Obesity represents one of the most significant global health challenges, with a profound impact on quality of life and the incidence of numerous conditions, such as type 2 diabetes, hypertension, and cardiovascular diseases. Over recent decades, the rising prevalence of obesity has driven the medical community to develop innovative solutions for managing this condition, particularly for patients who do not respond effectively to conventional therapies. Bariatric surgery is now considered one of the most effective approaches for achieving sustainable weight reduction in patients with severe obesity. Among various techniques, Laparoscopic Sleeve Gastrectomy (LSG) and Endoscopic Sleeve Gastrectomy (ESG) have shown promising results, while presenting distinct characteristics and indications. In this context, selecting the optimal surgical technique for each patient is a matter of considerable clinical importance. Biomedical image segmentation, facilitated by advanced software, can aid in creating a groundtruth database useful for improving patient selection and refining intervention protocols. This research aims to segment the gastric region of patients who are candidates for bariatric surgery using advanced image processing techniques, with the objective of studying preoperative gastric volumes and exploring differences among patients potentially more suited to LSG or ESG. The creation of a groundtruth database of annotated images can be utilized in the future to train neural networks, contributing to the development of artificial intelligence models for the automatic recognition of gastric features. Additionally, statistical tests will be conducted to evaluate the impact of variables such as age, BMI, and the presence of comorbidities on gastric volume, to determine whether these factors may influence stomach capacity and, consequently, the effectiveness of the surgical techniques employed.
L’obesità rappresenta una delle maggiori sfide sanitarie a livello globale, con un impatto profondo sulla qualità della vita e sull’incidenza di numerose patologie, quali diabete di tipo 2, ipertensione e malattie cardiovascolari. Negli ultimi decenni, la crescente prevalenza dell’obesità ha portato la comunità medica a sviluppare soluzioni innovative per il trattamento di questa condizione, in particolare per i pazienti che non rispondono efficacemente alle terapie convenzionali. La chirurgia bariatrica è oggi considerata uno degli approcci più efficaci per ottenere una riduzione sostenibile del peso in pazienti affetti da obesità grave. Tra le varie tecniche, la Laparoscopic Sleeve Gastrectomy (LSG) e l'Endoscopic Sleeve Gastrectomy (ESG) hanno dimostrato risultati promettenti, pur presentando caratteristiche e indicazioni differenti. In questo contesto, la scelta della tecnica chirurgica ottimale per ciascun paziente è un fattore di rilevante importanza clinica. La segmentazione delle immagini biomediche, tramite l’uso di software avanzati, può contribuire a sviluppare un database groundtruth utile a migliorare la selezione dei pazienti e ad affinare i protocolli di intervento. La ricerca si propone di segmentare il distretto gastrico dei pazienti candidati a intervento bariatrico, utilizzando tecniche avanzate di elaborazione delle immagini, con l’obiettivo di studiare i volumi gastrici pre-operatori e di esplorare le differenze tra pazienti potenzialmente più idonei a LSG o ESG. La creazione di un database groundtruth di immagini annotate potrà essere utilizzato in futuro per l’addestramento delle reti neurali, contribuendo a sviluppare modelli di intelligenza artificiale per il riconoscimento automatico delle caratteristiche gastriche. Parallelamente, verranno eseguiti test statistici per valutare l’effetto di variabili come l’età, il BMI e la presenza di patologie concomitanti sul volume gastrico, al fine di comprendere se questi fattori possano influire sulla capacità dello stomaco e, di conseguenza, sull’efficacia delle tecniche chirurgiche utilizzate.
Segmentazione del distretto gastrico da immagini biomediche di pazienti bariatrici prima di intervento chirurgico per lo sviluppo di un database groundtruth
GIRLANDA, RICCARDO
2023/2024
Abstract
Obesity represents one of the most significant global health challenges, with a profound impact on quality of life and the incidence of numerous conditions, such as type 2 diabetes, hypertension, and cardiovascular diseases. Over recent decades, the rising prevalence of obesity has driven the medical community to develop innovative solutions for managing this condition, particularly for patients who do not respond effectively to conventional therapies. Bariatric surgery is now considered one of the most effective approaches for achieving sustainable weight reduction in patients with severe obesity. Among various techniques, Laparoscopic Sleeve Gastrectomy (LSG) and Endoscopic Sleeve Gastrectomy (ESG) have shown promising results, while presenting distinct characteristics and indications. In this context, selecting the optimal surgical technique for each patient is a matter of considerable clinical importance. Biomedical image segmentation, facilitated by advanced software, can aid in creating a groundtruth database useful for improving patient selection and refining intervention protocols. This research aims to segment the gastric region of patients who are candidates for bariatric surgery using advanced image processing techniques, with the objective of studying preoperative gastric volumes and exploring differences among patients potentially more suited to LSG or ESG. The creation of a groundtruth database of annotated images can be utilized in the future to train neural networks, contributing to the development of artificial intelligence models for the automatic recognition of gastric features. Additionally, statistical tests will be conducted to evaluate the impact of variables such as age, BMI, and the presence of comorbidities on gastric volume, to determine whether these factors may influence stomach capacity and, consequently, the effectiveness of the surgical techniques employed.File | Dimensione | Formato | |
---|---|---|---|
Girlanda_Riccardo.pdf
embargo fino al 19/11/2027
Dimensione
9.29 MB
Formato
Adobe PDF
|
9.29 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/20.500.12608/76480