The left atrial appendage (LAA) plays a key role in thrombus formation in patients with atrial fibrillation (AF) due to its unique anatomy and low blood flow velocity. Computational fluid dynamics (CFD) is used to evaluate LAA hemodynamics and assess thrombotic risk, but traditional approaches rely on rigid-wall models or fluid-structure interaction methods. The latter are computationally expensive and require left atrial wall mechanical properties that are difficult to obtain. This study introduces a technique that combines ECG-gated computed tomography (CT) and mesh morphing to construct patient-specific CFD models with dynamic wall motion. Using CT data from two subjects, 3D models of the left atrium (LA) were extracted for ten phases of the cardiac cycle. A reference-phase mesh was used as a template to register all other phase models, maintaining consistent node count and connectivity throughout the cardiac cycle. The wall displacement, obtained by temporal interpolation of the node coordinates extracted from the isotopological meshes, was employed to set up the moving wall CFD simulations. The results were then compared with those generated by rigid-wall simulations. The proposed method allowed the generation of deforming volumetric meshes that follow the LA wall motion while preserving grid topology and overall mesh quality. Simulations using moving walls showed notable differences in flow patterns and hemodynamic indices compared to rigid-wall models. These results highlight how different modeling strategies can substantially influence simulated LAA hemodynamics. The observed discrepancies underscore the complexity of the problem and motivate further investigation, including the use of more advanced modeling frameworks and synthetic benchmarks.