Evaluating the Correlation Between Cardiac CT-defined Wall Thickness and Ventricular Tachycardia Substrate

Purpose: cardiac CT has shown promise in characterizing arrhythmogenic substrates in patients with scar-related ventricular tachycardia (VT). Previous studies indicated that wall thickness (WT) maps can effectively predict critical areas for catheter ablation (CA) and stereotactic arrhythmia radioablation (STAR). This study aimed to replicate and extend previous findings by evaluating the effectiveness and robustness of a novel pre-procedural workflow for identifying VT substrates using WT maps derived from cardiac CT. Specifically, we sought to assess the correlation between left ventricular wall thinning and the presence of VT substrates.

Methods: Ten ECG-gated contrast enhance cardiac CT datasets in diastole were used. The left ventricular (LV) epicardium and endocardium were segmented using TotalSegmentator and a threshold-based method, followed by manual refinement. WT was estimated using a Laplace-based method and validated by comparing WT maps with those from MUSIC platform. Wall thinning regions were categorized from 1 mm to 5 mm, and comparisons were made using Dice scores and Hausdorff Distances (HD). Three blinded electrophysiologists independently identified critical VT isthmus locations, which were correlated with VT activation maps to assess accuracy.

Results: Table 1 presents clinical characteristics and validation outcomes. The mean Dice for wall-thinning regions across all thickness categories was 0.46 ± 0.17, with a HD of 28.00mm ± 19.46mm. A systematic +2 mm bias was observed in WT estimation relative to the MUSIC reference. Upon applying bias correction, the Dice increased to 0.73 ± 0.03, and the HD decreased to 1.12mm ± 2.38mm, signifying enhanced concordance WT maps from MUSIC platform. 

Conclusion: This study demonstrates the robustness of a cardiac CT-derived WT mapping workflow in identifying arrhythmogenic substrates in scar-related VT. Despite moderate alignment with the MUSIC platform and a minor WT estimation bias, the workflow reliably identified VT-critical regions consistent with activation maps. These findings highlight WT mapping as a valuable tool for enhancing pre-procedural planning accuracy for CA and STAR treatments.