It is widely accepted in the cardiovascular field that the presence of turbulent flow, particularly in the region of the carotid bifurcation, is a major factor in the development of atherosclerotic plaques. The presence of such plaques results in a progressive reduction of the vascular lumen, increasing the likelihood of cerebral ischemic events. The aim of this thesis is to analyze the carotid bifurcation in a sample of 129 patients using DICOM format images obtained by computed tomography (CT). The main objective is to perform a radiomic analysis aimed at developing a machine learning algorithm for the classification of carotid plaques into symptomatic and asymptomatic. This classification is intended to provide decision support for the early identification of patients who are candidates for preventive surgery to reduce the risk of ischemic stroke. To complement the radiomic analysis, computational fluid dynamics (CFD) simulations and geometric analyses of the vessel lumen were performed for each subject. These additional studies were designed to identify possible correlations between the geometric and fluid dynamic characteristics of the vessel and plaque vulnerability, understood as the propensity for rupture and subsequent occurrence of ischemic events. The results obtained highlight the potential of an integrated approach based on radiomic analysis, CFD and morphological study of the vessel, although they highlight some methodological limitations. In light of these considerations, it is considered desirable to further investigate the different methodologies used in order to refine the accuracy and reliability of the results.
In ambito cardiovascolare, è ampiamente riconosciuto che la presenza di flusso turbolento, in particolare nella regione della biforcazione carotidea, rappresenta un fattore determinante per l’insorgenza di placche aterosclerotiche. La presenza di tali placche comporta una progressiva riduzione del lume vascolare e un incremento del rischio di eventi ischemici cerebrali. Il presente elaborato di tesi si propone di analizzare la biforcazione carotidea in un campione di 129 pazienti, utilizzando immagini in formato DICOM acquisite tramite tomografia computerizzata (TC). L’obiettivo principale consiste nell’ effettuare un’analisi radiomica finalizzata allo sviluppo di un algoritmo di apprendimento automatico per la classificazione delle placche carotidee in sintomatiche e asintomatiche. Tale classificazione si prefigge di fornire un supporto decisionale utile per l’identificazione precoce di pazienti che potrebbero beneficiare di interventi chirurgici preventivi al fine di ridurre il rischio di ictus ischemico. A completamento dell’analisi radiomica, sono state condotte simulazioni di fluidodinamica computazionale (CFD) e analisi geometriche del lume vascolare per ciascun paziente. Queste ulteriori indagini sono volte ad individuare possibili correlazioni tra le caratteristiche geometriche e fluidodinamiche del vaso e la vulnerabilità delle placche, intesa come propensione alla rottura e alla conseguente insorgenza di eventi ischemici. I risultati ottenuti evidenziano le potenzialità di un approccio integrato basato su analisi radiomica, simulazioni di fluidodinamica computazionale (CFD) e valutazioni morfologiche del vaso. mettendo tuttavia in luce alcune limitazioni metodologiche. Alla luce di tali considerazioni, in prospettiva futura, si ritiene auspicabile un approfondimento delle diverse metodologie impiegate per affinare l’accuratezza e l’affidabilità dei risultati.
Identificazione di placche carotidee vulnerabili mediante analisi radiomica, computazionale e geometrica
ACRI, ALBERTO
2025
Abstract
It is widely accepted in the cardiovascular field that the presence of turbulent flow, particularly in the region of the carotid bifurcation, is a major factor in the development of atherosclerotic plaques. The presence of such plaques results in a progressive reduction of the vascular lumen, increasing the likelihood of cerebral ischemic events. The aim of this thesis is to analyze the carotid bifurcation in a sample of 129 patients using DICOM format images obtained by computed tomography (CT). The main objective is to perform a radiomic analysis aimed at developing a machine learning algorithm for the classification of carotid plaques into symptomatic and asymptomatic. This classification is intended to provide decision support for the early identification of patients who are candidates for preventive surgery to reduce the risk of ischemic stroke. To complement the radiomic analysis, computational fluid dynamics (CFD) simulations and geometric analyses of the vessel lumen were performed for each subject. These additional studies were designed to identify possible correlations between the geometric and fluid dynamic characteristics of the vessel and plaque vulnerability, understood as the propensity for rupture and subsequent occurrence of ischemic events. The results obtained highlight the potential of an integrated approach based on radiomic analysis, CFD and morphological study of the vessel, although they highlight some methodological limitations. In light of these considerations, it is considered desirable to further investigate the different methodologies used in order to refine the accuracy and reliability of the results.File | Dimensione | Formato | |
---|---|---|---|
TESI DOTTORATO -Alberto ACRI.pdf
accesso aperto
Dimensione
6.14 MB
Formato
Adobe PDF
|
6.14 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/210364
URN:NBN:IT:UNIME-210364