技術論文とプレゼンテーション

Problem and use of COMSOL Multiphysics The left atrium (LA) plays an important role in the normal function of the left side of the heart. The aim of this study is to present a physiologically consistent and computationally efficient LA model for evaluation of LA function. Existing computational models often require highly detailed patient-specific geometries, which increases complexity and computational cost. To overcome this, the study developed a simplified benchmark model of the LA using COMSOL Multiphysics. The LA was represented as a spherical chamber with four pulmonary vein inlets and a simplified mitral valve outlet. COMSOL was used to implement a fluid-structure interaction (FSI) model, coupling blood flow and tissue mechanics through the Arbitrary Lagrangian-Eulerian (ALE) method. In-vivo pulmonary vein flow measured by MRI and physiological material parameters were applied as boundary conditions, allowing the model to simulate realistic LA hemodynamics and wall motion while keeping the geometry simple.

Summary of results The model successfully reproduced several key physiological markers of normal LA function. The simulated LA pressure waveform showed realistic a- and v-peaks, with values of 13.6 mmHg and 12.2 mmHg, respectively, and a mean LA pressure of 6.90 mmHg, all within reported physiological ranges. The volume curve also matched normal atrial behavior: the minimum LA volume was 56 mL and the maximum was 100 mL, both within normal ranges, while the active stroke work (43.5 mL/mmHg) was close to published in-vivo data. Transmitral flow was represented by three peaks corresponding to early filling, mid-diastolic flow, and atrial contraction. The simulated E-peak (0.63 m/s), A-peak (0.43 m/s), and E/A ratio (1.47) were all within physiological limits.

Conclusion The study concludes that a simplified LA geometry implemented in COMSOL Multiphysics can still reproduce major physiological features of LA function when realistic in-vivo boundary conditions are used. Despite its geometric simplifications, the model produced accurate pressure, volume, and transmitral velocity patterns under normal conditions. The overall findings suggest that detailed patient-specific atrial geometry may not always be necessary for clinically relevant LA flow simulations. This makes the model promising as a computationally efficient benchmark and a potential foundation for future patient-specific or disease-related cardiac simulations.