Left Atrial Ejection Fraction Assessed by Prior Cardiac CT Predicts Recurrence of Atrial Fibrillation after Pulmonary Vein Isolation

Figures

• 

  Figure 1. Automatic left atrial segmentation (red) as proposed by Wolf et al. [20] (electrophysiology/EP planning tool, Intellispace Portal, Philips, Best, The Netherlands) with additional manual segmentation of the left atrial appendage (LAA) (white). (A,B) demonstrate the automatic segmentation of the left atrium (LA) in atrial systole and diastole. (C,D) demonstrate the 3D surface rendering of the LA and LAA in systole and diastole.

• 

  Figure 2. Logistic regression model for risk of recurrence (%) with left atrial ejection fraction (LAEF) (%) to left atrial volume (LAVmax) (mL) as outcome predictors. For example, a patient with an LAEF of 40% and LAVmax of 100 mL had a risk of recurrence of less than 10% (dark green area). However, a patient with lower values of LAEF, e.g., 10%, and the same value of LAVmax of 100 mL had a high risk of recurrence (>80%). The figure also illustrates that LAEF had more predictive power than LAVmax in this study population.

• 

  Figure 3. Box-plots of LAEF and left atrial appendage ejection fraction (LAAEF) (%) for non-recurrence (group A) vs. AF recurrence (group B), separated into the subgroups paroxysmal and persistent AF.

• 

  Figure 4. Scatterplots of LAAEF (%), left atrial appendage stroke volume (LAASV) (ml), LAEF (%) and LASV (ml) for CT readers 1 and 2. Green lines indicate the agreement of both readers; red solid lines indicate regression lines with 95% confidence limits (dotted lines).