Programme

Keynote speakers

We are very pleased that the distinguished scientists, thought leaders and excellent speakers Denis Noble, Tom Kirkwood, Peter Hunter and Yoram Rudy will each open one of the four plenary sessions of the conference:
 

Denis NobleIs the digital patient possible: What are the roadblocks and how do we negotiate them?

Professor Denis Noble, University of Oxford


 

Tom KirkwoodUntangling the complexity of ageing

Professor Tom Kirkwood, Newcastle University
 


 

Peter HunterReproducible modeling: Standards, databases and software tools

Professor Peter Hunter, University of Auckland


 

Yoram RudyMulti-scale Integration of Cardiac Excitation: From Molecular Structure to the Human Heart

Professor Yoram Rudy, Washington University in St. Louis
 

Workshops

Cardiovascular modelling

Organizer:  Arun Holden,  Multidisciplinary Cardiovascular Research Centre, University of Leeds
Co-organizer: Molly Maleckar, Simula Research Laboratory, Oslo, Norway

Cardiovascular modelling

Organizer:  Arun Holden,  Multidisciplinary Cardiovascular Research Centre, University of Leeds
Co-organizer: Molly Maleckar, Simula Research Laboratory, Oslo, Norway

Rationale:

Grodins  ( Quart Rev Biol 34 93-116 1959) began the integrated modelling of cardiovascular dynamics  by volumes, pressures and flows  using linear equations and analytical methods.  However, nonlinearity  is essential for explaining the phenomenology myocardial and vascular smooth muscle electrophysiology, calcium dynamics, and tissue mechanics.

Voltage-dependent channels  produce cell threshold behaviour and  allow autorhymicity. Myocytes  types are widely modelled by multiple parameter, high order systems of ordinary differential equations.  Propagation is modelled as nonlinear waves by partial differential systems.  Dysrhythmias emerge  via bifurcations (e.g. into alternans)  or by conduction blocks, and unidirectional conduction block can lead to re-entrant tachyarrhythmias and fibrillatory activity.   Detailed  3D reconstruction of cardiac activity can be partially validated by optical mapping of V and Ca++ on the surface of isolated hearts, and by clinical endo-, epi-cardial and body surface recordings, or cine- and tagged MRI. Cardiovascular tissue also includes the extracellular matrix, endothelial cells,  and fibrocytes which may be coupled to myocytes.  Vascular endothelial cells modulate vascular tone, mediate responses to oxidative stress, and endothelial dysfunction can lead to the development  and progression of  atherosclerosis. Endothelial damage leads to vascular smooth muscle proliferation and extracellular matrix synthesis  and plaque formation.

The current  challenge is to integrate these detailed electrophysiology, mechanics, fluid dynamics and biochemical signalling models into an overall model of the  cardiovascular system under autonomic control, during normal sinus rhythm, ageing, and the development of pathologies.

Presentations:

  • Eleftheria Pervolaraki, Barrie Hayes-Gill, Andrew Hogarth, Arun Holden, Hannah Law, Craig Russell, Muzahir Tayebjee: Cardiac arrhythmia from prenatal to elderly: coupling sparse clinical recordings with dense spatio-temporal models
  • Siri Kallhovd, Valeriya Mezzano, Sjur Gjerald, Jørg Saberniak, Farah Sheikh, Kristina Haugaa, Molly Maleckar: Localization and not extent of fibrofatty infiltration is the primary factor determining conduction disturbance in a computational model of arrhythomogenic cardiomyopathy
  • Simone Rivolo, Lucas Hadjilucas, Matthew Sinclair, Nicolas P. Smith, Jack Lee: Impact of coronary bifurcations morphology on wave propagation
  • Vinzenz Eck, Jonathan Feinberg, Hans Petter Langtangen, Leif Rune Hellevik: Effects of parametric uncertainty in blood flow simulations
  • Ismail Adeniran, Henggui Zhang: A 3D electromechanical model of the human atria for the study of atrial fibrillation
  • Pablo Lamata, Anastasia Nasopoulou, Manav Sohal, Aldo Rinaldi, Nic Smith, Steven Niederer: Improving the assessment of diastolic performance

Summary of discussion:

The scope of this session was to highlight new and notable contributions from diverse VPH research streams all of which, nonetheless, focus on the cardiovascular system. These included organ-level electromechanical approaches, arrhythmia electrophysiology, vessel modelling, and approaches to computational fluid dynamics, among others.

The primary deliverable of the cardiovascular system is local perfusion of extracellular tissue space, via flow and diffusion from the capillaries. Within most of the vascular system flow is laminar, with non-Newtonian effects due to the RBCs apparent only in vessels < ~20 μm diameter, and flow is turbulent only in and around the heart where flow velocities exceed 1 m/s, or at arterial branch points or constrictions. While the physics of blood flow is simple, the geometry of the cardiovascular system is complicated. Organ vascular trees can be approximated by fractals; constructing a computational model of the cardiovascular system is essentially applied computational fluid dynamics in a branching network.
 
When Grodins (Quart Rev Biol 34 93-116 1959) began the integrated modelling of cardiovascular dynamics by volumes, pressures and flows, he used linear equations and analytical methods. Guyton extended this model via a network approach, to include neural and hormonal factors: a model of the cardiovascular system needed to include its own regulation. The resultant control map was too large to be printed on a page, and so was published within the Annual Review of Physiology as an inserted fold out (as especially instructive to understanding the complexity of the physiological relationships involved in cardiovascular modelling, the webpage can be referenced [here] Guyton C, Coleman TG, Grancer HJ. (1972 ) Circulation: overall regulation Annual Review of Physiology 34 13-44).

Cardiovascular modelling's scope also naturally includes consideration of myocardial contraction, the muscular engine driving the fluid flow. A contemporary re-incarnation of Guyton's 1970s map could be a computer program, with the core focus shifted from the linear dynamics of fluid flow to the nonlinear dynamics of myocardial electrophysiology. This nonlinearity is essential for explaining the phenomenology of myocardial and vascular smooth muscle electrophysiology, calcium dynamics, and tissue mechanics.

Myocytes types can be widely modelled by multiple parameters in high order systems of ordinary differential equations. Propagation can then be modelled as nonlinear waves by partial differential systems. Dysrhythmias may emerge via functional bifurcations in deterministic behavior (e.g. into alternans) or by conduction blocks, and unidirectional conduction block can lead to re-entrant tachyarrhythmias and fibrillatory activity. Detailed 3D reconstruction of cardiac activity can be partially validated by optical mapping of V and Ca++ on the surface of isolated hearts, and by clinical endo-, epicardial and body surface recordings, or via cine- and tagged MRI.

Of course, cardiovascular tissue is more than just its myocytes. It also includes endothelial cells and fibroblasts, which in addition to producing the extracellular matrix which surrounds and supports myocytes, may be coupled to each other as well as to myocytes and almost certainly apportion a variety of auto- and paracrine effects. Vascular endothelial cells modulate vascular tone, mediate responses to oxidative stress, and endothelial dysfunction can lead to the development and progression of atherosclerosis.

A current challenge is to integrate these detailed electrophysiological, mechanical, fluid dynamical and biochemical signalling models into cardiovascular systems models which may both help to answer specific questions about the development of cardiovascular pathologies as well as complement clinical cardiological practice. As illustrated by the range of presentations at the Cardiovascular modelling workshop, there are research and clinical imaging technologies that provide detailed information on structure and behaviours, a portfolio of cardiac cellular and tissue electrophysiology and electromechanics models, and vascular geometries that are all currently being applied for specific problems.

The increasing standardisation of data and model formats means that these models could potentially be coupled into a larger computational model of the whole human cardiovascular system. However, current models do not provide compact support for the entire cardiovascular system – for instance, while there is a plethora of competing cardiac myocyte electrophysiology models, there is only sparse information on vascular smooth muscle cell electrophysiology and its local control mechanisms. A key activity in constructing integrated cardiovascular system models is thus clearly identifying areas where sufficiently high quality and resolution data is lacking.
An initial strategic step towards obtaining such data could be to construct a model of the isolated, retrograde perfused Langendorff heart, that coupled tissue electrophysiology, mechanics and fluid dynamics for a beating heart. Experimental validation of such a model for would allow a virtual human Langendorff heart to be constructed and potentially marketed as a test-bed for quantitatively predicting the effects of pharmacological agents on cardiac performance, and as an experimental tool for investigating arrhythmogenic mechanisms.Such an electromechanical heart including relevant fluid dynamics could then be coupled to simple or detailed vascular tree models. Vascular smooth muscle electrophysiology, its local control, and how it changes throughout the vascular tree, would be needed if systemic pharmacological effects on local flow and perfusion are to be explored (e.g. of Sildenafil citrate). The construction of such a cardiovascular system model is a quantitative, predictive embodiment of physiology, pharmacology and anatomy, and is conceptually satisfying. It produces a general tool of broad applicability, but one that is looking for a killer application.

An alternative approach to the modular, connected modelling of the physiology of the cardiovascular system in its entirety is to continue focus on specific questions that will be answered by custom models of parts of the cardiovascular system. Of course, pragmatic focus on specific application areas helps to more easily identify funding in the current climate, but also produces a goal-oriented work program and results that can be prospectively used immediately by the rest of the community. Basic research in cardiology, for instance, can be augmented via electromechanical models of contraction of the human ventricles, which may offer information regarding intramural stresses that clinical imaging modalities currently cannot, paving the way for e.g. in silico biomarkers to leverage impact on diagnostics and prognostics. It is clear that despite the many technical, practical, and cultural challenges posed to wide adoption of cardiovascular modelling in a clinical setting, that enormous strides have been made in the last decades. Approaches which favour the address of pragmatic scientific and clinical queries and are likewise underpinned by biophysics are well- poised to make substantial impact on human health.


Computational integration of organ physiology

Organizer: Dan Beard, Department of Molecular and 
Integrative Physiology, University of Michigan, Ann Arbor
Co-organizer: Dirk Drasdo, Institut National de Recherche en Informatique et en Automatique (INRIA), Rocquencourt/Paris

Computational integration of organ physiology

Organizer: Dan Beard, Department of Molecular and 
Integrative Physiology, University of Michigan, Ann Arbor
Co-organizer: Dirk Drasdo, Institut National de Recherche en Informatique et en Automatique (INRIA), Rocquencourt/Paris

Rationale:

Every organ in the body performs a set of specialized physiological functions that determine its contributions to whole-body function in health and disease. Yet although the organs and organ systems of the body perform distinct and diverse physiological functions, these functions emerge from a common core of basic biological/biophysical phenomena that may be captured by sophisticated multi-scale computational models. Basic cellular metabolic, mechanical, electrical, signaling, and molecular genetics processes are specialized within every organ and organ system, in which a number of types of cells are arranged in a distinct spatial pattern or patterns that determine how the cells interact chemically, mechanically, and electrically. Afferent and efferent nerves communicate with parenchymal cells via electrical, mechanical, and chemical transduction of signals between cells. Blood vessels perfuse organ tissues, transporting substrates, metabolic wastes, and endocrine factors. Simulating the combined operation of all of these processes in the individual organs of the body provides the basis for simulation virtual organisms at the level of mechanistic fidelity necessary to map from molecular function to whole-body phenotype.

Presentations:

  • Tim Ricken, Daniel Werner, Hermann Georg Holzhütter, Uta Dahmen, Olaf Dirsch: On a two-scale function-perfusion model for fatty liver
  • Laura Cooper, Joshua Scallan, James Heppell, Geraldine Clough, Bharathram Ganapathisubramani, Tiina Roose:
  • Modelling the mechanical behaviour of collecting lymphatic vessels
  • Filippo Castiglione, Teresa Colombo, Vinca Prana, Paolo Tieri: Modelling inflammation: A step toward the simulation of type-2 diabetes
  • Andy Olivares, Jerome Noailly:
  • Atherosclerosis explored with an agent-based model
  • Ahmad Diab, Mahmoud Hassan, Jeremy Laforet, Brynjar Karlsson, Catherine Marque: EHG source localization using signals from a uterus electrophysiological model
  • Dirk Drasdo: How quantitative modelling can inform on disease pathogenesis: lessons from liver

 


Interoperability infrastructures bridging molecular- to organ-level data and models

Organizer: Bernard de Bono, CHIME Institute, University College London & Auckland Bioengineering Institute, University of Auckland
Co-organizer: Peter Hunter

Interoperability infrastructures bridging molecular- to organ-level data and models

Organizer: Bernard de Bono, CHIME Institute, University College London & Auckland Bioengineering Institute, University of Auckland
Co-organizer: Peter Hunter

Rationale:

The practice and research of biomedicine generates considerable quantities of data and model resources. In particular, the organized archiving of high-throughput genomic and biochemical assays, as well as large scale biobanking and electronic health anonymization efforts, are starting to build a detailed, if patchy, multiscale record of tissue biology at both population and individual level. The VPH community is now faced with the challenge of interpreting this wealth of new information, and specifically to detect physiologically meaningful correlates between molecular-, subcellular-, cellular- and organ-level measurements of relevance to model construction and validation. What strategies are to be adopted to create a coherent VPH modeling framework that consistently interoperates across these levels? Do we need new forms of knowledge representation? Do we need new forms of knowledge management? This workshop draws upon real community efforts, and their solutions, to build interoperability infrastructures that address data-model integration, as well as the modular assembly of complex models, across the scale divide.

Presentations:

  • Clemens Wittwehr, Hristo Aladjov, Steve Edwards:
  • The Adverse Outcome Pathway Knowledge Base (AOP-KB)
  • Tommy Yu, David Nickerson, Michiel Helvensteijn, Bernard de Bono, Peter Hunter: Knowledge management of semantic metadata for multiscale modeling: Physiology-based PK- PD modelling within the PMR infrastructure 
  • João Ferreira, Bernard de Bono, Francisco M Couto: From data to knowledge: A tool for clustering multiscale resources for physiology research 
  • Ernesto Coto, Juan Arenas, Alfredo Saglimbeni, Debora Testi, Alejandri Frangi: The VPH-Share plugin for workflow composition and execution 
  • Marek Kasztelnik, Marian Bubak, Maciej Malawski, Piotr Nowakowski, Ernesto Coto, Juan Arenas: Support for Taverna Workflows in VPH-Share Cloud Platform 
  • Spiros Koulouzis, Dmitry Vasyunin, Adam Belloum, Marian Bubak:
  • Data Storage Federation for VPH applications

Make them run: modelling challenges identified by exercise physiology

Organizer: Trine Karlsen, 
Department of Circulation and Medical Imaging, NTNU
Co-organizer: Allen Kelly, Department of Circulation and Medical Imaging, NTNU

Make them run: modelling challenges identified by exercise physiology

Organizer: Trine Karlsen, 
Department of Circulation and Medical Imaging, NTNU
Co-organizer: Allen Kelly, Department of Circulation and Medical Imaging, NTNU

Rationale

Regular exercise training and high cardiovascular fitness are strongly associated with lifelong health, in particular protection towards cardiovascular disease. Data on epigenetic markers, mRNA expression, protein expression, metabolic signatures, and tissue and organ phenotypes show that the effects of exercise training manifest at all phenotypic levels. We are still far from a quantitative understanding of the regulatory biology underlying the observed changes; changes that are intimately linked to cellular phenotypic plasticity. If computational physiology becomes capable of describing the dynamics underlying phenotypic plasticity as a function of exercise, this will most likely have a tremendous impact on organismal biology in general. The workshop will outline state-of-the-art knowledge in exercise physiology, identify key questions being addressed in the experimental community and challenge the computational physiology community to take explanatory responsibility for exercise-related phenomena that are currently highly enigmatic and which appear to demand the construction of sophisticated mathematical representations

Presentations:

  • Jeroen Jeneson, Remco Renken, Marco van Brussel, Dan Beard, Bert Groen: Model that! in vivo observations on muscle energetics in whole body exercise
  • Vegard Malmo, Allen Kelly, Ulrik Wisløff, Godfrey Smith, Jan-Pål Loennechen: Interval training increases AF resistance in aged Rats by increasing conduction velocity and electrical stability
  • Kari Jørgensen, Morten Andre Høydal, Anne Berit Johnsen, Svein Erik Gaustad, Eirik Skogvoll, Godfrey L. Smith, Ulrik Wisløff: Exercise training reduces life-threatening arrhythmias in heart failure rats: the importance exercise training intensity
  • Nina Zisko: Generation 100: A randomized controlled study of the effects of long-term exercise training on mortality in elderly people. Study protocol and design 
  • Micaela Morettini, Massimo Sacchetti, Aurelio Cappozzo, Claudia Mazzà: Effects of intensity and duration of physical exercise described by a mathematical model of Interleukin-6 dynamics
  •  Johannes van Beek, Farahaniza Supandi, Hannes Hettling: Modelling muscle energy turnover and whole body heat transport during a Tour de France stage

Summary of discussion:

Cardiovascular fitness is strongly linked with health status throughout life, and it is now widely accepted that regular exercise can confer significant health benefits to individuals, particularly those suffering from cardiovascular complications. Specificity of a prescribed exercise program is now emerging as an important factor in determining the success of exercise as a therapeutic approach. Divising an optimal exercise regime for a variety of conditions experimentally can take years, and as such may be a task better suited to computational modelling. During this organised session both experimentalists and modellers presented up-to-date knowledge on current physiological challenges involving exercise training either as a test modality, or as a therapeutic intervention. Presentations from researchers at NTNU demonstrated the effectiveness of exercise training in restoring contractile function (Kari Jørgensen) and normalizing cardiac electrophysiology (Allen Kelly). The initial phase of a large clinical trial known as Generation 100, aimed at elderly individuals between 65 and 74, was described in detail (Nina Zisko). Speakers also presented data generated from different compartmentalized modelling studies, used to investigate a variety of exercise related factors, including interleukin inflammatory response and heart rate to various exercise lengths and intensities (Claudia Mazza), intramuscular metabolism and phosphate cycles during one and two leg cycling using MRI (Jeroen Jenesen) and heat dissipation during maximal endurance exercise in elite cyclists cycling the Alpe D'Huez stage in Tour De France (Johannes van Beek).

During the session it was clear that a number of obstacles broadly govern the effectiveness of successfully modelling the in vivo response to exercise. Availability of experimental data, particularly patient data, can be sparse and non-specific, forcing modellers to use datasets from test subjects who may not necessarily represent the patient demographic they wish to describe. Given the multitude of exercise protocols currently in the literature, the question of which data set to use also becomes an issue when a model is designed "after the fact". These issues highlight the need for better communication and direct collaboration between experimentalists and modellers to design studies where datasets are generated from appropriate populations. This lack of experimental data also extends to the study of elite athletic performance, particularly when frequent complications with doping heavily influence results. For experimental researchers, understanding the fundamental signaling pathways involved in the beneficial response to exercise training remains a key obstacle in designing an appropriate exercise regimen, as well as effectively informing future computational modelling studies. Determining the underlying physiological differences between responders and non- responders to exercise training, as well as understanding the interaction between exercise and inactivity will be critical avenues of experimental research for the future.

Mathematical challenges of multiscale modelling

Organizer: Merryn Tawhai, Auckland Bioengineering 
Institute, University of Auckland
Co-organizer: Blanca Rodriguez, Department of Computer Science, University of Oxford

Mathematical challenges of multiscale modelling

Organizer: Merryn Tawhai, Auckland Bioengineering 
Institute, University of Auckland
Co-organizer: Blanca Rodriguez, Department of Computer Science, University of Oxford

Rationale:

Multi-scale models attempt to simulate emergent function of a system from the interaction of their subcomponents, which themselves may operate over vastly different length and time scales. Physiological and biological models employ a wide range of mathematical approaches that are each generally suited for a particular scale of interest or a specific task. For example network approaches for metabolic models, systems of ordinary differential equations for single cell function, partial differential equations to describe fluid and solid mechanics, and cellular automota and agent-based models that can address complexity whilst retaining physiological meaning. Now as we attempt to measure, describe, conceptualise, and model functions that occur over ranges of spatial and/or temporal scales, we generally require the simultaneous use of different types of mathematics within a single multi-scale modelling framework. The key challenge is to determine consistent and physically realistic mathematical descriptions of the coupling between scales, which is necessary for robust predictions of emergent function and synergistic interactions. Associated challenges are those of model design and size, and methods to transfer parameters both up and down between the scales. Of equal importance to the conceptualisation and mathematics of function is the mathematical description of the physical domain in which function is predicted: studying structure-function interactions requires models that capture the important geometric features of e.g. a tissue or organ. This introduces a further challenge of trade-off between geometric complexity and the ability to model function with high precision. Appropriate model reduction and parameterisation therefore remains key, alongside the need to keep models amenable to classic mathematical analysis. Approaches to address these challenges in different organ systems and across different scales will be considered.

Presentations:

  • Pablo Lamata, Sander Land, Nicolas Smith, Steven Niederer: Moving from patients to populations through robust personalised cardiac modelling
  • Oliver Röhrle, Michael Sprenger, Syn Schmitt: Using nested models to achieve a multiscale, continuum-mechanical, forward dynamics simulation framework for musculoskeletal systems
  • John Rice, Slava Gurev: High-resolution models of human ventricles: electrophysiology and mechanics
  • Paris Perdikaris, Leopold Grinberg, George Karniadakis: An effective fractal-tree closure model for simulating blood flow in large arterial networks
  • Alessio Gizzi, Christian Cherubini, Annamaria Altoma:
  • Three-dimensional thermo-viscoelasticity for electroactive intestine modeling
  • Blanca Rodriguez, Oliver Britton, Kevin Burrage, Esther Pueyo, Alfonso Bueno-Orovio: Multiscale mathematical models to capture and investigate sources and modulators of variability in cardiac electrophysiology

Metamodelling methodology for easing model construction and validation

Organizer: Kristin Tøndel, Simula Research, Oslo
Co-organizer: Gunnar Cedersund, Department of biomedical engineering, Linköping University

Metamodelling methodology for easing model construction and validation

Organizer: Kristin Tøndel, Simula Research, Oslo
Co-organizer: Gunnar Cedersund, Department of biomedical engineering, Linköping University

Rationale:

Computational physiology models do in general contain a large number of parameters, numerous state variables, and intricate functional relationships between these state variables. Thus the parameter-to-phenotype map, i.e. the multidimensional mapping between model inputs and model outputs, can possess a very complex topography. Robust development and use of models require an explication of this topography. A metamodel of a given computational physiology model is a statistical prediction model that provides a description of the parameter-to-phenotype map. It is becoming increasingly clear that metamodels may become very useful for computational physiology in connection with model construction, model validation, parameter estimation, sensitivity analysis, model comparisons and computational compaction facilitating clinical use. The workshop will provide a state-of-the-art overview of metamodelling methodology and associated methodology for making use of metamodels in these application areas.

Presentations:

  • Kristin Tøndel, Andrew Crozier, Steven Niederer, Nic Smith: Insight into model mechanisms and more efficient model development and validation by multivariate metamodelling
  • Harald Martens, Anders Lyngvi Fougner, Hodjat Rahmati, Øyvind Stavdahl, Ole Morten Aamo: From measurements to models and back – and forth
  • Kristin McLeod, Kristin Tøndel, Samuel Wall, Jørg Saberniak, Kristin Haugaa: Metamodelling of structural abnormalities in the ARVC heart
  • Victor Martins dos Santos, Willem de Vos, Edoardo Saccenti: The Virtual Intestine: Metamodelling and systems biology of host– food– microbe interactions in the mammalian gut
  • Jeroen Feher, Susheel Varma, Martina Sciola, Paul Morris, Rod Hose: Characterising uncertainty of VPH-related multiscale models
  • Gunnar Cedersund, Martin Gollvik, Markus Karlsson, Mikael Forsgren, Olof Leinhard-Dahlqvist, Harald Martens, Peter Lundberg: Inverse metamodelling makes a mechanistic model for liver diagnosis robust and clinically fast

Model-guided medical device design and assessment

Organizer: Marco Viceconti, Department of 
Mechanical Engineering, University of Sheffield
Co-organizer: Claudio Cobelli, Department of Information Engineering, University of Padova

Model-guided medical device design and assessment

Organizer: Marco Viceconti, Department of 
Mechanical Engineering, University of Sheffield
Co-organizer: Claudio Cobelli, Department of Information Engineering, University of Padova

Rationale:

Traditionally, bioengineers use computational approaches in the early phases of the design of a new medical device, primarily for exploratory purposes.  Once a basic design is defined, prototypes are realised, and then tested first pre-clinically and then clinically, with multiple cycles of design revision as the product shortcomings become evident.  But in the last few years pioneering efforts have been made to explore the possibility of making much more extensive use of computer simulation to improve the medical device design process. This opens up a number of very interesting possibilities, but also raises a number of concerns on the accuracy, reliability, and certification of these in silico medicine methods when used in the development and assessment of medical devices.  Like many other industrial sectors virtual prototyping can improve the products and reduce the costs, but for medical devices this must be achieved with an extreme attention to the reliability of these simulations. The workshop will provide an overview of this pioneering work, explore innovative uses of advanced patient-specific models to guide device design, to improve the pre-clinical assessment phase and to reduce, refine and to some extent replace clinical experimentation, and discuss how to address associated model reliability issues. 

Presentations:

  • Claudio Cobelli: The FDA accepted type 1 diabetes simulator: In silico desing and test of the artificial pancreas
  • Alfons Hoekstra, Charles Bona-Casas, Hannan Tahir, Joris Borgdorff: A 3D Multiscale In-Stent Restenosis Model: A milestone for the next generation of in-stent restenosis modeling 
  • Nenad Filipovic, Dalibor Nikolic, Igor Savelijic, Branislav Jeremic, Slobodan Arsenijevic: Modeling of cervical dilator device 
  • Esra Neufeld, Niels Kuster: Methodology for assessing MRI implant safety and designing MR- safe implants using functionalized anatomical models 
  • Tristan Belzacq, Vit Novácek, Gaëtan Guérin, Frédèric Turquier: Patient-specific computational abdominal wall modelling: Application to abdominal wall hernia repair 
  • Bertrand Fréchède, Stéphane Howley, Kristy Tan, Yoann Lafon: Subject-specific approach and active muscle implementation in a 3D FE model of the neck for orthopaedics applications

Model-guided tissue engineering and stem cell therapy

Organizer: Liesbet Geris, Biomechanical 
Engineering, University of Liege
Co-organizer: Jose Manuel Garcia-Aznar, University of Zaragoza

Model-guided tissue engineering and stem cell therapy

Organizer: Liesbet Geris, Biomechanical 
Engineering, University of Liege
Co-organizer: Jose Manuel Garcia-Aznar, University of Zaragoza

Rationale:

The creation of man-made living implants is the holy grail of tissue engineering (TE). As basic science advances, one of the major challenges in TE is the translation of the increasing biological knowledge of complex cell and tissue behavior into a predictive and robust engineering technology of clinical relevance. Considering the biological complexity involved, computational modeling is likely to become a key factor in the maturation of the TE field and its clinical impact. This workshop aims to give an overview of computational modeling efforts in the field of tissue engineering across the whole span of biomedically relevant tissues and organs. Specific attention will be given to which modelling formalisms are most appropriate to use for a given medical application and how these formalisms are connected to data-driven or hypothesis-driven approaches. Furthermore, attention will be given to issues related to specific model building, training and validation, and to issues related to how in silico work should be connected to in vitro and in vivo work in the TE field in order to ensure maximum progress.

Presentations:

  • Janine Post, Stefano Schivo, Jetse Scholma, Jaco Van der Pol, Johan Kerkhofs, Liesbet Geris, Rom Langerak: An ECHO in biology: validating the executable chondrocyte
  • Hans Van Oosterwyck, Tim Odenthal, Bart Smeets, Herman Ramon: Quantifying cell mechanical forces as determinants of stem cell fate
  • Jose Manuel Garcia-Aznar, Ismael Gonzalez-Valverde, Thomas Rüberg: Multiscale Simulation of Cell Migration for the Guidance of Tissue Growth
  • Adrien Baldit, Ana Campos, Marzia Brunelli, Cécile Perrault, Damien Lacroix: Multiscale modelling in tissue engineering: a Virtual Physiological approach
  • Yann Guyot, Ioannis Papantoniou, Jan Schrooten, Liesbet Geris:
  • A Multiphysics model of neotissue growth in a perfusion bioreactor
  • Aurélie Carlier, Nick van Gastel, Geert Carmeliet, Hans Van Oosterwyck, Liesbet Geris:
  • To heal or not to heal: designing successful bone tissue engineering strategies with a multiscale bioregulatory model

Models for surgical decision support

Organizer: Kerstin Denecke, Innovation Center Computer Assisted Surgery (ICCAS), Medical Faculty, University of Leipzig
Co-organizer: Gabriel Kiss, Department of Circulation and Medical Imaging, NTNU

Models for surgical decision support

Organizer: Kerstin Denecke, Innovation Center Computer Assisted Surgery (ICCAS), Medical Faculty, University of Leipzig
Co-organizer: Gabriel Kiss, Department of Circulation and Medical Imaging, NTNU

Rationale:

Improved clinical examination methods and developments towards personalized medicine lead increasingly to complex patient data to be considered within clinical decision making processes. Clinical decision support systems aim at making the optimum use of patient data and are supposed to support in this process of information management and interpretation. They should lead to high percentages of appropriate treatment and reduce mortality and complications. However, such systems did not yet arrived sufficiently in clinical practice, since approaches often lack relationships to scientific evidence and are poorly integrated with clinician's workflow. In the last years, the idea of model-based decision support came up that bases upon two assumptions:

  1. Medical knowledge can be modelled including diagnosis, treatment, and decision making processes, i.e. it can be formally described which parameter characterize a specific diagnosis or which steps are performed within a decision-making process (domain theory).
  2. The observations made during physical or other clinical assessment of a patient can be described and instantiates the formal patient model (situation description).

The workshop aims at providing an excellent opportunity for the presentation and discussion of state of the art in model-based decision support as well as for the presentation of models and concepts  specifically designed for surgical decision support. The workshop further intends to collect requirements and key challenges to be addressed in future with respect to digital patient modeling for surgical decision support.

Presentations:

  • Kerstin Denecke: Model-based decision support: A multi-layer approach to Digital Patient modeling
  • Jörg Sabczynski, Maria João Cardoso, Jaime S. Cardoso, David Hawkes, Gerrit-Jan Liefers, Mohammed Keshtgar, Ralph Sinkus, Björn Eiben: PICTURE: Predicting the cosmetic outcome of breast cancer surgery
  • Francesco Migliavacca, Lorenza Petrini, Debora Testi, Michel Rochette, Gordon Clapworthy, Gabriele Dubini, Giancarlo Pennati: Real time fatigue behaviour of peripheral stents
  • Gabriel Kiss, Hans Torp: Multi-modal visualization and augmentation for cardiac applications
  • Thomas Langø, Sinara Vijayan, Frank Lindseth, Sebastien: Feasibility of 4D ultrasound-based motion tracking in focused ultrasound therapy of tumors in moving abdominal organs
  • Sjur Urdson Gjerald, Sebastian Sarvari, Hans Henrik Odland, Samuel Wall: A generic right ventricle model for simulating patient-specific pacing procedures from echocardiography data


Summary of discussion:

Key challenges of integrating the models with clinical work were discussed.
Some of these are listed below:

  • Realism: the developed models need to have a certain degree of realism in order for them to be accepted by the clinical partners.
  • Reuse of methods: Models should be made available between various research groups, this would reduce the amount of overhead (same techniques implemented at multiple institutions) and would allow for more accurate comparisons among the proposed models
  • Standardized evaluation: would be desirable to have standardized evaluation tools
  • Databases: Consistent and well annotated databases are desirable, this would greatly simplify the validation process
  • Computational time: Some assumptions and simplifications are always made to reduce the computation time. An inherent tradeoff needs to be made between the computation time and the realism of the models. Real-time applications are often requested for applications.
  • Clinically predefined errors: Most clinical partners define an acceptable clinical error and most likely they will not use the methods if they are above the error limit.
  • Balance between technical innovation and clinical needs: Technical partners always push for the newest methods. The question is if they are required for a given application?
  • Evaluation and validation: A new model should be proven to be better than existing methods to justify its clinical use.
  • Collaboration: Several groups are involved in big European projects involving up to 15 different partners. Good collaborations and exchange of ideas are possible even if many different institutions participate in a given project.

Multiscale modelling of cancer

Organizer: Georgios S. Stamatakos, In Silico Oncology Group, Institute of Communication and Computer Systems, National Technical University of Athens
Co-organizer: Holger Stenzhorn

Multiscale modelling of cancer

Organizer: Georgios S. Stamatakos, In Silico Oncology Group, Institute of Communication and Computer Systems, National Technical University of Athens
Co-organizer: Holger Stenzhorn

Rationale:

Cancer is a highly complex disease and natural phenomenon. It is manisfested at virtually all spatiotemporal scales pertinent to life, spanning from the atomic to the whole body spatial scale and from the nsec to the year temporal scale. The plethora of interdependent mechanisms jointly constituting the natural phenomenon of cancer and its response to treatment dictates the development of complex mathematical and computational multiscale models aiming at both the quantitative understanding of the phenomenon and the optimization of cancer treatment in the patient individualized context. Clinical adaptation and validation are two sine qua non processes in view of the clinical translation of such models (in silico oncology). Clinically validated cancer models  are expected to serve as platforms for performing in silico experiments by exploiting the patient's own multiscale data such as imaging, histological and molecular data in order to select the most appropriate treatment scheme. The workshop aims at providing an excellent opportunity for the presentation and discussion of state of the art multiscale cancer modelling efforts along with clinical translation activities. Since the development of cancer hypermodels of which the component models may be developed by different  modeling groups is becoming a more and more realistic scenario, aspects of the joint development of cancer hypermodels on a global scale will also be addressed.

ePoster presentations

TAILORED-treatment – developing algorithms for diagnosing the septic patient
J. Hays

Ultrasound-guided regional anaesthesia assisted by patient specific VPH models
F. Lindseth, I. Reinertsen, E. Smistad, C. Askeland, K. Johansen

Drug synergies in a gastric cancer cell line discovered by signal transduction modeling
Å. Flobak, A. Baudot, E. Remy, L. Thommesen, D. Thieffry, M. Kuiper, A. Lægreid

Modeling of the human induced pluripotent stem cell-derived cardiomyocyte action potential
M. Paci, S. Severi, J. Hyttinen

The inter-sample structural variability of regular tissue-engineered scaffolds affects significantly the local cell environment
A. Campos Marin, D. Lacroix

Mathematical modelling of bone formation based on local Ca2+ concentrations in calcium phosphate scaffolds
V. Manhas, Y. Guyot, Y. Chai, G. Kerckhofs, A. Carlier, L. Geris

The GRECO consortium: Taking care of gene regulation knowledge
M. Kuiper, A. Lægreid, T. Greco

Logic synthesis methods for medical data mining
G. Borowik, A. Grzanka, T. Łuba, A. Kra´sniewski

s-core network decomposition: A generalization of k-core analysis to weighted networks
M. Eidsaa, E. Almaas

Ultrasound-mediated delivery of nanoparticles to tumour tissue
C. Davies, S. Eggen, A. Åslund, R. Hansen, B. Angelsen, Y. Mørch, S. Berg

Long-term liver motion tracking with 4D ultrasound
S. Muller, S. Vijayan, T. Langø

Reinterpreting b-adrenergic sensitivity from time-resolved data via minimal mechanistic modelling
W. Lövfors, E. Nyman, K. Lundengård, J. Altimiras, G. Cedersund

A cellular potts model of a stented arterial vessel to study the role of initial smooth muscle cells migration in neointimal development
H. Tahir, I. Niculesco, R. Merks, A. Hoekstra

Three-layer-isotropic skull conductivity representation in the EEG forward problem using spherical head models
H. Hallez, E. Cuartas-Morales, G. Castellanos-Gonzales, B. Vanrumste

Multi-level modeling of insulin signaling: Merging of islands of high-quality models with a background systems-level model for the entire proteome
R. Magnusson

Temporal control and activation of inflammatory genes in endo thelial cells in response to wall shear stress.
D. Baeriswyl, R. Shipley, Y. Ventikos

A 3-D electrophysiological model of the human left ventricle
S. Pravdin, H. Dierckx, A. Panfilov, V. Berdyshev, L. Katsnelson, O. Solovyova,
V. Markhasin

Generating patient-specific finite-element meshes by parametric constrained image registration
D. Silva, D. Barber, R. Hose

Verification and validation of computational modelling of medical devices and treatment: Critical considerations in view of multi-physics modeling of in vivo processes
E. Neufeld, N. Kuster

Creating patient specific models to optimise multipolar lead pacing
protocols following reverse remodelling in cardiac resynchronisation therapy patients

A. Lee, M. Sohal, G. Plank, R. Razavi, P. Lamata, C. Rinaldi, S. Niederer

A generalized model of tumor growth and response to treatment using the PUN approach
C. Guiot, I. Stura

Nano network of neurons from communication engineering perspective
M. Veletic, P. Floor, F. Mesiti, I. Balasingham

INEX – A simple but effective computational model to simulate neuronal-astrocytic activity
K. Lenk, I. Vornanen, E. Räisänen, J. Hyttinen

The digital salmon: Virtual physiology in aquaculture
J. Vik, A. Gjuvsland, J. Torgersen, S. Lien, S. Omholt

3D model of a heart in blender
M. Michalska, A. Grzanka

A flexible one dimensional simulation framework for blood flow in arterial networks
V. Eck, L. Hellevik

VPH-share: Patient-centred multiscale cloud-enabled computational  workflows
S. Varma, M. Villa-Uriol, E. Manchini, E. Schileo, P. Lamata, N. Smith, R. Hose

MR-Elastography in nonlinear deformable materials: Towards cardiac MRE
D. Nordsletten, O. Holub, S. Lambert, L. Bilston, R. Sinkus

Mechanistic modelling investigates the neural basis behind the hemodynamic response in fMRI
K. Lundengård, B. Svensson, S. Sten, J. Schyberg, G. Cedersund, F. Elinder,
M. Engström

Visual features in phonemes classification
M. Sobotka,A. Grzanka

Computational prediction the effect of V241F KCNQ1 mutation on human atrial fibrillation
K. Lim, R. Imaniastuti, J. Yeom, E. Shim

Determination of the role of AQP-4 channels in neuromyelitis optica through mathematical modelling
S. Laranjeira, M. Symmonds, J. Palace2, S. Payne, P. Orlowski

A first truly systems level mechanistic model – unravelling the gene regulation of Th2 differentiation
M. Köpsén, W. Lövfors, G. Cedersund, M. Benson, M. Gustafsson

‘E-SPINE', The virtual physiological Korean spine project
Y. Kim

Privacy-preserving scalable data federation with applications to research patient registries
S. Varma

Introduction of Korean human project; Digital Korean and visible Korean
S. Lee, S. Lee, S. Lee, H. Par

Asthma sputum inflammatory phenotypes, a transcriptome analysis using taverna a knowledge discovery framework
A. Maleki-Dizaji, C. Newby, R. Berair, R. Smallwood, C. Brightling

Differential gene co-expression networks in human tissue
A. Voigt, K. Nowick, E. Almaas

The gene eXpression knowledge base
A. Venkatesan, S. Tripathi, A. Sans De Galdeano, W. Blondé, A. Lægreid, V. Mironov, M. Kuiper

Mathematical model of ammonia handling in the rat renal medulla
L. Noiret, S. Thomas

Electrotonic and structural effects of fibrosis on the genesis of atrial fibrillation: Insights from 3D human atria model
R. Morgan, M. Colman, M. Kruger, G. Seemann, K. Rhode, O. Aslanidi

Practical identifiability of cardiac constitutive laws using 3D tagged MRI
D. Nordsletten, M. Hadjicharalambous, R. Chabiniok, L. Asner, N. Smith, R. Razavi

Human dentofacial system as an element of project "Virtual Physiological Human"
Y. Nyashin, V. Lokhov

GPU accelerated finite element analysis of trabecular bone tissue
C. Nita, Y. Chen, L. Lazar, V. Mihalef, L. Itu, M. Viceconti, C. Suciu

Independent stress and deformation control at medical treatment
V. Lokhov, Y. Nyashin

A biophysically detailed model of cardiac contraction predicts the cooperative interactions between crossbridges and tropomyosin
S. Land, N. Smith, S. Niederer

Neuronal stimulation scenarios at nanoscale
F. Mesiti, P. Floor, M. Veletic, I. Balasingham

Non-invasive relative pressure estimates from fluid energy using PCMRI
D. Fabrizio, P. Lamata, C. Figueroa, N. Smith, D. Nordsletten

Physiology of human muscle oxygenation after treatment of patients suffering ischemic heart disease
A. Paiziev, N. Sayfiev, S. Payzieva

A simple finger tracking computerized test: Identification of parametric VPH motor model
F. Kharaman, M. Da Lio, L. Zaccarian, M. De Cecco, M. Confalonieri

Modelling the functional impact of KCNA5 mutations on the physiological and mechanical activities of the human atria
H. Ni, M. Colman, H. Zhang

P300 generators in Parkinson's disease according to symptoms: A LORETA modelling study
O. Ivanenko, S. Kryzhanovskyi, A. Cherninskyi, I. Zyma, I. Karaban

NoTremor: Virtual, physiological and computational neuromuscular models for the predictive treatment of Parkinson's disease.
K. Moustakas, K. Votis, D. Tzovaras, K. Gurney, P. Brown, M. Hu, M. Da Lio

A computational framework for describing the saccadic eye movement system of the Parkinsonian digital patient
S. James, A. Blenkinsop, S. Anderson, C. Papapavlou, K. Moustakas, K. Gurney

Intelligent predictive artificial pancreas by biomimicry of the native beta cells – concept, prototype and the future
N. Skjaervold, A. Fougner, N. Elvemo

Modelling insulin and glucose dynamics in diabetes mellitus type 1: Intravenous, subcutaneous and intraperitoneal approach
K. Kölle, A. Fougner, H. Frøyen, Ø. Stavdahl

Mathematical modelling of glucose homeostasis for usage in an intensive care unit
A. Hjelm, G. Cedersund, F. Sjöberg, M. S Chew, L. Bergenzaun, R. Johansson, H. Naga

Comprehensive laboratory animal monitoring system
H. Johannessen, M. Olsen, C. Zhao,D. Chen

Systems biology based analysis of metalloproteins of Mycobacterium tuberculosis
S. Soumi, E. Almaas
 

Hands on demonstrations of VPH-compliant software

Share, discover and access biomedical resources:  VPH-Share in action

Debora Testi, VPH-Share Platform

The VPH-Share presented the software platform which allows biomedical researchers and clinicians to share data, and tools and to compose them in complex workflows to build new knowledge. The system is accessible via an easy to use web interface and allows searching for resources, the upload/creation of new resources, the remote execution of tools and workflows relying on an efficient cloud platform.

VPH tools from the Auckland Bioengineering Institute

Poul Michael Fonss Nielsen

This tutorial highlighted three VPH-compliant tools, PMR2, OpenCOR, and MAP Client. PMR2 is a software framework for providing data and knowledge repositories, with the data encapsulated in version controlled workspaces. OpenCOR is a tool for model editing, annotation, and simulation, initially based on CellML but extensible to other model encoding standards. The MAP Client is a workflow editor and execution tool being developed in order to capture the semantics of workflows.

The Flower of Life: Vessels in the human lungs (red) and first generations of branches starting from the trachea (green).

Photo: Solveig Fadnes, MI Lab and Department of Circulation and Medical Imaging, NTNU