Fatigue Analysis of Novel Aortic Valve Stent Using Finite Element Analysis

Aortic stenosis and Regurgitation are the common heart valve diseases that about 20% of the US population get diagnosed with each year. Recent advances in minimally invasive surgical techniques like transcatheter aortic valve replacement (TAVR) have paved the way for high-risk patient treatment with minimal hospital stay and use of anesthetic. While the patient data on prosthetic stent based aortic valve durability from the last decade has instilled confidence, this data is still considered inadequate for large scale adoption of TAVR therapy for low and medium risk patients. As one of the leading failure modes for the Nitinol based prosthetic stents is material fracture due to cycling fatigue, it is critical to identify a method to accurately predict it. While the historic scientific literature available indicates the presence of stress based, strain based and damage-tolerance analysis-based assessments of fatigue behavior of Nitinol material, there still exists a gap specifically for accurate prediction of fatigue life of Nitinol based prosthetic stents using computational methods like finite element analysis. This article presents a case study of estimating the fatigue life of novel aortic valve when subjected to physiological loading conditions using FEA while using minimal computational resources. This FEA and strain-based approach using constant-life diagram method of fatigue estimation will allow for preliminary evaluation of fatigue life and make design and process changes before fabrication and testing of prosthetic aortic.


Introduction
The conventional open aortic valve replacement using cardiopulmonary bypass/open heart surgery remains standard of care in patients with successful results reported in the literature [1], [2].However, elderly patients with high comorbidities are considered highrisk candidates for this type of surgery [3].Percutaneous valve replacements like transcatheter aortic valve replacement (TAVR) have emerged as the preferred treatment option for these high-risk patients.This minimally invasive treatment method through transluminal procedures drastically reduced the risk of open-heart surgery and significantly reduced duration of intensive care unit stay post procedure.Although the transcatheter procedures have witnessed excellent outcomes, certain engineering, and considerations in designing and testing the aortic valves should be addressed for further improving the patient outcomes and to inspire confidence in the physician community to steer towards TAVR for low-risk patients.The design considerations include suitability of the implant to the patient anatomy, hemodynamics, structural dynamics of the implanted valve and the prosthetic fatigue durability [4].Specifically, fatigue fracture is a highly recognized complication following stent implantation [5]and should be judiciously assessed during the design phase [6].The stresses/strains leading to stent fracture are caused by a combination of factors such as oversizing, cyclic hemodynamic pressures and leaflet forces exerted on the stent [7].While there is extensive literature on fatigue life prediction for linear elastic materials [8], these are deemed invalid for super-elastic Nitinol.The ability to predict fatigue behavior of Nitinol subjected to cyclic mechanical motion during the systolic and diastolic phases of the heart is critical for an adequate design of the aortic implant.A review of available Nitinol fatigue prediction methods did not provide a coherent insight into a comprehensive understanding of a practical approach to this problem.Therefore, this paper focuses on presenting a finite element analysis method of conducting a preliminary assessment of prosthetic Nitinol stent fatigue life based on the constant-life fatigue data presented by Pelton et.Al [9].To this effect, a novel prosthetic aortic valve was designed and subjected to physiological loading conditions that a native aortic valve is subjected to using FEA.The resulting mean and amplitude strains in the stent frame are then analyzed to compare against the fatigue endurance limit data reported in the literature to assess the ability of this novel design to endure fatigue from cyclic loading.

Methods Nitinol stent Model
The critical geometrical parameters of the stent designed for this study are reported in Table 1.The optimal ranges for these parameters were initially assessed through experimental FEA runs to fit the chosen annular size and possess optimal radial resistive force to minimize paravalvular leakage and migration risks.The geometry of the stent design has numerous repeatable units-namely diamonds-as shown in figure 1.The critical parameters identified on these diamonds were altered among other stent design parameters in choosing the optimal design to be assessed.For modeling this stent design, initial 2D Lasercut pattern is drafted using Solidworks 2022, which is then wrapped around a tube with an outer diameter (OD) of 9mm and inner diameter (ID) of 8 mm using ABAQUS 2023 to create a narrow tube stent.A series of rigid mandrels were then used to expand this narrow stent to the required final stent dimensions in gradual steps.Each expansion step of the stent was designed to not allow the total logarithmic strain (LE) to exceed the 8% strain limit.At the beginning of each expansion step, the pre-strains from the previous step are removed to mimic the stress-relieving process.The uniaxial material properties of the Nitinol (Ni50.8Ti49.2) used for this stent is based on the work reported by Senthilnathan et.al [10].The material properties used for this analysis are listed in Table 2.

Leaflet Model
The geometry of the leaflets is conceived based on its primary requirements to be able to co-apt with each other at the time of valve closing, being able to optimally conform to transvalvular hemodynamics while minimizing cusp stress during systolic and diastolic phases of the cardiac cycle.Several experimental FEA runs were performed with varying geometrical parameters of the leaflet design to optimize the valve closing and opening with minimal induced stresses and strains.The CAD model of the leaflets was used to mimic the aortic leaflets sewed to the stent frame and any tissue present in the skirt region of a typical prosthetic valve model is excluded from the model used in this study.

Figure 3: Leaflet Geometric Profile
The uniaxial mechanical properties of porcine pericardial sac fixated for 5 days using 0.2% HDI (hexamethylene di-isocyanate) solution was identified to closely mimic those of aortic leaflets [11].Hence, these material properties were used to model the aortic valve leaflets in this study.While care was taken to adequately model the leaflets and its material, this was only done to best represent the geometry and the material stiffness of the leaflets being sewn to the prosthetic stent as it has an impact on the stent stresses and strains when subjected to cyclic loading.This article's focus remains on assessing the fatigue life of the stent, while the assessment of leaflet design is outside the scope of this article.

Aortic Root Model
The aortic root is a complex anatomical unit consisting of ventriculo-aortic junction, aortic leaflets, the interleaflet triangles, Valsalva sinuses with coronary ostia, sinotubular junction (STJ), and the ascending aorta [12].The stent design is required to integrate adequately into the native aortic root post deployment for optimal post-operative performance, like minimizing paravalvular leakage (PVL), and minimizing stresses on the native anatomy which otherwise could lead to anatomical ruptures.For this stent design, an aneurysmal aorta with root dilatation as reported in Morganti et.al article [12], with ventricular-aortic junction diameter (Da) of 30 mm was chosen.The detailed anatomical parameters of this aortic root are presented in Figure 2. Based on these parameters, the CAD model for the aortic root is developed to be used in conjunction with the aortic stent and leaflet design.This model chosen was only used as representative geometry for the aortic root in contact with the stent near the basal plane, and the analysis of the aortic root is not included in the scope of the study reported in this article.

Figure 8: Aortic Root Anatomical Parameters
Meshing-FEA Initial experimental runs informed the choice of meshing for each of the model elements used in this analysis.A 4-node quadrilateral surface reduced integration element "SFM3D4R" was used for aortic root, an 8-node linear brick reduced integration element "C3D8RH" was used for the Nitinol stent frame with quantity of four elements across its thickness.For the leaflets, a 4-node reduced shell element "S4R" was used for meshing.A surface-to-surface contact with friction factor of 0.1 and a penalty stiffness factor of 0.01 is defined between the stent and the aortic root.

Figure 10: Valve assembly in Aortic Root
Loading-FEA The stent model with an outer diameter of 33mm is crimped to intermediate crimp of 25mm diameter and is deployed in the aortic root model with an annular diameter of 30 mm.This represents an oversizing of 10% by perimeter which falls in the category of large oversizing feasible group reported in this published study [13].To represent the physiological loading conditions during different cardiac phases, the geometric changes in the annular size diameters are applied as radial displacements to the prosthetic valve [14].As for the pressure gradients, ISO5840-3:2021 standard's stipulation for hypertensive conditions for valve testing were used.Therefore, for the systolic phase, using quasi-static simulation, 40mm of Hg pressure gradient is applied across the annulus in 0.4 seconds.Similarly, for the diastolic phase, 144mm of Hg of pressure gradient is applied while also applying a distention radial displacement of 0.42mm which corresponds to 2.8% reduction in mean annular diameter from systolic to diastolic phase of cardiac cycle.These systolic and diastolic loads are applied for a total of 3 cycles.

Methods
The stent frame was assessed for radial displacements as a result of applied loading during systolic and diastolic phases of cardiac cycle.As Figure 11 indicates, the maximum radial displacement in the stent frame is seen in the bottom end of the frame which is in contact with the aortic root and is under compressive load due to oversizing besides the pressure gradient loads.Stent frame radial displacements are summarized in which are closely assessed for estimating the stent frame strains.The stent frame was then assessed for the mean strains and the strain amplitudes derived from the max principal strains observed in the systolic and diastolic cardiac phases.The highest strain amplitude, as seen in Figure 12, is 0.68% and majority of the high strain amplitudes are observed near the transition region of diamond to its connection point with the next diamond.
Similarly, the highest mean strain on the stent frame is 6.76% which is in locations similar to where high strain amplitudes were observed (see Figure 13).The constant-life diagram presented by Pelton et.Al.
[9] plots mean strain against strain amplitude to showcase the data from Gen-1 Nitinol (Ni50.8Ti49.2) based diamond stent samples and micro-dogbone samples that have varying degree of fatigue life.As shown in the diagram, all samples with strain amplitudes under 0.4% were reported to have survived 10 7 cycles of cyclic loading, while a few samples with strain amplitudes between 0.4% to 0.6% to have survived the 10 7 cycles if their mean strains are in a 3-7% range.The diagram also indicates that a majority of the samples with strain amplitudes over 0.6% were reported to not have survived the 10 7 cycles of cyclic loading.The solid symbols from the diagram represent samples that didn't endure 10 7 cycles of loading, and the open symbols represent samples that run out under 10 7 cycles (see Figure 14).The current stent frame fatigue life assessment was attempted with these values of amplitude strain and mean strain at the end of 3 cycles of cyclic loading.This chart was used to plot the mean and amplitude strain data from the current analysis to make a preliminary assessment of durability of this stent frame using FEA.
Based on the stent frame maximum strains noted above, and using the constant-life plot, the plotted stent frame's strain values indicate that it is unlikely to survive the 10 7 cycles of fatigue cycling as it falls above the trend line (Figure 14).To further assess the current stent frame's fatigue endurance, all the integration points on the stent frame are plotted on the constant-life diagram to illustrate its fatigue behavior (Figure 15).The dot plot indicates that 99.97% of the integration points are under the trend line and therefore survive the 10 7 cycles of fatigue loading, while the rest 0.03% fall over the trend line and represent areas of stent frame that likely will be areas of concern for fracture.Therefore, it is clear from this plot that only a few areas of the stent frame like the connection joints between the diamonds, potential sharp corners, and sharp edges, need to be further optimized to bring down the strain amplitude below the 0.6% level.These slightly high strain values for the current stent frame are likely due to the lack of representation of electropolishing, deburring of edges in the FEA model used.In the physical world, at the time of fabrication, the stent frame will be subjected to electro-polishing to enhance its surface finish and corrosion resistance.Therefore, some of these high strain values as shown in the FEA stent frame model will remain as test artifacts of this analysis.Other improvement methodologies include minor modification of stent frame geometry parameters as noted in Table 1, breaking sharp edges on the stent frame etc.However, these optimization options are not covered under the scope of work presented in this article.

Conclusion
The novel stent design presented in this article is subjected to systolic and diastolic hydrodynamic pressure gradients while implanted in undersized native annulus for adequate anchoring.This implant, subjected to 3 systole-diastole cyclic loads in FEA yielded mean and amplitude strain values, were plotted in the constant-life diagram for fatigue life reported in the literature.The plot was able to yield valuable insights into preliminary assessment of optimizations necessary on the stent design before conducting the actual physical tests which tend to be expensive and time consuming.While the data presented in the constant-life diagram in the literature is based on stent diamonds and uniaxial load data of micro-dogbone structures, more evaluations are necessary for the suitability of such data in assessing the fatigue life of whole stent structure Also, the fatigue life depends on the thermal and mechanical stress profile of the stent frame resulting from the specific fabrication process used which may not be accurately represented in the FEA model.Therefore, the fatigue life prediction using FEA and constant-life diagram based on generic diamond data may not be accurate.Nevertheless, this relatively inexpensive FEA method used for preliminary assessment of fatigue life of stent design provides valuable feedback for its design before manufacturing and conducting real-world fatigue tests.

Conflict of Interests
Authors declare no conflict of interest.

Figure 2 :
Figure 2: Nitinol Stent Shape set process in FEA

Figure 4
Figure 4 Uniaxial Material Properties-0.2%HDI 5-day fixated Porcine Pericardium ABAQUS software was used to model the leaflets using incompressible hyper-elastic material with a Poisson ratio of 0.475.Several material calibration runs with the available hyper-elastic material models present in the ABAQUS software were done to determine an ideal option for material stability and curve fitting with the chosen experimental data of the porcine pericardium.The Marlow material model was identified as the optimal fit and material calibration was done to minimize the residual fit error between the material properties of Marlow model and the experimental data (Figure6).