668Principal component analysis can identify ventricular regions with highest variability in the arrhythmogenic substrate of ventricular tachycardia patients. (18th June 2020)

• Record Type: Journal Article
• Authors: Thibault, B; Richer, L P; Relan, J; Mcspadden, L; Ryu, K; Rivard, L; Dyrda, K; Dubuc, M; Mondesert, B; Cadrin-Tourigny, J; Tadros, R; Macle, L; Khairy, P; Gregoire, J; Harel, F
• Is Part Of: Europace. Volume 22(2020)Supplement 1
• Publication Date: 2020-06-18
• Subjects: Arrhythmia -- Treatment -- Periodicals; Cardiac pacing -- Periodicals; Catheter ablation -- Periodicals; Heart -- Physiology -- Periodicals; Electrophysiology -- Periodicals; 617.4120645
• DOI: 10.1093/europace/euaa162.233
• Languages: English
• ISSNs: 1099-5129

Abstract

Abstract: Introduction: Left ventricle (LV) substrate can be characterized by imaging modalities providing physiological variables (tissue perfusion, ischemia, etc.) grouped in a complex dataset. Principal Component Analysis (PCA) can identify the dataset variables that best explain its variance. Variables linked to substrate arrhythmogenicity will lead PCA to identify LV regions most influenced by the variables on a LV 3D geometry. Purpose: To evaluate whether regions identified by PCA correspond to regions targeted by physicians in a ventricular tachycardia (VT) ablation procedure. Methods: Ischemic VT subjects underwent SPECT/CT perfusion imaging (rest and stress) prior to LV voltage mapping with the cardiac mapping system. Co-registration allowed projection of ablation sites onto SPECT/CT geometries. PCA retrospectively analyzed the following dataset: tissue perfusion (rest and stress), tissue ischemia and the local (1cm2 sub-regions) stress-rest difference for: 1) standard deviation (STD) and 2) skewness (Skew). PCA components (PCAc) explaining ≥85% of the total variance were plotted on the LV 3D geometry to display regions highly influenced by the dataset variables (see figure below). Results: Ten subjects (9 males, 66 ± 8 years old, LVEF 37 ± 11%) underwent co-registration. In 7/10 subjects, tissue perfusion (in PCAc#1) and ischemia (in PCAc#2) were most influential on LV regions variance. Ischemia and low perfusion (≤40%) areas equaled 32 ± 19 % of the LV with 21 ± 23% of these areas highly involved in the arrhythmogenic substrate (≥75% of explained LV variance). The location of 63 ± 18% of ablation sites were <1 cm from areas explaining ≥50% of LV variance. Conclusion: Preliminary results showed that PCA can synthesize the influence of different perfusion derived variables, like tissue perfusion and ischemia, used in SPECT/CT imaging of the LV. Further analysis is needed to confirm whether this could be used to select VT ablation targets.