Catheter Ablation -> Ventricular Arrhythmias -> Clinical Trials / Outcomes D-PO05 - Poster Session V (ID 39) Poster

D-PO05-193 - Target Selection For Noninvasive Cardiac Radioablation Using A 17-segment Model Structured Approach (ID 1499)

 P. Samson: Nothing relevant to disclose.


Background: Noninvasive cardiac radioablation has been used for treatment refractory VT. Noninvasive target selection is new for electrophysiologists (EPs) and requires synthesis of multiple cardiac images detailing the electrophysiologic, anatomic, and metabolic properties of the heart.
Objective: To examine variability in target selection and assess agreement when employing a structured approach using a 17-segment LV model.
Methods: During the Symposium for Noninvasive Radioablation (SNORAD) in 2019, EPs were given 2 cases for analysis. Participants selected virtual targets using a 17-segment model (Figure). Pre-test targets from clinically available data at the time of presentation were selected first (12-lead ECGs of VT and previous endocardial voltage maps), and then serial target selection was done with 1) ECGI of induced VT, 2) anatomic imaging (CT/MRI) and 3) metabolic imaging (PET/SPECT). A final post-test target was then selected once the full data set had been reviewed.
Results: There were 32 responses for Case 1 and 34 responses for Case 2. Overall agreement for the single most commonly selected segment was 47% for Case 1 and 65% for Case 2 for the pretest phase. For the post-test phase, that agreement increased to 75% and 95% for Case 1 and Case 2, respectively. Of the three imaging types, anatomic imaging had the highest agreement (73% and 97%). Metabolic imaging had the fewest number of segments chosen (mean 2.7 and 2.8). Anatomic imaging had most segments chosen (4.3 and 4.0).
Conclusion: A structured approach to noninvasive targeting using a 17-segment model improves inter-observer agreement. Future trials may minimize variability using this method.