Endovascular Device Design in the Future: Transformation From Trial and Error to Computational Design

Transcription

1 I12 J ENDOVASC THER REVIEW Endovascular Device Design in the Future: Transformation From Trial and Error to Computational Design Christopher K. Zarins, MD, and Charles A. Taylor, PhD Stanford University School of Medicine and School of Engineering, Stanford, California, USA. Endovascular devices have been designed by trial and error, with bench and animal testing followed by human clinical trials to determine whether the devices are safe and effective. Despite remarkable advances over the past 15 years, there are persistent concerns regarding the long-term durability of endovascular devices. This may be due to deficiencies in device design, which has lagged behind other industries in adopting computational methods that are now routinely used to design, develop, and test new aircraft and automobiles. Similar computational design and failure mode simulations that evaluate performance under stress conditions have not been widely applied in the development of endovascular devices. Advances in medical imaging and computational modeling now allow simulation of physiological conditions in patient-specific 3- dimensional vascular models, which can provide a framework to design and test the next generation of endovascular devices. This modeling will allow the prospective design of devices that can withstand the force variations in the cardiovascular system that occur during bending, coughing, and varying degrees of exercise, as well as the extremes encountered during sudden impact in contact sports. Utilization of computational design methodology that takes into consideration the physiology of the cardiovascular system will improve future endovascular devices so that they are safer and more effective and durable. J Endovasc Ther. Key words: stent-graft, endograft, device design, computational design, computational fluid dynamics, finite element analysis The remarkable advances in endovascular therapy during the past 15 years have occurred as a result of creative ideas and innovative medical devices that allow less invasive treatment of cardiovascular disease Building on a wealth of experience and engineering expertise, endovascular devices have been designed and tested using the time-honored trial-and-error strategy. 16 During the past 30 years, we have also witnessed incredible advances in information technology and computer power, along with the development of sophisticated computational tools for modeling, simulating, and solving heretofore unsolvable complex problems. 17 Computational device design strategies are now widely used by most safetycritical industries, such as the airline and automotive manufacturers, to design and evaluate new engineering products using simulation tools to examine potential failures without risk or consequences. Such compu- Solicited reviews published in the Journal of Endovascular Therapy reflect the opinions of the author(s) and do not necessarily represent the views of the Journal or the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS. The authors have no commercial, proprietary, or financial interest in any products or companies described in this article. Address for correspondence and reprints: Christopher K. Zarins, MD, Division of Vascular Surgery, Stanford University Medical Center, 300 Pasteur Drive, Room H-3642, Stanford, CA USA; zarins@stanford.edu ß 2009 by the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS Available at

2 J ENDOVASC THER I13 tational tools have had limited use in medical device design because of the complexity of living systems and the lack of investment in developing suitable computational models for designing and testing new medical devices. Such models must account for the unique features of the human circulation with appropriate 3-dimensional (3D) anatomical and physiological input data to define relevant boundary conditions. 18,19 During the past decade, we have developed a simulation-based model for the cardiovascular system, which allows modeling and study of the hemodynamic and biomechanical forces acting on implanted devices during rest and varying degrees of exercise This method integrates the advances in medical imaging with computational mechanics to provide a model system that can be used to test devices implanted in the cardiovascular system. AORTIC ENDOGRAFT DESIGN Trial and Error Aortic endografts are designed to prevent aneurysm rupture by providing a conduit for aortic blood flow while excluding flow from the aneurysm sac. The first successful endovascular aortic aneurysm repairs were accomplished by modifying a standard prosthetic graft used in open aneurysm repair so that it could be attached to the aorta by a balloon-expandable stent delivered through the femoral artery. 24,25 A variety of endovascular stent-grafts have been developed since that time, each seeking to emulate the success of open aneurysm repair while using the strategy of less invasive transfemoral device delivery. A number of initially promising design iterations have fallen by the wayside or required modification because of structural, fabric, or delivery system failures There are currently 5 Food and Drug Administration (FDA) approved endografts for endovascular aneurysm repair (EVAR) of abdominal aortic aneurysm (AAA) and 3 FDA-approved endografts for endovascular repair of thoracic aortic aneurysm (TAA) Each is unique in design, with differing device and delivery system characteristics. All incorporate a selfexpanding metallic structural framework with a fabric graft. Unique design features include infrarenal or suprarenal fixation mechanisms, radial force or penetrating hook and/or barb fixation, and modular or unibody endograft design. Each device was developed with rigorous bench testing of stent and fabric durability, animal implantation testing, and clinical trials in patients with comparison to standard open surgery. While it is widely recognized that the pulsatile nature of blood flow exerts significant force on endovascular devices, particularly in the aorta, actual measurement of invivo displacement forces has been difficult, especially during extremes of exercise, such as bending, straining, running, and sudden impact. In-vitro 52 and in-vivo 53 testing has been performed to determine linear pullout forces or resistance to downward displacement force presumed to be exerted by pulsating blood flow. However, aortic flow is characterized by complex 3D patterns of flow that vary with exercise, 19,54 56 and displacement force vectors are not linear and axisymmetrical. While trial-and-error device design has resulted in a variety of fixation mechanisms to resist device displacement, migration remains an issue for each of the currently available endografts The Dawn of Computational Design Aircraft design began with the same trialand-error strategy as aortic aneurysm repair. Early aviation pioneers devised a variety of strategies to fly, with success measured largely by a safe landing. It was back to the drawing board if the flight was unsuccessful, oftentimes requiring an effort to find a new test pilot. This design and test strategy led to a continual improvement in aircraft, culminating in the first successful flight of a heavier-than-air flying machine by Wilbur and Orville Wright in Kitty Hawk, North Carolina, in Progress in aviation has gone hand-in-hand with developments in mathematics, materials science, and aeronautical, mechanical, and electrical engineering. Staggering levels of government and private funding have produced remarkable accomplishments, including supersonic aircraft,

3 I14 J ENDOVASC THER space flight, and a passenger airline industry with an admirable safety record. The design and evaluation of airplanes rests on the pillars of science: theory, observation, and experiment. The fundamental equations of elasticity and fluid mechanics, which were known prior to the dawn of aviation, provided a theoretical foundation for airplane designers. Empirical methods were used to observe the interactions between airflow and airfoils. Experimental methods, including wind tunnels, were used to evaluate aircraft components and airplanes themselves before flight. However, airplane design was still largely an empirical and experimental science due to reliance on analytical methods for solving the equations governing the theoretical models of complex systems. Specifically, analytical solutions of nonlinear, partial differential equations, such as those pertaining to the mechanics of materials and fluid dynamics, are possible only in the most ideal cases and are entirely infeasible for an entire wing or an airplane itself. The airline industry took a quantum leap with the development of the space program and unmanned flight. This required computation and computer design, planning, and control, which translated to computational design of fixed-wing aircraft as well as space vehicles. An example is the Boeing 777 aircraft. In 1995, it became the first aircraft to be computationally designed, tested, and certified airworthy before ever having been flown. Over a 5-year time frame ( ), Boeing applied a new and visionary computing technology to engineering and manufacturing processes. A fundamentally new approach was used to design and build the airplane. Each element of the airplane s airframe, parts, and systems were built and linked by computer. Using 3D digital software, designers could see parts of solid images and then simulate the assembly of those parts on the screen, easily correcting misalignments and other fit or interference problems. As a result of these new innovative processes, the 777 program exceeded its goal of reducing change, error, and rework by 50%. Parts and systems fit together in much better alignment than other airplanes. Digital preassembly was used and significantly improved precision and accuracy and reduced errors, with improved overall quality at a lower cost. COMPUTATIONAL FLUID DESIGN (CFD) FOR CARDIOVASCULAR DISEASE Just as aircraft must withstand the rigors of flight, stents and stent-grafts must withstand the biomechanical forces applied to them in the cardiovascular system. The forces are not uniform in the vascular tree. For example, the coronary, carotid, renal, and peripheral vessels, where the key requirement is preserving vessel patency and preventing stenosis and occlusion, each have different flow conditions and biomechanical forces applied to them. In the thoracic and abdominal aorta, on the other hand, the issue is primarily one of preventing vessel enlargement and rupture. The mechanical forces applied to aortic devices are extreme, and device displacement and migration are primary considerations. Stenosis and occlusion are infrequent phenomena, unless device movement results in kinking. Thus, design considerations must be vessel-specific, and a thorough knowledge of the flow and biomechanical forces are needed, both in resting and exercise conditions. In addition, the cardiovascular system changes over time, with enlargement, elongation, and development of tortuosity, stenoses, and occlusions; these changes are exacerbated by changes in position and function sedentariness versus exercise. Design considerations are different for aneurysmal and occlusive disease and for vessels of different diameter: the aorta versus medium vessels (iliac, carotid, superficial femoral artery) versus small vessels (coronary, tibial). A decade ago, we first introduced the idea of cardiovascular simulation for treatment planning. At the Society for Vascular Surgery meeting in 1998, we presented a computational framework for designing and testing a variety of vascular surgery treatment options (Fig. 1), including open surgical and endovascular treatments of peripheral occlusive disease. 62 Since that time, we have focused on developing the framework for a realistic cardiovascular modeling system. In August 2002, with the support of the National Science

4 J ENDOVASC THER I15 Figure 1 Computational framework for designing and testing surgical and endovascular treatments in peripheral vascular disease. Presented at and utilized by a panel of expert vascular surgeons at the Critical Issues Forum during the Society for Vascular Surgery meeting in June Reprinted with permission from Taylor CA et al. 62 Copyright 1999 by the International Society for Computer Aided Surgery. Foundation, we embarked on a 5-year project entitled Simulation-Based Medical Planning for Cardiovascular Disease. The goal was to develop, implement, and validate an integrated problem-solving environment for simulation-based medical planning for cardiovascular disease, incorporating accurate and efficient image-based geometrical solid modeling, 63 operative planning, automatic finite element mesh generation, 64 finite element fluid mechanics, and scientific visualization methods. Computational Methodology Computational methodology for the cardiovascular system begins with construction of a computer model based on human imaging studies using contrast computed tomography (CT) or magnetic resonance (MR) The relevant vascular anatomy is imaged (for aortic aneurysms, the abdominal or thoracoabdominal aorta; for cerebrovascular disease, the aortic arch, neck, and intracranial arteries; for coronary artery disease, gated 64-slice chest CT; for peripheral occlusive disease, the aortoiliofemoral and lower extremity arteries), and a computer model of the anatomy is constructed (Fig. 2). Using the anatomical 3D geometric model, a finite element analysis (FEA) is performed to model interactions between the device and the circulatory system. Computer simulation of stents and stentgrafts in anatomically and physiologicallyaccurate patient models requires 3 subsystems: (1) the blood flow in the circulatory system, (2) the mechanical behavior of the

5 I16 J ENDOVASC THER Figure 2 Schematic of solid model construction of abdominal aorta from imaging data. (A) Volume-rendered image of a contrast-enhanced magnetic resonance angiogram. (B) Centerline paths were created along the vessels of interest. (C) Two-dimensional segmentations of vessel lumen were taken perpendicularly to the vessel path. Segmentations were found using a level set method. (D) Two-dimensional segmentations were lofted to form solid models for each vessel, which were then joined to form a complete 3D solid model of the aorta and its branches. (E) The solid model was discretized into a finite element mesh (gold) and is shown with the original volume-rendered magnetic resonance angiogram. Reprinted with permission from Tang et al. 21 Copyright 2006 by the American Physiological Society. vessel wall and surrounding tissues under pulsatile and non-pulsatile loading, and (3) the mechanical behavior of the device and delivery system. The interactions of endovascular stents or stent-grafts with the arterial wall and the blood flow are very complex. These interactions must be studied in detail in order to avoid the mechanical failure of the devices and to preserve their long-term efficacy and durability. Recent studies show that the failure of stent-grafts may occur in different ways: from endograft migration, endoleaks, 33,40,41,66 and metal stent fracture or graft wear. 27,28,29,31 38,40 43 Stent migration is due to mechanical factors, such as the induced displacement force on the bifurcated grafts and proximal fixation site, as well as the pressure differentials between the stentgraft and the aneurysm sac. 67 The fatigue of the metal and/or fabric is due to the presence of areas of higher stress and strain concentration, as a result of excessive friction between the stent components. 68 Significant research has been done performing in-vitro studies to evaluate the fixation forces acting on stent-grafts However, the nature of these studies is often too simplistic, since they usually consider rigid wall models of the vessel with steady flow conditions and make the assumption that migration forces act in the downstream direction of blood flow. In-vitro and in-vivo experimental models have measured linear downward force needed to displace endografts 52,53,72,73 but have not considered other potential force vectors acting on endografts. Furthermore, these in-vitro and in-vivo studies inherently lack the patient-specificity and the 3D perspective that only virtual models can provide. Advances have also been made in the surveillance of the performance of these devices over time. 74,75 While very important to ensure stent-graft integrity, these studies only provide an a posteriori assessment of the performance of the devices. Model-based computer simulations can supply the necessary tools to better design and test the performance of endovascular stent-grafts in a 3D patient-specific basis. Finite Element Analysis FEA of stent-graft performance has been focused on CFD analysis of the flow patterns before and after implementation of the device and calculation of the resulting shear forces and estimation of device migration. Published reports on simulation of the mechanical behavior of the stent-graft are almost nonexistent. 76 A few studies have been per-

6 J ENDOVASC THER I17 formed where the interactions between blood flow, the stent-graft, intraluminal thrombus, and the aneurysm wall are considered However, in all previous studies, the effect of the proximal branches on hemodynamic forces in the infrarenal aorta and realistic pressure waves were missing. Furthermore, models of the interactions between the solid mechanics of the device and the fluid mechanics of the bloodstream lacked full volumetric analyses and were limited to idealized planar configurations. Two of the main challenges related to modeling the interactions of stents-grafts and the aorta involve including realistic anatomy and matching the physiological conditions. For example, previous studies do not include the celiac, superior mesenteric, and renal arteries, even though hemodynamics of the infrarenal aorta are dominated by the recirculation vortices introduced by the branches. Previous studies are also idealized further by neglecting out-of-plane curvature of the aorta. Forces on aortic stent-grafts are comprised of a time varying hydrostatic blood pressure, the dynamic pressure related to the inertia of blood impacting the device, and the shear forces. As a result, any computational simulation of hemodynamics in the aorta needs to result in physiological flow splits between the branch vessels and aorta and realistic pressure waveforms. Simply put, the outlet boundaries of the computational domain need to incorporate information about the downstream vasculature since it is the microcirculatory beds of tissues and organs that determine the flow distribution and pressure of the major arteries including the aorta. Boundary conditions are specified to include and reflect central and regional, organ-specific blood flow of the aortic branch vessels. Time varying loads exerted by the blood flow are calculated at rest and for varying flow conditions. These flow conditions can be modified to reflect a variety of physiological conditions, such as at rest and with varying degrees of exercise, pre- and postprandial visceral flow conditions, variations in degree of hypertension, heart rate, and cardiac output. The computational model can also be modified for variations in movement due to both pulsatility and compliance of the vessel wall and movement due to bending and stretching, such as flexing the knee and hip or movement of the renal arteries with respiration. In addition, sudden extremes of force, such as might be experienced in contact sports or trauma, can be simulated. Our research has enabled the development of several important technologies, including a new software framework for patient-specific cardiovascular modeling 63 ; new methods for direct 3D geometric modeling 65 ; methods for anisotopic, adaptive mesh generation 64 ; and 3D finite element methods for simulating blood flow, including methods for defining lumped parameter and impedance outflow boundary conditions 22 and a novel method for large-scale fluid structure interaction. 20 The tools for realistic computational models of blood flow in the human aorta are now available. 21 The computational solution produces realistic 3D images of pulsatile blood flow patterns with variations in flow velocity and flow patterns showing vortices and recirculation zones. Areas of stenosis with pressure gradients across the stenosis are clearly seen, and the significance of a stenosis can be determined both at rest and with varying degrees of exercise. Aortic wall motion can be seen and aneurysms are clearly depicted. Shear stress can be calculated, as well as wall stress exerted on the aneurysm sac. Displacement forces acting on endografts placed to treat the aneurysm can be calculated for both abdominal and thoracic stent-grafts. The influence of angulation and tortuosity can be determined. We have recently applied these methods to perform 3D analysis of displacement forces acting on aortic endografts. These studies have shown, for the first time, that the primary vector of force acting on an abdominal aortic endograft is in the sideways direction, rather than in the downward direction of blood flow as has been previously assumed (Fig. 3). This is consistent with recent clinical observations that lateral movement of endografts within the aneurysm sac at 1 year is an indicator of late endograft failure. 82 Given the availability of these computational methodologies, future endovascular devices will be tested in 3D computational models that accurately reflect the expected in-vivo

7 I18 J ENDOVASC THER Figure 3 Computational model of an endograft in an abdominal aortic aneurysm. This displacement force analysis shows the magnitude and vector of migration forces acting on the endograft. Note the primary vector of force is sideways relative to the direction of blood flow. Computational modeling provided the first demonstration of true in-vivo forces acting on endografts. Prior to this analysis, migration forces were assumed to be acting in the downward direction of blood flow, and endografts were designed based on this assumption. Future endograft designs will be based on knowledge of the true forces acting on the endograft. Presented by C. K. Zarins at International Congress XXI on Endovascular Interventions held in Scottsdale, Arizona, USA, on February 10 14, flow conditions. Extremes of loads and stresses that may be experienced by the device will be simulated, and long-term durability testing will be conducted in simulation models before human implantation and clinical trials. Failure modes will be anticipated and tested. This will result in improved device designs with improved safety and effectiveness and better long-term results. CONCLUSION Future endovascular devices will be designed using computational methods that take into consideration the variations in hemodynamic and biomechanical conditions occurring in the cardiovascular system (during rest and exercise). This modeling will result in enhanced endovascular devices with improved function and better long-term results. REFERENCES 1. Biamino G. The excimer laser: science fiction fantasy or practical tool? J Endovasc Ther. 2004;11(Suppl II):II-207 II Chuter TA, Buck DG, Schneider DB, et al. Development of a branched stent-graft for endovascular repair of aortic arch aneurysms. J Endovasc Ther. 2003;10: Chuter TA, Gordon RL, Reilly LM, et al. An endovascular system for thoracoabdominal aortic aneurysm repair. J Endovasc Ther. 2001;8: Donas KP, Kafetzakis A, Umscheid T, et al. Vascular endostapling: new concept for endovascular fixation of aortic stent-grafts. J Endovasc Ther. 2008;15: Fogarty TJ, Arko FR, Zarins CK. Endograft technology: highlights of the past 10 years. J Endovasc Ther. 2004;11(Suppl II):II-192 II Henry M, Amor M, Ethevenot G, et al. Percutaneous peripheral atherectomy using the Rotablator: a single-center experience. J Endovasc Surg. 1995;2: Hinchliffe RJ, Hopkinson BR. Development of endovascular stent-grafts. Proc Inst Mech Eng [H]. 2007;221: Kasirajan K, Schneider PA, Kent KC. Filter devices for cerebral protection during carotid angioplasty and stenting. J Endovasc Ther. 2003;10: Katzen BT, MacLean AA. Past, present, and future endograft devices. Tech Vasc Interv Radiol. 2005;8:16 21.

8 J ENDOVASC THER I Murphy EH, White RA, Rosenthal D, et al. Evaluation of the Crux IVC Filter in an animal model. J Endovasc Ther. 2008;15: Palmaz JC. Intravascular stents in the last and the next 10 years. J Endovasc Ther. 2004; 11(Suppl II):II-200 II Parodi JC. Endoluminal treatment of arterial diseases using a stent-graft combination: reflections 20 years after the initial concept. J Endovasc Surg. 1997;4: Ramaiah V, Del Santo PB, Rodriguez-Lopez JA, et al. Trellis thrombectomy system for the treatment of iliofemoral deep venous thrombosis. J Endovasc Ther. 2003;10: Rastan A, Sixt S, Schwarzwälder U, et al. Initial experience with directed laser atherectomy using the Clirpath Photoablation Atherectomy System and Bias sheath in superficial femoral artery lesions. J Endovasc Ther. 2007;14: Stanley BM, Semmens JB, Lawrence-Brown MM, et al. Fenestration in endovascular grafts for aortic aneurysm repair: new horizons for preserving blood flow in branch vessels. J Endovasc Ther. 2001;8: Abel DB, Dehdashtian MM, Rodger ST, et al. Evolution and future of preclinical testing for endovascular grafts. J Endovasc Ther. 2006;13: Hughes T. The Finite Element Method. Princeton, NJ: Prentice-Hall; Taylor CA, Draney MT. Experimental and computational methods in cardiovascular fluid mechanics. Ann Rev Fluid Mech. 2004; 36: Taylor CA, Hughes TJ, Zarins CK. Finite element modeling of three-dimensional pulsatile flow in the abdominal aorta: relevance to atherosclerosis. Ann Biomed Eng. 1998; 26: Figueroa CA, Vignon-Clementel IE, Jansen KC, et al. A coupled momentum method for modeling blood flow in three-dimensional deformable arteries. Comput Methods Appl Mech Engrg. 2006;195: Tang BT, Cheng CP, Draney MT, et al. Abdominal aortic hemodynamics in young healthy adults at rest and during lower limb exercise: quantification using image-based computer modeling. Am J Physiol Heart Circ Physiol. 2006;291:H668 H Vignon-Clementel IE, Figueroa CA, Jansen KC, et al. Outflow boundary conditions for threedimensional finite element modeling of blood flow and pressure in arteries. Comput Methods Appl Mech Engrg. 2006;195: Wilson NM, Arko FR, Taylor CA. Predicting changes in blood flow in patient-specific operative plans for treating aortoiliac occlusive disease. Comput Aided Sur. 2005;10: Volodos NL, Shekhanin VE, Karpovich IP, et al. A self-fixing synthetic blood vessel endoprosthesis [in Russian]. Vestn Khir Im I I Grek. 1986;137(11): Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg. 1991; 5: Arko FR, Lee WA, Hill BB, et al. Increased flexibility of AneuRx stent-graft reduces need for secondary intervention following endovascular aneurysm repair. J Endovasc Ther. 2001; 8: Becquemin JP, Poussier B, Allaire E, et al. Endograft fabric disintegration simulating a type II endoleak. J Endovasc Ther. 2002;9: Böckler D, von Tengg-Kobligk H, Schumacher H, et al. Late surgical conversion after thoracic endograft failure due to fracture of the longitudinal support wire. J Endovasc Ther. 2005;12: Guidoin R, Marois Y, Douville Y, et al. Firstgeneration aortic endografts: analysis of explanted Stentor devices from the EUROSTAR registry. J Endovasc Ther. 2000;7: Haider SE, Najjar SF, Cho JS, et al. Sac behavior after aneurysm treatment with the Gore Excluder low-permeability aortic endoprosthesis: 12-month comparison to the original Excluder device. J Vasc Surg. 2006;44: Heintz C, Riepe G, Birken L, et al. Corroded nitinol wires in explanted aortic endografts: an important mechanism of failure? J Endovasc Ther. 2001;8: Idu MM, Reekers JA, Balm R, et al. Collapse of a stent-graft following treatment of a traumatic thoracic aortic rupture. J Endovasc Ther. 2005;12: Lange C, Ødegård A, Lundbom J, et al. Type III endoleak from a thoracic aortic stent-graft. J Endovasc Ther. 2002;9: Major A, Guidoin R, Soulez G, et al. Implant degradation and poor healing after endovascular repair of abdominal aortic aneurysms: an analysis of explanted stent-grafts. J Endovasc Ther. 2006;13: Melissano G, Tshomba Y, Civilini E, et al. Disappointing results with a new commercially available thoracic endograft. J Vasc Surg. 2004;39:

9 I20 J ENDOVASC THER 36. Najibi S, Steinberg J, Katzen BT, et al. Detection of isolated hook fractures 36 months after implantation of the Ancure endograft: a cautionary note. J Vasc Surg. 2001;34: Riepe G, Heilberger P, Umscheid T, et al. Frame dislocation of body middle rings in endovascular stent tube grafts. Eur J Vasc Endovasc Surg. 1999;17: Rumball-Smith A, Wright IA, Buckenham TM. Strut failure in the body of the Zenith abdominal endoprosthesis. Eur J Vasc Endovasc Surg. 2006;32: Rutherford RB. Structural failures in abdominal aortic aneurysm stentgrafts: threat to durability and challenge to technology. Semin Vasc Surg. 2004;17: Umscheid T, Stelter WJ. Time-related alterations of shape, position, and structure in selfexpanding modular aortic stent-grafts. J Endovasc Surg. 1999;6: Wanhainen A, Nyman R, Eriksson MO, et al. First report of a late type III endoleak from fabric tears of a Zenith stent graft. J Vasc Surg. 2008;48: Zarins CK, Arko FR, Crabtree T, et al. Explant analysis of AneuRx stent grafts: relationship between structural findings and clinical outcome. J Vasc Surg. 2004;40: Ghanim K, Mwipatayi BP, Abbas M, et al. Late stent-graft migration secondary to separation of the uncovered segment from the main body of a Zenith endoluminal graft. J Endovasc Ther. 2006;13: Arko FR, Murphy EH. Endovascular aneurysm repair utilizing the AneuRx and Talent stent grafts. Perspect Vasc Surg Endovasc Ther. 2008;20: Ricotta JJ, Oderich GS. The Cook Zenith AAA endovascular graft. Perspect Vasc Surg Endovasc Ther. 2008;20: Alterman DM, Stevens SL. The Excluder aortic endograft. Perspect Vasc Surg Endovasc Ther. 2008;20: Wang GJ, Carpenter JP, Endologix Investigators. The Powerlink system for endovascular abdominal aortic aneurysm repair: six-year results. J Vasc Surg. 2008;48: e Hughes CG, Sulzer CF, McCann RL, et al. Endovascular approaches to complex thoracic aortic disease. Semin Cardiothorac Vasc Anesth Nov 24.[Epub ahead of print] 49. Qu L, Raithel D. Two-year single-center experience with thoracic endovascular aortic repair using the EndoFit Thoracic Stent-Graft. J Endovasc Ther. 2008;15: Makaroun MS, Dillavou ED, Wheatley GH, et al. Five-year results of endovascular treatment with the Gore TAG device compared with open repair of thoracic aortic aneurysms. J Vasc Surg. 2008;47: Morales JP, Greenberg RK, Morales CA, et al. Thoracic aortic lesions treated with the Zenith TX1 and TX2 thoracic devices: intermediateand long-term outcomes. J Vasc Surg. 2008;48: Resch T, Malina M, Lindblad B, et al. The impact of stent design on proximal stent-graft fixation in the abdominal aorta: an experimental study. Eur J Vasc Endovasc Surg. 2000; 20: Murphy EH, Johnson ED, Arko FR. Devicespecific resistance to in vivo displacement of stent-grafts implanted with maximum iliac fixation. J Endovasc Ther. 2007;14: Cheng CP, Herfkens RJ, Taylor CA. Abdominal aortic hemodynamic conditions in healthy subjects aged at rest and during lower limb exercise: in vivo quantification using MRI. Atherosclerosis. 2003;168: Taylor CA, Cheng CP, Espinosa LA, et al. In vivo quantification of blood flow and wall shear stress in the human abdominal aorta during lower limb exercise. Ann Biomed Eng. 2002;30: Taylor CA, Hughes TJ, Zarins CK. Effect of exercise on hemodynamic conditions in the abdominal aorta. J Vasc Surg. 1999;29: Conners MS, Sternbergh WC, Carter G, et al. Endograft migration one to four years after endovascular abdominal aortic aneurysm repair with the AneuRx device: a cautionary note. J Vasc Surg. 2002;36: Dalainas I, Nano G, Bianchi P, et al. Aortic neck dilatation and endograft migration are correlated with self-expanding endografts. J Endovasc Ther. 2007;14: Zarins CK. Stent-graft migration: how do we know when we have it and what is its significance? [commentary] J Endovasc Ther. 2004;11: Sampaio SM, Panneton JM, Mozes G, et al. AneuRx device migration: incidence, risk factors, and consequences. Ann Vasc Surg. 2005;19: Tonnessen BH, Sternbergh WC, Money SR. Mid- and long-term device migration after endovascular abdominal aortic aneurysm repair: a comparison of AneuRx and Zenith endografts. J Vasc Surg. 2005;42:

10 J ENDOVASC THER I Taylor CA, Draney MT, Ku JP, et al. Predictive medicine: computational techniques in therapeutic decision-making. Comput Aided Surg. 1999;4: Wilson NM, Wang KC, Dutton RW, et al. A software framework for creating patient specific geometric models from medical imaging data for simulation based medical planning of vascular surgery. Lect Notes Comput Sci. 2001;2208: Sahni O, Müller J, Jansen KE, et al. Efficient anisotropic adaptive discretization of the cardiovascular system. Comput Methods Appl Mech Engrg. 2006;195: Bekkers EJ, Taylor CA. Multiscale vascular surface model generation from medical imaging data using hierarchical features. IEEE Trans Med Imaging. 2008;27: Buth J, Harris PL, van Marrewijk C, et al. The significance and management of different types of endoleaks. Semin Vasc Surg. 2003;16: Greenberg RK, Turc A, Haulon S, et al. Stentgraft migration: a reappraisal of analysis methods and proposed revised definition. J Endovasc Ther. 2004;11: Zarins CK, Arko FR, Crabtree T, et al. Explant analysis of AneuRx stent grafts: relationship between structural findings and clinical outcome. J Vasc Surg. 2004;40: Volodos SM, Sayers RD, Gostelow JP, et al. An investigation into the cause of distal endoleaks: role of displacement force on the distal end of a stent-graft. J Endovasc Ther. 2005;12: Volodos SM, Sayers RD, Gostelow JP, et al. Factors affecting the displacement force exerted on a stent graft after AAA repair an in vitro study. Eur J Vasc Endovasc Surg. 2003;26: Hinchliffe RJ, Natarajan S, Hopkinson BR. In vitro analysis of modular aortic stent-graft failure. J Endovasc Ther. 2006;13: Arko FR, Heikkinen MA, Lee ES, et al. Iliac fixation length and resistance to in-vivo stent graft displacement. J Vasc Surg. 2005;41: Lambert AW, Williams DJ, Budd JS, et al. Experimental assessment of proximal stentgraft (Intervascular) fixation in human cadaveric infrarenal aortas. Eur J Vasc Endovasc Surg. 1999;17: van Prehn J, van der Wal MB, Vincken K, et al. Intra- and interobserver variability of aortic aneurysm volume measurement with fast CTA postprocessing software. J Endovasc Ther. 2008;15: Diehm N, Dick F. Commentary: Aneurysm sac diameter measurement versus volume analysis in EVAR surveillance: out with the old and in with the new. J Endovasc Ther. 2008;15: Molony DS, Callanan A, Morris LG, et al. Geometrical enhancements for abdominal aortic stent-grafts. J Endovasc Ther. 2008;15: Howell BA, Kim T, Cheer A, et al. Computational fluid dynamics within bifurcated abdominal aortic stent-grafts. J Endovasc Ther. 2007;14: Pacanowski JP, Stevens SL, Freeman MB, et al. Endotension distribution and the role of thrombus following endovascular AAA exclusion. J Endovasc Ther. 2002;9: Liffman K, Lawrence-Brown MM, Semmens JB, et al. Analytical modeling and numerical simulation of forces in an endoluminal graft. J Endovasc Ther. 2001;8: Li Z, Kleinstreuer C. Analysis of biomechanical factors affecting stent-graft migration in an abdominal aortic aneurysm model. J Biomech. 2006;39: Li Z, Kleinstreuer C. Fluid-structure interaction effects on sac-blood pressure and wall stress in a stented aneurysm. J Biomech Eng. 2005; 127: Rafii BY, Abilez OJ, Benharash P, et al. Lateral movement of endografts within the aneurysm sac is an indicator of stent-graft instability. J Endovasc Ther. 2008;15: