Introduction:
Modern endocardial contact-mapping provides high-resolution electro-anatomical (EA) maps of the left atrium (LA), thus allowing to display myocardial substrate based on impaired signal amplitude. Previous methods for characterizing LA myocardium rely on arbitrary thresholds and manual tracing of low-voltage areas (LVA). The amplitude and distribution of signals above these thresholds remain unexamined. The aim of this study is to develop novel parameters for characterizing LA myocardium and LA cardiomyopathy (AC) based on the analysis of the full distribution of signal amplitudes across the entire LA.
Methods:
We retrospectively identified 50 adult patients who underwent primary pulmonary vein isolation (PVI) for paroxysmal or persistent AF between 08/22 and 05/23 fulfilling the selection criteria: i) EA mapping with the multipolar mapping catheter Octaray®; ii) acquisition of voltage maps in sinus rhythm (SR) with ≥5000 points/map; iii) transthoracic echocardiography acquired in SR ≤48 hours before PVI. Exclusion criterion was previous LA ablation. The pulmonary veins were manually excised from the EA maps and the signal amplitude (mV) of each point was plotted. We calculated total LVA applying the voltage threshold <1.0 mV. A novel set of variables describing the voltage distribution across the entire LA (mean, median, variance, inter quartile range (IQR), skewness, kurtosis) was analyzed. Correlations with total LVA and independent AC markers (NT-proBNP, CHA₂DS₂-VA score) were assessed. Differences in variables between previously defined and published groups mild and severe AC were calculated. Clustering will be applied to voltage distribution variables and the predictive performance of the variables will be tested using a machine-learning approach.
Results:
The mean age of the studied sample (n=50) was 63±11 years, 62% were men, 64% showed persistent AF, median CHA2DS2-VA score was 1.5 (1, 3) and NT-proBNP was 190 (71, 391) pg/ml. A median of 5771 (5217, 6988) points/map were acquired. In this all-comer PVI cohort, mean resp. median of signal amplitude (SA) was 2.4 ± 0.96 mV resp. 2.3 (1.5-2.8) mV, variance was 2.5 (1.3-3.1), IQR was 1.7 (1.4-2), skewness was 1.3 (1-1.7) and kurtosis was 5.0 (4-7.4). Variables did not differ between gender and subtypes of AF. The distinct characteristics of the variables were clearly associated with low resp. high LVA values. Strongest correlations with NT-proBNP, CHA₂DS₂-VA and LVA were found for the variables mean SA (-0.61, -0.64, -0.89), median SA (-0.58, -0.62, -0.87), SA variance (-0.58, -0.51, -0.72) and SA IQR (-0.55, -0.55, -0,84). All variables tested differed between previously defined groups mild and severe AC: mean SA (3.1 vs 1.5, p<0.0001), median SA (2.8 vs 1.3, p<0.0001), SA variance (2.9 vs 1.3, p<0.0001), SA IQR (2 vs 1.3, p<0.0001), SA skewness (1 vs 1.9, p<0.0001) and SA kurtosis (4.1 vs 8.6, p<0.0001). Results from clustering and the predictive performance of the variables will be available at the congress.
Conclusion:
For the first time, we identified variables to characterize LA myocardium not only based on the full distribution of signal amplitudes of the LA and but also avoiding arbitrary thresholds. The variables differed significantly between groups mild and severe AC. This novel approach could form the basis for an automated, investigator-independent analysis of LA myocardium and characterization of AC.
