A reduced order model formulation for left atrium flow: an atrial fibrillation case

Area of Science:

• Biomedical Engineering
• Computational Fluid Dynamics
• Medical Physics

Background:

• Atrial fibrillation (AF) significantly alters blood flow dynamics, leading to potential complications like thrombosis.
• Accurate hemodynamic analysis is crucial for understanding AF-related risks.
• Patient-specific modeling offers personalized insights into disease progression.

Purpose of the Study:

• To develop and validate a data-driven reduced order model (ROM) for patient-specific blood flow analysis in AF.
• To efficiently reconstruct key hemodynamic indices related to blood stasis.
• To compare the performance of the ROM against a full order model (FOM) in terms of accuracy and computational speed.

Main Methods:

• A proper orthogonal decomposition-radial basis function (POD-RBF) approach was used to build the ROM.
• The full order model (FOM) was based on incompressible Navier-Stokes equations discretized using the finite volume (FV) method.
• Both Newtonian and Casson's non-Newtonian constitutive laws were employed to model blood viscosity.

Main Results:

• The POD-RBF ROM accurately reconstructed hemodynamic indices, including those related to blood stasis.
• The model demonstrated significant computational speed-up compared to the FOM.
• The framework allowed for parametric studies including cardiac output, plasma viscosity, and hematocrit.

Conclusions:

• The data-driven POD-RBF ROM provides an efficient and accurate computational tool for patient-specific blood flow analysis in AF.
• This approach facilitates the investigation of hemodynamics under various physiological conditions.
• The developed model has potential applications in personalized medicine and risk assessment for AF patients.