Mechanical Thrombectomy (MT) is a promising therapy for the treatment of cerebral stroke. MT consists of a minimally invasive interventional procedure which allows the removal of a blood clot (thrombus) obstructing intracranial vessels. MT aims at restoring physiological blood flow by using a device to remove the thrombus blocking the artery. The most innovative thrombectomy devices are the stent retrievers. These self-expandable devices are deployed within a micro-catheter and positioned across the blood clot; once the thrombus is embedded within the device, it can be retrieved into the guide catheter and safely removed from the brain. Several clinical trials and in vitro studies have been performed to prove their effectiveness, safety and superiority against other techniques, such as intra-arterial and intra-venous recombinant tissue plasminogen activator (rt-PA); however, lack of in silico models can easily be assessed in the literature and the aim of this work is the development of reliable in silico models to simulate the thrombectomy procedure. Three different stent retrievers available on the market are considered: EmboTrap Revascularization Device (Cerenovus), Trevo XP ProVue Stentriever (Stryker Neurovascular) and Solitaire Revascularization Device (Medtronic). The first one is developed starting from the CAD model provided by the Cerenovus Company (Galway, Ireland) whereas the latter two are computationally generated implementing a dedicated MATLAB (MathWorks, Natick, Massachusetts) code. The devices are then modelled with one-dimensional beam finite elements in the ANSA pre-processor (BETA CAE Systems). Experimental tests of uniaxial traction are performed and replicated with computational simulations to identify the material model parameters; the beam equivalent section is evaluated by means of confocal laser scanning microscopy (CLSM) analyses. Simulations to evaluate the interaction between the devices and the clot are carried out. These latter are performed with an idealized cylindrical vessel and with clots characterized by different length, diameter, and percentage of red blood cells (RBCs) composition. In this way, the influence of these parameters and the effects of their variation on the outcome of the thrombectomy procedure can be assessed. All the simulations are run in the finite elements solver LS-DYNA (Livermore Software Technology Corporation, Livermore, USA). The described procedure is computationally reproduced as truthfully as possible like the real clinical procedure in order to analyse and compare the behaviour of the three modelled stents. The capability of the device to retrieve the clot is assessed as a function of different parameters such as damping factors, friction coefficients, composition, length and diameter of the clot and diameter of the vessel. Stresses and strains of both the clot and the devices are evaluated by means of the META post-processor (BETA CAE Systems). Based on these analyses, the optimal numerical model of each stent is highlighted, and considerations are carried out. The computational models are reliable, and they accurately reproduce the real devices in terms of shape, geometry, sizes, beam equivalent section and material properties. In particular, the mechanical behaviour of the computational models during the uniaxial traction test is strongly influenced by the variation of the equivalent section of the beam elements whereas variations of the Austenite Young Modulus do no significatively alter it. The crimping and release simulations allow to evaluate the most suitable damping factor for each device, which should be maintained as low as possible in the computational simulations in order not to alter their kinematic. The damping factor influences the crimp and release simulations in terms of kinematic system damping energy. In detail, the optimal damping factor results to be equal to 0.1 for the Trevo and Solitaire models and to 0.05 for the EmboTrap stent retriever. Regarding the simulations of the thrombectomy procedure with the clot, these are influenced by several parameter: • The increase and the decrease of the percentage of RBCs composition influence the stresses and strains the clot is subjected to; as percentage increases, the deformation increases too. However, it does not affect the outcome of the procedure. • As the friction coefficients increase, the stresses and strains on the clot increase too. • As the dimension of the vessel decreases, the stresses and strains on the device increase. • The length of the clot influences only the simulation with the Trevo model. As it decreases, the clot is able to enter inside the strut of the device and slide away. All simulations are successful in terms of effective capture and retrieval of the clot with exception of the simulation with the Trevo device and the shortest clot. The other trials do not fail, although there are cases in which a delay in the retrieval of the clot occurs. This is mainly associated to a high friction coefficient between the clot and the vessel and between the clot and the stent. In conclusion, this thesis demonstrates the dependence of the computational thrombectomy procedure on several parameters and therefore the potential impact of these parameters on the procedure outcome.
La Trombectomia Meccanica (MT) è una promettente terapia per il trattamento degli ictus cerebrali. Essa consiste in una procedura di intervento mininvasiva che ha l’obiettivo di ripristinare un flusso sanguigno fisiologico per mezzo di un dispositivo che consente la rimozione di coaguli di sangue che ostruiscono i vasi intracranici. Diverse soluzioni sono state proposte e tra queste gli stent retrievers sono i dispositivi più promettenti. Questi dispositivi auto-espandibili vengono posizionati in prossimità del coagulo grazie all’ausilio di un micro-catetere; una volta che il trombo è intrappolato nel dispositivo questo viene condotto nel catetere guida e rimosso in sicurezza dal cervello. Diversi studi in vivo e in vitro sono stati realizzati con il fine di dimostrare la loro efficacia, sicurezza, e superiorità rispetto ad altre tecniche proposte, quali la somministrazione intra-arteriosa e l’intra-venosa dell’attivatore del plasminogeno tissutale ricombinante (rt-PA). Tuttavia, analizzando la letteratura ciò che si evince è la mancanza di modelli in silico, ovvero di modelli computazionali. Lo scopo di questo lavoro è quello di sviluppare modelli in silico che siano in grado di simulare affidabilmente la procedura di trombectomia. Tre differenti stent retrievers disponibili sul mercato sono stati inclusi in questo progetto: EmboTrap Revascularization Device (Cerenovous), Trevo XP ProVue Stentriever (Stryker Neurovascular) e il Solitaire Revascularization Device (Medtronic). Il primo di questi è ricavato a partire dal modello CAD fornito dalla compagnia Cerenovus (Galway, Irlanda) mentre gli altri due, in assenza del modello CAD, sono generati implementando un apposito codice Matlab (MathWorks, Natick, Massachusetts). I dispositivi così ottenuti sono discretizzati con beam monodimensionali ad elementi finiti nel preprocessore ANSA (BETA CAE Systems). In una prima fase, test di trazione uniassiale sperimentali sono eseguiti e replicati con simulazioni computazionali per identificare i parametri del modello; la sezione equivalente dei beams viene valutata per mezzo di un’analisi tramite microscopia confocale a scansione laser (CLSM). Successivamente, diverse simulazioni per valutare l’interazione tra i dispositivi e il coagulo vengono effettuate. Queste ultime sono realizzate con un vaso cilindrico idealizzato e con coaguli caratterizzati da differente lunghezza, diametro, e percentuale di globuli rossi (RBCs). Così facendo, l’influenza di questi parametri e gli effetti della loro variazione sulla procedura di trombectomia possono essere stimati. Tutte le simulazioni sono state eseguite per mezzo del risolutore ad elementi finiti LS-DYNA (Livermore Software Technology Corporation, Livermore, USA). La procedura descritta viene riprodotta a livello computazionale in modo più verosimile possibile alla reale procedura clinica per analizzare e confrontare il comportamento dei tre modelli di stent. La capacità del dispositivo di catturare il coagulo e di consentirne la rimozione dal vaso è valutata in funzione di diversi parametri quali il fattore di damping, i coefficienti di frizione, la composizione, la lunghezza e il diametro del coagulo e il diametro del vaso. Sforzi e deformazioni sia sul coagulo che sul dispositivo sono valutati per mezzo del post-processore META (BETA CAE Systems). Basandosi su queste analisi, il modello numerico ottimale per ogni stent viene identificato con le opportune e dovute considerazioni. I modelli computazionali sono affidabili e riproducono accuratamente i dispositivi in termini di forma, geometria, dimensioni, sezione equivalente dei beams e proprietà del materiale. In particolare, il comportamento meccanico dei modelli computazionali durante il test di trazione uniassiale è fortemente influenzato dalla variazione della sezione equivalente degli elementi monodimensionali, mentre le variazioni del Modulo di Young dell’Austenite non lo alterano in modo significativo. Le simulazioni di crimp e release consentono di valutare il fattore di damping più adatto per ciascun dispositivo; il suo valore dovrebbe essere mantenuto il più basso possibile nelle simulazioni computazionali al fine di non alterarne la cinematica. In generale, il fattore di damping influenza le simulazioni di crimp e release in termini di cinematica e system damping energy. Il valore di damping ottimale è pari a 0.1 per i modelli computazionali di Trevo e Solitaire e pari a 0.05 nel caso dell’EmboTrap. Per quanto riguarda le simulazioni della procedura di trombectomia con il coagulo, queste sono influenzate da diversi parametri: • L’aumento e la diminuzione della percentuale della composizione dei globuli rossi influenzano le sollecitazioni e le deformazioni a cui è soggetto il coagulo, all’aumentare di tale percentuale anche la deformazione aumenta. Tuttavia, non influisce sull'esito della procedura. • All'aumentare dei coefficienti di attrito, aumentano anche le sollecitazioni e le deformazioni sul coagulo. • Al diminuire delle dimensioni del vaso, aumentano le sollecitazioni e le deformazioni sul dispositivo. • La lunghezza del coagulo influenza solo la simulazione con il modello del Trevo. Una diminuzione della lunghezza del coagulo consente a questo di entrare all'interno delle maglie dello stent e di sfuggire dal dispositivo. Tutte le simulazioni hanno esito positivo in termini di cattura e rimozione del coagulo, ad eccezione della prova con il modello del Trevo e con il coagulo di lunghezza minore. Le altre simulazioni non sono fallite, sebbene in alcuni casi si verifica un ritardo nella cattura del coagulo. Ciò è principalmente associato a un elevato coefficiente di attrito tra il coagulo e il vaso e tra il coagulo e lo stent. In conclusione, questa tesi dimostra la dipendenza della procedura di trombectomia meccanica da diversi parametri e quindi l'impatto di questi stessi sugli esiti delle simulazioni.
Computational models of stent retrievers for the mechanical thrombectomy procedure
VILLA, ANITA;PORTA, CAROLINA
2018/2019
Abstract
Mechanical Thrombectomy (MT) is a promising therapy for the treatment of cerebral stroke. MT consists of a minimally invasive interventional procedure which allows the removal of a blood clot (thrombus) obstructing intracranial vessels. MT aims at restoring physiological blood flow by using a device to remove the thrombus blocking the artery. The most innovative thrombectomy devices are the stent retrievers. These self-expandable devices are deployed within a micro-catheter and positioned across the blood clot; once the thrombus is embedded within the device, it can be retrieved into the guide catheter and safely removed from the brain. Several clinical trials and in vitro studies have been performed to prove their effectiveness, safety and superiority against other techniques, such as intra-arterial and intra-venous recombinant tissue plasminogen activator (rt-PA); however, lack of in silico models can easily be assessed in the literature and the aim of this work is the development of reliable in silico models to simulate the thrombectomy procedure. Three different stent retrievers available on the market are considered: EmboTrap Revascularization Device (Cerenovus), Trevo XP ProVue Stentriever (Stryker Neurovascular) and Solitaire Revascularization Device (Medtronic). The first one is developed starting from the CAD model provided by the Cerenovus Company (Galway, Ireland) whereas the latter two are computationally generated implementing a dedicated MATLAB (MathWorks, Natick, Massachusetts) code. The devices are then modelled with one-dimensional beam finite elements in the ANSA pre-processor (BETA CAE Systems). Experimental tests of uniaxial traction are performed and replicated with computational simulations to identify the material model parameters; the beam equivalent section is evaluated by means of confocal laser scanning microscopy (CLSM) analyses. Simulations to evaluate the interaction between the devices and the clot are carried out. These latter are performed with an idealized cylindrical vessel and with clots characterized by different length, diameter, and percentage of red blood cells (RBCs) composition. In this way, the influence of these parameters and the effects of their variation on the outcome of the thrombectomy procedure can be assessed. All the simulations are run in the finite elements solver LS-DYNA (Livermore Software Technology Corporation, Livermore, USA). The described procedure is computationally reproduced as truthfully as possible like the real clinical procedure in order to analyse and compare the behaviour of the three modelled stents. The capability of the device to retrieve the clot is assessed as a function of different parameters such as damping factors, friction coefficients, composition, length and diameter of the clot and diameter of the vessel. Stresses and strains of both the clot and the devices are evaluated by means of the META post-processor (BETA CAE Systems). Based on these analyses, the optimal numerical model of each stent is highlighted, and considerations are carried out. The computational models are reliable, and they accurately reproduce the real devices in terms of shape, geometry, sizes, beam equivalent section and material properties. In particular, the mechanical behaviour of the computational models during the uniaxial traction test is strongly influenced by the variation of the equivalent section of the beam elements whereas variations of the Austenite Young Modulus do no significatively alter it. The crimping and release simulations allow to evaluate the most suitable damping factor for each device, which should be maintained as low as possible in the computational simulations in order not to alter their kinematic. The damping factor influences the crimp and release simulations in terms of kinematic system damping energy. In detail, the optimal damping factor results to be equal to 0.1 for the Trevo and Solitaire models and to 0.05 for the EmboTrap stent retriever. Regarding the simulations of the thrombectomy procedure with the clot, these are influenced by several parameter: • The increase and the decrease of the percentage of RBCs composition influence the stresses and strains the clot is subjected to; as percentage increases, the deformation increases too. However, it does not affect the outcome of the procedure. • As the friction coefficients increase, the stresses and strains on the clot increase too. • As the dimension of the vessel decreases, the stresses and strains on the device increase. • The length of the clot influences only the simulation with the Trevo model. As it decreases, the clot is able to enter inside the strut of the device and slide away. All simulations are successful in terms of effective capture and retrieval of the clot with exception of the simulation with the Trevo device and the shortest clot. The other trials do not fail, although there are cases in which a delay in the retrieval of the clot occurs. This is mainly associated to a high friction coefficient between the clot and the vessel and between the clot and the stent. In conclusion, this thesis demonstrates the dependence of the computational thrombectomy procedure on several parameters and therefore the potential impact of these parameters on the procedure outcome.File | Dimensione | Formato | |
---|---|---|---|
TESI.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
5.39 MB
Formato
Adobe PDF
|
5.39 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/164557