Population-based priors in cardiac model personalisation for consistent parameter estimation in heterogeneous...

Area of Science:

• Computational Biology
• Biomedical Engineering
• Cardiovascular Research

Background:

• Personalised cardiac models require accurate parameter values for clinical applications.
• Limited patient data restricts unique parameter estimation in complex cardiac models.
• Current methods face obstacles in achieving precise parameter identification.

Purpose of the Study:

• To introduce and evaluate the Iteratively Updated Priors (IUP) algorithm for personalised cardiac model parameter estimation.
• To address the challenge of non-unique parameter identification in personalised cardiac models.
• To demonstrate the utility of IUP for handling missing or varied clinical measurements.

Main Methods:

• Developed an algorithm, Iteratively Updated Priors (IUP), for successive personalization of cardiac models.
• Employed maximum a posteriori (MAP) estimation, using prior distributions from previous iterations.
• Validated the approach on a heterogeneous database of 811 patient cardiac models.

Main Results:

• The IUP algorithm converges to a reduced-dimension linear subspace for parameter estimation.
• Identified a relevant parameter subspace for effective model personalization.
• Demonstrated consistent parameter estimation even with missing or varied clinical data.

Conclusions:

• The IUP algorithm provides a robust method for unique parameter estimation in personalised cardiac models.
• Leveraging population statistics within a reduced parameter subspace enhances model accuracy and consistency.
• This approach facilitates improved clinical applicability of personalised cardiac modeling.