Effect of Wall Motion Sampling on CFD-Derived Left Atrial Flow Metrics

Area of Science:

• Cardiovascular Imaging and Simulation
• Medical Physics
• Computational Fluid Dynamics

Background:

• Limited temporal resolution in medical imaging (10-20 frames/cycle) necessitates interpolation for patient-specific computational fluid dynamics (CFD) simulations.
• The impact of this temporal interpolation on hemodynamic predictions, particularly in the left atrium (LA), remains undercharacterized.

Purpose of the Study:

• To systematically investigate the effect of temporal undersampling on hemodynamic predictions in patient-specific LA models.
• To determine the minimum required temporal resolution for accurate assessment of blood stasis in the left atrial appendage (LAA).

Main Methods:

• Generated high-temporal-resolution LA wall motion data using electromechanical simulations.
• Created downsampled datasets (10, 20, 40 frames/cycle) to emulate clinical imaging.
• Performed CFD simulations using an immersed boundary method and passive scalar transport to compute blood residence time (TR).

Main Results:

• Global LA hemodynamic indices showed <15% error with varying frame rates.
• Left atrial appendage (LAA) residence time (TR) was significantly affected by temporal undersampling, with 10-frame data overestimating TR by up to 43%.
• 10-frame reconstructions showed elevated stasis and missed physiological features; 20-40 frames/cycle yielded errors ≤9%.

Conclusions:

• Temporal undersampling significantly impacts LAA blood stasis metrics, with 10 frames/cycle being unreliable.
• Reconstructions with ≥20 frames/cycle provide sufficiently reliable LAA blood stasis indices for clinical use.
• Lower temporal resolutions require cautious interpretation for patient stratification and accurate ranking.