Computational modelling applied to the biomedical field constitutes an innovative and fundamental support tool for various clinical decisions. In recent history, in silico medicine has proven to be very useful in multiple applications, such as prevention, diagnosis, prescription of therapies, and simulations of surgical procedures. Despite the huge number of benefits associated with the use of computational models, they inherently lead to risks due to approximations and the propagation of uncertainties on the parameters used as inputs in the model. The assessment of these risks and the study of the influence that a given input can have on the output are called Uncertainty Quantification (UQ) and Sensitivity Analysis (SA), and they constitute a fundamental step in studying the reliability of simulations. The TEVAR procedure is an endovascular repair technique applied in contexts related to various vascular diseases, particularly of the aorta. It entails several advantages and better results compared to the standard open repair approach. Computational simulations of this procedure have led to efficient predictions of possible complications and constitute an excellent decision-making tool. The objective of this thesis is to conduct a UQ and SA study of two different inputs through the creation of a sample space of TEVAR procedure simulations using a Design Of Experiments approach, in which the thickness and Young modulus of the aortic wall were chosen as factors affected by uncertainty. A patient-specific simulation of a TEVAR procedure for the repair of an aortic arch aneurysm was repeated 121 times, each time pairing two different values of thickness and elastic modulus. In the post-processing phase, the focus was the analysis of the variability related to outputs considered particularly significant because they are strongly influenced by the stiffness of the system given by the coupling of the two input parameters. In particular, data concerning the stress on the stent, the stress on the aortic wall, and the distance between the graft and the vessel were extracted. The study revealed a different influence of the two parameters on the behaviour of the vessel and the interaction between the device and the aorta. Consequently, depending on the purpose of the simulation, uncertainties about the thickness or Young modulus are acceptable or not. In general, this analysis allows to quantify the noise related to the numerical resolution of the problem, which is inherent to the model. This leads to more reliable and predictable results.
La modellizzazione computazionale applicata all’ambito biomedico costituisce un innovativo e fondamentale strumento di supporto per diverse decisioni cliniche. Nella storia recente, la medicina in silico si è dimostrata molto utile in molteplici applicazioni, come ad esempio la prevenzione, la diagnosi, la prescrizione di terapie e le simulazioni di interventi chirurgici. Nonostante i numerosi benefici legati all’impiego dei modelli computazionali, essi comportano intrinsecamente dei rischi associati alle approssimazioni e alla propagazione delle incertezze sui parametri inseriti nel modello. La valutazione di questi rischi e lo studio dell’influenza che un determinato input può avere sull’output, prendono il nome di Uncertainty Quantification (UQ) e Sensitivity Analysis (SA), e costituiscono un passaggio fondamentale nello studio dell’affidabilità delle simulazioni. La procedura TEVAR è una tecnica di riparazione endovascolare applicata in contesti legati a più patologie vascolari, in particolare dell’aorta. Essa comporta svariati vantaggi e migliori risultati rispetto alla riparazione standard a cielo aperto. Le simulazioni computazionali di questa procedura hanno portato ad un’efficiente previsione delle possibili complicazioni e costituiscono un ottimo strumento decisionale. L’obiettivo di questa tesi è condurre uno studio di UQ e SA di due diversi input attraverso la creazione di uno spazio campionario di simulazioni di procedura TEVAR adottando un approccio Design Of Experiments, in base al quale sono stati scelti come fattori affetti da incertezza lo spessore e il modulo di Young della parete aortica. Una simulazione paziente-specifica di una procedura TEVAR atta alla riparazione di un aneurisma dell’arco aortico è stata ripetuta 121 volte accoppiando in ogni occasione due diversi valori di spessore e modulo elastico. Nel post-processing ci si è concentrati sull’analisi della variabilità di output considerati particolarmente significativi in quanto fortemente influenzabili dalla rigidezza del sistema data dall’accoppiamento dei due parametri di input. In particolare, sono stati estratti dati riguardanti lo sforzo nello stent, lo sforzo nella parete aortica e la distanza tra graft e vaso. Quanto emerso dallo studio è una diversa influenza dei due parametri sul comportamento del vaso e sull’interazione tra dispositivo e aorta. Di conseguenza, in base allo scopo della simulazione, un’incertezza sullo spessore o sul modulo di Young risulta accettabile o meno. In generale, questa analisi permette di quantificare il rumore intrinseco nella risoluzione numerica del modello stesso, portando a risultati più affidabili e prevedibili.
Uncertainty quantification in simulazioni di procedura TEVAR
Velcich, Vittoria
2023/2024
Abstract
Computational modelling applied to the biomedical field constitutes an innovative and fundamental support tool for various clinical decisions. In recent history, in silico medicine has proven to be very useful in multiple applications, such as prevention, diagnosis, prescription of therapies, and simulations of surgical procedures. Despite the huge number of benefits associated with the use of computational models, they inherently lead to risks due to approximations and the propagation of uncertainties on the parameters used as inputs in the model. The assessment of these risks and the study of the influence that a given input can have on the output are called Uncertainty Quantification (UQ) and Sensitivity Analysis (SA), and they constitute a fundamental step in studying the reliability of simulations. The TEVAR procedure is an endovascular repair technique applied in contexts related to various vascular diseases, particularly of the aorta. It entails several advantages and better results compared to the standard open repair approach. Computational simulations of this procedure have led to efficient predictions of possible complications and constitute an excellent decision-making tool. The objective of this thesis is to conduct a UQ and SA study of two different inputs through the creation of a sample space of TEVAR procedure simulations using a Design Of Experiments approach, in which the thickness and Young modulus of the aortic wall were chosen as factors affected by uncertainty. A patient-specific simulation of a TEVAR procedure for the repair of an aortic arch aneurysm was repeated 121 times, each time pairing two different values of thickness and elastic modulus. In the post-processing phase, the focus was the analysis of the variability related to outputs considered particularly significant because they are strongly influenced by the stiffness of the system given by the coupling of the two input parameters. In particular, data concerning the stress on the stent, the stress on the aortic wall, and the distance between the graft and the vessel were extracted. The study revealed a different influence of the two parameters on the behaviour of the vessel and the interaction between the device and the aorta. Consequently, depending on the purpose of the simulation, uncertainties about the thickness or Young modulus are acceptable or not. In general, this analysis allows to quantify the noise related to the numerical resolution of the problem, which is inherent to the model. This leads to more reliable and predictable results.File | Dimensione | Formato | |
---|---|---|---|
2024_07_Velcich_Tesi.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
3.32 MB
Formato
Adobe PDF
|
3.32 MB | Adobe PDF | Visualizza/Apri |
2024_07_Velcich_ExecutiveSummary.pdf
non accessibile
Descrizione: Executive Summary
Dimensione
1.11 MB
Formato
Adobe PDF
|
1.11 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/10589/223385