Background

Substrate-based ablation of ventricular tachycardia (VT) is associated with reduced mortality and VT recurrence. Preprocedural substrate modeling facilitates ultra-high-definition mapping (HDM). However, discordance in scar delineation between cardiac magnet resonance imaging and HDM remains a limitation. Furthermore, detection of electrophysiological substrate by HDM is activation wavefront-dependent and further surrogates for myocardial substrate characterization are needed. The aim of this study was to analyze the relationship of intraprocedural integrated multidetector computed tomography (MDCT)-derived left ventricular (LV) wall thinning with multipolar basket catheter electrograms and local tissue impedance (LI) as surrogate for myocardial scar.

Methods
In two centers, nine patients with structural heart disease (SHD) scheduled for VT ablation underwent MDCT for structural substrate modeling. Generated 3D isosurfaces of the LV, including wall thickness (WT) maps, were integrated into the electroanatomic mapping system, and HDM was performed using a multipolar basket catheter. Severe LV wall thinning was defined as WT ≤2 mm. In case of substrate-based VT ablation, the relationship of MDCT-derived LV wall thinning with multipolar basket catheter electrograms, generator impedance (GI) and LI measurements from three incorporated mini-electrodes in the ablation catheter before and during radiofrequency (RF) ablation was analyzed offline using custom-made software for segmentation according to the AHA 17-segment model and single EGM analysis.

Results

Analysis of 110,000 endocardial surface points from HDM from five patients (median age 60 [31-69] years, 3 male, LV ejection fraction 25 [23-52] %, median 1 [0-3] previous VT ablations) in conjunction with MDCT image-integration revealed gradually reduced bipolar and unipolar voltages in areas with wall thinning compared to normal myocardium (p<0.001). The multipolar basket catheter detected low bipolar voltage potentials <1.0 mV in 74.6% of areas with severe wall thinning ≤2 mm vs. 32.1% in areas without wall thinning (p<0.001), compared to 100% vs. 53.5% of ablation catheter mini electrode potentials <1.0 mV (p=0.018).

In 97 RF ablations, baseline LI positively correlated with WT (r = 0.40, R² 0.16, p=0.001) and was lowest in areas with severe wall thinning (mean baseline LI 97±11 Ω in WT ≤2 mm, 101±14 Ω in WT 3-5 mm and 116±16 Ω in WT >5 mm, p=0.004). Over 73±59 sec mean ablation time LI dropped to 85±11 Ω, 91±13 Ω and 103±13 Ω, respectively (all p<0.05 compared to baseline LI and p=0.003 for group comparison). Absolute LI drops (12±8 Ω, 11±5 Ω and 13±10 Ω, p=0.956) and relative LI drops (12±8 %, 11±4 % and 11±8 %, p=0.865) did not differ between groups. Mean baseline GI was independent of the underlying WT (107±9 Ω in WT ≤2 mm, 106±9 Ω in WT 3-5 mm and 104±6 Ω in WT >5 mm, p=0.248) and did not differ after ablation (108±6 Ω, 103±18 Ω and 104±4 Ω, respectively, all p<0.05 compared to baseline LI and for group comparison).

Conclusion

Intraprocedural integration of MDCT-derived LV wall thinning for structural scar modeling facilitated HDM and identified areas of interest for substrate-based VT ablation. Areas of LV wall thinning were characterized by low voltage multipolar basket catheter and ablation catheter mini electrode potentials. Local myocardial tissue impedance correlated with LV WT and appeared as surrogate for LV myocardial wall thinning.