Effects of in situ fenestration stent-graft of left subclavian artery on the hemodynamics after thoracic endovascular aortic repair

Abstract

Objectives

The left subclavian artery during thoracic endovascular aortic repair could be reconstructed by in situ fenestration. This study aims to evaluate the effects of thoracic endovascular aortic repair with in situ fenestration thoracic endovascular aortic repair on the hemodynamics.

Methods

A male patient suffering from aortic dissection is treated by in situ fenestration thoracic endovascular aortic repair and the fenestration stent implanted in the left subclavian artery is partially protruding in the aortic arch for the stability. Two-phase non-Newtonian blood model is applied and three-element Windkessel model is implemented to reproduce physiological pressure waves. Simulations are carried out in three postoperative models to analyze different in situ fenestration thoracic endovascular aortic repair strategies; Case A: the protrusion length of fenestration stent is 23.2 mm representing the clinical postthoracic endovascular aortic repair aorta; Case B: the protrusion length is reduced by half simulating the improved surgery; Case C: the protruding portion is removed to simulate the ideal fenestration.

Results

In Case A, a pressure difference is found on the fenestration stent surface and a blood acceleration phenomenon around the stent is observed. Only 2.36% of the inlet blood flow is assigned to the left subclavian artery. In the improved surgery, the blood supply to the left subclavian artery is elevated to 4.01%. As for the ideal fenestration, a further improvement is observed (6.14%). Moreover, the aortic arch surface exposed to low time-averaged wall shear stress expands significantly when the protrusion length is shortened.

Conclusions

Overall, we conclude that appropriately shortening the protrusion length of the stent-graft may improve the efficacy of in situ fenestration thoracic endovascular aortic repair from the perspective of hemodynamics.

Keywords

Aortic dissection left subclavian artery reconstruction Windkessel model atherosclerosis computational fluid dynamics

Introduction

Thoracic aortic dissection is a catastrophic cardiovascular disease with a high mortality rate. Thoracic endovascular aortic repair (TEVAR) has been the preferred treatment approach as it is minimally invasive and is able to remodel the dissected aorta.¹ However, when the left subclavian arteries (LSAs) are involved in the aortic dissection, conventional TEVAR is not suitable. Fenestrated stent-graft technique is introduced to overcome the limitation of complex morphological features in clinical application.² Standard TEVAR with in situ fenestration thoracic endovascular aortic repair (ISF-TEVAR) consists of two steps. First, the origin of the involved supra-aortic branches could be intentionally occluded by the main endograft for adequate proximal landing zones. Then the blocked branches would be revascularized by in situ fenestration stent-graft, which is generally performed by puncture needles, radiofrequency probes, or laser.^3–5

Over the past decade, some clinical researchers have applied fenestrated stent-graft technique to aortic artery diseases. McWilliams et al.² described the first in situ graft fenestration clinical application which preserved the LSA during endovascular repair of a thoracic aortic aneurysm. Bicknell et al.⁶ reported their experience of treating abdominal aortic aneurysms using fenestrated stent-graft technology and the postoperative complication and mortality rates are low. Notably, Qin et al.⁷ successively performed in situ laser fenestration during TEVAR on 24 patients suffering from aortic dissection, aneurysm, or mural thrombus, and they demonstrated that fenestration is a feasible and safe technique for the reconstruction of the branches on aortic arch. Glorion et al.⁸ reviewed the development of fenestration and found that postoperative short-term results are satisfactory while long-term data remain scarce.

However, the hemodynamic consequences of differing stent-graft implantation in the LSA are largely unknown. Additionally, computational fluid dynamics (CFD) study about the efficacy of ISF-TEVAR is scarce and the hemodynamic parameters of the postoperative aorta are conducive to provide better insight into the role of fenestration.

In the present study, we aim at evaluating the effects of in situ fenestration on the hemodynamics after ISF-TEVAR. For this purpose, a two-phase non-Newtonian blood model is applied to simulate blood flow in postoperative aortas. To our knowledge, this is the first time to use CFD methodology to analyze patient-specific hemodynamics after ISF-TEVAR. Simulations are carried out in three postoperative models to compare three feasible surgical strategies in the clinic. The distribution of blood flow, wall shear stress-related indices, and energy loss (EL) are quantitatively explored to reveal the impact of the protrusion length of the stent-graft implanted in the LSA.

Methods

This study is approved by the Ethics Committee of Zhongshan Hospital, Fudan University, Shanghai, China (Ref No. Y2015-193).

Operative details

A 54-year-old male patient suffering from acute type B aortic dissection is analyzed with his written approval (Figure 1). The aortic arch in this study is type II,⁹ which is quite common and typical in the clinical vascular intervention.¹⁰ The preoperative proximal entry tear extends to the root of the LSA and the landing distance is too small to carry out standard TEVAR. For adequate proximal landing zones, the LSA is supposed to be intentionally covered by the main endograft during TEVAR. Therefore, the patient is treated with ISF-TEVAR to remodel the dissected aorta and rebuild the blocked LSA. Specifically, the main endograft is deployed to cover the proximal tear and LSA, then an ostium located at the root of the LSA is created by a new adjustable puncture device.¹¹ Finally, the fenestration stent is implanted in the LSA with partial body protruding into the main endograft for stability (23.2 mm). The detail procedure of the ISF-TEVAR is shown in Figure 2. After six-month follow-up, partial thrombosis is developed and the proximal false lumen gradually disappears.

Figure 1.

Three-dimensional reconstruction of the aortic geometries. (a) Postoperative CTA image, (b) postoperative MRI image, (c) postoperative geometry with in situ fenestration, (d) postoperative geometry with improved fenestration, (e) postoperative geometry with ideal fenestration, (f) inlet flow rate waveform, and (g) three-element Windkessel model.

Figure 2.

(a) Preoperative angiography; (b) development of aortic stent-graft, covering the LSA ostium; (c) a steerable sheath and an adjustable puncture needle were positioned perpendicular to the aortic stent-graft; (d) the needle pierces the fabric of the stent-graft; (e) enlarge the fenestration with a 4 mm × 30 mm balloon; (f) enlarge the fenestration with an 8 mm × 40 mm balloon; (g) deployment of a 10 mm × 50 mm covered stent (Viabahn; W.L. Gore & Associates, Flagstaff, AZ, USA) bridging the fabric hole toward the LSA; and (h) final angiography.

Geometry reconstruction

The original CTA images are collected to reconstruct clinical postoperative aorta (Case A) by using Mimics 19.0 (Materialise, Belgium), with supra-aortic branches retained and extended upward by 20 mm for the full development of the blood flow in GeoMagic Studio (GeoMagic Inc., USA). Other vascular branches of the descending aorta and residual false lumen are excluded as the focus of this study is aortic arch. The other two postoperative geometries are constructed by artificially reducing the protrusion length of fenestration stent for comparison. Specifically, the protrusion length is reduced by half (11.6 mm) to simulate the improved surgery (Case B). The protruding portion of the fenestration stent is totally removed to simulate the ideal fenestration (Case C). It could be implemented by a new covered stent, whose membrane is partially removed while retaining the bare stent structure. All the three fluid domains are meshed using ANSYS-ICEM (ANSYS Inc., USA) and each domain has more than 3,500,000 elements with five boundary layers near the wall. Grid-independence tests are performed for each geometry and the differences in peak wall shear stress and time-averaged wall shear stress (TAWSS) are all below 3%.

Numerical model

The blood is assumed as a two-phase non-Newtonian system, in which the particulate red blood cells (RBCs) suspend in the continuous plasma flow.¹² The platelets and other components with negligible volume fractions are ignored. The conservation equations of mass and momentum are solved separately for each phase and the Schiller and Naumann model is applied to determine the interphase momentum exchange.¹³

The blood is treated as an incompressible fluid with a mixture density of 1080 kg/m³.¹⁴ The plasma is Newtonian with a dynamic viscosity of 0.001 Pa s and a density of 1000 kg/m³.¹² The density of RBCs is set to 1178 kg/m³ according to the initial hematocrit of 0.45.¹⁵ A modified Carreau–Yasuda viscosity model is adopted to capture the non-Newtonian viscosity of the blood flow,¹² which results from the aggregation and deformation of the RBCs.^16,17 The Saffman-Mei inertial lift force model is used to describe the lateral migration of the RBCs.¹⁸ The vessel wall is assumed to be rigid with no-slip conditions. Notably, the protruding portion of the fenestration stent implanted in the LSA is almost stationary throughout the cardiac cycle according to our clinical MRI observation, so the interaction between the blood and stent is reasonably ignored in this study.

Computational details

A realistic flow rate waveform extracted from a previous study¹⁹ is applied at the ascending aorta inlet. The shear–stress transport turbulence model is used for plasma while the dispersed phase zero equation is applied for RBCs.²⁰ Three-element Windkessel model is implemented at the outlets of the supra-aortic branches and the descending aorta to reproduce physiological pressure waves, where the model parameters, that is proximal resistance (R₁), distal resistance (R₂), and compliance (C) of the downstream vasculature, are calculated according to previous report.²¹ The hematocrit at all boundaries is set to 0.45.¹⁵

The simulations are carried out on ANSYS Workbench 16.1 (ANSYS Inc., USA) and the Windkessel model is defined with CFX expression language. A constant time step of 0.001 s is adopted and the maximum residual for convergence is set as 10⁻⁵. All the simulations are run for three cardiac cycles to reach a periodic solution and the last cycle data are postprocessed using CEI Ensight 10.1 (ANSYS Inc., USA).

Results

Pressure and velocity

The presence of the fenestration stent in the aortic arch may have a significant impact on the blood flow field and the region around the stent is of particular importance. Figure 3 shows the postoperative distribution of pressure and velocity at a cut plane in the aortic arch. At peak systole (Figure 3(a) and (b)), the pressure within the fenestration stent is relatively low, indicating that a pressure difference is exerted on the stent surface. Moreover, a blood acceleration phenomenon is observed around the stent and the velocity inside the fenestration stent is not uniform. At early diastole (Figure 3(c) and (d)), the magnitude of pressure is decreased and the highest pressure region shrinks to the aortic arch wall adjacent to the descending aorta which also coincides with the high velocity region. The overall wall pressure distribution is in agreement with previous studies²² and the difference is negligible when the protrusion length of fenestration stent gets shorter.

Figure 3.

Pressure and velocity magnitude at a cut plane of aortic arch at (a, b) peak systole and (c, d) early diastole.

Blood flow

Figure 4 shows the quantitative comparisons of the flow rate results. At the postoperative aorta (Case A), it can be seen that the amount of flow crossing the supra-aortic branches which supplies the cerebrovascular and upper limb accounts for 34.39% of the inflow and the majority of the flow (25.06%) goes through the brachiocephalic artery. However, the LSA only receives 2.36% of the total inlet flow. When the surgery is improved (Case B), a slight increase is observed on the blood flow going through the supra-aortic branches (35.51%) and the blood supply to the LSA is also strengthened and increases to 4.01%. It should be noted that the blood flowing through the brachiocephalic artery and left carotid artery is reduced slightly. Moreover, when the fenestration is ideal (Case C), the amount of aortic branches blood supply increases further (37.34%) and the blood supply to the LSA elevates to 6.14%. The quantitative comparisons of the flow rate indicate that the protrusion length of fenestration stent may have a significant effect on the blood flow distribution.

Figure 4.

The comparison of the blood flow proportion crossing each outlet among the three cases. BT: brachiocephalic trunk; DA: descending aorta; LCA: left carotid artery; LSA: left subclavian artery.

Wall shear stress

The distribution of TAWSS for each case is illustrated in Figure 5. The postoperative TAWSS distribution of the ascending aorta and the aortic arch is uniform while a region of high TAWSS can be observed at the descending aorta due to the significant constriction and a sudden change in flow direction. When the fenestration is improved or ideal, there is no considerable difference on the overall distribution of TAWSS. It should be noted that the maximum TAWSS is relatively lower as the protrusion length of the fenestration stent becomes shorter. The lower panel in Figure 5 shows the comparison of low TAWSS region (less than 0.4 Pa²³) where atherosclerosis is more prone to form. At the postoperative aorta, the surface exposed to low TAWSS appears at the ascending aorta and the aortic arch adjacent to the left carotid artery. Notably, the low TAWSS region expands significantly when the protrusion length is shortened.

Figure 5.

Top: The comparison of TAWSS characteristics. Bottom: The comparison of low TAWSS regions (<0.4 Pa).

Figure 6 investigates the effects of the fenestration stent on oscillatory shear index (OSI) distribution which is related to flow oscillation. The postoperative high OSI can be observed at the ascending aorta, arch branches, and the top of descending aorta (Case A). OSI values are similar in the entire aorta for the three cases, while the most notable difference is found in the LSA owing to the presence of fenestration stent. The comparison of high OSI (more than 0.25) is shown in the lower panel in Figure 6, where the blood flow is considerably disturbed. When the fenestration is improved or ideal, the region exposed to high OSI at the LSA gets smaller, indicating that the protruding stent could strengthen the oscillatory nature of the flow.

Figure 6.

Top: The comparison of OSI characteristics. Bottom: The comparison of high OSI regions (>0.25).

EL

The comparison of EL throughout a cardiac cycle, defined as the difference between import total energy and all outlets energy, is shown in Figure 7. In the background, the gray dashed waveform is the normalized inlet flow rate and four vertical dashed lines correspond to four important time points: mid-systolic acceleration (T1 = 0.6 s), peak systole (T2 = 0.16 s), mid-systolic deceleration (T3 = 0.22 s), and early diastole (T4 = 0.42 s). It can be seen that the EL waveforms for fenestration stents with different protruding length follow the similar trend, especially between the Case A and Case B. While the notable differences only occur at two time periods, that is from T1 to T2 and T3 to T4, respectively. The peak EL for all three cases is observed at t = 0.8 s, which is close to the time point of mid-systolic acceleration (T1). When the fenestration is improved, a small reduction is found on the magnitude of EL and the average value, which is depicted by the horizontal dashed line of the corresponding color. The protruding length has a profound effect on the EL, with the average EL for the ideal fenestration (Case C) decreasing by 9.4%, from 0.0893 to 0.0809 W.

Figure 7.

EL profiles during a cardiac cycle. The gray dashed waveform is the normalized inlet flow rate and four vertical dashed lines correspond to four important time points: mid-systolic acceleration (T1 = 0.6 s), peak systole (T2 = 0.16 s), mid-systolic deceleration (T3 = 0.22 s), and early diastole (T4 = 0.42 s). Horizontal dashed lines represent average value. EL: energy loss.

Discussion

ISF-TEVAR is characterized by one or more fenestrations created in the proximal portion of the main endograft. These fenestrations are used to house the supplemental stents, which could reconstruct the blocked aortic branches with partial body protruding into aortic arch to prevent migration. The high technical success, low morbidity and mortality, and good early patency significantly expand the clinical application of ISF-TEVAR.^7,24 However, the long-term reliability and efficacy of ISF-TEVAR still needs to be evaluated and few literatures have investigated the postoperative hemodynamics in patient-specific aortic dissection treated by ISF-TEVAR. Kandail et al.²⁵ reported a computational study assessing the hemodynamic outcomes of differing fenestrated stent-grafts for endovascular repair of abdominal aortic aneurysms, and the blood flow distribution depends on the configuration of stent-graft. van Bakel et al.²⁶ investigated the impact of differing endograft design on hemodynamic parameters after Zone 0 endovascular repair. The present study focuses on the efficacy of ISF-TEVAR for aortic dissection and the hemodynamic consequences of the differing protrusion length of the fenestration stent.

As blood flows through the fenestration stent, it exerts pressure on stent surface. From the simulation results, we can see that the pressure inside the fenestration stent is lower than outside at peak systole and the pressure difference may promote the stent contraction or migration. Moreover, it is interesting to note that the blood around the fenestration stent accelerates due to the reduced flow area and a relatively low velocity region is formed in front of the stent. In summary, the presence of fenestration stent has a significant effect on the aortic arch flow field.

The adequate perfusion is crucial for maintaining normal organ function. Dong et al.²⁷ reported that the amount of flow crossing the LSA varied from 3.52 to 5.28% in healthy young human. After ISF-TEVAR, only 2.36% of the total inlet flow is assigned to the reconstructed LSA, which is under the normal clinical value. When the fenestration is improved or even ideal, the blood supply to the LSA is further strengthened and increases to 4.01 and 6.14%, respectively, where the ratios are close to the normal range. In addition, slight reductions of the blood flowing through the left carotid artery and descending aorta are observed with the protrusion length of the fenestration stent shortening. These results demonstrated a strong dependence of flow distribution on the protrusion length of the fenestration stent into the main endograft, with shorter protrusion leading to more adequate perfusion to the reconstructed LSA.

The protrusion length of the fenestration stent also has a significant effect on the distribution of wall shear stress-related indices. The postoperative aortic arch is exposed to low levels of TAWSS, making the region potentially susceptible to atherosclerosis formation. Notably, there is almost no change in this region as the protrusion length of the stent is shortened by half. However, when the fenestration stent is totally removed, the low TAWSS region expands significantly which means an increased probability of atherosclerosis, indicating the TAWSS distribution is more sensitive to the upper half of the stent. A high OSI in the LSA and descending aortic stenotic region is observed owing to the diameter reduction. As the protruding length of the stent is shortened, the high OSI region at the LSA gradually gets smaller, suggesting that the blood flowing through the LSA tends to be stable. This is because the flow rate is improved and blood flowing at a high velocity tends to keep flowing in the same direction. These results indicate that the reduction of protrusion length of the stent could weaken the oscillatory nature of the flow.

With respect to the EL during a cardiac cycle, the waveforms for different protruding length of the stents are similar and the difference between the Case A and Case B is negligible. Two time periods with notable differences are observed, which is mainly attributed to the rapid change in the inlet velocity magnitude. It should be emphasized that the EL profiles have four peaks for all three cases during the peak systole. The postoperative maximal EL is 0.973 W, which is found at t = 0.8 s after the mid-systolic acceleration time point (T1). The average EL decreases by 9.4%, when the fenestration stent is totally removed. It can be deduced that the upper half of the stent has a more profound influence on the EL than the lower half.

The first limitation of this work is that only a single patient is analyzed in this paper. More patient-specific studies and clinical follow-ups are warranted to confirm our simulation conclusion in follow-up work. Second, the arterial wall is assumed to be rigid. Previous study has verified that the interaction between the vessel wall and blood flow is of particular importance. Finally, the inlet flow rate waveform is not patient specific and MRI data would be used as boundary conditions and to verify simulation results.

Conclusion

The present study investigates the patient-specific blood flow in a type B aortic dissection undergoing ISF-TEVAR with LSA reconstructed by in situ fenestration. The pressure difference between the inner and outer surfaces of the fenestration stent might promote the stent contraction or migration. When the fenestration is improved or even ideal, the blood supply to the LSA strongly depends on the protrusion length of the fenestration stent and the flow gradually tends to be stable. However, ideal fenestration may expose the aortic arch to low TAWSS, making the region more susceptible to atherosclerosis formation. Hence, we conclude that the fenestration stent has a strong influence on the hemodynamics and the protrusion length potentially needs to be shortened properly. Further large population-based studies are needed to confirm these preliminary findings. Overall, our study preliminarily demonstrates that the CFD has the potential to quantitatively evaluate the efficacy of ISF-TEVAR and aid clinicians to optimize the surgery.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a National Natural Science Foundation of China (grant number 51576049).

References

Nienaber

Divchev

Palisch

et al. Early and late management of type B aortic dissection. Heart 2014; 100: 1491–1497.

McWilliams

Murphy

Hartley

et al. In situ stent-graft fenestration to preserve the left subclavian artery. J Endovasc Ther 2004; 11: 170–174.

Sonesson

Dias

Resch

et al. Laser generated in situ fenestrations in Dacron stent grafts. Eur J Vasc Endovasc Surg 2016; 51: 499–503.

Katada

Kondo

Tsuboi

et al. Endovascular total arch repair using in situ fenestration for arch aneurysm and chronic type A dissection. Ann Thorac Surg 2016; 101: 625–630.

Tse

Lindsay

Roche-Nagle

et al. Radiofrequency in situ fenestration for aortic arch vessels during thoracic endovascular repair. J Endovasc Ther 2015; 22: 116–121.

Bicknell

Cheshire

Riga

et al. Treatment of complex aneurysmal disease with fenestrated and branched stent grafts. Eur J Vasc Endovasc Surg 2009; 37: 175–181.

Qin

Zhao

Wang

et al. In situ laser fenestration is a feasible method for revascularization of aortic arch during thoracic endovascular aortic repair. J Am Heart Assoc 2017; 6: e004542.

Glorion

Coscas

McWilliams

et al. A comprehensive review of in situ fenestration of aortic endografts. Eur J Vasc Endovasc Surg 2016; 52: 787–800.

Madhwal

Rajagopal

Bhatt

et al. Predictors of difficult carotid stenting as determined by aortic arch angiography. J Invasive Cardiol 2008; 20: 200–204.

10.

Wang

Zhang

Xin

Morphologic features of the aortic arch and its branches in the adult Chinese population. J Vasc Surg 2016; 64: 1602–1608.e1.

11.

Wang

Zhou

Guo

et al. A new adjustable puncture device for in situ fenestration during thoracic endovascular aortic repair. J Endovasc Ther 5 June 2018. DOI: 10.1177/1526602818776623.

12.

Jung

Lyczkowski

Panchal

et al. Multiphase hemodynamic simulation of pulsatile flow in a coronary artery. J Biomech 2006; 39: 2064–2073.

13.

Schiller

A drag coefficient correlation. Zeit Ver Deutsch Ing 1933; 77: 318–320.

14.

Gijsen

Allanic

Van de Vosse

et al. The influence of the non-Newtonian properties of blood on the flow in large arteries: unsteady flow in a 90 curved tube. J Biomech 1999; 32: 705–713.

15.

Dill

Costill

DL.

Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J Appl Physiol 1974; 37: 247–248.

16.

Chien

Usami

Dellenback

et al. Blood viscosity: influence of erythrocyte aggregation. Science 1967; 157: 829–831.

17.

Chien

Usami

Dellenback

et al. Blood viscosity: influence of erythrocyte deformation. Science 1967; 157: 827–829.

18.

Xie

Zhang

The effect of red blood cells on blood heat transfer. Int J Heat Mass Transfer 2017; 113: 840–849.

19.

Gallo

Lefieux

Morganti

et al. A patient-specific follow up study of the impact of thoracic endovascular repair (TEVAR) on aortic anatomy and on post-operative hemodynamics. Comput Fluids 2016; 141: 54–61.

20.

Hou

Tsagakis

Wendt

et al. Three-phase numerical simulation of blood flow in the ascending aorta with dissection. In: Paper presented at proceedings of the 5th European conference on computational fluid dynamics (ECCOMAS CFD’10), 2010.

21.

Pirola

Cheng

Jarral

et al. On the choice of outlet boundary conditions for patient-specific analysis of aortic flow using computational fluid dynamics. J Biomech 2017; 60: 15–21.

22.

Cheng

Wood

Gibbs

et al. Geometric and flow features of type B aortic dissection: initial findings and comparison of medically treated and stented cases. Ann Biomed Eng 2015; 43: 177–189.

23.

Malek

Alper

Izumo

Hemodynamic shear stress and its role in atherosclerosis. JAMA 1999; 282: 2035–2042.

24.

XiaoHui

Wei

et al. Comparison of supra-arch in situ fenestration and chimney techniques for aortic dissection involving the left subclavian artery. Vascular 14 October 2018. DOI: 10.1177/1708538118807013.

25.

Kandail

Hamady

XY.

Comparison of blood flow in branched and fenestrated stent-grafts for endovascular repair of abdominal aortic aneurysms. J Endovasc Ther 2015; 22: 578–590.

26.

van Bakel

Arthurs

van Herwaarden

et al. A computational analysis of different endograft designs for Zone 0 aortic arch repair. Eur J Cardiothorac Surg 2018; 54: 389–396.

27.

Dong

Chen

Xiong

Ultrasonic determination of normal value of proportional distribution of aortic blood flow in Chinese youths. Chin J Gen Surg 2017; 10: 1633–1636.