Abstract 2741: Multi-Modality, Four-Dimensional Ultrasound-Based Image Guidance for Left Ventricular Ablation
Introduction: Ablation of unstable, peri-infarct ventricular tachycardia requires identification of the infarcted zone, often inconsistently disclosed by voltage mapping or 2-dimensional (2-D) ultrasound. The addition of multi-modality, 4-D dynamic strain rate imaging to identify an infarct and its border zone has the potential to improve ablative accuracy.
Methods: Five dogs underwent infarct creation confirmed by delayed enhancement (DE) magnetic resonance imaging (DE-MRI). In all dogs (> 1 month after infarct), left ventricular (LV) endo-chamber volumes were rendered from 2-D phased array intracardiac ultrasound images. Scar was identified using 4 methods:
thinning, hyperechoicity, and wall motion abnormality by 2-D ultrasound,
strain rate imaging,
voltage mapping, and
Low strain rate < 20% in combination with low voltage < 4mV was used to direct ablation along the infarct border zone. In a 12-segment model of the LV, the degree of fit between each modality and gross pathology was determined.
Results: Global left ventricular myocardial strain rate varied from 4% to 85% with normal values > 20%. Infarcted tissue demonstrated systolic strain rates of 4 –25% (mean nadir 10.3±2.3%). Tissue adjacent to border zones demonstrated a relative step-up in strain by > 30%, the degree of which was location-dependent. Noninfarcted areas of low strain included the basal left entricle at the level of the mitral annulus. On a 0 – 4 degree of fit scale (0 = no fit, 1 = < 50%, 2 = 50 –70%, 3 = 70 –90%, 4 = > 90% fit with scar on gross pathology), the mean degree of fit was 4.0±0.0 by DE-MRI > 2.2±0.8 by strain = 2.0±0.8 by 2-D ultrasound > 1.4±0.5 by voltage. In all dogs, any complementary imaging modality provided better predictive accuracy for scar than voltage mapping. In 2/5 dogs, scar assessment by strain was more accurate than 2-D ultrasound or voltage mapping. Integration of multi-modality data resulted in ablation accuracy of 3.0±0.7 by degree of fit.
Conclusion: In ventricular scar mapping, direct imaging provides more robust scar localization than voltage mapping alone. While DE-MRI is an optimal, complementary modality, it is not always possible. In these cases, dynamic strain rate imaging can facilitate scar border zone detection and ablation.