Micromechanical and Ultrastructural Properties of Abdominal Aortic Aneurysms

Abdominal aortic aneurysms are a common condition of uncertain pathogenesis that can rupture if left untreated. Current recommended thresholds for planned repair are empirical and based entirely on diameter. It has been observed that some aneurysms rupture before reaching the threshold for repair whilst other larger aneurysms do not rupture. It is likely that geometry is not the only factor influencing rupture risk. Biomechanical indices aiming to improve and personalise rupture risk prediction require, amongst other things, knowledge of the material properties of the tissue and realistic constitutive models. These depend on the composition and organisation of the vessel wall which has been shown to undergo drastic changes with aneurysmal degeneration, with loss of elastin, smooth muscle cells, and an accumulation of isotropically arranged collagen. Most aneurysms are lined with intraluminal thrombus, which has an uncertain effect on the underlying vessel wall, with some authors demonstrating a reduction in wall stress and others a reduction in wall strength. The majority of studies investigating biomechanical properties of ex vivo abdominal aortic aneurysm tissues have used low-resolution techniques, such as tensile testing, able to measure the global material properties at the macroscale. High-resolution engineering techniques such as nanoindentation and atomic force microscopy have been modified for use in soft biological tissues and applied to vascular tissues with promising results. These techniques have the potential to advance the understanding and improve the management of abdominal aortic aneurysmal disease.


Background
Abdominal aortic aneurysms (AAAs) are found in 1. [3][4][5].7% of the male population over the age of 65 [1]. They are usually asymptomatic prior to rupture, following which the mortality is 50-90% [2]. Open surgical repair (OSR), which involves replacement of the diseased segment with a prosthetic graft, has an early mortality of 3.2%, whilst following endovascular aneurysm repair (EVAR), in which a self-expanding stent graft is deployed under imaging guidance via the femoral arteries, it is 1.2% [3]. Planned repair is usually delayed until the risk of aneurysm rupture exceeds the risks associated with repair. Necropsy and clinical studies have indicated that the risk of rupture accelerates with increasing diameter [4]. The current recommended and most widely utilised threshold for repair is based on geometry. That is, when the maximum aneurysm diameter exceeds 5.5 cm in men and 5 cm in women or increases by > 1 cm/year [5]. Whilst delaying repair until these empirical thresholds are reached has been shown to be as safe as early repair for a population [6], autopsy studies have shown that small AAAs do rupture [7], and in certain instances large aneurysms will not rupture given the life-expectancy of the patient. Furthermore, not all aneurysms enlarge at the same rate [8], with the potential for faster growing

Open Access
Artery Research *Correspondence: m.hossack@liverpool.ac.uk 1 Mechanical, Materials and Aerospace Engineering, University of Liverpool, Liverpool, UK Full list of author information is available at the end of the article aneurysms to rupture whilst under surveillance or awaiting repair.
Biomechanically, AAA rupture occurs when stress acting upon the wall overcomes the strength of the wall to resist this stress. The predominant source of stress is systemic pressurisation, which is orders of magnitude larger than the shear stress due to blood flow across the wall. The theoretical basis for utilising the maximum diameter to predict wall stress is the Law of Laplace. This states that the wall stress of a thin-walled sphere is proportional to its diameter. However, this cannot account for complex geometry as is seen in AAA. Using realistic geometry, it can be seen that different parts of a given AAA's wall will display significantly different wall stresses [9,10] and two aneurysms with identical maximum diameters can have significantly different wall stress distributions (Fig. 1).
To improve and personalise AAA rupture risk prediction, biomechanically based models have been developed which include a quantification of individual wall stress using the finite element analysis (FEA) method [11]. This requires information regarding the material properties of the tissues, as well as aortic geometry, wall thickness, applied forces, boundary conditions, and use of appropriate constitutive models [2]. Using these methods, it has been demonstrated that peak wall stress (PWS) is significantly higher in symptomatic and ruptured, than electively repaired AAAs [9]; PWS is superior to maximum diameter in predicting rupture or development of symptoms [10]; and PWS correlates to the site of AAA rupture [12]. Improving our understanding of the material properties of AAA tissue has the potential to improve rupture risk prediction. These properties are determined by the composition and organisation of the tissue. In Sect. 2, we summarise the structural changes in the vessel wall occurring with aneurysmal dilatation. In Sect. 3 we review the mechanical testing of AAA wall tissue at different length scales, focusing on novel applications of high-resolution engineering techniques. Intraluminal thrombus tissue (ILT) lines most AAAs and has been shown to affect the underlying aortic wall in a variety of ways. In Sect. 4, we review the structure and mechanical testing of ILT.

Biochemistry
The aorta comprises of cellular components embedded in an extracellular matrix (ECM), which provides the structural and mechanical function. The cellular components consist of endothelial cells, which line and interact with the flow lumen; smooth muscle cells (SMCs) and fibroblasts, which are responsible for synthesising and organising the ECM. The ECM consists mainly of the fibrillar proteins elastin and collagen, which are organised into fibres. They are predominantly responsible for the mechanical properties of the aorta, providing passive elastic recoil and tensile strength, respectively. They constitute ~ 50% of the dry weight of large arteries such as the aorta [13]. The ECM also contains proteoglycans (PGs), glycoproteins and glycosaminoglycans (GAGs). The principle PG in the aorta is versican, which interacts with hyaluronan-one of the three GAGs found in the aorta-to form aggregates that trap water, creating viscoelastic and turgor pressure, possibly preventing deformation of the vessel wall and compression of the cellular components, by the pulsatile pressure within the lumen [14].
The composition of the aorta changes as it becomes aneurysmal. There is a well-established drastic loss of elastin and SMCs, and a compensatory increase in collagen concentration [15]. Reductions in the amount of PG [14], GAG [16], and numerous other ECM proteins [17] have also been demonstrated. These changes are mediated by unregulated proteinase activity, an explanation of which is beyond the scope of this review.

Healthy Aorta
The aorta is a hierarchical structure that can be divided into three layers-intima, media and adventitia-separated by internal and external elastic lamellae. In healthy tissue, each of these layers is distinct. The intima (innermost) consisting of a single layer of endothelial cells, is mechanically insignificant in young healthy vessels but stiffens with age [18][19][20]. The media is the thickest and most mechanically significant layer. It consists of a series of concentric subunits-medial lamellar units (MLUs)-comprising elastin, SMCs and collagen [21]. Elastin is found in three forms: thick continuous circumferentially orientated sheets (lamellae), which are the main load-bearing layer at physiological loads [22]; dense interlamellar fibres, which protrude obliquely from lamellae to surround the SMCs, facilitating mechanotransduction [23]; and radially orientated struts connecting adjacent lamellae, providing support for radial loads, preventing delamination or dissection (Fig. 2).
Scanning electron microscopy (SEM) images of healthy ascending aortic tissue reveal the elastic lamellae to be perforated by irregular pores (Fig. 3A), which may allow radial passage of metabolites across the aortic wall. The lamellae are themselves formed from a coalescence of fibrillary components running in small bundles, and their surface is characterised by protrusions of fibrous fascicles extending into the interlamellar space (Fig. 3B) [24].
Collagen is found in all layers of the aortic wall. The orientation of the fibres is layer-specific with no fibres passing in the radial direction. In the intima and media, fibres are predominantly circumferential (Fig. 4, Healthy, panes I, M); and in the adventitia, axial (Fig. 4, Healthy, pane A). Transition zones between intimamedia and media-adventitia represent the internal and external elastic laminae, respectively [18].
At unloaded and physiological pressures adventitial collagen fibres appear wavy [18], straightening at higher pressures, preventing over-stretch and rupture, producing the characteristic non-linear stress-strain curve seen in healthy arteries. Collagen is organised in thin uniform fibrils gathered into bundles, with individual fibrils connected by small proteoglycans (Fig. 5) [24]. . The interlamellar zones contain smooth muscle cells whose elliptical nuclei (N) are orientated circumferentially and whose cytoplasm (Cyt) abuts interlamellar elastin fibres (black arrows), which transmit mechanical forces to the cells. Collagen fibres (white arrows) are adjacent to lamellar surfaces and are orientated predominantly circumferentially. Reprinted from O'Connell et al. [21] with permission from Elsevier

Aneurysmal Aorta
AAA tissue is characterised by loss of the normal threelayer distinction. There is substantial variability in fibre architecture, even between samples from the same aneurysm, and a marked reorientation of collagen fibres into the circumferential direction. Towards the abluminal side, there is adipocyte deposition and cystic medial degeneration (Fig. 4, AAA panes) [18].
Niestrawska et al. measured collagen fibre orientation in AAA tissue, grouping samples according to their  relationship with a healthy 'inflection point' , indicating rapid stiffening and therefore collagen recruitment, established by tensile testing [25]. All samples demonstrated drastic reductions in elastin and SMC content. In stage 1 (healthy inflection point) samples there was a marked reorientation of collagen fibres to a circumferential direction (Fig. 6, left). Stage 2 (more distensible) samples were characterised by an accumulation of isotropically dispersed collagen peripherally (Fig. 6, centre). Stage 3 (less distensible) samples demonstrated a significantly thickened layer of isotropically dispersed collagen comprising a median 66.22% of the wall thickness ( Fig. 6, right). Nerves normally seen in the periadventitial tissue could be found embedded in this 'neo-adventitia' . Within stage 3, two distinct morphologies were noted: The 'safely remodelled' subgroup consisted of a thick layer of collagen entrapping nerves, with no visible adipocytes. Whereas the 'vulnerable' group demonstrated significant amounts of inflammation and adipocytes. No difference in the median diameters of safe and vulnerable AAAs was seen, reinforcing the theory that maximum diameter alone is insufficient to characterise an individual's rupture risk.
The presence of triglyceride (TG) molecules in the adventitia has been shown to be much higher in AAA sac than neck tissue, as demonstrated by Tanaka et al. using a matrix-assisted laser desorption/ionisation imaging mass spectrometry (MALDI-IMS) imaging system [26]. This system allows characterisation of molecules on vascular tissue sections without requiring antibodies or staining. TG's were found within ectopic adipocytes, which were much larger in the sac region. The collagen fibres here were thinner and displayed a more irregular structure ( Fig. 7). High circulating TG levels have been implicated as a risk factor for aneurysm rupture [27]. The accumulation of TG within adipocytes located in the adventitia may lead to adipocyte hypertrophy and disruption of the ECM collagen network, so crucial to resist over-stretching and rupture.
Majewski et al. examined the ultrastructure of tissue from the non-dilated neck and sac of infra-renal AAA patients with transmission electron microscopy [28]. Patients were grouped according to subsequent neck Representative intensity plots for three stages of aneurysm wall tissue based on results from tensile testing, demonstrating collagen fibre orientation through the aortic wall. Depth of 0 denotes luminal side. Red colour indicates fibre angles at which there are fibres at that orientation; blue, absence of fibres. 0° fibre angle represents the circumferential direction, 90°, axial. Reprinted from Niestrawska et al. [25] with permission from Elsevier dilatation during follow-up. Tissue from patients with subsequent minor dilatation (median 1 mm increase) demonstrated minimal disorganisation in the aortic neck wall. The media demonstrated a normal arrangement of elastin fibres, collagen and SMCs. The adventitia was composed of regular fibroblasts surrounded by collagen fibres, proteoglycans and glycosaminoglycans. Tissue from patients with increased neck dilatation (median 5.2 mm increase) demonstrated partially disrupted elastic fibres, haphazardly organised collagen bundles and inflammatory cell infiltrates. Ultrastructural findings correlated with pre-operative distensibility measured by motion-mode ultrasound (US)-neck tissue from more distensible aorta demonstrated less severe ultrastructural changes and underwent less subsequent dilatation, and vice versa. Unsurprisingly aneurysm sac tissue displayed the most severe changes. Large areas without endothelial lining were seen. Elastic fibres were either, fragmented and surrounded by necrotic cells, or totally absent. Areas of the media contained haphazardly arranged collagen fibres and glycosaminoglycans. Other areas were acellular with no visible SMCs. Monocyte/macrophages, lymphoid cells and erythrocytes were seen, with some macrophages containing lipid. The fibrillary structures were completely without organisation [28].

Macroscale
A detailed account of the macroscale mechanical testing of aortic tissue is beyond the scope of this review. We have summarised the most important findings from the main modalities of in vivo and ex vivo mechanical testing.

In Vivo Testing
Ejection of blood from the left ventricle induces a wave that propagates through the arterial tree, causing distensibility of the vessel wall. The speed of this wave, pulse wave velocity (PWV), has been used to measure arterial stiffness in patients with AAA, finding they had significantly higher PWV (higher stiffness), with a less uniform pulse wave propagation [29]. Van Disseldorp et al. used time-resolved three-dimensional (3D) US to track aortic wall motion during systole and diastole in healthy and AAA patients. Simulated displacements generated from FEA were fitted to measured axial displacements by varying the aortic stiffness until convergence occurred, resulting in the subject specific aortic stiffness, which was found to correlate with diameter [30].
Magnetic resonance elastography (MRE) examines the propagation of shear waves through soft tissues. It has been used to investigate the correlation between aortic stiffness and aneurysm diameter finding significantly higher stiffness in AAA, than in both non-dilated vessel (in AAA cases) and age-matched healthy patients [31]. Patients with similar diameters had very different stiffness values, reinforcing the theory that maximum diameter alone is insufficient to differentiate between individual AAA's.

Ex Vivo Testing
Uniaxial tensile tests (UTTs) have been widely used to characterise the mechanical properties of aortic tissue ex vivo. Measurements are frequently performed in longitudinal and circumferential orientations to allow for anisotropy. There are well documented inconsistencies in the reporting of stress-stain behaviour derived from UTTs in the literature due to the different methods employed to calculate 'stress' and 'strain' . This can make  Table 1 summarises experimental studies on human tissue since 1994. Aneurysmal tissue has been shown to be stiffer [33] and weaker [34,35] than control. When separated into layers the adventitia was found to be significantly weaker in AAA, whilst the media was not significantly different [36]. Larger AAAs (> 5.5 cm maximum diameter) were found to have higher failure properties than smaller AAA s (< 5.5 cm) [37]; tissue from ruptured AAAs has been found to be both weaker [38] and not significantly different [39] from unruptured tissue; and AAA wall underlying thick ILT has been found to be weaker than that underlying thin ILT [40].
UTT can provide useful information regarding strength and stiffness but it is unable to replicate in vivo loading conditions. A fundamental goal in constitutive modelling is to be able to predict the mechanical behaviour of a tissue under its loading state [41]. Biaxial tensile testing (BTT) allows a 2D stress state, and is sufficient to develop constitutive relations. A description of BTT of AAA tissue is beyond the scope of this review but further details can be found in [42,43].

Microscale
Macroscale testing is able to provide data regarding the global material properties of a tissue. By implementing higher-resolution testing methods, it is possible to measure localised mechanical properties. This is especially useful for heterogeneous biological tissues, where they can be used to measure the mechanical properties of discrete tissue components.

Nanoindentation
Nanoindentation was originally developed to characterise engineering materials. The testing set up and principles are discussed in detail elsewhere [44,45]. In brief, a rigid indenter, usually made of diamond, is used to probe the material under examination. The load and displacement are continuously recorded producing a load-unload curve from which the mechanical properties, usually hardness and elastic modulus, are calculated. Use of a sharp tip allows the elastic properties of hard (calcified) biological materials to be measured at the micron-length scale. However, soft biological materials present a challenge due to a high surface roughness and low stiffness making it difficult for the apparatus to register specimen contact prior to penetration. Thus the original technique has been modified to allow examination of soft biological materials, with use of a larger radius tip (100 µm) recommended to obtain reproducible results [46]. This necessarily compromises on spatial resolution. Nanoindentation has been used to characterise the mechanical properties of non-human healthy aortic tissue specimens demonstrating that the specimen stiffness' increases from the inner to outer portion of the vessel in both ferret [47] and porcine aorta [48]; stiffness increases with distance from the heart in ovine aorta [49] and is higher in the circumferential than axial orientation in porcine aorta [50] (Table 2).
Nanoindentation has been shown to be useful when used to probe aneurysmal human thoracic tissue [51,52]. It has also been used on aneurysmal abdominal aorta.  [53]. Contrary to previous findings from macroscale mechanical testing using imaging [30,31], aneurysmal degeneration was associated with a reduced elastic modulus (reduced stiffness) in each layer except the adventitia, where it was non-significantly stiffer. This discrepancy may be because the aforementioned imaging studies are only able to measure a proxy of tissue stiffness, and even then, only of the whole composite structure including ILT and calcification, which were removed by Meekel et al. [53].

Scanning Acoustic Microscopy (SAM)
SAM is an imaging technique capable of providing both histological and mechanical information of in vitro biological specimens. The image contrast is derived from differences in the speed of propagation of acoustic waves through the tissue, which is related to the Young's modulus. It is distinct from PWV and is able to measure the stiffness of separate tissue components. Detailed account of SAM is beyond the scope of this review but SAM has been applied to AAA tissue, and compared with atherosclerotic tissue. The only significant finding was of a slower wave speed and therefore lower elasticity in the intima of aneurysmal tissue [54].

Atomic Force Microscopy (AFM)
AFM involves use of a nanometre-scale sharp probe mounted on a compliant cantilever. This tip is brought to the sample surface, with forces between sample and tip resulting in deflection of the cantilever. The deflection is measured by a laser light reflected off the cantilever upper surface to a photodetector. AFM is not only able to measure the mechanical properties at the nanoscale but it can also provide high spatial resolution images of tissue ultrastructure.
AFM has been used on atherosclerotic human abdominal [55] and healthy porcine thoracic [56] aorta, demonstrating that the elastin-rich lamellae are significantly stiffer than the interlamellar zones when a low strain compressive force is applied. Lindeman et al. applied AFM to human control and aneurysmal tissue from degenerative AAA [57]. AFM was used to probe the adventitia using both a sharp (20 nm) and blunter ball (10 µm) tip, theoretically allowing testing of individual fibres and tissue composite respectively. In control aorta, the response to indentation was the same regardless of tip diameter. In AAA, the sharp tip

Intraluminal Thrombus (ILT)
ILT is present in 75% of AAA [58]. It is a three-dimensional fibrin structure consisting of blood proteins, blood cells, platelets and cellular debris, which accumulates in AAA in regions of low oscillatory shear index [59]. Its accumulation maintains physiological blood pressure and velocity in the expanding lumen. In patients undergoing AAA repair without evidence of frank aneurysm leak, the presence of a high attenuating peripheral crescent on pre-operative unenhanced CT (so-called 'crescent sign') was 77% sensitive and 93% specific for aneurysm 'complications' discovered during repair [60]. Thrombus from patients with the crescent sign and AAA rupture was characterised by clefts of seeping blood, which were not identified in unruptured AAA not demonstrating crescent sign [61]. Thus, ILT failure is thought to precede AAA rupture. Indeed, circumferential stress in a pressurised thick-walled tube is highest and lowest at the inner and outer edges, respectively [62].

Structure
Initially thrombus is characterised by erythrocytes, platelets, thrombocytes and leucocytes embedded within a loose fibrin network. As it evolves, a secondary fibrin network of thick fibrin fibres forms through loss of erythrocytes and accumulation of platelets. A well-organised ILT structure can be described by three individual layers: luminal (red), medial (white) and abluminal (brown), which can be differentiated histologically. The innermost and freshest 'luminal layer' is characterised by fibrin fibres arranged into thick bundles with fine secondary structures cross-linking them. In the medial layer, some fibres are degenerated, whereas in the abluminal layer, all fibres are degenerated with no coherent structure [63]. Not all thrombus is well-organised and three distinct morphologies have been described based on visual inspection and mechanical properties (Fig. 8) [64]. The ILT is traversed by canaliculi allowing passage of fluid and cells [65]. Gasser et al. used automatic image processing to examine the micro structure of histological slices from 7 ILT specimens harvested from human AAA [66]. They noted that the ratio of pore to ligament area reduced from the luminal (23.28% ± 4.03%) to the abluminal (19.59% ± 5.31%) layer. The abluminal layer exhibited more large pores and fewer small pores compared to the luminal and medial layers. Furthermore, pore orientation in the luminal layer tended to be towards the circumferential direction, whilst in the abluminal layer there was no preferred orientation, explaining the anisotropic and isotropic mechanical properties of these respective layers seen in some studies [64,67].

Mechanical Testing
ILT is thought to affect the underlying aortic wall in multiple ways. Tanios et al. demonstrated that amount of types I, III and total collagen correlated negatively with increasing ILT thickness. They concluded that ILT may act to cushion the underlying aortic wall, thereby reducing wall stress and the need for compensatory collagen synthesis [68]. This was based on extrapolation from the observation that collagen content was ILT has been shown to reduce wall stress in numerous computational studies [68][69][70][71][72] with the size of the effect depending on the shape of the aneurysm and the material properties of the thrombus. However, in one of the few experimental studies, Schurink et al. demonstrated that the pressure measured within ILT of AAA correlated with that measured in the radial artery, implying that ILT does not reduce the pressure (and therefore wall stress) on the underlying aortic wall [73]. This would account for the clinical observation that AAAs treated by iliac ligation and extra-anatomical bypass-inducing thrombosis in the aneurysm-can still rupture post-operatively [74][75][76]. One explanation for a reduction in wall stress despite transmitting systemic pressure to the aortic wall is provided by Thubrikar et al. who found experimentally that the longitudinal and circumferential strains induced by pressurising explanted AAA were significantly higher after removal of the thrombus [77], indicating that adherent ILT reinforces the AAA wall. Polzer et al. demonstrated this computationally by modelling the ILT as a poroelastic material [78].
Growth of ILT has been associated with aneurysm growth and rupture [79,80]. It has been shown to be a reservoir for proteases with known activity against ECM proteins (albeit, the majority of the active proteases were found in the luminal layer, whilst the abluminal layer contained predominantly inactive forms due to presence of inhibitors) [81]. The wall adjacent to ILT has been shown to be thinner, contain fewer elastin fibres, fewer SMCs, and more apoptotic nuclei with clusters of inflammatory cells than wall adjacent to free flowing blood [82]. Vorp et al. demonstrated that aortic wall adjacent to thick ILT (> 4 mm) was significantly weaker than tissue adjacent to thin ILT (< 4 mm) in the same patient. Tissue from thick ILT areas showed a higher degree of inflammation and neovascularisation [40]. These changes were attributed to relative hypoxia by demonstrating a significantly lower partial pressure of oxygen (PaO 2 ) in the wall of patients with thick ILT [40]. Koole et al. demonstrated that ILT accumulation is positively correlated with MMP (matrix metalloproteinase)-2 content in the aortic wall, even after multivariable analysis [83], whilst Erdozain et al. demonstrated that SMCs secrete more MMP-2 under hypoxic conditions [84]. Thus, although ILT has been shown to reduce computational wall stress, it also has a deleterious effect on the underlying wall strength. Whether this is a direct effect due to hypoxia, an indirect effect due to reduction in wall stress, or a combination of both, remains uncertain.
Studies using FEA for patient specific rupture risk indices, such as PWS and rupture potential index (RPI), necessarily make several modelling assumptions, either because it is impossible or impractical to use patient-specific values. Often, ILT is either excluded entirely or given homogenous material properties. Martufi et al. demonstrated the importance of including realistic material parameters for ILT, as well as aortic wall thickness, in an FEA model of rupture risk [85]. Tensile testing was used to establish the tissue strength and material properties of tissue taken from AAA patients. This tissue was categorised into 4 groups depending on wall and ILT thickness (thick or thin). The strongest tissue was found in the thin wall, thin thrombus group (772 ± 295 kPa); thick wall, thick thrombus was the weakest (334 ± 254 kPa). A patient-specific wall and ILT geometry was segmented creating a mesh for FEA. Each wall element was given an accurate wall thickness, and material properties based on the results from tensile testing. The generated PWS was 311 kPa. Further simulations disregarding ILT; using constant wall thickness; and using homogenous material properties for the wall resulted in significant changes in the location and magnitude of the PWS obtained, highlighting the importance of including ILT and its effect on the underlying aortic wall in FEA simulations.
Numerous studies have investigated the mechanical properties of human abdominal aortic ILT. A representative sample are summarised in Table 3. It is difficult to directly compare these due to differences in testing methods and analysis. In general, the elastic modulus obtained relates to the method of testing, with studies using rheometry [86] giving the lowest values, followed by studies using compression testing [87], and studies using tensile testing [69] giving the highest. However, the trends seen within the studies are similar. In summary, stiffness and strength decrease from the luminal to abluminal side (i.e. non-homogeneity), whereas the viscoelastic properties do not vary significantly [88]. ILT displays a degree of nonlinearity on biaxial tensile testing in both circumferential and longitudinal orientations, with this decreasing from luminal to abluminal layers [64]. The difference between the longitudinal and circumferential directions is minimal (i.e. isotropy) in the medial and abluminal layers but anisotropy has been demonstrated in the luminal layer [66,67]. Two authors have investigated the regional heterogeneity of ILT-Celi et al. using uniaxial testing [89] and O'Leary et al. [64] using biaxial testing. Both found that the posterior region was stiffer than anterior. Tong et al. provide a comprehensive review of the structure, mechanics and histology of ILT [65].
Slaboch et al. [90] first used nanoindentation to investigate the mechanical properties of thrombi. They created murine thrombi and performed nanoindentation with a 1.75 mm flat punch cylindrical tip. Elastic contact theory was used to calculate the elastic modulus from the linear loading curve, with the result (55.1 ± 12.1 kPa) being  UTT indicates uniaxial tensile testing, BTT biaxial tensile testing, DMA dynamic mechanical analysis, NA not applicable, -no data consistent with others authors using compression to establish the mechanical properties of thrombi [87]. As ILT structure varies histologically, from the luminal (newer) to the abluminal (older) surface, it is possible that the material properties vary throughout its thickness. Van Dam et al. measured the material properties of ILT taken from 7 AAA patients using rheometry. They found that the variations found within one thrombus are of the same order of magnitude as found between patients. Accordingly, they concluded that the same material parameters may be used to describe all thrombi [88].

Conclusions
The maximum diameter threshold for aneurysm repair has been shown to be inferior to biomechanically based models for rupture risk assessment. To date, these models have utilised average material properties obtained from macroscale mechanical testing. It has been demonstrated that aneurysmal degeneration involves drastic structural changes and remodelling at the microscale. It is possible, using novel applications of high-resolution engineering techniques such as nanoindentation and AFM, to measure the micromechanical properties of ex vivo abdominal aortic tissue. These measurements could be correlated with biochemical assays and histological assessment to develop more realistic constitutive relations and material parameters for use in finite element simulations, which could further improve and personalise rupture risk prediction.