic accuracy of the AI against an experienced electrophysiologist will be on the order of 95%. In PRICE phase 2, ECG diagnosis by the AI will be compared with the diagnosis by 3 general cardiologists of the same ECGs. It is expected that the AI will surpass the general cardiologists in terms of diagnostic accuracy. The final clinical phase of the PRICE project will comprise a randomized controlled community trial of risk patients to establish the superiority in stroke prevention of AI detection of AF on smart-watch ECGs vs. no AF detection.
Interventions
DIAGNOSTIC_TESTElectrocardiogram analysis by Artificial Intelligence
1-channel electrocardiograms are collected to train an Artificial Intelligence in the automatic diagnosis of regular and irregular heart rhythms
Sponsors
A-Rhythmik GmbH
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to 85 Years
Healthy volunteers
Yes
Inclusion criteria
* Heart rhythm of interest present on ECG
Exclusion criteria
* Patient incapable of or not willing to sign informed consent form
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| ECG T wave inversion | Immediate | Assessment of presence (Yes) or absence (No) of T wave inversion |
| ECG QRS-complex duration | Immediate | Measurement of width/duration of QRS complex; distinction between narrow (\<=110ms) and wide (\>110ms) |
| ECG QRS-complex fragmentation | Immediate | Assessment of presence (Yes) or absence (No) of QRS-complex fragmentation |
| ECG QTc interval | Immediate | Calculation of heart rate corrected QT interval (QTc) via Bazett formula from measured QT interval |
| Diagnostic accuracy of AI | 1 year | Overall diagnostic accuracy of the AI in the diagnosis of normal SR, AF, APBs, VPBs, and nonsustained VT (gold standard: diagnosis by experienced electrophysiologist) |
| ECG R-R interval | Immediate | 30-sec mean and standard deviation of R-R intervals |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| ECG PQ interval | Immediate | Measurement of PQ interval (onset of P wave to onset of Q wave) on ECG |
| ECG QT interval | Immediate | Measurement of QT interval (onset of Q wave to end of T wave) on ECG |
| ECG P wave | Immediate | Assessment of presence (Yes) or absence (No) of P wave on ECG; measurement of P-wave duration (in ms) |
Countries
Germany
Contacts
Primary ContactKarl-Heinz Kuck, MD
Backup ContactMichael Schlüter, PhD
Outcome results
None listed