Biological background and modelling workflow.

(A) Model. Gadoxetate, a gadolinium ion (Gd3+)-containing MRI-contrast agent, is taken up by hepatocytes via the OATP1B1/B3 transport proteins. After uptake, gadoxetate is actively excreted into the bile via the MRP2 transport proteins. Gadoxetate can also to some extent be transported back to the blood via MRP3/4 transporters. (B) Modelling workflow. After acquiring MR images before and after gadoxetate injection, the images are processed to get timeseries of gadoxetate concentrations. Finally, the timeseries are combined with pharmacokinetic models in order to estimate the gadoxetate transport rates, which are used as biomarkers for liver function.(C) Model structures for the Patlak, Perfusion, and Whole-Body models. Abbreviations: Hep, hepatocyte compartment; ES, liver extracellular compartment; Plasma, blood plasma compartment; EES, extracellular extravascular compartment; Spleen IS, intracellular spleen compartment; ki, hepatic uptake rate; kef, hepatic efflux rate; kback, Gadoxetate transport rate back to the blood; Fp, total blood flow to the liver; fa, arterial flow fraction; fv, portal venous flow fraction (i.e., 1-fa). (D) Purpose of the study. Previous research has focused on how liver function and gadoxetate pharmacokinetics can be described in humans as well as in rats (yellow arrow; see Figs 3–5). Other research has also studied how liver function can be affected by drugs (blue arrow, see Fig 5). However, the clinically interesting question of how a given drug affects liver function in humans is much more difficult to answer. We therefore here introduce a translational modelling framework in order to predict how a given drug can affect liver function in humans (green arrow, see Fig 6).