Sensitivity analysis and inverse uncertainty quantification for the left ventricular passive mechanics.

Area of Science:

• Computational mechanics
• Cardiovascular physiology
• Biomaterials science

Background:

• Personalized computational cardiac models integrate physiology and mechanics for improved cardiac patient care.
• Quantifying parameter uncertainties is crucial for clinical decision support using these models.
• The Holzapfel-Ogden model is a widely used constitutive law for passive myocardium.

Purpose of the Study:

• To address challenges in quantifying parameter perturbations and estimation uncertainty in personalized cardiac models.
• To apply global sensitivity analysis (SA) and inverse uncertainty quantification (I-UQ) to the Holzapfel-Ogden model.
• To improve the reliability and clinical applicability of computational cardiac models.

Main Methods:

• Global sensitivity analysis (SA) using Gaussian process (GP) surrogate models to assess parameter impact.
• Inverse uncertainty quantification (I-UQ) with GP surrogate models for parameter estimation uncertainty.
• Bayesian inference via Markov Chain Monte Carlo (MCMC) for full uncertainty quantification.

Main Results:

• SA identified key parameters influencing left ventricle (LV) model predictions.
• I-UQ quantified the uncertainty in material parameter estimates from noisy data.
• The study revealed relationships between SA, I-UQ, and external factors like LV cavity pressure.

Conclusions:

• The study provides a framework for robust uncertainty quantification in cardiac models.
• Findings offer insights into cardio-mechanic model formulation and the Holzapfel-Ogden model.
• Improved model reliability can enhance risk prediction and therapeutic strategies in cardiology.