Article

Computational Fluid Dynamics and Aortic Dissections: Panacea or Panic?

Register or Login to View PDF Permissions
Permissions× For commercial reprint enquiries please contact Springer Healthcare: ReprintsWarehouse@springernature.com.

For permissions and non-commercial reprint enquiries, please visit Copyright.com to start a request.

For author reprints, please email rob.barclay@radcliffe-group.com.
Information image
Average (ratings)
No ratings
Your rating

Abstract

This paper reviews the methodology, benefits and limitations associated with computational flow dynamics (CFD) in the field of vascular surgery. Combined with traditional imaging of the vasculature, CFD simulation enables accurate characterisation of real-time physiological and haemodynamic parameters such as wall shear stress. This enables vascular surgeons to understand haemodynamic changes in true and false lumens, and exit and re-entry tears. This crucial information may facilitate triaging decisions. Furthermore, CFD can be used to assess the impact of stent graft treatment, as it provides a haemodynamic account of what may cause procedure-related complications. Efforts to integrate conventional imaging, individual patient data and CFD are paramount to its success, given its potential to replace traditional registry-based, population-averaged data. Nonetheless, methodological limitations must be addressed before clinical implementation. This must be accompanied by further research with large sample sizes, to establish the association between haemodynamic patterns as observed by CFD and progression of aortic dissection.

Disclosure:The authors have nothing to disclose.

Received:

Accepted:

Published online:

Citation: Vascular & Endovascular Review 2018;1(1):27–9.

Correspondence Details:Andrew MTL Choong, Department of Cardiac, Thoracic and Vascular Surgery, National University Heart Centre, Singapore, Level 9, NUHS Tower Block, 1E Kent Ridge Road, Singapore 119228. E: suramctl@nus.edu.sg

Copyright Statement:

This work is open access under the CC-BY-NC 4.0 License which allows users to copy, redistribute and make derivative works for non-commercial purposes, provided the original work is cited correctly.

The traditional approach to investigation and management of aortic dissections has revolved around clinical examination, laboratory tests and a slew of imaging modalities including CT, chest X-ray, MRI, ultrasound and transoesophageal echocardiography.1

However, an inherent limitation of these techniques is that they do not consider the temporal dynamicity of aortic blood blow; they capture only a snapshot of the blood flow at single points in time.

Recent research has highlighted the use of computational fluid dynamics (CFD) as a complementary tool to improve our limited understanding of the complex biomechanical behaviour of blood flow in both normal aortas and those with pathology.

The potential application of CFD is widespread, spanning from technological development of new devices to routine clinical decision-making.2,3

The amalgamation of engineering and medical disciplines has allowed computer simulation to be used to solve numerical equations related to fluid flow. Since CFD’s inception in the 1950s by researchers from the Massachusetts Institute of Technology,4 several studies have attempted to employ CFD techniques to analyse blood flow in different aortic pathologies including aortic aneurysms,5–7 aortic dissections5,8–12 and differences before and after endovascular aortic repair (EVAR).13

A combination of technological advancements in computing software (ANSYS FLUENT,14 Open Foam,15 SIMVascular,16 ADINA,17 in-house coding18 and the falling cost of supercomputing have paved the way for the use of computing resources to solve mathematical equations in medicine. The Navier–Stokes equation, for instance, allows modelling of intravascular pressure and flow parameters. The use of appropriately framed model parameters gives a realistic picture of blood flow and pressure waveforms in real time, enabling the investigation of blood flow velocity in relation to pressure and density as well as a myriad of other stresses and forces, including ones that cannot be measured, such as wall shear stress (WSS).

Haemodynamics is considered to play a paramount role in the development and progression of all types of aortic dissection aortic dissection but unfortunately remains poorly understood. Hitherto, no clinical consensus has been established as to whether medical, surgical or endovascular treatment is most appropriate for the management of aortic dissection. This has plausibly been attributed to the lack of an imaging criteria to determine the best treatment for individual patients. Therefore, there is burgeoning interest in the use of patient-specific CFD in clinical decision-making.

This review aims to provide an overview of the benefits and challenges of CFD in the management of aortic dissections.

Haemodynamic Changes in True and False Lumens

Imaging techniques such as CT angiography and MRI have allowed clinicians to accurately visualise the vasculature, which can then be reconstructed by employing various software packages such as Materialise Mimics (Materialise NV) and 3D Slicer (open source). When these are used with a suitable meshing algorithm, CFD software can be applied to the vascular geometry to run simulation tests, along with 4D flow MRI and other imaging modalities such as 2D PC-MRI to provide realistic boundary parameters.19 It is important to select the correct boundary condition for CFD models as this will improve CFD outcome as highlighted in the literature.20-22 As stated in a recent publication,23 it is important to consider the peripheral vascular network as a boundary condition, and model it through different elements of Windkessel Model. This will ensure a comprehensive analysis of blood flow dynamics, which should be useful in the future.

CFD could be employed to investigate haemodynamic changes in both the true lumen (TL) and false lumen (FL), where geometrical changes as a result of the dissection may change the entire flow field significantly. This may provide clues on when to treat an uncomplicated type B dissection. It remains a challenge to identify patients who are at the greatest risk of developing aneurysmal changes and should be given priority for treatment. This could be attributed to the unique geometrical features of the true and false lumen in every patient, which means that changes in the haemodynamic field vary between individuals. Karmonik et al. demonstrated that occlusion of the exit tear can cause an increase in FL pressure, since the geometry is altered after the occlusion. In addition, several studies have showed FL dilation causes a reduction in pressure within the FL.24 However, some studies have demonstrated that pressure in the TL is generally higher than in the FL. Recent publications with state-of-the-art CFD models have shown that the pressure difference between the TL and the FL is strongly affected by the distensibility of the aortic wall,25, 26 which should be given consideration when modelling the pressure in aortic dissection.

Cheng et al. showed that altered flow patterns in FL and TL may affect disease progression, and this is best explained by changes in wall shear stress (WSS). WSS exerted on the cell surface causes morphological deformation of the cells in the direction of blood flow, triggering rapid cytoskeletal remodelling and activating signalling cascades with the consequent acute release of nitric oxide and prostacylin followed by activation of transcription factors including NF-κB, c-fos, c-jun and SP-1.27

Low WSS is also associated with endothelial dysfunction, reduced nitric oxide production, increased oxidative stress, atheroma/neointima formation and a propensity for vasoconstriction rather than vasodilatation.28 In contrast, high and moderate WSS is associated with good endothelial function, reduced expression of adhesion molecules, increased expression of endothelial nitric oxide synthase and reduction in oxidative stress.29,30 However, the threshold for low and high WSS appears controversial, and varies between studies. While Cheng et al. showed that WSS can go up to 17.98 Pa in the true lumen,31 Karmonik et al. showed that maximum WSS can decrease from 0.9 to 0.4Pa,32 and low WSS was determined to be less than 0.4 Pa. However, the authors believe that WSS can be geometry dependent, and might serve as an invaluable marker of vessel wall health, and thus, may help surgeons to prioritise patients for treatment.

Haemodynamics of Exit and Re-entry Tears

Wan Ab Naim and colleagues showed that a re-entry tear can provide a return path for blood flow back to the TL during systole and an extra outflow path into the FL during diastole, which may alter the progression of a dissected aorta.33 A high velocity profile located at the entry tear may result in high WSS. On the one hand, high time-averaged WSS (TAWSS) values have been found8,34 to increase the progression of the entry tear. Elevated WSS depends on the site of entry. On the other, a reduction in shear stress can minimise the propagation of dissection. However, because each patient has a unique anatomical structure, there is a large range of WSS values across various types of tears. For example, the TAWSS exceeded 5 Pa in a study by Alimohammadi et al.8 but twice this value was given (10 Pa) in a study by Karmonik et al.

Haemodynamic Differences Before and After Endovascular Aortic Repair

Improvements in haemodynamic patterns within the aorta are expected after endovascular aortic aneurysm repair but this varies from patient to patient depending on the specific pathology and boundary conditions.

Unfortunately, some patients develop thrombosis in the false lumen after EVAR, and this is postulated to be due to haemodynamic factors. Menichini et al., for instance,35 showed how turbulent flow in the aorta may promote thrombus formation in the FL, particularly following thoracic endovascular aneurysm repair (TEVAR). In addition, Wan Ab Naim et al. demonstrated that geometrical factors such as a re-entry tear and abdominal branches may cause the development of complete and incomplete FL thrombosis after stent graft repair.36 These studies show that haemodynamic changes should be monitored closely to assess the risk of thrombus formation in the FL. The use of computational flow dynamics may accurately provide crucial information about WSS and change in velocity patterns, allowing clinicians to assess the risk of thrombus formation in every patient.

Stent Design

In addition, CFD offers a platform for stent design optimisation, with the primary aim of reducing the haemodynamic impact (reduced sscillatory shear index, renal replacement therapy and TAWSS) of the stent on the vessel. Simulation tests allow for assessment of the stent’s mechanical and hemodynamic parameters that influence its performance. Strut thickness, for instance, has been found to be an important factor in predicting a stent’s performance.37–39

A benefit of CFD is it makes it possible to accurately analyse the myriad of factors that cause potentially devastating stent-related complications, including malpositioning, neointimal hyperplasia and collapse.40,41 Vascular surgeons are then able to identify patients at high risk of such complications, and can decide to implement prophylactic interventions.

Clinical Integration of Computational Flow Dynamics

CFD may prove to be an invaluable tool across the different stages of clinical management for patients who present with aortic dissection. However, much needs to be done to integrate CFD in virtual treatment planning and patient-specific risk prediction. Ideally, there should be smooth integration of a patients’ vasculature (cardiovascular imaging) with patients’ clinical data (baseline characteristics) before running a CFD simulation. However, once this has been attained, surgeons would have a comprehensive understanding of the condition they are dealing with, and can decide on the optimal treatment option.

Research wise, this may also represent a paradigm shift from population-based data to digital patient representations,42,43 the former of which is severely limited because it requires large participant numbers and clinical trials to establish evidence. Instead, the combination of (Bayesian) machine-learning methods and CFD virtual data would enable continuous predictions of outcomes, thereby reducing the cost, time and resources associated with large-scale clinical trials. However, data are insufficient at present to establish a multidimensional database for machine-learning methods to be conducted appropriately.44

Limitations

The benefits of CFD must be viewed in the context of known limitations. First, the sample sizes in published studies are small, given that most analyse a cohort of fewer than 30 patients. Ideally, a large cohort of patients should be recruited with long-term follow-up (1 year), to establish the association between progression of a dissected aorta and haemodynamic factors such as disturbed flow and elevated WSS.

Secondly, the CFD technique itself is limited by its failure to consider biochemical interactions, although this is understandable because it was first used to model kinetics.45 Therefore, CFD should never be used in isolation and improvements are warranted in terms of setting the boundary conditions.

Finally, CFD can never be entirely accurate in modelling the actual aortic environment, including pulsatile blood flow and vascular structure. Moreover, simulations may not be that specific to the individual patient given the continuous physiological fluctuations, which are affected by a host of factors such as lifestyle, medication or genetic predisposition. Integration of patient-specific data is lacking and should be addressed.

Conclusion

The adoption of CFD modelling is a new era in vascular surgery. While potentially highly useful in the diagnosis, prediction and prognostication of aortic dissections, the application remains in its infancy. Addressing methodological and logistical challenges are paramount before implementation into clinical practice.

References

  1. Erbel R, Aboyans V, Boileau C, et al. 2014 ESC Guidelines on the diagnosis and treatment of aortic diseases: Document covering acute and chronic aortic diseases of the thoracic and abdominal aorta of the adult. The Task Force for the Diagnosis and Treatment of Aortic Diseases of the European Society of Cardiology (ESC). Eur Heart J 2014;35(41):2873–926.
    Crossref | PubMed
  2. Sun Z, Chaichana T. A systematic review of computational fluid dynamics in type B aortic dissection. Int J Cardiol 2016;210:28–31.
    Crossref | PubMed
  3. Numata S, Itatani K, Kanda K, et al. Blood flow analysis of the aortic arch using computational fluid dynamics. Eur J Cardiothorac Surg 2016;49(6):1578–85.
    Crossref | PubMed
  4. Jones MR, Attizzani GF, Given CA 2nd, et al. Intravascular frequency-domain optical coherence tomography assessment of carotid artery disease in symptomatic and asymptomatic patients. JACC Cardiovasc Interv 2014;B(6):674–84.
    Crossref | PubMed
  5. Karmonik C, Bismuth JX, Davies MG, Lumsden AB. Computational hemodynamics in the human aorta: a computational fluid dynamics study of three cases with patient-specific geometries and inflow rates. Technol Health Care 2008;16(5):343–54
    PubMed
  6. Chen CY, Anton R, Hung MY, et al. Effects of intraluminal thrombus on patient-specific abdominal aortic aneurysm hemodynamics via stereoscopic particle image velocity and computational fluid dynamics modeling. J Biomech Eng 2014;136(3):031001;PMCID: PMC5101028.
    Crossref | PubMed
  7. Tse KM, Chiu P, Lee HP, Ho P. Investigation of hemodynamics in the development of dissecting aneurysm within patient-specific dissecting aneurismal aortas using computational fluid dynamics (CFD) simulations. J Biomech. 2011;44(5):827–36.
    Crossref | PubMed
  8. Alimohammadi M, Sherwood JM, Karimpour M, et al. Aortic dissection simulation models for clinical support: fluid-structure interaction vs. rigid wall models. Biomed Eng Online 2015;14:34; PMCID: PMC4407424.
    Crossref | PubMed
  9. Karmonik C, Muller-Eschner M, Partovi S, et al. Computational fluid dynamics investigation of chronic aortic dissection hemodynamics versus normal aorta. Vasc Endovascular Surg 2013;47(8):625–31.
    Crossref | PubMed
  10. Karmonik C, Bismuth J, Shah DJ, et al. Computational study of haemodynamic effects of entry- and exit-tear coverage in a DeBakey type III aortic dissection: technical report. Eur J Vasc Endovasc Surg 2011;42(2):172–7.
    Crossref | PubMed
  11. Karmonik C, Partovi S, Muller-Eschner M, et al. Longitudinal computational fluid dynamics study of aneurysmal dilatation in a chronic DeBakey type III aortic dissection. J Vasc Surg 2012;56(1):260–3.e1.
    Crossref | PubMed
  12. Cheng Z, Tan FP, Riga CV, Bicknell CD, Hamady MS, Gibbs RG, et al. Analysis of flow patterns in a patient-specific aortic dissection model. J Biomech Eng 2010;132(5):051007.
    Crossref | PubMed
  13. Karmonik C, Bismuth J, Davies MG, Shah DJ, Younes HK, Lumsden AB. A computational fluid dynamics study pre- and post-stent graft placement in an acute type B aortic dissection. Vasc Endovascular Surg 2011;45(2):157–64.
    Crossref | PubMed
  14. Ong CW, Ho P, Leo HL. Effects of Microporous stent graft on the descending aortic aneurysm: a patient-specific computational fluid dynamics study. Artif Organs 2016;40(11):E230–e40.
    Crossref | PubMed
  15. Kelly S, O’Rourke M. Fluid, solid and fluid-structure interaction simulations on patient-based abdominal aortic aneurysm models. Proc Inst Mech Eng H 2012;226(4):288–304.
    Crossref | PubMed
  16. Figueroa CA, Taylor CA, Yeh V, et al. Preliminary 3D computational analysis of the relationship between aortic displacement force and direction of endograft movement. J Vasc Surg 2010;51(6):1488–97; PMCID: PMC2874723.
    Crossref | PubMed
  17. Chandra S, Raut SS, Jana A, Biederman RW, Doyle M, Muluk SC, et al. Fluid-structure interaction modeling of abdominal aortic aneurysms: the impact of patient-specific inflow conditions and fluid/solid coupling. J Biomech Eng 2013;135(8):81001; PMCID: PMC3705803.
    Crossref | PubMed
  18. Filipovic N, Milasinovic D, Zdravkovic N, et al. Impact of aortic repair based on flow field computer simulation within the thoracic aorta. Comput Methods Programs Biomed 2011;101(3):243–52.
    Crossref | PubMed
  19. Stankovic Z AB, Garcia J, Jarvis KB, Markl M. 4D flow imaging with MRI. Cardiovasc Diagn Ther 2014;4:173–92; PMCID: PMC3996243.
    Crossref | PubMed
  20. Madhavan S, Kemmerling EMC. The effect of inlet and outlet boundary conditions in image-based CFD modeling of aortic flow. Biomed Eng Online 2018;17(1):66; PMCID: PMC5975715.
    Crossref | PubMed
  21. Moon JY, Suh DC, Lee YS, Kim YW, Lee JS. Considerations of Blood Properties, Outlet Boundary Conditions and Energy Loss Approaches in Computational Fluid Dynamics Modeling. Neurointervention 2014;9(1):1–8; PMC3955817.
    Crossref | PubMed
  22. Du T, Hu D, Cai D. Outflow boundary conditions for blood flow in arterial trees. PLoS ONE 2015;10(5):e0128597; PMCID: PMC4441455.
    Crossref | PubMed
  23. Morris PD, Narracott A, von Tengg-Kobligk H, et al. Computational fluid dynamics modelling in cardiovascular medicine. Heart. 2015;102(1):18–28; PMCID: PMC4717410.
    Crossref | PubMed
  24. Karmonik C, Bismuth J, Shah D, et al. Computational study of haemodynamic effects of entry-and exit-tear coverage in a DeBakey type III aortic dissection: technical report. Eur J Vasc Endovasc Surg 2011;42(2):172–7.
    Crossref | PubMed
  25. Bonfanti M, Balabani S, Greenwood JP, et al. Computational tools for clinical support: a multi-scale compliant model for haemodynamic simulations in an aortic dissection based on multi-modal imaging data. J R Soc Interface 2017;14(136). pii: 20170632; PMCID: PMC5721167.
    Crossref | PubMed
  26. Rudenick PA, Segers P, Pineda V, et al. False lumen flow patterns and their relation with morphological and biomechanical characteristics of chronic aortic dissections. Computational model compared with magnetic resonance imaging measurements. PLoS ONE 2017;12(1):e0170888; PMCID: PMC5270334.
    Crossref | PubMed
  27. Konner K, Nonnast-Daniel B, Ritz E. The arteriovenous fistula. J Am Soc Nephrol 2003;14(6):1669–80
    PubMed
  28. Krishnamoorthy MK, Banerjee RK, Wang Y, Z et al. Hemodynamic wall shear stress profiles influence the magnitude and pattern of stenosis in a pig AV fistula. Kidney Int 2008;74(11):1410–9.
    Crossref | PubMed
  29. Lehoux S, Castier Y, Tedgui A. Molecular mechanisms of the vascular responses to haemodynamic forces. J Intern Med 2006;259(4):381–92.
    Crossref | PubMed
  30. Harrison DG, Widder J, Grumbach I, Chen W, Weber M, Searles C. Endothelial mechanotransduction, nitric oxide and vascular inflammation. J Intern Med 2006;259(4):351–63.
    Crossref | PubMed
  31. Cheng Z, Tan FPP, Riga CV, et al. Analysis of flow patterns in a patient-specific aortic dissection model. J Biomech Eng 2010;132(5):051007.
    Crossref | PubMed
  32. Karmonik C, Partovi S, Müller-Eschner M, et al. Longitudinal computational fluid dynamics study of aneurysmal dilatation in a chronic DeBakey type III aortic dissection. J Vasc Surg 2012;56(1):260–3.e1.
    Crossref | PubMed
  33. Wan Ab Naim WN, Ganesan PB, Sun Z, et al. The impact of the number of tears in patient-specific Stanford type B aortic dissecting aneurysm: CFD simulation. J Mech Med Biol 2014;14(02):1450017.
    Crossref
  34. Chen D, Müller-Eschner M, von Tengg-Kobligk H, Barber D, Böckler D, Hose R, et al. A patient-specific study of type–B aortic dissection: evaluation of true-false lumen blood exchange. Biomedical engineering online. 2013;12(1):65; PMCID: PMC3734007.
    Crossref | PubMed
  35. Menichini C, Cheng Z, Gibbs RG, Xu XY. A computational model for false lumen thrombosis in type B aortic dissection following thoracic endovascular repair. J Biomech 2018;66:36–43.
    Crossref | PubMed
  36. Wan Ab Naim WN, Ganesan PB, Sun Z, et al. Flow pattern analysis in Type B aortic dissection patients after stent‐grafting repair: comparison between complete and incomplete false lumen thrombosis. Int J Numer Method Biomed Eng 2018;34(5):e2961.
    Crossref
  37. Gundert TJ, Marsden AL, Yang W, LaDisa JF, Jr. Optimization of cardiovascular stent design using computational fluid dynamics. J Biomech Eng 2012;134(1):011002.
    Crossref | PubMed
  38. Murphy JB, Boyle FJ. A full-range, multi-variable, CFD-based methodology to identify abnormal near-wall hemodynamics in a stented coronary artery. Biorheology 2010;47(2):117–32.
    Crossref | PubMed
  39. Martin D, Boyle F. Sequential structural and fluid dynamics analysis of balloon-expandable coronary stents: a multivariable statistical analysis. Cardiovasc Eng Technol 2015;6(3):314–28.
    Crossref | PubMed
  40. Keller BK AC, Hose DR, Gunn J, et al. Contribution of mechanical and fluid stresses to the magnitude of in-stent restenosis at the level of individual stent struts. Cardiovasc Eng Technol 2014;5:164–75.
    Crossref
  41. Pasta S, Cho JS, Dur O, et al. Computer modeling for the prediction of thoracic aortic stent graft collapse. J Vasc Surg 2013;57(5):1353–61.
    Crossref | PubMed
  42. Bonnici T, Tarassenko L, Clifton DA, Watkinson P. The digital patient. Clin Med (Lond) 2013;13(3):252–7.
    Crossref | PubMed
  43. Morris PD, Narracott A, von Tengg-Kobligk H, et al. Computational fluid dynamics modelling in cardiovascular medicine. Heart 2016;102(1):18–28; PMCID: PMC4717410.
    Crossref | PubMed
  44. Cheng Z, Riga C, Chan J, et al. Initial findings and potential applicability of computational simulation of the aorta in acute type B dissection. J Vasc Surg 2013;57(2 Suppl):35s–43s.
    Crossref | PubMed
  45. Karmonik C, Partovi S, Davies MG, et al. Integration of the computational fluid dynamics technique with MRI in aortic dissections. Magn Reson Med 2013;69(5):1438–42.
    Crossref | PubMed