Obesity represents one of the major health challenges of the 21st century, with a growing impact on public health, healthcare systems, and patients’ quality of life. From both an anatomical and functional perspective, the small intestine plays a crucial role in digestion and absorption processes, and its length can influence the effectiveness of therapeutic strategies, particularly in bariatric surgery. Traditional techniques, such as intraoperative or post-mortem measurement, are invasive, error-prone, and difficult to apply on a large scale due to their high time consumption. Therefore, there is a need to identify alternative measurement methods capable of increasing acquisition speed while improving the efficiency of metabolic/bariatric bypass surgery. The introduction of a preoperative, non-invasive, accurate, and effective method for the assessment of TSBL would allow for the analysis of interindividual variability, a factor that can significantly influence the choice of surgical strategy. In this context, cross-sectional imaging techniques, such as computed tomography (CT) and magnetic resonance imaging (MRI), may play a key role by enabling precise measurements in a non-invasive way. The aim of this thesis is to develop and validate an automated approach for the segmentation and estimation of small intestine length from MRI and CT images in obese patients, using a neural network model. The goal is to evaluate the accuracy of the two imaging modalities, comparing results from manual measurement, MRI and CT, and the automated assessment provided by the neural network based on bioimaging.
L’obesità rappresenta una delle principali sfide sanitarie del XXI secolo, con un impatto crescente sulla salute pubblica, sui sistemi sanitari e sulla qualità della vita dei pazienti. A livello anatomico e funzionale, l’intestino tenue gioca un ruolo cruciale nei processi di digestione e assorbimento, e la sua lunghezza può influenzare l’efficacia di strategie terapeutiche, in particolare nella chirurgia bariatrica. Le tecniche tradizionali, come la misurazione intraoperatoria o post-mortem, sono invasive, soggette a errori e poco applicabili su larga scala perché molto dispendiose in termini di tempo. Si rende quindi necessaria l’identificazione di metodiche di misurazione alternative, capaci di aumentare la velocità di acquisizione della misura migliorando l’efficienza dell’intervento chirurgico di bypass metabolico/bariatrico. L’introduzione di una metodica preoperatoria, non invasiva, accurata ed efficace per la valutazione della TSBL consentirebbe di analizzarne la variabilità interindividuale, elemento che può influenzare significativamente la scelta della strategia chirurgica. In questo contesto, le tecniche di imaging sezionale, come la tomografia computerizzata (TC) e la risonanza magnetica (RM), potrebbero rivestire un ruolo chiave, offrendo la possibilità di effettuare misurazioni precise in modo non invasivo. L’obiettivo di questa tesi è sviluppare e validare un approccio automatizzato per la segmentazione e la stima della lunghezza dell’intestino tenue a partire da immagini di RM e TC in pazienti obesi, utilizzando un modello di rete neurale. L’intento è valutare l’accuratezza delle due modalità di imaging, confrontarne i risultati tra la misurazione manuale, quella da RM e TC e quella automatizzata dalla rete neurale basata sulle bioimmagini.
Valutazione automatica della lunghezza dell’intestino tenue mediante rete neurale convoluzionale su immagini di RM e TC
SCHIOCHETTO, FILIPPO
2024/2025
Abstract
Obesity represents one of the major health challenges of the 21st century, with a growing impact on public health, healthcare systems, and patients’ quality of life. From both an anatomical and functional perspective, the small intestine plays a crucial role in digestion and absorption processes, and its length can influence the effectiveness of therapeutic strategies, particularly in bariatric surgery. Traditional techniques, such as intraoperative or post-mortem measurement, are invasive, error-prone, and difficult to apply on a large scale due to their high time consumption. Therefore, there is a need to identify alternative measurement methods capable of increasing acquisition speed while improving the efficiency of metabolic/bariatric bypass surgery. The introduction of a preoperative, non-invasive, accurate, and effective method for the assessment of TSBL would allow for the analysis of interindividual variability, a factor that can significantly influence the choice of surgical strategy. In this context, cross-sectional imaging techniques, such as computed tomography (CT) and magnetic resonance imaging (MRI), may play a key role by enabling precise measurements in a non-invasive way. The aim of this thesis is to develop and validate an automated approach for the segmentation and estimation of small intestine length from MRI and CT images in obese patients, using a neural network model. The goal is to evaluate the accuracy of the two imaging modalities, comparing results from manual measurement, MRI and CT, and the automated assessment provided by the neural network based on bioimaging.| File | Dimensione | Formato | |
|---|---|---|---|
|
Schiochetto_Filippo.pdf
Accesso riservato
Dimensione
7.19 MB
Formato
Adobe PDF
|
7.19 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/99268