Personalized Fluoroscopic Angles in Watchman™ Left Atrial Appendage Closure Landing Zone Assessment: A Three-Dimensional Printed Simulation Study

Abstract

Background Atrial fibrillation causes ischemic stroke when thrombi dislodge from a cardiac outpouching, the left atrial appendage (LAA), and embolize to the brain. LAA occlusion with the Watchman™ device (Boston Scientific Corporation, MA, USA), which prevents stroke, requires accurate LAA measurements for device sizing. We explore whether standard fluoroscopic LAA measurements improve when obtained at CT-derived viewing angles personalized to LAA anatomy while concurrently referring to three-dimensional (3D) CT. Methods Left atrial 3D reconstructions created from contrast CT (n=28) were analysed to identify personalized viewing angles wherein LAA dimensions (LAA maximum landing zone diameter and LAA length) were best observed. The 3D-CT reconstructions were then 3D printed with stands. Fluoroscopy of anatomically oriented models in the catheter lab simulated LAA angiography. Fluoroscopic images were acquired at standard (caudal 20˚/right anterior oblique 30˚) and personalized viewing angles. Repeated measurements of LAA dimensions were taken from CT (Control), fluoroscopy at standard angles (Standard), personalized angles (Blinded), and personalized angles while concurrently referring to 3D CT (Referred). Results Control measurements correlated and agreed better with Referred and Blinded measurements than with Standard measurements (diameter correlation and agreement: Control/Standard r=.554, limits of agreement [LOAs]=6.83/-5.91; Control/Blinded r=.641, LOA =5.67/-5.54; Control/Referred r=.741, LOA=4.69/-4.14; length correlation and agreement: Control/Standard rs=.829, LOA=9.61/-3.02; Control/Blinded rs=0.789, LOA=7.13/-4.94; Control/Referred rs=.907, LOA=4.84/-4.13). Personalized angles resulted in hypothetical device size predictions more consistent with Control (device size correlation: Control/Standard rs=.698, Control/Blinded rs=.731, Control/Referred rs=.893, P<0.001). False ineligibility rates were Standard=6/28, Blinded=6/28, and Referred=2/28. Conclusion This simulation suggests that personalized fluoroscopic viewing angles with in-procedural reference to 3D CT may improve the accuracy of LAA maximum landing zone diameter and length measurements at the Watchman landing zone. This improvement may result in more consistent device size selection and procedural eligibility assessment. Further clinical research on these interventions is merited.