Compartmental analysis of nuclear imaging data for the quantification of FDG liver metabolism

TL;DR

This study develops a compartmental modeling approach combined with optimization techniques to quantify FDG liver metabolism from nuclear imaging data, validated with synthetic data simulations.

Contribution

It introduces a novel compartmental model for liver glucose metabolism integrated with a statistical optimization method for parameter estimation.

Findings

Successful modeling of glucose transport in liver

Effective parameter estimation using Ant Colony Optimization

Validation with synthetic data confirms model accuracy

Abstract

This paper utilizes compartmental analysis and a statistical optimization technique in order to reduce a compartmental model describing the metabolism of labelled glucose in liver. Specifically, we first design a compartmental model for the gut providing as output the tracer concentration in the portal vein. This quantity is then used as one of the two input functions in a compartmental model for the liver. This model, in turn, provides as output the tracer coefficients quantitatively describing the effectiveness with which the labelled glucose is transported between the different compartments. For both models, the computation of the solutions for the inverse problems is performed by means of an Ant Colony Optimization algorithm. The validation of the whole process is realized by means of synthetic data simulated by solving the forward problem of the compartmental system.