The virtual physiology of the human artery

From virtual lab to clinical bedside: proposal to develop a multiscale computational model of the virtual human artery and evaluate pathophysiology prior to clinical intervention.
The virtual physiology of the human artery

Cardiovascular diseases remain the predominant cause of death worldwide, with recent reports projecting heart failure to increase dramatically in the next few years. In contrast to previous articles on laboratory bench-work, this post is on experiments conducted with computer-aided simulations in ‘virtual labs’ to address patient-specific cardiovascular disease. The virtual procedures primarily aim to evaluate pathophysiological mechanisms contributing to disorders of the cardiovascular system and facilitate accurate early clinical intervention.

One such complex vascular disorder, atherosclerosis, the primary cause of blood clotting leading to death (thrombosis)1, originates with accumulation of blood-borne lipids in the inner-most artery wall (intima), followed by a detailed immune response and progressive creation of a lesion in the vessel wall2. Mature lesions then form atherosclerotic plaques, restrict blood vessels (stenosis), which in turn effect local flow dynamics causing high shear stress, plaque rupture, platelet activation and thromboembolism to completely block upstream coronary arteries leading to heart attack and stroke3 (Figure 1)2.

Figure 1: Step-wise atherosclerosis progression to form thrombosis (Figure modified and reproduced from reference2).

The deployment of a stent (a mechanical scaffold)4, serves as a non-invasive clinical intervention procedure to restore stenosed (narrowed) artery to its original dimensions (Video 1). However, restenosis (re-narrowing) and thrombosis occur post-stenting to cause further clinical complications5. Drug-eluting stents (DES) were developed to actively inhibit smooth muscle cell (SMC) hyperproliferation responsible for re-narrowing artery (restenosis)6, and dual antiplatelet therapy (DAPT) was established on a case-by-case basis post-stenting to prevent thrombosis7,8. However, these medical interventions are not definite solutions and require regular optimization.

Video 1: Short animation of balloon assisted deployment of a coronary stent in a stenosed artery with plaque deposition (yellow substance), to restore normal blood flow (Video available via YouTube).

Similar to atherosclerosis, aneurysms are another major vascular disorder, mainly located in the abdominal aorta, thoracic aorta and vessels of the brain9. In all vascular disease mechanisms and their clinical intervention procedures, an intricate play exists between pulsating blood pressure and shear stresses exerted on the arterial wall, resulting in adverse biological responses3.

Hemodynamics is therefore a major field of study, due to its essential role in healthy vessel maintenance and arterial disease development. Since accurate measurement of in-patient (in vivo) blood-flow is arduous even with advanced medical imaging systems, computational hemodynamics takes precedence as a research tool to understand in vivo biological processes3. Many studies have been conducted by modelling arterial (patho)physiology to understand hemodynamics in vascular disease with computational assumptions made on the physics of blood rheology3. A few simulations are listed below (table 1).

Hemodynamics modeled

Computational technique integral to the virtual study


Patient-specific blood flow in coronary arteries

Three-Dimensional Finite Element Model of blood flow and vessel wall dynamics


Blood flow in stenotic vessels

Steady and unsteady Two-Dimensional and Three-Dimensional flow in stenotic vessels


Patient-specific arterial stiffness in normal, hypertensive and aneurysmal aortas.

Pulse Wave Imaging (non-invasive ultrasound imaging based technique)


Blood flow in end-to-side anastomoses (connection between two blood vessels)

Computational Fluid Dynamics (CFD) under laminar and weakly turbulent flow conditions.


Biomechanical stress in coronary atherosclerosis

Optical Coherence Tomography (OCT) angiography fusion-based Computational Fluid Dynamics (CFD)


Impact of arterial wall stress/strain and stiffness on atherosclerosis plaque formation at coronary bifurcations

OCT and magnetic resonance imaging (MRI) based computational structural analyses using finite element method.


Characterizing human coronary artery deformation to design coronary stents

Cardiac-gated computed tomography data based development of 3-D surface geometries


Table 1: Computational hemodynamics simulated in virtual labs using patient-specific medical images to understand blood flow in diseased and healthy arteries.

The physics-based models3 listed in table 1 were created using patient-specific medical images, while the modelling systems detailed next aim to include both biology and chemistry to more accurately represent pathophysiology of arterial disease3. For example, in-stent restenosis (ISR) is a clinical complication that occurs when smooth muscle cells (SMC) proliferate and form neointima (thickened artery) as part of an adverse biological response to mechanical stress exerted by the stent (Figure 2 A). The multiscience and multiscale nature of ISR after drug eluting stent (DES) implantation was modeled by a paradigm known as complex automata (CxA) that combined three different subprocesses operating on different time scales: 1) hemodynamics (blood flow), 2) Cell cycle dynamics of SMC (growth/apoptosis) and 3) drug diffusion mechanisms of the stent (Figure 2B). Preliminary results show that the subsequent two-dimensional model accurately reproduced tissue geometry and stent dimensions in agreement to those observed experimentally in vivo (Figure 3), facilitating hypothetical tests to comprehensively study ISR in-lab17. To evaluate realistic stent designs, however, it is necessary to run three-dimensional simulations3. Computational paradigms here are referenced from17 and highly simplified for brevity.

Figure 2: Simulating in-stent restenosis (ISR) for DES. A) Schematic process of stent implantation followed by in-stent restenosis. B) The simplified single-scale model depicts three simulated biological processes and their coupling. (Figures modified and reproduced from and ref 17).

Figure 3: Two-Dimensional simulation model of artery. Left: initial cell configuration after stent deployment, the struts are visibly embedded in upper/lower wall (black arrow) alongside intimal SMC agents in blue (white arrow). Right: simulation run for 28 days, developing neointima has reduced lumen diameter and increased wall shear stress (WSS) similar to in vivo studies. Fluid shear stress is colour coded (red high, blue low) (Figure modified and reproduced from ref 17).

Further algorithms can be included to model processes such as thrombus formation, endothelial loss and regrowth, in three-dimensions for a more realistic domain 18. The proposed next step forward in this research is to couple cell and molecular level physiological simulations across time scales, to create a multiscale model representing healthy and pathological arteries in their entirety (Figure 4)3.

Figure 4: Proposed virtual artery: a combination of models on several scales make up the virtual artery. These include whole-body one-dimensional models, three-dimensional fully resolved hemodynamics and cell-based models of blood, arterial wall and intracellular response to understand interaction between stent-artery (Figure modified and reproduced from reference3)

The projected vision will couple physiological components simulated over the years, including; 1) whole body models of blood flow19, 2) 3D hemodynamics based on numerical methods20, 3) cellular (Potts) models for arterial wall21, 4) models of tunica media17 and 5) cell based models for blood, platelet aggregation and clot formation22, in a multiscale model to form the composite virtual artery3. The proposal will rely on the multiscale modelling and simulation framework (MMSF), based on the concept of CxA3. Completion of the project will be a community-wide effort, to build the virtual artery in alignment to the Virtual Physiological Human (VPH) initiative. The virtual artery is expected to have a strong impact on clinical cardiology, enabling early treatment/intervention with greater accuracy23. Key limitations of the technique include computational issues related to simulations both conceptually and in practice3. For instance, some scale-bridging methods such as 1) coupling cell-based blood flow models to 3-D continuous fluid models and 2) coupling cell-based models of arterial wall to continuous solid mechanical models, are not well developed nor yet existent3. The multiscale simulations also require high-performance computing at petascale, to simulate hierarchies of physical processes3. Emerging exascale supercomputers in development by the ComPat project aim to lift such limitations 24.

The proposed virtual artery is a multiscale model of physiology that aims to bridge the gap between physics, chemistry and biology creating a powerful paradigm for the treatment of cardiovascular disease. The research will facilitate exciting possibilities, including combined evaluation of hemodynamics, biochemical and cellular interactions at the artery wall. In future, the virtual human artery will enable accurate early diagnosis of vascular disease and predict outcomes of planned clinical interventions such as stent implantation, in advance, to minimize risk and improve patient health.

Poster image: stented vessel fragment via news article: ‘Scientists design complete computer simulation of human artery’, December 27 2016.


  1. Domanski, M., Lloyd-Jones, D., Fuster, V. & Grundy, S. Can we dramatically reduce the incidence of coronary heart disease? Nat Rev Cardiol 8, 721-725, (2011).
  2. Libby, P., Ridker, P. M. & Hansson, G. K. Progress and challenges in translating the biology of atherosclerosis. Nature 473, 317-325, (2011).
  3. Hoekstra, A. G. et al. Towards the virtual artery: a multiscale model for vascular physiology at the physics–chemistry–biology interface. Phil.Trans.R.Soc.A. 374, 1-11, (2016).
  4. Sigwart, U., Puel, J., Mirkovitch, V., Joffre, F. & Kappenberger, L. Intravascular stents to prevent occlusion and restenosis after transluminal angioplasty. NEJM 316, 701-706, (1987)
  5. Jeewandara, M. T., Wise, G. S. & Ng, K. M. Biocompatibility of Coronary Stents. Materials 7,769-786,(2014).
  6. Stefanini , G. G. & Holmes , D. R. J. Drug-Eluting Coronary-Artery Stents. NEJM 368, 254-265, (2013).
  7. Mauri, L. et al. Twelve or 30 Months of Dual Antiplatelet Therapy after Drug-Eluting Stents. NEJM 371, 2155-2166, (2014).
  8. Palmerini, T. et al. Three, six, or twelve months of dual antiplatelet therapy after DES implantation in patients with or without acute coronary syndromes: an individual patient data pairwise and network meta-analysis of six randomized trials and 11 473 patients. Eur.Heart.J. 0,1-10, (2017).
  9. Isselbacher, E. M. Thoracic and Abdominal Aortic Aneurysms. Circulation 111, 816-828, (2005).
  10. Kim, H. J. et al. Patient-Specific Modeling of Blood Flow and Pressure in Human Coronary Arteries. Ann.Biomed.Eng 38, 3195-3209, (2010).
  11. Berger, S. A. & Jou, L. D. Flows in stenotic vessels. Annu.Rev.Fluid Mech. 32, 347-382, (2000).
  12. Li, R. X. et al. Pulse Wave Imaging (PWI) in Normal, Hypertensive and Aneurysmal Human Aortas In Vivo: A Feasibility Study. Phys.Med.Biol. 58, 4549-4562,(2013).
  13. Loth, F., Fischer, P. F. & Bassiouny, H. S. Blood flow in end-to-side anastomoses. Annu.Rev.Fluid Mech. 40, 367-393, (2008).
  14. Thondapu, V. et al. Biomechanical stress in coronary atherosclerosis: emerging insights from computational modelling. Eur.Heart.J. 38, 81-92, (2016).
  15. Ohayon, J. et al. Is arterial wall-strain stiffening an additional process responsible for atherosclerosis in coronary bifurcations?: an in vivo study based on dynamic CT and MRI. Am J Physiol Heart Circ Physiol. 301, H1097-H1106, (2011).
  16. Choi, G., Xiong, G., Cheng, C. P. & Taylor, C. A. Methods for Characterizing Human Coronary Artery Deformation From Cardiac-Gated Computed Tomography Data. IEEE Trans.Biomed.Eng. 61, 2582-2592, (2014).
  17. Caiazzo, A. et al. A Complex Automata approach for in-stent restenosis: Two-dimensional multiscale modelling and simulations. J.Comput.Sci 2, 9-17, (2011).
  18. Tahir, H. et al. Endothelial repair process and its relevance to longitudinal neointimal tissue patterns: comparing histology with in silico modelling. J.R.Soc.Interface 11, 1-11, (2014).
  19. van de Vosse, F. N. & Stergiopulos, N. Pulse Wave Propagation in the Arterial Tree. Annu.Rev.Fluid Mech. 43, 467-499, (2011).
  20. Axner, L. et al. Simulations of time harmonic blood flow in the Mesenteric artery: comparing finite element and lattice Boltzmann methods. Biomed Eng Online. 8, 1-8, (2009).
  21. Tahir, H., Niculescu, I., Bona-Casas, C., Merks, R. M. H. & Hoekstra, A. G. An in silico study on the role of smooth muscle cell migration in neointimal formation after coronary stenting. J.R.Soc.Interface 12, 1-10, (2015).
  22. Mountrakis, L., Lorenz, E. & Hoekstra, A. G. Scaling of shear-induced diffusion and clustering in a blood-like suspension. Epl-Europhys Lett 114, 1-7, (2016).
  23. Fenner, J. W. et al. The EuroPhysiome, STEP and a roadmap for the virtual physiological human. Phil.Trans.R.Soc.A. 366, 2979-2999, (2008).
  24. Borgdorff, J. et al. Performance of distributed multiscale simulations. Phil.Trans.R.Soc.A. 372, 1-17, (2014).