Clin Res Cardiol (2021). 10.1007/s00392-021-01933-9

3D left atrial volume quantification by cardiac magnetic resonance as a predictor of left atrial substrate during electroanatomic mapping
T. R. Schmidt1, S. Ulbrich2, M. Wagner2, T. Gaspar2, A. Linke1
1Klinik für Innere Medizin, Kardiologie und Intensivmedizin, Herzzentrum Dresden GmbH an der TU Dresden, Dresden; 2Rhythmologie, Herzzentrum Dresden GmbH an der TU Dresden, Dresden;
Introduction:
There is growing evidence that atrial fibrillation (AF) is the result of atrial remodeling caused by an inflammatory reaction and infiltration leading to fibrotic wall changes. Left atrial (LA) dilatation represents part of the remodeling process. Estimation of LA volume from cardiac magnetic resonance (CMR) cine images before recording of an electroanatomic map could be used for prediction of low voltage areas (arrhythmogenic substrate) in the LA wall. Knowledge of left atrial substrate probability can contribute to improved individualized treatment.
 
Objective:
To establish whether 3D-LA angiography derived volume of a preprocedural CMR scan predicts low voltage zones (LVZ).
 
Methods:
A prospective observational study is performed, where patients at University Heart Center Dresden admitted for primary catheter ablation of AF underwent a routine preprocedural CMR for ablation planning. During CMR a LA angiography using Gadolinium contrast was performed to create a 3D LA volume model. Dedicated software (ADAS 3D® and 3D Slicer) was used for 3D model creation and volume quantification. In the subsequent pulmonary vein isolation procedure, an electrophysiologic study with electroanatomic mapping (CARTO® or NAVX® software) of the LA was completed, to asses for low voltage areas (defined as point measurements <0,5mV) on the LA wall. Computed variables include total LVZ area [cm²] and the number blocks containing low voltage areas after division of each LA wall segment (inferior, anterior, septal, lateral, posterior) into 9 blocks. IBM SPSS® was used for statistical analysis.
 
Results:
Complete data was collected for 49 patients (N=49), 57,1% males, mean age 68,9 years ±9,124, median BMI 27.5 (IQR 5,7) and mean BSA 2m² ±0.25. LVZ were present in 34,7% patients (n=17). Collected variables with normal distribution (as assed by Shapiro-Wilk-Test, p >0,05) include BSA, 3D-LA Volume. Variables with non-normal distribution include 3D-LA Volume/BSA, LVZ Area, number of LVZ blocks. Mean 3D-LA Volume in LVZ negative and LVZ positive groups was 121.06ml ±32.92 n=32 and 149.59ml ±33.97 n=17 respectively. There was a statistically significant difference in mean LA Volume, which was 28.53ml (CI 95%-[49.08,7.98]) lower in LVZ negative patients, t(31.853)=2.828, p<0.05 using Welch test.
Correlation analysis for variables with non-normal distribution was performed using Spearman-correlations coefficient [ρ], with a moderate correlation between LA Volume and LVZ Area (ρ=.386, p<.05, N=44) and number of LVZ blocks (ρ=.402, p<.05, N=46). A strong correlation is reported for LA Volume/BSA with LVZ Area (ρ=.493, p=.01, N=44) and number of LVZ blocks (ρ=.534, p<.001, N=46). A binomial logistic regression analysis was used to determine the effect of BSA and LA volume to predict the presence of LVZ. The regression model was statistically significant, χ²(2) = 19.945, p <.001. Overall percentage of accuracy in classification was 77.6%, with a sensitivity of 64.7% and a specificity of 84.4%.
 
Summary:
3D-left atrial modelling and volume quantification from CMR data can routinely be performed with dedicated software. A strong correlation of LA Volume/BSA with the total size of low voltage zones presence in the left atrium can be found. Additionally, LA Volume together with BSA can be used to predict the presence of low voltage zones during electroanatomic mapping.
 













LA wall segment subdivision on the 3D Model

https://dgk.org/kongress_programme/ht2021/P775.htm