IntroductionDespite fundamental improvements in surgical treatment of Congenital Heart Defects, there are still challenges related to premature failure of the material used for such corrections, thus resulting in repeated operations during a patient’s life. This is particularly the case for complex defects with Right Ventricular Outflow Tract (RVOT) obstruction, such as in Tetralogy of Fallot/Pulmonary Atresia, whereby the pulmonary valve reconstruction remains problematic due to short-term durability of the currently used replacement solutions. We set out to test, for the first time, the suitability of amniotic membrane derived from human placenta for use in cardiovascular replacement of pulmonary valve.MethodsThe decellularized and preserved amniotic membrane, obtained through our optimised protocol, was characterised for mechanical and hydrodynamic properties in vitro, and then implanted in the RVOT position of two Landrace piglets for in vivo feasibility and performance evaluation.ResultsBoth the in vitro and in vivo assessments showed favourable outcomes. The decellularized amniotic membrane had mechanical properties comparable to the native porcine pulmonary valve leaflets. In hydrodynamic testing, the decellularized amniotic membrane-made valve exhibited favourable opening dynamics, with smooth and coordinated leaflet motion throughout the cycle. In vivo, the decellularized amniotic membrane-based valved conduit showed patency in the short- and long-term with no sign of stenosis or regurgitation.DiscussionThis study provides an in vivo proof of concept that the decellularized amniotic membrane can be implanted and perform as functional pulmonary valve in a porcine animal model mimicking the clinical scenario of Tetralogy of Fallot surgical correction in infants.

Fluid-structure interaction (FSI) is crucial in the numerical simulation of cardiovascular phenomena, where pulsatile blood flow dynamically interacts with highly deformable tissues. High-fidelity FSI approaches have become essential to enhance the understanding of potentially lethal pathologies, assisting diagnosis and development of novel therapeutic solutions. This work presents and experimentally validates a new, totally meshless FSI approach, specifically designed for cardiovascular applications. The method is based on the Lagrangian smoothed particle hydrodynamics (SPH), employing a unified physics to represent both blood and deformable walls, avoiding FSI interfaces. A key advantage of this method lies in its ability to overcome the SPH complex issue in contour management, a common challenge that typically increases the complexity of this methodology in FSI applications. Deformable walls are immersed in the fluid domain, and a buffer region of fluid is defined to handle the structural deformation. For validation, a new FSI benchmark is proposed and analyzed with the particle image velocimetry technique. Tailored to entail the typical complexities of relevant cardiovascular situations, the benchmark involves pulsatile flow interacting with a chamber with deformable curved walls, moving through both filling and emptying phases. Despite its simplified geometry, designed to allow a reliable experimental validation, the structure experiences a field of three-dimensional strains and large volume variations, thereby replicating complexities often associated with more intricate models. Numerical and experimental results show good agreement in terms of fluid velocity field and structural deformation, establishing the proposed totally meshless FSI approach as a reliable tool for complex cardiovascular modeling.

The morphological analysis of the left atrial appendage is crucial to ensure that the most appropriate treatment is selected for patients suffering from atrial fibrillation. However, the process is susceptible to three sources of subjectivity: identification of the region that separates the left atrial appendage from the left atrium, computation of the appendage features, and morphological classification. We developed an automated method for detecting the left atrial appendage from 3D reconstruction from clinical images, overcoming the need for human intervention. Our method operates on a triangulated surface mesh containing both the left atrium and the left atrial appendage. The method exploits skeletonisation, distance analysis, affinity evaluation, and plane identification techniques for accurate appendage detection. Furthermore, it computes the left atrial appendage size and performs network analysis of the appendage skeleton to characterize the auricula using specific metrics. In addition, we introduced a metric-based classification approach, wherein clusters are formed based on the extracted features. This approach allows for more objective and data-driven classification, enhancing the accuracy and reliability of the morphological analysis. By integrating these advancements, our method represents a significant step towards reducing subjectivity in the analysis of the left atrial appendage, thus facilitating more consistent and standardized investigations of its morphologies and therapeutic selection.

Subaortic stenosis, a heart disease characterised by a narrowing of the left ventricular outflow tract, is frequently caused by the presence of a subaortic membrane (SAM) located at the aortic valve inlet. This anatomical obstruction leads to significant haemodynamic alterations and leaflets fluttering, whose mechanisms are not yet fully understood. This research investigates, through computer simulations, the SAM's haemodynamic impact and the mechanism behind leaflets fluttering.A mono-physics fluid-structure interaction approach, based on the meshless smoothed particle hydrodynamics method, was employed. This approach represents both blood and structures with particles without defining interfaces, efficiently capturing large deformations and dynamic phenomena. Two common types of SAMs were investigated - a discrete thin SAM layer (flexible) and a thick fibromuscular ridge SAM (stiff) - and compared with a healthy aortic valve.Projected dynamic valve area (PDVA) was used as a reference parameter to quantify leaflet oscillation. While the PDVA in the healthy aortic valve stabilised at 283 mm2 without oscillation, both pathological cases exhibited self-sustained periodic fluctuations. In the presence of discrete thin SAM layer, the mean PDVA decreased by 3% compared to the healthy control. This reduction was more pronounced for thick fibromuscular ridge configurations, where the mean PDVA was 9% lower than the healthy case. Notably, stiffer SAM configurations more than doubled the oscillation amplitude (from 3.12 mm2 to 6.77 mm2) and increased the oscillation frequency by 8% relative to flexible membranes. Vortices dynamics was analysed, determining the phases of their formation, growth and migration. Through the analysis of velocity, vorticity, and shear stress maps, this study provides critical insights into the origin of fluttering and its influence over these key haemodynamic parameters.Findings demonstrate that the oscillatory leaflet motion is the result of vortices formation and shedding. The stiffness of the SAM significantly modulates the fluttering behaviour. While structural damage and haematological complications were not directly simulated, the identified oscillations represent haemodynamic conditions associated in literature with such pathologies. The observed alterations in wall shear stress magnitude and direction provide a physical basis for the mechanical environment that could contribute to endothelial cell dysfunction in the presence of SAM.

The impact of recovered atrial dynamics on pre-existing thrombi within the left atrial appendage (LAA) remains a poorly understood phenomenon, despite its potential clinical implications during spontaneous or induced restoration of sinus rhythm (SR) in patients with atrial fibrillation (AF). While computational studies have investigated thrombus formation during AF, none have examined the fate of established thrombi following SR restoration. This work addresses that gap by modelling thrombus formation and embolisation in the LAA under AF and after the return to normal rhythm. A meshless particle-based thrombus model was employed to simulate the formation of clots in a patient specific LAA, under AF conditions, and then analyse haemodynamic changes and thrombus behaviour under persistent AF and recovered SR. Results show that SR restoration profoundly modifies thrombus growth and erosion compared with AF. The resumption of atrial contractility increases local shear strain rate, especially in the proximity of anatomical features such as knees and trabeculae, driving progressive thrombi erosion and detachment from secondary lobes. These processes are amplified during the atrial systolic phase, when mechanical forces promote thrombus detachment. Conversely, if AF persists, the associated reduction in blood wash-out and shear stresses promote thrombus growth. Clots located in regions of higher stagnation remain stable under both restored SR and persistent AF. By providing the first direct computational comparison of thrombus behaviour in AF and after SR restoration, this work offers novel insights with potential to inform personalised therapeutic strategies.

Nitinol superelasticity is a crucial property for collapsible/expandable cardiovascular implants. The high class of risk associated with these devices requires an accurate mechanical characterisation as starting point for a reliable design methodology. Actual standards (ASTM-F2516) are based on standard metal testing and recommend the use of extensometers for the measurement of average nominal strains during tensile tests. However, measurement provided by classic extensometers may not be adequate to capture the strain evolution during the material phase transformation. These limitations can be overcome by full-field optical techniques such as Digital Image Correlation (DIC). This work presents a comprehensive comparison of various techniques for measuring strain on nitinol specimens. These include crosshead position, single and averaging (applied on both sides of the specimen) measurements using physical extensometers, virtual extensometer, and 1D and 2D DIC. Results are compared on the basis of the values determined for the upper/lower plateau stress, as defined by ASTM-F2516. Results show that the use of extensometers on nitinol strips subjected to tensile testing is inaccurate during phase transformation. Physical extensometers also introduce spurious local concentrated pressures at the knives contact region, which may alter the material stress plateaus values. DIC resulted more adequate to provide an accurate evaluation of strain localisation.

Atrial Fibrillation (AF) is a cardiac disease altering the human heart rate. It is posing an increasing burden to society, with complications that lead to stroke and ischemic events from thromboembolisms, originating in the left atrial appendage (LAA). Percutaneous LAA occlusion (LAAO) is becoming an increasingly adopted preventive treatment option due to its minimally invasive nature. However, this treatment faces complex challenges: the heterogeneity of LAA morphologies limits the pre-operative planning and several procedures are associated with peri-device leakage from malposition and device-related thrombi. One of the two most commonly deployed LAAO devices (LAAODs) is the Amulet Amplatzer (AA), a mesh-like pacifier device. In-silico models have demonstrated their potential to serve as supporting tools for clinical planning, providing insight able to enhance the efficacy and safety of the intervention. Most of the computational studies approximate the AA to a closed surface model. In this work, we aimed to develop a more realistic and detailed structural model of the AA, capturing the mesh of wires. Experimental tests on the physical device were conducted to compare the behaviour of simplified closed surface models and the newly developed braided geometry. The results have demonstrated how closed surface models of the AA fail to capture the real deformation mechanism of the physical device. Conversely, the more realistic braided characterisation mimics more closely the changes in shape of the physical AA, by capturing the change in angles of the wires. Finally, the virtual deployment of the intertwined model into a patient-specific LAA resulted in a configuration similar to the clinically implanted AA.

Various biomaterials are currently used in clinical settings for heart valve repair and replacement. However, the optimal tissue preparation technique remains elusive. In this study, a non-crosslinked tissue obtained from bovine pericardium, developed by Adeka Corporation, was compared with three commercially-available tissues: two fixed tissues obtained from crosslinked bovine pericardium, CryoLife PhotoFix and LeMaitre CardioCel, and an unfixed one obtained from swine intestinal submucosa, CorMatrix CorPatch. The four biomaterials were used to produce aortic valve prostheses, and their hydrodynamic performance and durability were evaluated according to the ISO5840 standards. Resistance to calcification was evaluated by exposing the tissue to simulated body fluids, followed by SEM and Micro-CT analysis. Thrombogenicity was investigated by exposing the tissue to fresh ovine blood, followed by imaging with SEM and quantifying platelet deposition with lactate dehydrogenase assay. All constructed valves were compliant with the ISO5840 for hydrodynamic assessment. Non-crosslinked tissues, Adeka and CorPatch, showed lower durability but exhibited improved in vitro performances in the simulated biological environments compared to fixed ones. The Adeka tissue demonstrated significantly lower calcium (p < 0.05, Adeka vs. CardioCel, PhotoFix) and platelet deposition (p < 0.05, Adeka vs. CardioCel, CorPatch, PhotoFix), along with improved durability compared to the other unfixed tissue (Adeka 357,198 vs. CorPatch 0 cycles). In vitro experiments indicate that the crosslinked tissues, CardioCel and PhotoFix, provided mechanical strength adequate to withstand the operating conditions required for heart valve tissue applications. Whilst the Adeka material, a non-crosslinked tissue surrogate, provides improved properties in terms of resistance to calcification and thrombus formation.

Rheumatoid Arthritis (RA) and Osteoarthritis (OA) are among the most impactful musculoskeletal disorders causing articular cartilage degradation, ultimately leading to loss of the joint functionality. Matrix-assisted Autologous Chondrocytes Implantation (MACI) is one of the most promising reconstructive techniques to treat chondral defects (CDs). MACI relies on a matrix cellularized with autologous chondrocytes implanted directly onto cartilage defects. Despite MACI's effectiveness, post-surgery rehabilitation remains a challenge, as it fails to induce the optimal mechanobiology necessary for an effective cartilage regeneration. Additionally, there is a significant patient-to-patient variability and the local loads occurring during rehabilitation might consequently vary greatly. We propose a personalized approach focused on the delivery local pro-regenerative mechanobiological cues to dramatically improve the cartilage restoration after MACI. We developed an innovative scaffold to be used as matrix in MACI, capable to enhance the cartilage repair by delivering in situ controlled, and personalized, mechanical cues triggering pro-regenerative cellular responses to embedded human articular chondrocytes (hACs). The scaffold relies on an electrospun matrix made of aligned fibers composed of PVDF-TrFE, a piezoelectric polymer, enriched with ferromagnetic Fe3O4 nanoparticles capable to confer magnetic properties to the scaffold. MNPs were simultaneously dispersed in the polymeric solution, and microfibers were collected onto a high-speed rotating collector to obtain an aligned micropattern, capable to give mechanical anisotropy to the scaffold. After cellularization with hACs, we subjected the scaffold to daily magnetic stimulation up to 14 days. The scaffold was highly responsive to external magnetic stimuli. In addition, hACs produced a type II collagen-rich extracellular matrix when cultured within the scaffolds subjected to magnetic stimulation. Remarkably, we observed an increase of cell viability, and of type II/type X collagen ratio. Our scaffold was able to provide pro-regenerative cues to hACs after mechanical cyclic deformations induced by repeated magnetic stimulations. Such an approach paves the way to an effective, and definitive therapeutic procedure for the treatment of chondral defects.

During crystallization, crystals nucleate and grow within materials, often impinging and interacting in a stochastic manner. This complexity has long hindered accurate reconstructions of a material’s crystallization history. By considering a representative material region with a finite crystal population, we derive equations that accurately predict crystal size and free surface evolution throughout the crystallization process. These equations, paired with a numerical solver, enable reconstructing nucleation events and crystallinity progression using the crystal size distribution and growth rates. We demonstrate this method by pinpointing the nucleation and crystallinity timelines of simulated, manufactured, and ancient geological materials, entirely without real-time observation. Our model offers unprecedented insights into extreme crystallization environments that are difficult to mimic, such as volcanic magma chambers, and supports the design of advanced materials.

Background: Subaortic stenosis is an aortic disease characterised by the presence of a membrane located at the aortic valve inlet, that causes a sudden reduction of the inflow lumen. The membrane develops as a tissue growth of variable thickness that can cause a major increase in the pressure gradient. In this case, when diagnosed, it is removed by surgical resection. Methods: To investigate the haemodynamic alteration introduced by subaortic membranes, an in vitr o study was designed and performed. Stiff and flexible membranes were implanted at the inlet of a bioprosthetic control valve. These mock membranes had different radial and angular alignment, modelling concentric and eccentric orifice positions. For each configuration, a range of different membrane extensions was studied, progressively reducing the orifice area at the inlet of the control valve. Results: Analysis of the hydrodynamic performances indicates that the detrimental effect of subaortic membranes becomes significant when the membrane orifice areas reduce below 75% of the unobstructed inflow lumen. Video analysis of the valve leaflets dynamics indicates that, together with a worsening in the systolic pressure gradient, the presence of subaortic membranes increases cusps fluttering. As the membrane orifice area reduces, leaflets experience faster oscillation frequencies at decreasing amplitudes. Conclusions: The fibromuscular or thin nature of the membrane has a significant role on the severity of the pathology, with higher stiffnesses generally producing worse hydrodynamics. The orifice dimension and position are also important on the systolic performance and can determine potential structural degradation and haematic damage. (c) 2025 AGBM. Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Calcium phosphates are the predominant type of mineral in humans. They are essential for the formation of bones and teeth but can also deposit in diseased tissues. Whitlockite, a calcium phosphate mineral, is almost exclusively found in diseased calcified lesions. Despite its significance, the crystal growth mechanisms of whitlockite under physiological conditions are poorly understood. To investigate this, we conducted a constant composition experiment and applied a growth kinetics model to study the growth of whitlockite and hydroxyapatite nanoparticles in simulated body fluid. Contrary to our expectations, whitlockite did not dissolve in the undersaturated simulated body fluid. Instead, whitlockite nanoparticles served as a substrate for the formation of poorly-crystalline apatite, resulting in increased particle size. Our findings may help explain the colocalisation of whitlockite particles and large poorly-crystalline apatite lesions in human diseases.

ABSTRACTCardiovascular diseases (CVDs) were responsible for approximately 19 million deaths in 2020, marking an increase of 18.7% since 2010. Biological decellularized patches are common therapeutic solutions for CVD such as cardiac and valve defects. The preparation of biomaterials for cardiac patches involves two main processing methods: glutaraldehyde or photooxidation cross‐linking (fixation) and noncross‐linked (nonfixation) processing. Despite the variety of products available in the market, cardiac patches still suffer from significant limitations, failing to adequately mimic the properties of biological tissue and restore its function. This study assesses the impact of different processing methodologies on the biological and biomechanical outcomes of three commercially available cardiac patches (CorPatch, CardioCel, PhotoFix) and one newly developed decellularized cardiac patch (Adeka) when implanted as right ventricular outflow tract (RVOT) repair material on a rat model. Four different patches for cardiovascular repair were selected based on their processing approaches and included: photooxidation crosslinked (PhotoFix), glutaraldehyde crosslinked (CardioCel), noncross‐linked small intestine submucosa (CorPatch) or enzyme, and hydrostatic pressure (Adeka) processed decellularized biomaterials. Structure and function were characterized prior to implantation via thickness mapping, cross‐section morphology, 2D surface topography, 3D volume microstructure, biaxial testing, uniaxial tensile testing, ball burst, and suture retention. Their host–biomaterials response was assessed in vivo using a relevant model for cardiovascular repair: a rat (RVOT) reconstruction with 8 and 16‐week timepoints. Topological analysis showed that the crosslinked cardiac patches had a more homogeneous thickness distribution when compared to the noncrosslinked patches. This agreed with histological evaluation, where cross‐linking processed materials better preserved collagen content than noncrosslinked patches who were also more delaminated. Biaxial data demonstrated that all patches, except CorPatch, recapitulated the anisotropic behavior of healthy left ventricle tissue. The Adeka patch in‐plane mechanics at 16 weeks was the one who better resembled the mechanics of healthy cardiac tissue. All patches showed appropriate biocompatibility and function at 8‐ and 16‐week timepoints for RVOT patching. This included echocardiographic assessment, biomechanics, macrophage infiltration and polarization, and angiogenesis. Consistently with a more porous laminae structure, explants histology showed higher cell infiltration in non‐crosslinked Adeka when compared to the crosslinked PhotoFix. Overall, both in vitro and in vivo tests indicate that the material processing does not impact the function, biomechanical performance, and the host response of the patches that can be considered as equally effective as materials based cardiac repair solutions.

Balloon aortic valvuloplasty (BAV), a minimally invasive procedure to alleviate aortic valve stenosis, commonly employs rapid ventricular pacing (RVP) for balloon stabilization. However, the repeated and extended operation time associated with this technique poses potential complications. This letter introduces a novel approach to mitigate these concerns by employing a dilatation mechanism that is synchronized with the cardiac frequency, wherein the balloon catheter is fully inflated and deflated to a safe, low volume during the decrement of the ventricular pressure. The synchronized pacing was tested at a heart rate of 60 bpm. To experimentally validate the performance of this new approach, mock aortic roots reproducing different calcification patterns were used to compare the leaflets' mobility after the dilatation test with traditional BAV. Results confirm successful balloon pacing, maintaining low volume before the ventricular pressure increases. The dilatation performance assessment underscores that the proposed methodology resulted in a higher improvement in terms of the transvalvular pressure gradient and opening area. Optimal performance occurs at 60 bpm, yielding a 30.28% gradient decrease and a 21.35% opening area increase. This research represents a notable step forward toward the development of BAV devices capable of autonomous stabilization, eliminating the need for RVP and its related complications. Furthermore, the use of calcified aortic root (AR) phantoms contributes to an enhanced understanding of hemodynamic implications during BAV procedures.

Left atrial appendage (LAA) thrombosis is a frequent and serious complication associated with atrial fibrillation (AF), which significantly increases the risk of ischemic events. The presence of a thrombus within the LAA can disrupt normal blood flow and eventually embolise, leading to impaired circulation and severe complications. An innovative fluid-structure interaction model, based on the smoothed particle hydrodynamics method, is used to simulate the formation and behaviour of thrombi in a patient-specific LAA morphology under AF condition. The clotting process is modelled by tracking the concentrations of key components involved in the coagulation cascade. The approach transforms fluid particles into a solid phase by applying internal spring forces when specific biochemical and haemodynamic conditions occur. The findings shed light on the haemodynamic and biochemical mechanisms driving thrombus formation and migration within the LAA, highlighting regions of clot growth and subsequent embolisation. Thrombus formation begins at the LAA tip, gradually expanding through lobes and trabeculae, and ultimately progressing internally to high-recirculation regions, increasing the risk of embolisation and of potential clinical complications. By accurately predicting thrombus formation, growth and fragmentation, this study offers valuable insight into the mechanisms associated with the thromboembolic risk in AF. These are crucial for developing targeted therapeutic strategies to minimise the risk of thromboembolic events in patient with AF.

Aortic valve replacements, both surgical and transcatheter, are nowadays widely employed treatments. Although clinically effective, these procedures are correlated with potentially severe clinical complications which can be associated with the non-physiological haemodynamics that they establish. In this work, the fluid dynamics changes produced by surgical and transcatheter aortic valve replacements are analysed and compared with an ideal healthy native valve configuration, employing advanced fluid-structure interaction (FSI) simulations. The aim of the study is to investigate how existing treatments may affect the aortic valve function, and giving indications about how to improve current therapies.Simulations were performed using the commercial software LS-DYNA, where the FSI strategy is based on the coupling of a Lagrangian approach for the structures and a Eulerian approach for the fluid, whilst the coupling between the two domains is reached through a hybrid arbitrary-Lagrangian-Eulerian algorithm. Idealised geometries are used for the aortic root and leaflets. The aortic wall was modelled as linear elastic material, whilst leaflets were modelled as hyperelastic incompressible, using an Ogden's constitutive model. A combination of physiological flow velocity and pressure differences are applied as boundary conditions to model realistically the whole cardiac cycle.Results are analysed throughout the cardiac cycle in terms of leaflets kinematics, flow dynamics, pressure and valve performance parameters. Globally, surgical valves presented worse performance than transcatheter counterparts (reduced effective orifice area, increased transvalvular pressure drop and increased opening and closing times). The clinical parameters of transcatheter devices were improved and closer to those of the healthy native valve, although the vortical activity within the Valsalva's sinuses was substantially altered. Here, the presence of the partition obstructed the washing out, resulting in higher degree of blood stasis and potential blood damage.The implantation of prosthetic devices produces major haemodynamic changes which alters the valve dynamics and leads to diminished performance. Currently, the design of these substitutes is not optimised to mimic realistic native conditions, particularly in terms of valve opening behaviour. Although transcatheter devices provide systolic performance similar to that estimated for the healthy native aortic model, none of the prosthetic solutions appeared to be able to fully restore healthy physiological conditions.

Biological soft tissues and polymers used in biomedical applications (e.g. in the cardiovascular area) are hyperelastic incompressible materials that commonly operate under multi-axial large deformation fields. Their characterisation requires biaxial tensile testing. Due to the typically small sample size, the gripping of the specimens commonly relies on rakes or sutures, where the specimen is punctured at the edges of the gauge area. This approach necessitates of an apron, excess of material around the gauge region. This work analyses the apron influence on the estimated mechanical response of biaxial tests performed by using a rakes gripping system, with the aim of verifying the test accuracy and propose improved solutions. In order to isolate the effect of the apron, avoiding the influence of anisotropy and inhomogeneity typical of most soft tissues, homogeneous and isotropic hyperplastic samples made from a uniform sheet of casted silicone were tested. The stress-strain response of specimens with different apron sizes/shapes was measured experimentally by means of biaxial testing and digital image correlation. Tests were replicated numerically, to interpret the experimental findings. The apron surrounding the gauge area acts as an additional annular constraint which stiffens the system, resulting in a significant overestimate in the stress values. This error can be avoided by introducing specific cuts in the apron. The study quantifies, for the first time, the correlation between the apron size/shape and the experimental stress overestimation, proposing a research protocol which, although identified on homogeneous hyperelastic materials, can be useful in providing more accurate characterisation of both, synthetic polymers and soft tissues.

IntroductionStatistical shape analysis (SSA) with clustering is often used to objectively define and categorise anatomical shape variations. However, studies until now have often focused on simplified anatomical reconstructions, despite the complexity of studied anatomies. This work aims to provide insights on the anatomical detail preservation required for SSA of highly diverse and complex anatomies, with particular focus on the left atrial appendage (LAA). This anatomical region is clinically relevant as the location of almost all left atrial thrombi forming during atrial fibrillation (AF). Moreover, its highly patient-specific complex architecture makes its clinical classification especially subjective.MethodsPreliminary LAA meshes were automatically detected after robust image selection and wider left atrial segmentation. Following registration, four additional LAA mesh datasets were created as reductions of the preliminary dataset, with surface reconstruction based on reduced sample point densities. Utilising SSA model parameters determined to optimally represent the preliminary dataset, SSA model performance for the four simplified datasets was calculated. A representative simplified dataset was selected, and clustering analysis and performance were evaluated (compared to clinical labels) between the original trabeculated LAA anatomy and the representative simplification.ResultsAs expected, simplified anatomies have better SSA evaluation scores (compactness, specificity and generalisation), corresponding to simpler LAA shape representation. However, oversimplification of shapes may noticeably affect 3D model output due to differences in geometric correspondence. Furthermore, even minor simplification may affect LAA shape clustering, where the adjusted mutual information (AMI) score of the clustered trabeculated dataset was 0.67, in comparison to 0.12 for the simplified dataset.DiscussionThis study suggests that greater anatomical preservation for complex and diverse LAA morphologies, currently neglected, may be more useful for shape categorisation via clustering analyses.

BACKGROUND:Complications of atrial fibrillation (AF) include ischemic events originating within the left atrial appendage (LAA), a protrusion of the left atrium with variable morphological characteristics. The role of the patient specific morphology and pathological haemodynamics on the risk of ischemia remains unclear.METHODS:This work performs a comparative assessment of the hemodynamic parameters among patient-specific LAA morphologies through fluid-structure interaction computational analyses. Three LAA models per each of the four commons patient-specific morphological families (chicken wing, cactus, windsock, and cauliflower) were analysed. Mechanical properties of the tissue were based on experimental uniaxial tests on a young pig's heart. Boundary conditions were imposed based on clinical assessments of filling and emptying volumes. Sinus rhythm and atrial fibrillation operative conditions were simulated and analysed.RESULTS:For each model, the effect of morphological and functional parameters, such as the number of trabeculae and LAA stroke volume, over the hemodynamics established into the appendage was analysed. Comparison between results obtained in healthy and diseased conditions suggested the introduction of a new parameter to quantify the risk of thrombosis, here called blood stasis factor (BSF). This is defined as the LAA surface area which permanently experiences levels of shear strain rate inferior to a threshold value, set to 5 s-1 (BSF5).CONCLUSIONS:This work suggests that the current morphological classification is unsuitable to evaluate the probability of thrombus formation. However, hemodynamic parameters easy to determine from clinical examinations, such as normalised stroke volume, LAA orifice flow rate and presence of extensive trabeculations can identify departures from healthy hemodynamics in AF and support a more systematic stratification of the thromboembolic risk.