Accessory Pathway, Artifical Intelligence, ECG
Conditions
Keywords
AI-ECG, AI, ECG, Accessory pathway, Accessory pathway localisation
Brief summary
This study seeks to validate the real-world accuracy of an AI-based algorithm for identifying the location of an accessory pathway from the 12-lead electrocardiogram
Detailed description
Silent validation study of an AI-ECG (artificial intelligence applied to electrocardiography) accessory pathway localisation algorithm, applied to prospective and consecutive cases in clinical practice, to determine its true accuracy and performance. A pre-existing AI-ECG algorithm will be applied to participant ECG data, collected at the time of their clinical electrophysiology study (EPS) for ablation of their accessory pathway. This will be compared to the ground truth of the successful ablation location, determined by fluoroscopy and/or 3D electroanatomical mapping from their procedure.
Interventions
None listed
Sponsors
Imperial College London
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
13 Years to 100 Years
Healthy volunteers
No
Inclusion criteria
* Referred for EPS procedure as part of their clinical care, with a finding of pre-excitation on their ECG * Manifest pre-excitation on their ECG any time prior to their procedure * Able to give consent * Minimum age 13 years old * Maximum age 100 years old
Exclusion criteria
* Unable to give consent * Adults \> 100 years old * Children \< 13 years old * Patients with known location of their accessory pathway from a previous EP study
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Performance and accuracy of the AI-ECG accessory pathway localisation algorithm | At completion of recruitment, anticipated at 18 months | Performance metrics of the AI-ECG accessory pathway localisation algorithm, including accuracy, F1-score, sensitivity, specificity, positive and negative predictive values. Benchmarked against the ground truth of human operator assessment from fluoroscopy and/or 3D electroanatomical mapping. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Accuracy of the ground truth locations from the human operator compared to the successful ablation location | At completion of recruitment, anticipated at 18 months | The ground truth of successful ablation location determined by operator assessment of fluoroscopy ± 3D mapping will be compared to the true ablation location on a complete 3D electroanatomical annular map |
| Relative performance of the AI-ECG algorithm compared to human estimation | At completion of recruitment, anticipated at 18 months | Difference in performance/accuracy between the AI-ECG accessory pathway localisation algorithm and human estimation from the 12-lead ECG |
| Relative performance of the AI-ECG algorithm compared to manual localisation algorithms | At completion of recruitment, anticipated at 18 months | Difference in performance/accuracy between the AI algorithm and pre-specified, established manual localisation algorithms (Arruda, Milstein, Pambrun, Boersma, D'Avila and Chiang) |
Countries
United Kingdom
Contacts
Primary ContactKeenan Saleh, MBBS
Outcome results
None listed