Reproducibility analysis of the computerized tomography angiography-derived left atrial wall thickness maps.

Area of Science:

• Cardiovascular Imaging
• Electrophysiology
• Medical Device Technology

Background:

• Left atrial wall thickness (LAWT) derived from computed tomography angiography (CTA) enables personalized pulmonary vein isolation (PVI).
• Adapting the ablation index (AI) to LAWT improves PVI safety and outcomes.

Purpose of the Study:

• To assess the reproducibility of LAWT segmentation from CTA.
• To evaluate the impact of LAWT segmentation reproducibility on the personalized AI for PVI.

Main Methods:

• Three observers with varying experience performed LAWT analysis on CTA scans from 30 patients.
• Intra- and inter-observer reproducibility of LA endocardial and epicardial surface segmentation was assessed.
• The impact of LAWT segmentation on the derived AI for PVI was evaluated.

Main Results:

• High geometric congruence was observed for both endocardial (99.4% within 1mm) and epicardial (82.4% within 1mm) surfaces.
• LAWT measurements demonstrated good intra- and inter-observer reproducibility, with increased concordance correlating with user experience.
• The adaptation of AI to LAWT maps resulted in a negligible average difference (<25 units) in the derived AI.

Conclusions:

• LAWT segmentation from CTA is geometrically accurate and reproducible.
• The method's reproducibility is enhanced by user experience.
• This approach has a minimal impact on the target AI, supporting its clinical utility for personalized PVI.