Comparative analysis of CT and CMR-derived ventricular models using ADAS 3D and inHEART for VT substrate assessment

Advances in post-processing software for cardiac magnetic resonance imaging (CMR) and cardiac computed tomography (CCT) have enhanced procedural guidance of catheter ablation procedures. However, till date a direct comparison between two widely used platforms (ADAS 3D LV (ADAS LV Medical, Spain) and inHEART (IHU LIRYC and Inria, France) is lacking.

This study aims to compare CCT- and CMR-derived ventricular substrate models in ischemic and non-ischemic VT patients using ADAS 3D and inHEART to assess inter-platform differences.

Patients having undergone both CCT and CMR imaging prior to VT ablation were retrospectively identified. ADAS 3D was used to create 3D patient-specific substrate models of the ventricles. CMR images were post-processed by inHEART and the resulting substrate models were provided in the online platform.

A custom visual score was employed to assess these substrate models, incorporating scar metrics alongside qualitative assessments (regional variations in wall thickness, scar core, and corridor visualization). Scar presence was scored from 0 to 3, with 0 indicating no scar and 3 indicating scar covering >50% of the left ventricle. Scar transmurality was scored from 0 to 4, reflecting increasing depth, with 4 indicating >75% wall thickness involvement. Qualitative scores (0 to 4) assessed differences in detail, with higher scores indicating progressively clearer visualization in one model over the other.

10 patients were included in the analysis (3 ischemic and 7 non-ischemic). ADAS 3D provided better visualization of the conduction corridors compared to inHEART (figure 1). The Wilcoxon signed-rank test revealed no significant differences in borderzone measurements between ADAS 3D and inHEART for both ischemic and non-ischemic (ischemic: Z = 0, p > 0.05; non-ischemic: Z = -0.81, p > 0.05). A comparable agreement was observed for scar core in ischemic patients (Z = -0.54, p > 0.05) and non-ischemic cardiomyopathy (Z = -0.67, p > 0.05).

In the non-ischemic patient group, ADAS 3D achieved a higher score, with a trend towards significance (figure 2). This difference was primarily caused due to cases with limited spatial resolution and image-registration issues.

This comparative analysis of inHEART and ADAS 3D highlights distinct strengths of each platform in pre-VT ablation image analysis. inHEART offers the advantage of offsite post-processing with minimal user input, making it accessible to a broader range of clinicians. ADAS 3D, while requiring on-site processing and greater user expertise, delivers more nuanced results. The detailed visualization of conduction corridors could make ADAS a valuable tool for targeted ablation planning. Selection between the two platforms should consider factors like clinician experience, equipment availability, and specific visualization needs, as each software could offer unique benefits tailored to different clinical contexts.Figure 1  Figure 2