Date of Award

2026

Document Type

Thesis

Degree Name

Master of Science in Artificial Intelligence

Department

Computer Science

Committee Chair and Members

Nicolas Gallo

Keywords

APAP hepatotoxicity, Computational fluid dynamics, Digital twins, Drug transport modelling, Hepatic microcirculation, Virtual organ simulations

Abstract

The liver’s highly structured vascular microarchitecture governs blood perfusion, metabolic zonation, and the spatial distribution of xenobiotic toxicity. Current computational models of hepatic drug metabolism often oversimplify this geometry, limiting their ability to capture realistic flow dynamics and cellular injury patterns. This study develops a multiscale computational framework to predict acetaminophen-induced hepatotoxicity using image-derived, three-dimensional hepatic lobule geometries. The model integrates computational fluid dynamics (CFD) with a mechanistic cellular injury module to simulate the interplay between perfusion, metabolism, and hepatocellular viability.

Realistic vascular reconstruction was achieved from histopathology liver slices, and the resulting geometry was meshed and solved using ANSYS Fluent. Blood was modeled as an incompressible Newtonian fluid with constant viscosity flowing through a porous parenchymal domain, while acetaminophen metabolism kinetics followed the Mechanistic Acetaminophen Liver Damage (MALD) structure. The model reproduces physiologically plausible centrilobular patterns of NAPQI accumulation and GSH depletion, consistent with known enzyme zonation and oxygen gradients in the hepatic lobule. Spatial predictions of hepatocyte injury aligned with experimental trends, demonstrating centrilobular necrosis consistent with clinical observations.

This work contributes to the Virtual Hepatic Lobule by providing a modular and physiologically grounded framework for virtual liver modeling. The integration of image-based geometry and mechanistic biochemistry establishes a pathway toward a predictive, patient-specific “digital liver” capable of simulating blood flow, metabolism, and drug-induced cellular injury.

Share

COinS