Feasibility of preprocedural LAA occluder sizing

Background Implantation of occluders in the left atrial appendage (LAA) has been established to prevent patients with atrial fibrillation (AF) from stroke and embolisation. Currently, the individual occluder sizing will be done intraprocedurally using fluoroscopy and/or transesophageal echo (TEE) imaging. Purpose Development of a new approach allowing preprocedural occluder sizing and its validation by conventional intraprocedural methods. Methods The CARTO electro-anatomical mapping system was used in 62 consecutive patients (40 m, 22 f, age: 69.8±8.4 years) to preprocedurally predict the size of Watchman (n=34) or Amplatzer Amulet (n=28) devices. During device implantation, 23 of the patients (37.1%) were in AF. LAA morphologies were Chicken Wing (40.0%), Cauliflower (31.7%), Windsock (20.0%) and Cactus (8.3%). CHA2DS2-VASc Score was 3.7±1.9 and HAS-BLED-Score was 3.6±1.3. Preprocedurally, routine cardiac CT scans were imported into the CARTO segmentation tool for segmentation of the left atrium. The resulting volume images (VI) and slice images (SI) were adjusted in their 3-dimensional orientation to fit the manufacturer's sizing recommendations. Subsequently, the match between pre- and intraprocedural sizing was compared with the actually implanted device size. Results In the Watchman group, preprocedural VI corresponded with 38.1% (8/21) and SI with 42.3% (11/26), while intraprocedural fluoroscopy corresponded with 29.6% (8/27) and TEE with 40.0% (12/30) of all actually implanted devices. According to clinical routine an aberration of one size is common practice. After including one additional size to the estimation, VI corresponded with 81.0% (17/21), SI with 84.6% (22/26), while fluoroscopy corresponded with 85.2% (23/27) and TEE with 80.0% (24/30) of all actually implanted occluders. In the Amplatzer Amulet group, VI corresponded with 37.0% (10/27), SI with 57.1% (16/28) and fluoroscopy with 71.4% (20/28) of all actually implanted devices. After including one additional size to the estimation, VI corresponded with 66.7% (18/27), SI with 85.7% (24/28) and fluoroscopy with 100% (28/28) of all actually implanted occluders. The remaining were selected empirically. Conclusion In this study cohort of 62 patients, feasibility of a preprocedural sizing using the CARTO segmentation tool was demonstrated for Watchman occluders providing results comparable with conventional intraprocedural fluoroscopy and/or TEE based intraprocedural sizing. Possibly due to the different manufacturer measuring recommendation, 10 to 12 mm away from the LAA ostium, results for the Amplatzer Amulet system were found to be less accurate. Nevertheless, the proposed approach offers the advantage of preoperative visualisation of the individual LAA anatomy and morphology. Funding Acknowledgement Type of funding sources: None.