peter wriggers thomas lenarz editors biomedical technology
TRANSCRIPT
Lecture Notes in Applied and Computational Mechanics 84
Peter WriggersThomas Lenarz Editors
Biomedical TechnologyModeling, Experiments and Simulation
Lecture Notes in Applied and ComputationalMechanics
Volume 84
Series editors
Peter Wriggers, Leibniz Universität Hannover, Hannover, Germanye-mail: [email protected] Eberhard, University of Stuttgart, Stuttgart, Germanye-mail: [email protected]
About this Series
This series aims to report new developments in applied and computationalmechanics—quickly, informally and at a high level. This includes the fields of fluid,solid and structural mechanics, dynamics and control, and related disciplines. Theapplied methods can be of analytical, numerical and computational nature.
More information about this series at http://www.springer.com/series/4623
Peter Wriggers ⋅ Thomas LenarzEditors
Biomedical TechnologyModeling, Experiments and Simulation
123
EditorsPeter WriggersInstitute of Continuum MechanicsLeibniz Universität HannoverHannoverGermany
Thomas LenarzHals-Nasen-OhrenklinikMedical School HannoverHannoverGermany
ISSN 1613-7736 ISSN 1860-0816 (electronic)Lecture Notes in Applied and Computational MechanicsISBN 978-3-319-59547-4 ISBN 978-3-319-59548-1 (eBook)DOI 10.1007/978-3-319-59548-1
Library of Congress Control Number: 2017941488
© Springer International Publishing AG 2018This book was advertised with a copyright holder “The Editor(s) (if applicable) and The Author(s)” inerror, whereas the publisher holds the copyright.This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or partof the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmissionor information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilarmethodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exempt fromthe relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, express or implied, with respect to the material contained herein orfor any errors or omissions that may have been made. The publisher remains neutral with regard tojurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer NatureThe registered company is Springer International Publishing AGThe registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
A major objective of medical research is related to the development of improvedmedication and implants. Due to the individual anatomy of each human being, theresearch direction points more and more towards a patient specific medicine. This inturn requires a better understanding of biological systems and of the performance ofimplants in humans. In engineering disciplines the application of virtual processdesign has originated many important innovations. Virtual modelling helpsunderstand and control processes. Furthermore, virtual testing is fast and flexible.Hence, many new products can be efficiently designed and verified by numericalapproaches.
In recent years these concepts were successfully applied in the field ofbiomedical technology. Based on the tremendous advances in medical imaging,modern CAD systems, high-performance computing and new experimental testdevices, engineering can provide a refinement of implant design and lead to saferproducts. Computational tools and methods can be applied to predicting the per-formance of medical devices in virtual patients. Physical and animal testing pro-cedures can be reduced by use of virtual prototyping of medical devices. Theseadvancements enhance medical decision processes in many areas of clinicalmedicine.
In this book, scientists from different areas of medicine, engineering and naturalsciences are contributing to the above research areas and ideas. The book provides agood overview of new mathematical models and computational simulations as wellas new experimental tests in the field of biomedical technology.
In the first part of the book the virtual environment is used in studying biologicalsystems at different scales and under multiphysics conditions. Modelling schemesare applied to human brain tissue, blood perfusion and metabolism in the livinghuman, investigation on the effect of mutations on the spectrin molecules in redblood cells and numerical strategies to model transdermal drug delivery systems.
The second part is devoted to modelling and computational approaches in thefield of cardiovascular medicine. The contributions start with an overview of cur-rent methods and challenges in the field of vascular haemodynamics. This is fol-lowed by new methods to accurately predict heart flow with contact between
v
the leaflets, estimation of a suitable zero stress state in arterial fluid structureinteraction, solution strategies for stable partitioned fluid-structure interactionsimulations, methods for stable large eddy simulation of turbulence in cardiovas-cular flow, a demonstration of the importance using non-Newtonian models inspecific hemodynamic cases, a multiscale modelling of artificial textile reinforcedheart valves, and new strategies to reduce the computational cost in fluid-structureinteraction modelling of haemodynamics. The part closes with a method to com-putationally assess the rupture risk of abdominal aortic aneurysm.
A parameter study of biofilm growth based on experimental observations andnumerical test as well as a multiscale modelling approach to dental enamel arecontributions that face current challenges in dentistry.
The part related to orthopaedics starts with an overview of challenges in total hiparthroplasty and is followed by a concept for a personalized orthopaedic traumasurgery based on computational simulations.
The last part addresses otology and shows that an off-the-shelf pressure mea-surement system can be successfully used for intrachochlear sound pressuremeasurements. The second contribution is a user-specific method for the auditorynerve activity, leading to a better understanding of the electrode nerve interface inthe case of cochlear implants.
All contributions highlight the state-of-the-art in biotechnology research andthus provide an extensive overview of this subject.
Hannover, Germany Peter WriggersJanuary 2017 Thomas Lenarz
vi Preface
Contents
Part I Biological Systems
Multiscale Aspects in the Multiphasic Modelling of Human BrainTissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Wolfgang Ehlers and Arndt Wagner
Simulation of Steatosis Zonation in Liver Lobule—AContinuummechanical Bi-Scale, Tri-Phasic, Multi-ComponentApproach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Tim Ricken, Navina Waschinsky and Daniel Werner
Nano-Mechanical Tensile Behavior of the SPTA1 Gene in the Presenceof Hereditary Hemolytic Anemia-Related Point Mutations . . . . . . . . . . . 35Melis Hunt
The Choice of a Performance Indicator of Release in TransdermalDrug Delivery Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Giuseppe Pontrelli and Laurent Simon
Part II Cardiovascular Medicine
Multiscale Multiphysic Approaches in Vascular Hemodynamics . . . . . . . 67Michael Neidlin, Tim A.S. Kaufmann, Ulrich Steinseiferand Thomas Schmitz-Rode
Heart Valve Flow Computation with the Space–Time Slip InterfaceTopology Change (ST-SI-TC) Method and Isogeometric Analysis(IGA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Kenji Takizawa, Tayfun E. Tezduyar, Takuya Teraharaand Takafumi Sasaki
Estimation of Element-Based Zero-Stress State in Arterial FSIComputations with Isogeometric Wall Discretization . . . . . . . . . . . . . . . . 101Kenji Takizawa, Tayfun E. Tezduyar and Takafumi Sasaki
vii
Fluid-Structure Interaction Modeling in 3D Cerebral Arteriesand Aneurysms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Yue Yu
Large-Eddy Simulation of Turbulence in Cardiovascular Flows . . . . . . . 147F. Nicoud, C. Chnafa, J. Siguenza, V. Zmijanovic and S. Mendez
Computational Comparison Between Newtonian and Non-NewtonianBlood Rheologies in Stenotic Vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169Bruno Guerciotti and Christian Vergara
Artificial Textile Reinforced Tubular Aortic Heart Valves—Multi-scale Modelling and Experimental Validation . . . . . . . . . . . . . . . . . . . . . . 185Deepanshu Sodhani, R. Varun Raj, Jaan Simon, Stefanie Reese,Ricardo Moreira, Valentine Gesché, Stefan Jockenhoevel, Petra Mela,Bertram Stier and Scott E. Stapleton
Preliminary Monolithic Fluid Structure Interaction Model forVentricle Contraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217D. Cerroni, D. Giommi, S. Manservisi and F. Mengini
The Biomechanical Rupture Risk Assessment of Abdominal AorticAneurysms—Method and Clinical Relevance . . . . . . . . . . . . . . . . . . . . . . 233T. Christian Gasser
Part III Dentistry
A Deeper Insight of a Multi-dimensional Continuum BiofilmGrowth Model: Experimental Observation and ParameterStudies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257Dianlei Feng, Henryke Rath, Insa Neuweiler, Nico Stumpp,Udo Nackenhorst and Meike Stiesch
Multiscale Experimental and Computational Investigation of Nature’sDesign Principle of Hierarchies in Dental Enamel . . . . . . . . . . . . . . . . . . 273Songyun Ma, Ingo Scheider, Ezgi D. Yilmaz, Gerold A. Schneiderand Swantje Bargmann
Part IV Orthopaedics
Challenges in Total Hip Arthroplasty . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295Gabriela von Lewinski and Thilo Floerkemeier
Personalized Orthopedic Trauma Surgery by Applied ClinicalMechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313M. Roland, T. Tjardes, T. Dahmen, P. Slusallek, B. Bouillonand S. Diebels
viii Contents
Part V Otology
Measurement of Intracochlear Pressure Differences in HumanTemporal Bones Using an Off-the-Shelf Pressure Sensor . . . . . . . . . . . . . 335Martin Grossöhmichen, Rolf Salcher, Thomas Lenarz and Hannes Maier
Development of a Parametric Model of the Electrically StimulatedAuditory Nerve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349Waldo Nogueira and Go Ashida
Contents ix
Part IBiological Systems
Multiscale Aspects in the MultiphasicModelling of Human Brain Tissue
Wolfgang Ehlers and Arndt Wagner
Abstract The enormous microscopic complexity of the multicomponent brain-
tissue aggregate motivates the application of the well-known Theory of Porous
Media (TPM). Basically, a quaternary TPM-model is applied to brain tissue, cf.
Ehlers and Wagner, Comput Method Biomech Biomed Eng 18:861–879 (2015), [1].
Besides the model’s broad range of application, such as the simulation of brain-
tumour treatment, we focus in this article on a specific anatomical property of the
brain-tissue aggregate. Namely, its separated pore space which is concurrently per-
fused by two pore liquids. These are the blood in the blood-vessel system and the
interstitial fluid in the interstitial-fluid space. In this regard, the constitutive formu-
lation of evolving liquid saturations under certain loading conditions needs to be
found. In order to microscopically underlay and motivate such a macroscopic consti-
tutive relation within a thermodynamically consistent TPM approach, a microscopic
study of the interaction between the pore compartments is performed and discussed
in terms of scale-bridging aspects.
1 Introduction and Motivation
For various medical purposes regarding brain tissues, it could be of immense benefit
to use numerical simulations. The aim to simulate real applications on the length
scale of centimetres to metres requests a macroscopic modelling, for example, in
case of the simulation of a therapeutic infusion process within brain tissue during
brain-tumour treatment as it is shown in Fig. 1. Obviously, processes occurring on
the complex microscale have a certain impact on the overall behaviour and need
to be included. One example, on which we will focus in the present article, is the
stiffness of the blood vessel system in comparison to the surrounding tissue aggre-
gate (under physiological pressure conditions). This issue is still under controversial
W. Ehlers (✉) ⋅ A. Wagner
Institute of Applied Mechanics (CE), University of Stuttgart,
Pfaffenwaldring 7, 70569 Stuttgart, Germany
e-mail: [email protected]
A. Wagner
e-mail: [email protected]
© Springer International Publishing AG 2018
P. Wriggers and T. Lenarz (eds.), Biomedical Technology, Lecture Notes in Applied
and Computational Mechanics 84, DOI 10.1007/978-3-319-59548-1_1
3
4 W. Ehlers and A. Wagner
Fig. 1 Simulation of a therapeutic infusion process within brain tissue during brain-tumour treat-
ment. The supplied domain of the brain hemisphere is displayed in green and the second-order
permeability tensors, governing the patient-specific anisotropic perfusion behaviour, are visualised
via ellipsoids in light blue on an exemplary cross section, cf. [3] for a comprehensive overview
discussion. Therefore, the purpose of this contribution is to perform within the
commercial finite-element (FE) software tool Abaqus®
(Dassault Systèmes, Vélizy-
Villacoublay, France, cf. http://www.3ds.com) microscopic and locally single-phasic
computations of a representative elementary volume of the brain-tissue aggregate.
This allows for the assignment of microscopic material parameters for the individ-
ual components. In this regard, elastically deformable blood vessels, separating the
pore compartments, are assumed. Thus, the liquid with the higher pressure displaces
the liquid with the lower pressure. The subsequent geometrical evaluation of certain
adjusted pressure situations (applied as boundary conditions) within such a micro-
scopic model allow for a meaningful postulation of a macroscopic constitutive equa-
tion. Note in passing that the presented study of scale-bridging aspects differs from
a complete two-scale approach using the FE2-Method (where both the macroscopic
as well as the underlying microscopic structure are analysed with the finite-element
method, cf. [2] in terms of the TPM). The scale-bridging approach proposed here
rather proceeds from an one-time analysis of a specific process on the microstruc-
ture and a subsequent transfer to the macroscopic model obtaining a microscopically
underlaid constitutive equation.
Multiscale Aspects in the Multiphasic Modelling of Human Brain Tissue 5
tissue cell
cell body interstitial fluid space
intracellular spaceintravascular space
extravascular space
blood plasma
blood vessel(capillary)
capillary membrane
blood cellextracellular space
Fig. 2 Exemplary sketch of microscopic fluid spaces in brain tissues, according to [3]
2 Anatomic Elements of Human Brain Tissue
Amongst other things, brain tissue includes the components interstitial fluid and
blood, which are mobile in two basically separated pore compartments. In this
regard, the interstitial fluid is situated within the interstitial fluid space, whereas
the intravascular space contains the blood. Moreover, further tissue spaces are com-
monly distinguished, such as the intra- and extracellular space or the extravascular
space, cf. Fig. 2 for a qualitative sketch. In general, the capillary membrane (wall) is
highly selective and governs the passing of molecules from the intravascular to the
extravascular space. Thus, the so-called blood-brain barrier separates the individual
pore compartments.
The interstitial fluid typically occupies a volume fraction of 20% of the brain’s
bulk volume, cf. [4]. However, a variation between 15% and 30% is normal and even
may fall to 5% during global ischaemia. Typical values of the blood volume fraction
are in the range of approximately 3% for healthy tissues, cf. [4] based on, e.g., [5]. In
contrast, in tumour-affected domains, the blood amount may vary between 1% and
20%.
3 TPMModel of the Overall Brain-Tissue Aggregate
To provide a framework for the considerations in this paper, the basis of the macro-
scopic thermodynamically consistent drug-infusion model is briefly described in this
section. Using the TPM, a volumetrical averaging procedure (smearing) of the under-
lying microscopic structure over a representative elementary volume (REV) leads
to an idealised macroscopic model of superimposed and mutually interacting con-
stituents. For a detailed discussion of the model, the interested reader is referred to
the underlying articles of [1, 6, 7] and citations therein. Basically, the quaternary
TPM model includes four constituents. In particular, three immiscible constituents
are given by the solid skeleton 𝜑S, the blood 𝜑
Band the overall interstitial fluid
𝜑I, where 𝜑
Icontains a dissolved miscible therapeutic solute 𝜑
D. To account for the
local composition of the aggregate, scalar structure parameters, the volume fractions
6 W. Ehlers and A. Wagner
n𝛼 = dv𝛼dv
, (1)
are introduced defined as the local ratios of the partial volume elements dv𝛼 with
respect to the volume element dv of the overall aggregate. Assuming fully saturated
conditions (no vacant space within the domain), the well-known saturation condition∑𝛼n𝛼 = nS + nB + nI = 1 is obtained. As was already mentioned, the pore liquids
are mobile in their individual pore compartments. Therefore, the volumetric amount
(saturation) of the blood
sB = 1 − sI = nB1 − nS
(2)
is specified in relation to the overall pore space (porosity). The set of model equations
proceeds from the specific balance relations for multiphasic materials, cf., e.g., [8] or
[9], where preliminary assumptions are initially included, i. e. isothermal conditions,
materially incompressible constituents, negligible gravitational forces and accelera-
tion terms. Therefore, the governing balance relations are given by the momentum
balance (3)1 of the overall aggregate 𝜑, the volume balance (3)2 of 𝜑B, the volume
balance (3)3 of 𝜑I
and the concentration balance (3)4 of 𝜑D
, viz.:
div𝐓 = 𝟎 ,(nB)′S + div(nB𝐰B) + nB div(𝐮S)′S = 0 ,(nI)′S + div(nI𝐰I) + nI div(𝐮S)′S = 0 ,nI(cDm)
′S + cDm div(𝐮S)′S + div(nIcDm 𝐰D) + cDm div(nB𝐰B) = 0 .
(3)
Therein, the overall Cauchy stress 𝐓 in (3)1 can be derived for, e.g., finite and
anisotropic elastic deformation processes of the solid in combination with the overall
pore pressure as was described in [1]. The notion ( ⋅ )′S indicates the material time
derivative with respect to the solid motion. In terms of the volume fractions and
their temporal changes
nS = nS0S (det 𝐅S)−1 and (nS)′S = − nS div (𝐮S)′S ,
nB = sB (1 − nS) and (nB)′S = (sB)′S (1 − nS) − sB(nS)′S ,nI = 1 − nS − nB and (nI)′S = − (nS)′S − (nB)′S ,
(4)
it turns out that a constitutive relation for the blood’s saturation sB is required, since
the saturation condition is not sufficient to determine all volume fractions. Note in
passing that the temporal change (sB)′S is based on the chosen constitutive ansatz for
sB, cf. Sect. 4.2. In Eq. (4), the quantity nS0S denotes the initial solid volume fraction
and det 𝐅S is the determinant of the solid’s deformation gradient 𝐅S = 𝐈 + GradS𝐮S.
Therein, the operator GradS denotes the material gradient of 𝜑S
with respect to its
reference position. Finally, the set of equations in (3) contains the seepage velocities
𝐰B, 𝐰I and 𝐰D which are described via
Multiscale Aspects in the Multiphasic Modelling of Human Brain Tissue 7
nB 𝐰B = − 𝐊SB
𝜇BR ( grad pBR +pdif .sB
grad sB ) ,
nI 𝐰I = − 𝐊SI
𝜇IR grad pIR ,
nIcDm 𝐰D = −𝐃D grad cDm + nIcDm 𝐰I .
(5)
The perfusion of the solid skeleton by the mobile but separated pore liquids is con-
sidered by the permeability tensors 𝐊SBand 𝐊SI
, whereas 𝐃Dis the diffusivity of the
therapeutic agent. Moreover, 𝜇BR
and 𝜇IR
are the effective shear viscosities of the
liquids. The pressure difference pdif . = pBR − pIR indicates the difference between
the individual pore-liquid pressures. Finally, the TPM model is solved for the pri-
mary variables solid displacement 𝐮S, effective pore-liquid pressures pBR, pIR and
therapeutic concentration cDm after inserting the constitutive relations into the set
of equations given in (3). With regard to the corresponding numerical solution of
the coupled system of partial differential equations using the research code Pandas
(Porous media Adaptive Nonlinear finite element solver based on Differential Alge-
braic Systems, cf. http://www.get-pandas.com), the Bubnov-Galerkin mixed FEM is
applied as shown in [3] for several numerical examples.
4 Microscopically Underlaid Macroscopic ConstitutiveRelation
As was already mentioned, the TPM model includes two pore liquids. Therefore, a
constitutive equation to describe the division for the volume fractions is required,
since there is still no general agreement concerning the stiffness of the blood-vessel
system in comparison to the surrounding tissue aggregate. Therefore, the purpose
of this section is to perform a microscopic and locally singlephasic computation of
a representative elementary volume of the brain tissue aggregate. This is realised
using Abaqus and consequently allows for the assignment of microscopic material
parameters for the individual components. The evaluation of adjusted pressure sit-
uations within such a microscopic model allows for a meaningful foundation of the
developed macroscopic constitutive equation.
4.1 Microscopic Model Settings
In terms of a microscopic description, a tube-like geometry is considered for the
description of a single vascular vessel, cf. Fig. 3. In order to apply a pressure dif-
ference between the inside and the outside of a blood vessel, pdif . = pBR − pIR is
subjected at the inside of the blood-vessel wall, cf. Fig. 3a. For the sake of simplic-
ity, a linear elastic material behaviour is chosen. The corresponding parameters are
collected in Table 1. Due to the elastically deformable blood vessels, the liquid with
8 W. Ehlers and A. Wagner
vessel
Rin
Rout
pdif.
1
2
(a) (b)
Fig. 3 Boundary conditions on a single blood vessel (a) and meshed geometry (b) with an exem-
plary point at the inside (1) and the outside (2) of the vessel, respectively
Table 1 Collection of parameters for a single blood vessel and a REV
Quantity Value Reference/Remark
Blood vessel inner radius Rin 1.6 × 10−4 m [10]
Blood vessel outer radius Rout 2.6 × 10−4 m [10]
Edge length lREV of a single
REV
7.3 × 10−4 m Assumption to satisfy the initial
(physiologic) condition sB = 0.2Poisson’s ratio 𝜇 0.48 For the cases (i) − (iii), according to [11]
Young’s modulus E 2792N/m2
For the cases (i) and (ii), assumption of
soft vessel walls
Young’s modulus E 1.2 × 105 N/m2
For case (iii), according to [12]
the higher pressure displaces the liquid with the lower pressure. In particular, three
different cases are studied:
∙ case (i): “soft” vessel walls under unconstrained conditions,
∙ case (ii): “soft” vessel walls under constrained conditions (cf. Sect. 4.1.1) and
∙ case (iii): “stiff” vessel walls under unconstrained conditions.
In these microscopic computations, the pressure difference is applied as a boundary
condition and the arising volume changes are geometrically evaluated in a subse-
quent step. To bridge from the microscopic to the macroscopic scale, many single
REV are arranged in a regular manner, cf. Fig. 4. In this regard, the edge length of
a single REV is chosen such that the initial saturation under zero pressure differ-
ence constitutes to sB = 0.2. Physically, a variation of the volume fractions within
a single REV requires an inflow (of the blood for a positive pressure difference) and
an outflow (of the interstitial fluid for a positive pressure difference) of the liquids.
This is provided by free-flow boundary conditions for both liquids in longitudinal
direction of the chosen blood-vessel orientation.
Multiscale Aspects in the Multiphasic Modelling of Human Brain Tissue 9
Fig. 4 Exemplary sketch of the macroscopic overall aggregate composed of multiple REVs
4.1.1 Geometrical Evaluation
Within Abaqus, the displacement of each node of the FE grid is obtained by the
solution of the considered initial-boundary-value problems. Note in passing that an
analytical solution for the widening of a linear-elastic cylindrical shell under con-
stant internal pressure does exist, cf., e.g., [13], and corresponds to an unconstrained
microscopic computation. In a geometric evaluation, the deformation of the solid
vessel wall can be captured by the displacements of the nodal points under load-
ing conditions. However, the focus of this contribution lies on the evaluation of the
liquid’s saturations. In this regard, the geometrical evaluation of the arising volume
fractions in the unconstrained cases (i) and (iii), cf. Fig. 5a, is trivial, since the dis-
placements of the nodes 1 (located at the inside of the vessel wall) and 2 (cor-
responding point in radial direction at the outside, cf. Fig. 3) are known from the
computation and allow for the determination using basic geometrical relations. In
addition, the widening of a single vessel is at a certain state somehow constrained by
the surrounding vessels, which themselves wide in a similar manner, as it is sketched
in Fig. 5b. This is approximated by a linear pathway in the contact domain and a cir-
cular shape in the edges. For the sake of simplicity, a constant thickness t of the
blood vessel is assumed. Obviously, this represents a simplification, since (at least
in the case of mutual contact) a deformation of the vessel wall may occur. However,
this assumption allows a relatively simple evaluation of the arising volume fractions.
Then, the volume fractions can be computed for the constrained case (ii), viz.:
VB =[𝜋 (d − t)2 + 4 (d − t) (lREV − 2 d) + (lREV − 2 d)2
]hREV ,
VI = (4 d2 − 𝜋 d2) hREV .(6)
Therein, hREV is the height of the REV required to compute a volume. However,
hREV vanishes within the determination of the liquid’s volume fractions via
nB = VB
VREVand nI = VI
VREV(7)
10 W. Ehlers and A. Wagner
(a) (b)
Fig. 5 a Unconstrained vessel widening within the REV and b constrained vessel deformation
caused by interacting vessels
and, thus, has no influence. In conclusion, this allows for the evaluation of the liquid’s
saturations under applied pressure conditions. In particular, this is carried out for
typical pressure states and collected in Sect. 4.3.
4.2 Macroscopic Constitutive Relation
The aim to simulate bio-technical/engineering applications of human brain tissues
requires a thermodynamically consistent setting of the blood-saturation function in
the context of the TPM model. This is somehow comparable to the procedure within
unsaturated soil mechanics. Therein, the capillary pressure, defined as the difference
between the pressures of the non-wetting and the wetting fluids in a common pore
space is used to evaluate the saturations via capillary-pressure-saturation conditions,
cf., e.g., [14]. However, in the specific case of human brain tissue, the blood and
the interstitial fluid are not situated in the same pore compartment such that these
relations cannot be applied. Instead, further considerations are made here in order to
specify the blood saturation, while bearing the microscopic considerations in mind.
In general, the macroscopic constitutive equation should provide the flexibility to
capture several circumstances. For the case of sufficiently soft elastic arterial walls,
a mutually volumetrical interaction is induced by an upcoming pressure difference.
Therefore, a constitutively chosen ansatz for the Helmholtz free energy 𝜓B
of the
blood constituent is postulated according to [1], viz.:
𝜓B(sB) = �̃�
B
𝜌BR
((𝛽B + 1) ln (sB) + 1
sB− ln (1 − sB)
)+ �̃�
B0 . (8)
Therein, �̃�B
and 𝛽B
denote material parameters, which allow via �̃�B
for the adaption
of the pore-pressure difference to typical pressure values as they exist in the skull
and for the initial blood saturation via 𝛽B. Furthermore, �̃�
B0 denotes the constant
reference potential (standard state potential). This approach satisfies the thermody-
namical restrictions and consequently yields a relation for the pressure difference
Multiscale Aspects in the Multiphasic Modelling of Human Brain Tissue 11
Fig. 6 Thermodynamically
consistent constitutive
relation for the
blood-saturation function sB
pdif .(sB) = sB 𝜌BR𝜕𝜓
B
𝜕sB= �̃�
B(
1 − 2 sBsB (sB − 1)
+ 𝛽B)
,
where 𝜕𝜓B
𝜕sB= �̃�
B
𝜌BR
(𝛽B
sB− 1
sB (sB − 1)− 1
(sB)2
)
.
(9)
The inversion of (9) finally leads to the blood saturation function
sB(pdif .) = 1
2( pdif .
�̃�B − 𝛽B)
(( pdif .
�̃�B − 𝛽B − 2
)+√
4 +(pdif .�̃�B − 𝛽
B)2)
.(10)
It should be noted that Eq. (10) is found by rational investigations of a meaningful
inversion of (9), which is a second-order function in sB. The derived relation for the
blood saturation (10) allows for a proper determination of the volume fractions using
the relations in (4). To give an example, an equation satisfying typical initial volume
fractions of the brain tissue as well as typical pressure values as they exist in the skull
is adapted by 𝛽B = 3.75 and �̃�
B = 200.0 N/m2, cf. Fig. 6. Physically, this leads to a
replacement of the interstitial fluid if the pressure difference pdif . is positive. Note in
passing that a constant value of nB0S = 0.05 yields an initial value of sB0S = 0.2 and,
as a result, pdif . = 0.
4.3 Results and Discussion
In this section, the evaluated results of the microscopic computation are compared
with the macroscopic constitutive approach. For the pressure in the vascular sys-
tem (cerebral blood pressure depending on its hierarchical position), the range in
the capillary bed varies between 10 mm of mercury (mmHg) (corresponds to 1333
N/m2) at the venous end and 30 mmHg (4000 N/m
2) at the arterial end. Whereas
the interstitial-fluid pressure (tissue pressure) has a typical value of 6 mmHg (800
12 W. Ehlers and A. Wagner
Fig. 7 Comparison of different approaches for the blood-saturation function. Evaluated pres-
sure states from microscopic computations are displayed with diamonds which are connected with
dashed lines
N/m2). To capture such typical conditions, the adjusting blood saturation is evalu-
ated for pressure differences in the range of −500 N/m2 ≤ pdif . ≤ 2000 N/m
2. In
particular, the pressure difference is applied in steps of 500 N/m2
with a subsequent
evaluation of the arising volume fractions according to Sect. 4.1.1. The evaluated
pressure states from the microscopic computations are then displayed with diamonds
and interpolated with dashed lines, cf. Fig. 7. The evaluation of case (i) (soft ves-
sel, unconstrained) yields to reasonable results in the pressure range of −500 N/m2
≤ pdif . ≤ 1500 N/m2. However, for a further increase of the pressure difference, non-
physical results (i. e. a blood saturation greater than one) occur. This is caused by the
possibility of an unconstrained widening of the blood vessel. In contrast, in the case
(ii) (soft vessel, constrained) it is nicely identified that the microstructural compu-
tations yield comparable results in relation to the constitutively chosen pressure-
saturation relation (Sect. 4.2) in the domain of nearly fully blood saturated condi-
tions. Again, the constitutive formulation of the blood saturation function is not arbi-
trary but has to fulfil the thermodynamic restrictions discussed in Sect. 4.2. There-
fore, the pathway vary slightly from the graph of case (ii) obtained by microscopic
computations. However, this allows for a scale-bridging in terms of microscopic and
macroscopic material parameters. In the specific case (ii), the microscopic material
parameters, Young’s modulus E = 2792 N/m2
and Poisson’s ratio 𝜇 = 0.48, of the
blood vessel wall correspond to the macroscopic material parameters �̃�B = 200.0
N/m2
and 𝛽B = 3.75 of the constitutive function. Assuming a higher stiffness of
the vessel (corresponding to somehow realistic experimental values of the human
intracranial artery) in case (iii) yields within the considered pressure regime a nearly
constant blood saturation. This would justify to assume a constant blood-volume
fraction within macroscopic simulations.
In conclusion, a constitutive approach for the relation between the pressure differ-
ence pdif . of the liquids and the blood saturation sB was presented which satisfies both
Multiscale Aspects in the Multiphasic Modelling of Human Brain Tissue 13
the thermodynamic consistence as well as physical conditions within human brain
tissue. Moreover, this macroscopic saturation condition was motivated and studied
by a microscopic computation of a representative microstructure using Abaqus and
a subsequent geometrical evaluation of the arising saturations under applied liquid
pressure conditions. With reliable experimental results, this would provide the pos-
sibility to identify the macroscopic material parameter of the constitutive function
based on the microscopic material parameter in terms of scale-bridging aspects.
Acknowledgements The authors would like to thank the German Research Foundation (DFG) for
the financial support of the project within the Cluster of Excellence in Simulation Technology (EXC
310/2) at the University of Stuttgart.
References
1. W. Ehlers, A. Wagner, Multi-component modelling of human brain tissue: a contribution to the
constitutive and computational description of deformation, flow and diffusion processes with
application to the invasive drug-delivery problem. Comput. Method Biomech. Biomed. Eng.
18, 861–879 (2015)
2. F. Bartel, T. Ricken, J. Schröder, J. Bluhm, A twoscale homogenisation approach for fluid
saturated porous media based on TPM and FE2-method. Proc. Appl. Math. Mech.15, 447–448
(2015)
3. A. Wagner, Extended modelling of the multiphasic human brain tissue with application to
drug-infusion processes. Dissertation, Report No. II-27 of the Institute of Applied Mechanics
(CE), University of Stuttgart, 2014
4. E. Syková, C. Nicholson, Diffusion in brain extracellular space. Phys. Rev. 88, 1277–1340
(2008)
5. R.K. Jain, Determinants of tumor blood flow: a review. Cancer Res. 48, 2641–2658 (1988)
6. W. Ehlers, A. Wagner, Constitutive and computational aspects in tumor therapies of multi-
phasic brain tissue, in Computer Models in Biomechanics, ed. by G.A. Holzapfel, E. Kuhl
(Springer, Netherlands, Dordrecht, 2013), pp. 263–276
7. A. Wagner, W. Ehlers, Continuum-mechanical analysis of human brain tissue. Proc. Appl.
Math. Mech. 10, 99–100 (2010)
8. W. Ehlers, Foundations of multiphasic and porous materials, in Porous Media: Theory Exper-iments and Numerical Applications, ed. by W. Ehlers, J. Bluhm (Springer, Berlin, 2002), pp.
3–86
9. W. Ehlers, Challenges of porous media models in geo- and biomechanical engineering includ-
ing electro-chemically active polymers and gels. Int. J. Adv. Eng. Sci. Appl. Math. 1, 1–24
(2009)
10. K.L. Monson, W. Goldsmith, N.M. Barbaro, G.T. Manley, Axial mechanical properties of fresh
human cerebral blood vessels. J. Biomech. Eng. 125, 288–294 (2003)
11. T.E. Carew, N.V. Ramesh, J.P. Dali, Compressibility of the arterial wall. Circ. Res. 23, 61–68
(1968)
12. K. Hayashi, H. Handa, S. Nagasawa, A. Okumura, K. Moritake, Stiffness and elastic behavior
of human intracranial and extracranial arteries. J. Biomech. 13, 175–184 (1980)
13. Y. Basar, W.B. Krätzig, Mechanik der Flächentragwerke: Theorie, Berechnungsmethoden(Anwendungsbeispiele. Springer Fachmedien, Wiesbaden, 2013)
14. T. Ricken, R. de Boer, Multiphase flow in a capillary porous medium. Comput. Mat. Sci. 28,
704–713 (2003)
Simulation of Steatosis Zonation in LiverLobule—A Continuummechanical Bi-Scale,Tri-Phasic, Multi-Component Approach
Tim Ricken, Navina Waschinsky and Daniel Werner
Abstract The human liver is an important metabolic organ which regulates metabo-
lism of the body in a complex time depending and non-linear coupled function-
perfusion-mechanism. Harmful microstructure failure strongly affects the viability
of the organ. The excessive accumulation of fat in the liver tissue, known as a fatty
liver, is one of the most common liver micro structure failures, especially in west-
ern countries. The growing fat has a high impact on the blood perfusion and thus
on the functionality of the organ. This interaction between perfusion, growth of fat
and functionality on the hepatic microcirculation is poorly understood and many
biological aspects of the liver are still subject of discussion. The presented compu-
tational model consists of a bi-scale, tri-phasic, multi-component approach based
on the theory of porous media. The model includes the stress and strain state of the
liver tissue, the transverse isotropic blood perfusion in the sinusoidal micro perfusion
system. Furthermore, we describe the glucose metabolism in a two-scale PDE-ODE
approach whereas the fat metabolism is included via phenomenological functions.
Different inflow boundary conditions are tested against the influence on fat deposi-
tion and zonation in the liver lobules. With this example we can discuss biological
assumptions and get a better understanding of the coupled function-perfusion ability
of the liver.
1 Introduction
For a realistic simulation of the liver behavior it is important to take into account
a detailed description of the biological processes. We present a bi-scale, tri-phasic,
continuum multi-component model simulating blood perfusion and metabolism in
T. Ricken (✉) ⋅ N. Waschinsky ⋅ D. Werner
TU Dortmund, August-Schmidt-Straße 6, Dortmund, Germany
e-mail: [email protected]
N. Waschinsky
e-mail: [email protected]
D. Werner
e-mail: [email protected]
© Springer International Publishing AG 2018
P. Wriggers and T. Lenarz (eds.), Biomedical Technology, Lecture Notes in Applied
and Computational Mechanics 84, DOI 10.1007/978-3-319-59548-1_2
15
16 T. Ricken et al.
macroscopic structure mesoscopic structure microscopic structure
Fig. 1 Bi-scale approach; macroscopic structure: millimeter spatial resolution to observe the vas-
cular system from the in- and outflow of the liver to smaller and smaller branches of the vessels;
mesoscopic structure: micrometer spatial resolution to observe the perfusion through the sinusoids;
microscopic structure: biochemical description of the metabolism in the liver cells
the human liver based on the two-phasic approach, see Ricken et al. [13]. To get a
better understanding of the liver we start with a short physiological outline. For more
informations, see Hubscher et al. [8].
The liver is the main metabolic organ in the human body. Roughly, the structure
of the organ is segmented into four hepatic lobes of unequal size within the range
of centimeters; the main parts are the right lobe (lobus hepatic dexter) and left lobe
(lobus hepatis sinister). Each lobe consists of liver lobules which have the size of
about a few millimeters. They are hexagonally structured parts of the liver tissue, see
Fig. 1. Microscopic sized hepatocytes, the liver cells, are arranged in radial columns
embedded in the liver lobules. The hepatic vascular system is important for the blood
supply of the hepatocytes. For this, the vascular system consists of branches which
supply the organ from the macroscopic to microscopic anatomy of the liver. On the
macroscopic level branches of hepatic artery (oxygen rich) and portal vein (nutri-
ent rich) run together and supply the segments of the organ. Intrahepatic bile ducts
proceed together with the blood vessels and support the liver during detoxification
processes as the bile fluid is excreted with decomposed products. The microcircula-
tion takes place in the liver lobule. In each corner of the liver lobule a portal triad is
located which is composed of the three branches of hepatic artery, portal vein and
a bile duct in their smallest subdivision. The microcirculation is described by blood
vessels connecting the portal triad with the central vein which are called sinusoids.
The central vein drains the blood leading to a directed blood flow from the portal
triad to the central vein, see Fig. 1.
Under consideration of the physiological background it is unavoidable to classify
the liver in different scales to focus on scale depending processes.
We focus on a bi-scale approach; on the mesoscopic level we describe the blood
perfusion through the liver lobule and on the microscopic level the metabolism in
the cells.
The mesoscopic model is based on an ansatz for transverse isotropic permeabil-
ity relation describing the perfusion given in Ricken et al. [12]. Due to the complex
Simulation of Steatosis Zonation in Liver Lobule . . . 17
insu
lin
increase
decreaseffaext
lacext
glcext
VHGU VGS
VLP
VGLY
FFA TG
additional hypothesis:
fat metabolism
reduced metabolic model:
glucose metabolism
Fig. 2 Metabolism: the biochemical reactions are described with the metabolites of external glu-
cose (glcext ), external lactate (lacext ), external FFA (f faext ), internal triglyceride (TG) and internal
glycogen (glyc). External components are included in the fluid phase whereas internal components
are included in the solid tissue of the liver. Triglyceride describes the fat phase in the liver. Reduced
metabolic model: the glucose pathways (VHGU), (VGS) and (VGLY) are modeled by sets of ordi-
nary differential equations. WithVHGU < 0: hepatic glucose production,VHGU > 0: hepatic glucose
utilization, VGLY < 0: gluconeogenesis, VGLY > 0: glycolysis, VGS < 0: glycogenolysis, VGS > 0:
glycogenesis. Additional hypothesis: the fatty acid pathway (VLP) is modeled by assumptions con-
cerning the uptake rate of FFA. With VLP > 0: lipogenesis and VLP < 0: lipolysis
vascular system of sinusoids it is impossible to model the numerous capillaries
with an accurate discrete geometrical description. Therefore we apply a homoge-
nized multiphase approach, based on the theory of porous media (TPM) which is
a convenient tool in order to describe complex, multi-phasic and porous materials;
see de Boer [2, 3] and Ehlers [5]. This approach includes the description of three
main phases, namely the liver tissue, fat tissue and blood vessels. Additionally to the
main phases, the homogenized liver lobule includes miscible components which are
important for the biochemical description of the metabolism. A description of the
mesoscopic model is given in the appendix.
We describe the biochemical processes on the microscopic scale of our model
as the metabolism takes place in the smallest parts of the liver tissue, the cells. The
liver is crucial for metabolites regulation of glucose and free fatty acid (FFA) due
to its ability to store glucose in form of internal glycogen and FFA as internal fat
deposition. The cells can switch between glucose and FFA deposition and utilization
depending on the requirements of the rest of the body. After ingestions, glucose and
FFA concentrations coming from the digestive system increase (hyperglycemia). The
liver cells start to deposit them as internal metabolites glycogen and triglyceride.
Whereas, during fasting times (hypoglycemia) the internal metabolites are served
as a backup system and can be used to increase the low concentration of external
metabolites. We summarize the metabolism in a set of ordinary differential equations
which consider the conditions of hyperglycemia and hypoglycemia and the relations
of the internal and external metabolites. The microscopic model approach is derived
from a detailed pharmacokinetic model of hepatic glucose metabolism, described in
König et al. [10]. We presented a reduced model for glucose metabolism in Ricken
et al. [13] and add a phenomenological approach for the fat metabolism (Fig. 2).
18 T. Ricken et al.
2 Glucose and Fat Metabolism
The liver ensures biosynthetic metabolic pathways of immense complexity. In this
study we focus on the main metabolic pathway of glucose described by a simpli-
fied ordinary differential equation system. Furthermore, we add a phenomenological
approach for the fat metabolism.
Figure 2 summarizes the applied pathways for glucose and fat metabolism. The
main task of the liver is the maintenance of homeostasis for metabolites which
might be necessary for the organism. For this the liver regulates the glucose and fat
level in order to offer enough energy for the brain and muscles. The metabolites are
stored or utilized depending on hormone concentration levels. We distinguish two
conditions: the fed (hyperglycemia) and fasted state (hypoglycemia). During feed-
ing the organism is saturated of glucose and FFA leading to utilization of external
metabolites. Glycolysis and glycogenesis are two possible metabolic pathways of
glucose. The first metabolic pathway ensures the energy supply for cells and repro-
duction of lactate. The second describes energy storage in form of internal glycogen.
Lipogenesis is the fatty acid pathway describing the ester of FFA and glycerol to
triglyceride. The lipid accumulation is reinforced by high carbohydrate diets.
During the fasting state the organism utilizes hepatic glycogen, non-carbohydrate
carbon substrates and lipids. Gluconeogenesis is one pathway for the generation
of glucose. The utilization of certain non-carbohydrate carbon substrates provides
a balanced glucose level. Breaking down hepatic glycogen describes the second
mechanism to maintain glucose, namely glycogenolysis. The utilization of lipids is
part of the lipolysis pathway and the reverse of lipogenesis. Molecules of lipids are
hydrolyzed to glycerol and FFA.
For a realistic depiction of the liver functionalities, we take the above described
main metabolic pathways into account. We use a reduced metabolic model, which
has been derived from a detailed kinetic model of hepatic glucose metabolism, see
König et al. [10]. It fully represents the metabolic behavior of depletion and utiliza-
tion of glycogen. For further information see Ricken et al. [12, 13].
FFA play an important role in essential functions for the organism. They deliver
cells energy and are important constituents for esterifying lipids. For a homeostatic
circulation the organism can store, produce and consume FFA. This process is mainly
regulated in the liver by fat metabolism. Stored triglyceride in adipocytes and dietary
fat are sources, which ensure the availability of FFA in the liver. Beside the external
concentration we need to take important enzymes and insulin into account. Insulin
is a hormone which is distributed depending on the glucose concentration level in
blood. The task of insulin is to regulate glucose and fat metabolism, so the pathways
are coupled. Enzymes are responsible for hydrolysis of proteins, which bind FFA for
circulation in the blood plasma.
The effects of fatty acid metabolism are not part of the kinetic model. In the
following we present the assumptions concerning the FFA pathway. Equally to
the glucose metabolism we calculate the rates of changing concentrations for the
metabolites �̂�𝛼𝛽
contributing in the FFA pathway. We apply the influencing factors
Simulation of Steatosis Zonation in Liver Lobule . . . 19
ρ̂FFA
glc
[mm
ol/l/
s]
glucose [mmol/l]
(1) insulin dependency
Fig. 3 Insulin dependency �̂�FFAglc : uptake rate of FFA depending on the concentration of glucose
ρ̂
ρ̂FFA
ffa
[mm
ol/l/
s]
ffa [mmol/l]
FFA
nTG
[1/s
]
volume fraction fat [-]
(2a) FFA dependency: hypoglycemia (3a) fat dependency
Fig. 4 Threshold of external FFA concentration 0, 6 mmol/l which leads to a hypoglycemic con-
dition; 2a FFA dependency �̂�FFAffa : uptake rate of FFA depending on the external FFA concentration.
3a fat dependency �̂�FFAnTG : uptake rate of FFA depending on the volume fraction of fat for a stable
computation
{�̂�FFAglc ; �̂�FFAffa ; �̂�FFAnL ; �̂�FFAnTG
}. The terms are postulated via the differential equation of
growth function with
�̂�FFAglc = exp−log(2)
glc3
glcTP3
�̂�FFAffa = +∕−1 + exp−40∗ ((f fa − 0,6)∗ 2)4
�̂�FFAnL = − 1 ∗ exp−log(2)
(nL)3
0,153 + 1
�̂�FFAnTG = − 1 ∗ exp−log(2)
(nTG)3
0,23 + 1
(1)
We calculate the maximum uptake rate of FFA considering insulin influence
(�̂�FFAglc ) which couples the glucose metabolism to the fat metabolism, see Fig. 3.
Figure 4 summarizes the effects of hypoglycemic conditions which lead to
20 T. Ricken et al.
ρ̂
ρ̂
FFA
ffa
[mm
ol/l/
s]
ffa [mmol/l]
FFA
nL[1
/s]
volume fraction fluid [-]
(2b) FFA dependency: hyperglycemia (3b) fluid dependency
Fig. 5 Threshold of external FFA concentration 0, 6 mmol/l which leads to b hyperglycemic con-
dition; 2b FFA dependency �̂�FFAffa : uptake rate of FFA depending on the external FFA concentration.
3b fluid dependency �̂�FFAnL : uptake rate of FFA depending on the volume fraction of fluid for a stable
computation
production of FFA. The uptake rate (�̂�FFAffa ) is positive as the organism needs more
FFA for energy supply. The hypoglycemic condition results in phase transition from
the fat phase to the fluid phase. To get a stable computation considering the satu-
ration condition we take the influencing factor (�̂�FFAnTG ) into account which steers an
excessive increase of fluid. However, Fig. 5 shows the impact during hyperglycemic
conditions. The uptake rate (�̂�FFAffa ) is negative as the liver stores redundant FFA and
synthesizes lipids. In this case the fat phase gets additional mass and we steer an
excessive increase of fat with (�̂�FFAnL ).
The mass exchange must fulfill the restrictions of the entropy inequality (24); the
exchange can be summarized with
�̄�TG𝛽 = �̄�
TG ∶=[�̂�FFAglc ; �̂�FFAffa ; �̂�FFAnL ; �̂�FFAnTG
]. (2)
3 Numerical Example: Comparison of DifferentAssumptions for the Perfusion Coupledto the Metabolism
Many aspects of the liver are still subject of discussion. One example is the blood
supplying system of the hepatic microcirculation. In this case it is still uncertain how
the blood is emanating from the blood vessels into the liver lobule. One assump-
tion, defined by Rappaport in 1954 [11], assumes an idealized hexagonal lobule
which regards the portal triad as the ‘terminal portal venule’. This case regards
the blood supplying system of the microcirculation initiated at the corners of the
liver lobule, emanating from the portal triads. In a realistic framework, this point of
view is not rational and a different assumption has to be applied. For this case
Simulation of Steatosis Zonation in Liver Lobule . . . 21
Fig. 6 Schematic sketch of
the feeding cycle showing
duration and typical patterns
of glycogen in a single
lobule during deposition and
depletion. Figure reproduced
from Babcock and
Cardell [1]
Rappaport described additional portal venules which are derived from the portal vein
and deliver the base of the liver lobule with inflowing blood. So the blood supplying
system of the microcirculation is located at the surface of the liver lobule.
To get a better understanding of the different approaches and the influence on the
metabolism, we simulate both hypotheses for the initial point of the microcirculation:
∙ Boundary-Model A: Idealized Case—Blood is emanating from the portal triad
∙ Boundary-Model B: Realistic Case—Blood is emanating from derived portal
venules
In the numerical example we compute the microperfusion in one liver lobule cou-
pled to the metabolism similar to Ricken et al. [13]. We evaluate the storage of inter-
nal glycogen deposition resulting from glucose metabolism analog to the experi-
ments performed by Babcock and Cardell. In [1], Babcock and Cardell evaluated the
glycogen storage during 24 h in one liver lobule after a feeding period (see Fig. 6).
Furthermore we analyze the results of the fat metabolism which accrues in the hepa-
tocytes (see Figs. 11–13).
We apply external boundary conditions for the glucose and FFA concentration
which are applied in the feeding artery after one meal in 24 h (see Fig. 7). Fur-
thermore we add external conditions for FFA. (cf. Stanhope et al. and Yuen et al.
[14, 16])
The boundary conditions focus on the two different assumptions (see Fig. 8) for
the initial point of the microcirculation (see Fig. 9). Additionally to the time depend-
ing DIRICHLET boundary conditions for the external concentration glucose and
FFA, which depend on the food intake (see Fig. 7), we apply constant DIRICHLET
boundary conditions for lactate at the inflow. As the blood flow is orientated in the
direction of the pressure gradient (transverse isotropic permeability relation, Ricken
et al. [13]) we apply a pressure difference between the inflow and the outflow with
constant values (see Fig. 10).
22 T. Ricken et al.
Fig. 7 External boundary
conditions: glucose profile
corresponding to 2 h food
intake and 22 h fasting based
on 24 h profiles of plasma
glucose (red). Analog profile
for plasma FFA (blue).
Experimental data from
Stanhope et al. [14] and
Yuen et al. [16]. Evaluation
for numerical example after
6, 9 and 18 h
gluc
ose
[mm
ol/l]
time [h]
6 h 9 h 18 h
0.03
0.45
0.02
outflow
(∂pLR; ∂ μαβ ; uuu)outflow
(∂pLR; ∂ μαβ ; uuu)
inflow(∂pLR; ∂ μαβ ; uuu)
inflow(∂pLR; ∂ μαβ ; uuu)
0.5
[mm][mm]
(a) (b)
Fig. 8 Depiction of the applied geometry and boundary conditions. a Boundary-Model A
b Boundary-Model B
3.1 Discussion
The simulation provides the evaluation of metabolites (see Figs. 11, 12, 13) resulting
from glucose and fat metabolism. On the one hand we focused on the
glucose pathway which depends on the external condition of blood glucose concen-
tration. During hyperglycemic conditions the external concentration is high resulting
in storage of glycogen. Whereas, a low glucose concentration encourages glycogen
depletion. On the other hand we evaluated the phenomenological approach of the
fat metabolism which leads to accumulation of fat depending on external blood FFA
concentration. Based on the assumption that fat metabolism is inhibited by the pres-
ence of insulin we can observe the coupling between glucose and fat metabolism.