Validation and Calibration of an Electrical Analog Model of Human Liver Perfusion Based on Hypothermic Machine Perfusion Experiments

Abstract

Hypothermic machine perfusion (HMP) is reviving as a better preservation method for donor livers than the golden standard of cold storage, but still faces challenges such as the risk for endothelial damage and flow competition between the arterial and portal venous inflow. Therefore, we previously developed an electrical analog model to investigate the effect of HMP settings on the human liver hemodynamics. While the model provided plausible results, it is based on a number of assumptions and its performance was never subjected to experimental validation. To this end, we present a new methodology to validate and calibrate this model to a specific liver. HMP experiments were performed to capture the perfusion behavior of a human liver during varying perfusion settings. Simultaneous pressure and flow signals were acquired at the hepatic artery, portal vein, and vena cava inferior. The calculation of hydraulic input impedances enabled reduced Windkessel models to be fitted to the global hepatic perfusion properties as an intermediate step. Based on these Windkessel models, the extended electrical analog model was calibrated to the specific available liver. Results revealed that literature values of one of the critical model parameters (wall viscoelasticity) are a few orders of magnitude off, having important consequences for simulated (pulsatile) hemodynamic variables. A novel methodology, based on HMP experiments, signal processing and unconstrained nonlinear optimization was developed to validate and calibrate the liver-specific extended electrical model. Future research may focus on extending this approach to other applications (e.g. liver pathologies such as cirrhosis).