Diagnosing different forms of dementia is one of the most complex and pressing challenges in modern medicine. With increasing life expectancy, the number of people suffering from dementia is constantly growing. These diseases not only severely impair patients' quality of life, but also place a huge burden on healthcare systems and patients' families. Traditionally, the diagnosis of dementia has been based on clinical assessment, neuropsychological testing and brain imaging using techniques such as MRI and positron emission tomography. However, the diagnostic accuracy of these methods is not always satisfactory, mainly because of the overlap in clinical symptoms and radiological features between different forms of dementia. In recent years, the integration of hybrid PET/RM imaging has opened new perspectives to improve the differential diagnosis of dementia. PET images provide functional information on brain metabolism, while MRI images provide high-resolution anatomical details. The combination of these techniques allows a more complete and accurate assessment of brain changes associated with different dementias. A further advance in this field is the application of artificial intelligence algorithms, which can analyse large amounts of complex data, identify patterns not visible to the human eye, and provide decision support to clinicians. In recent years, several studies have demonstrated the usefulness of artificial intelligence in medicine to optimise the entire diagnostic and therapeutic process. This thesis is in support of this new frontier of medicine, promoting the idea that, although they may seem abstract and raise doubts in some, tools based on artificial intelligence are in fact valuable allies in improving and guaranteeing the highest level of care for patients at all times, assisting doctors and complementing and enhancing human capabilities.
La diagnosi delle diverse forme di demenza rappresenta una delle sfide più complesse ed urgenti della medicina moderna. Con l’aumento dell’aspettativa di vita, il numero di persone affette da demenze è in costante crescita. Queste patologie non solo compromettono gravemente la qualità della vita dei pazienti, ma pongono anche un enorme carico sui sistemi sanitari e sulle famiglie dei pazienti stessi. Tradizionalmente, la diagnosi delle demenze si basa su valutazioni cliniche, test neuropsicologici e imaging cerebrale tramite tecniche come la risonanza magnetica e la tomografia ad emissione di positroni. Tuttavia, l’accuratezza diagnostica ottenuta con questi metodi non è sempre soddisfacente, principalmente a causa della sovrapposizione dei sintomi clinici e delle caratteristiche radiologiche tra le diverse forme di demenza. Negli ultimi anni, l’integrazione delle immagini ibride PET/RM ha aperto nuove prospettive per migliorare la diagnosi differenziale delle demenze. Le immagini PET forniscono informazioni funzionali sul metabolismo cerebrale, mentre le immagini RM offrono dettagli anatomici ad alta risoluzione. La combinazione di queste tecniche consente una valutazione più completa e accurata delle alterazioni cerebrali associate alle diverse demenze. Un ulteriore progresso in questo campo è rappresentato dall’applicazione di algoritmi di intelligenza artificiale, che possono analizzare grandi quantità di dati complessi, identificare pattern non visibili all’occhio umano e fornire supporto decisionale ai medici. Negli ultimi anni diversi studi hanno dimostrato l’utilità dell’intelligenza artificiale nel campo della medicina per l’ottimizzazione di tutto l’iter diagnostico e terapeutico. Questa tesi si pone a sostegno di questa nuova frontiera della medicina, promuovendo l’idea che, nonostante possano sembrare astratti e suscitare dubbi in alcuni, gli strumenti basati sull’intelligenza artificiale, sono in realtà preziosi alleati per migliorare e garantire sempre il massimo livello di cura ai pazienti, fornendo supporto ai medici, integrando e potenziando le capacità umane.
Applicazione di Algoritmi di Intelligenza Artificiale su Immagini Ibride PET/RM per la Diagnosi Differenziale delle Demenze
CERONI, ELENA
2023/2024
Abstract
Diagnosing different forms of dementia is one of the most complex and pressing challenges in modern medicine. With increasing life expectancy, the number of people suffering from dementia is constantly growing. These diseases not only severely impair patients' quality of life, but also place a huge burden on healthcare systems and patients' families. Traditionally, the diagnosis of dementia has been based on clinical assessment, neuropsychological testing and brain imaging using techniques such as MRI and positron emission tomography. However, the diagnostic accuracy of these methods is not always satisfactory, mainly because of the overlap in clinical symptoms and radiological features between different forms of dementia. In recent years, the integration of hybrid PET/RM imaging has opened new perspectives to improve the differential diagnosis of dementia. PET images provide functional information on brain metabolism, while MRI images provide high-resolution anatomical details. The combination of these techniques allows a more complete and accurate assessment of brain changes associated with different dementias. A further advance in this field is the application of artificial intelligence algorithms, which can analyse large amounts of complex data, identify patterns not visible to the human eye, and provide decision support to clinicians. In recent years, several studies have demonstrated the usefulness of artificial intelligence in medicine to optimise the entire diagnostic and therapeutic process. This thesis is in support of this new frontier of medicine, promoting the idea that, although they may seem abstract and raise doubts in some, tools based on artificial intelligence are in fact valuable allies in improving and guaranteeing the highest level of care for patients at all times, assisting doctors and complementing and enhancing human capabilities.File | Dimensione | Formato | |
---|---|---|---|
Tesi Elena Ceroni.pdf
non disponibili
Dimensione
811.7 kB
Formato
Adobe PDF
|
811.7 kB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/17532