Back to search

University of Minnesota

Validation and Application of a Liver Hemodynamic Model for Cirrhosis, Resection, and Transplantation

Abstract

dc:description.abstract

The liver performs essential filtration, metabolic, detoxification, and synthetic functions that depend strongly on its unique dual-inflow vascular structure. Blood enters through the portal vein and hepatic artery, passes through the intrahepatic vasculature and sinusoidal channels (capillary-like microvessels), and returns through the hepatic veins to the vena cava. During cirrhosis, liver fibrosis distorts and narrows the sinusoids, leading to increased hydraulic resistance that elevates portal vein pressure and may lead to recanalization of vessels that bypass the liver, known as collaterals. Similarly, reduction of functional liver mass during resection and transplantation decreases the available intrahepatic vascular bed, altering flow distribution and resistance in ways that can elevate portal loading and increase sinusoidal pressure and wall shear stress in the remnant liver or graft. Motivated by these distinct but related hemodynamic disturbances, this dissertation develops a computationally efficient lumped-parameter modeling approach to study liver hemodynamics across cirrhosis, liver resection, and living donor liver transplantation. A central challenge in patient-specific hepatic flow modeling is that key vascular parameters, including diameters of the smallest microvessels that feed into the sinusoids and the effective resistance of the sinusoidal network, are not directly measurable even though they strongly control liver resistance, pressure gradients, and wall shear stress. To address this, the dissertation introduces a calibration methodology based on Bayesian optimization, using literature-based normal physiological reference values, normothermic machine perfusion (NMP) measurements from resected porcine livers, and clinical hemodynamic data from living donor liver transplantations (LDLT) recipients to infer case-specific parameter sets under physiological constraints. The model is further extended by incorporating diameter- and hematocrit-dependent apparent viscosity, revised boundary-condition formulations for cirrhotic and post-resection states, segment-wise sinusoidal remodeling for cirrhosis, and clinically relevant resection configurations. Applied to cirrhosis, the model predicts that progressive sinusoidal narrowing and increasing heterogeneity redistribute blood through an increasingly small subset of high-conductance sinusoidal pathways, while collateral dilation diverts a growing fraction of total flow away from the liver and contributes to portal hypertension. When validated against NMP measurements from 15 resected porcine livers, the model reproduces measured portal inflow and hepatic venous outflow with good overall agreement and enables estimation of sinusoidal wall shear stress and pressure gradients that cannot be measured directly. Across the nine LDLT cases studied, the patient-specific simulations reproduce recipient portal and arterial hemodynamics with overall good agreement and predict that graft resection consistently shifts the transplanted liver toward a higher sinusoidal loading state relative to the intact donor liver. Beyond producing fitted case-specific simulations, the dissertation also examines feasible parameter sets to assess model robustness and output sensitivity. This analysis shows that the available data more clearly identify the overall vascular resistance required to match each case. However, they do not uniquely determine how that resistance is distributed among the portal, hepatic venous, and sinusoidal vessels. Overall, this dissertation establishes a physiologically grounded and computationally tractable model-based approach for linking clinically accessible measurements to otherwise inaccessible sinusoidal-scale hemodynamics, providing a basis for future patient-specific graft evaluation, risk assessment, surgical planning, and experimental testing in perfusion or animal-model settings.

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Saini, Himanshi

Subjects

dc:subject × 6

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Citation
Saini, Himanshi. (2026). Validation and Application of a Liver Hemodynamic Model for Cirrhosis, Resection, and Transplantation. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/281809.

Chain of custody

source
Harvested from
University of Minnesota
Base URL
conservancy.umn.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Saini, Himanshi. Validation and Application of a Liver Hemodynamic Model for Cirrhosis, Resection, and Transplantation. 2026. https://hdl.handle.net/11299/281809