Deep learning-based insights on T:R ratio behaviour during prolonged screening for S-ICD eligibility

Area of Science:

• Cardiology
• Biomedical Engineering
• Medical Imaging

Background:

• Subcutaneous implantable cardiac defibrillators (S-ICD) eligibility relies on the T:R ratio.
• Current T:R ratio cut-offs include a safety margin for ECG signal amplitude fluctuations.
• A novel deep learning tool is introduced to quantify T:R ratio fluctuations for S-ICD screening.

Purpose of the Study:

• To assess the role of a deep learning tool in measuring T:R ratio fluctuations.
• To identify an optimal T:R ratio threshold for S-ICD screening.
• To minimize inappropriate shocks due to T-wave oversensing (TWO) while including true S-ICD candidates.

Main Methods:

• Utilized Holter recordings (24h) to capture S-ICD vectors in 37 patients.
• Applied a deep learning tool to analyze T:R ratio fluctuations throughout recordings.
• Evaluated multiple T:R ratio cut-offs to identify high-risk patients for TWO.

Main Results:

• 54% of patients passed screening at a T:R ratio of 1:3; all passed at 1:1.
• Patients with normal hearts were the only subgroup to pass screening with all tested ratios.
• The study included patients with heart failure, hypertrophic cardiomyopathy, congenital heart disease, and prior inappropriate S-ICD shocks.

Conclusions:

• Prolonged screening is proposed for S-ICD eligibility to reduce TWO risk.
• The optimal T:R ratio for screening likely falls between 1:3 and 1:1.
• Further research is needed to establish definitive screening thresholds.