A multi-dimensional CFD framework for fast patient-specific fractional flow reserve prediction

Area of Science:

• Cardiovascular Imaging and Physiology
• Computational Fluid Dynamics
• Medical Diagnostics

Background:

• Fractional flow reserve (FFR) is the gold standard for diagnosing coronary myocardial ischemia.
• Current 3D computational fluid dynamics (CFD) methods for noninvasive FFR prediction using coronary computed tomography angiography (CTA) have limitations in accuracy and efficiency.
• Accurate and efficient noninvasive FFR assessment is crucial for guiding cardiovascular interventions.

Purpose of the Study:

• To develop and validate a multi-dimensional CFD framework for improved noninvasive FFR prediction.
• To enhance the accuracy of FFR prediction by incorporating 0D patient-specific boundary conditions.
• To increase the efficiency of FFR prediction by generating optimized 3D initial conditions.

Main Methods:

• A multi-dimensional CFD model integrating 3D, 1D, and 0D vascular models was developed.
• 0D patient-specific boundary conditions were derived using clinical parameters and an optimization algorithm.
• 1D vascular models were used to evaluate convergence and generate 3D initial conditions for improved efficiency.

Main Results:

• The predicted FFR (FFR_C) showed strong linear correlation (r=0.80, p<0.001) and high consistency with invasive FFR.
• FFR_C demonstrated high diagnostic performance: accuracy 88.5%, sensitivity 93.3%, specificity 83.9%.
• The multi-dimensional CFD framework improved prediction efficiency by 71.3% compared to conventional methods.

Conclusions:

• The proposed multi-dimensional CFD framework significantly enhances both accuracy and efficiency of noninvasive FFR prediction.
• This approach offers a promising tool for noninvasive diagnosis of coronary myocardial ischemia.
• The method has the potential to improve clinical decision-making in cardiology.