Towards personalization of 1D blood flow models: sensitivity analysis and parameter calibration

Area of Science:

• Computational hemodynamics
• Biomedical engineering
• Cardiovascular modeling

Background:

• Individualizing cardiovascular models is crucial for personalized medicine but challenging with limited patient data.
• High-dimensional parameter spaces in computational hemodynamics hinder accurate model personalization.

Purpose of the Study:

• To develop and validate a method for constructing patient-specific 1D blood flow models from scarce data.
• To address the challenge of parameter space exploration in cardiovascular model personalization.

Main Methods:

• Utilized a simplified Anatomically Detailed Arterial Network model for systemic circulation.
• Employed Covariance Matrix Adaptation Evolutionary Strategy (CMA-ES) for parameter calibration.
• Conducted global sensitivity analysis to understand parameter influence.

Main Results:

• Demonstrated the feasibility of accurate patient-specific model derivation with limited data.
• Quantified the impact of physical parameters on waveform characteristics.
• Identified potential impacts on clinically relevant biomarkers.

Conclusions:

• The proposed approach enables effective personalization of cardiovascular models in data-scarce environments.
• Provides insights into parameter sensitivity and its clinical relevance for hemodynamics.