The morphology of the LAA is extremely heterogeneous. Acurate sizing of the landing zone is crucial for the optimal choice of the device diameter. Cardiac CT has been shown to provide acurate measurement of the LAA. TOE and selective LAA angiography can provide discordant results during the procedure, and there is no consensus about the most reliable imaging modality.

Compare the sizing of the LAA using a multimodality imaging approach.

We retrospectively included 45 patients (aged 67.6±6.2 years) who underwent LAA occlusion using the Amplatzer® cardiac plug (ACP). Mean diameter of the landing zone was determined using the different imaging modalities: MPR for cardiac-CT and 3D TOE, multiple angle view at 0, 45, 90 and 120° using 2D TOE, and monoplanar projection (RAO30° cranial 20°) for angiography. High intra and inter observer reproductibility were found for diameter (r=0.93 and 0.87, p<0.01 respectively with mean difference of –0.1mm [–1.9;+1.8] and –0.2mm [–2.9;+2.5] respectively) and area measurement (r 0.94 and 0.88, p<0.01 respectively, with a mean difference of 0cm² [–0.15;+0.16] and 0cm² [–0.9;+0.9], respectively)of the landing zone by CT. 2D TEE mean diameter was underestimated compared to CT (–2.6mm [–3.1;+8.3]) that was reduced to –1.8mm [–3.6;+ 7.1] when a 120° view was available. 3D TEE also underestimated LAA diameter (-2mm [–2.7;+6.7]). Angiographic diameter was weakly correlated with CT (r=0.35, –1.4mm [–5.7;+8.5]).

We found significant variability in the sizing of the LAA landing zone, using different imaging modalities. 3D data set provided by Cardiac-CT and 3D TOE looks interesting to get a complete overview of LAA anatomy. Multiple angle view from 0 to 120° are mandatory when using 2D TOE. Monoplanar LAA selective angiography seems not acurate for LAA sizing and may be avoided when non invasive imaging modalities are available.