A virtual population cohort approach for fetal cardiac valve modeling

Mathematical models of fetal cardiovascular physiology provide valuable insights when studying the fetal circulatory system. In 0D and 1D models, fetal cardiac valves are often represented as diodes, offering simplicity and scalability but failing to capture realistic valvular behavior and can result in unrealistic pressure drops. More accurate models based on the Bernoulli equation capture valvular dynamic behavior more realistically, but they require constant tuning for specific cases, challenging simulation of fetal cardiac growth. This study introduces a virtual population cohort approach informed by Bayesian inference as a solution to this challenge. By applying this method to a standardized aortic valve model of a 40-week-old fetus, it demonstrates its effectiveness in identifying input parameter distributions that reflect healthy fetal aortic valve behavior. The approach involves defining a template model and determining an appropriate parameter space to simulate physiological behavior. Bayesian inference method facilitates identification of these parameters, resulting in a virtual population cohort that closely represents real physiological relevant fetal aortic valve conditions. The findings show that this approach successfully identifies a virtual population cohort of the fetal aortic valve model, including uncertainty of model parameters and their correlations with model outcomes. This approach offers a widely applicable framework with potential for models that can adapt to the evolving physiological conditions of fetal growth.