Title. Artificial intelligence software for stroke imaging. Development of standardized CT imaging techniques with optimization of focus point. Introduction. The use of artificial intelligence (AI) can support the radiologist in early and precise diagnosis of acute ischemic stroke. iSchemaView Rapid is an AI software that analyses CT volumes quickly and is able to identify patients at high ischemic risk. All these things allow the implementation of therapy for patients, that can increase their survival and reduce deficit. Purpose. To analyse the features and how iSchemaView Rapid works. To define the CT imaging characteristics required for the software to function correctly. To create a standardized CT protocol to ensure reproducibility in CT imaging in patients with suspected stroke within a Hub and Spoke hospital setting. Methods and materials. Five scientific articles from Pubmed and the iSchemaView Rapid User Manual were analysed to identify the characteristics that CT imaging must have for correct recognition of the lesion by the algorithm. 21 CT series were analysed for suspected stroke performed with CT scanners from Siemens (models “Somatom Definition Edge” and “Somatom Definition Flash”), GE (models “Optima 660” and “Revolution 64”), Toshiba (“Aquilion Prime TSX 303A”) and Philips (“Brilliance iCT 256”) to identify the most frequent artifacts that don’t allow the software to highlight the presence of lesions. Finally, the focus points during a CT study for suspected stroke were defined and analysed. Results and Conclusion. CT imaging characteristics have been defined to optimize the use of iSchemaView Rapid. A standardized CT protocol was produced for patient preparation and positioning, technical execution and post-processing to obtain reproducible imaging. Focus point optimization allows to improve the quality of the imaging produced and reduce the presence of artefacts, ensuring better use of the software as well as a decrease in CT image processing time. The correct use of the software supports radiologist in formulating the diagnosis of stroke, guaranteeing better clinical outcomes. Keywords. Artificial intelligence, stroke imaging, focus point, CT techniques, image processing.
Introduzione. L’utilizzo dell’intelligenza artificiale (IA) è in grado di supportare il radiologo in una precoce e precisa diagnosi di ictus ischemico acuto. L’iSchemaView Rapid è un software di IA che analizza i volumi TC in tempi rapidi ed è in grado di identificare i pazienti ad alto rischio ischemico permettendo così l’attuazione di una terapia tempestiva che può aumentare la sopravvivenza degli stessi e ridurre eventuali deficit. Scopo. Analizzare le caratteristiche ed il funzionamento di iSchemaView Rapid. Definire le caratteristiche dell’imaging TC richiesto per il corretto funzionamento del software. Creare un protocollo TC standardizzato per garantire riproducibilità nell’imaging TC in pazienti con sospetto stroke all’interno di un presidio ospedaliero di tipo Hub e Spoke. Materiali e metodi. Sono stati analizzati cinque articoli scientifici provenienti da Pubmed e l’User Manual di iSchemaView Rapid per identificare le caratteristiche che l’imaging TC deve possedere per il corretto riconoscimento della lesione da parte dal software stesso. Sono state analizzate le immagini TC degli studi per sospetto stroke eseguiti nell'arco di due mesi per identificare gli artefatti più frequenti che non consentono al software di evidenziare la presenza di lesioni. Sono stati, inoltre, definiti e analizzati i focus point durante uno studio TC per sospetto stroke. Risultati e conclusioni. Sono state definite le caratteristiche dell’imaging TC per ottimizzare l’uso di iSchemaView Rapid. È stato prodotto un protocollo TC standardizzato per la preparazione e posizionamento del paziente, l’esecuzione tecnica e il post-processing per ottenere imaging riproducibile. L’ottimizzazione dei focus point permette di migliorare la qualità dell’imaging prodotto e di ridurre la presenza di artefatti, garantendo un miglior utilizzo del software nonché una riduzione del tempo di elaborazione delle immagini TC. Il corretto impiego del software supporta il radiologo nella formulazione della diagnosi di stroke, garantendo migliori outcome clinici.
SOFTWARE DI INTELLIGENZA ARTIFICIALE PER LO STROKE IMAGING. SVILUPPO DI TECNICHE TC STANDARDIZZATE CON OTTIMIZZAZIONE DEI FOCUS POINT
BARZAN, LUANA
2022/2023
Abstract
Title. Artificial intelligence software for stroke imaging. Development of standardized CT imaging techniques with optimization of focus point. Introduction. The use of artificial intelligence (AI) can support the radiologist in early and precise diagnosis of acute ischemic stroke. iSchemaView Rapid is an AI software that analyses CT volumes quickly and is able to identify patients at high ischemic risk. All these things allow the implementation of therapy for patients, that can increase their survival and reduce deficit. Purpose. To analyse the features and how iSchemaView Rapid works. To define the CT imaging characteristics required for the software to function correctly. To create a standardized CT protocol to ensure reproducibility in CT imaging in patients with suspected stroke within a Hub and Spoke hospital setting. Methods and materials. Five scientific articles from Pubmed and the iSchemaView Rapid User Manual were analysed to identify the characteristics that CT imaging must have for correct recognition of the lesion by the algorithm. 21 CT series were analysed for suspected stroke performed with CT scanners from Siemens (models “Somatom Definition Edge” and “Somatom Definition Flash”), GE (models “Optima 660” and “Revolution 64”), Toshiba (“Aquilion Prime TSX 303A”) and Philips (“Brilliance iCT 256”) to identify the most frequent artifacts that don’t allow the software to highlight the presence of lesions. Finally, the focus points during a CT study for suspected stroke were defined and analysed. Results and Conclusion. CT imaging characteristics have been defined to optimize the use of iSchemaView Rapid. A standardized CT protocol was produced for patient preparation and positioning, technical execution and post-processing to obtain reproducible imaging. Focus point optimization allows to improve the quality of the imaging produced and reduce the presence of artefacts, ensuring better use of the software as well as a decrease in CT image processing time. The correct use of the software supports radiologist in formulating the diagnosis of stroke, guaranteeing better clinical outcomes. Keywords. Artificial intelligence, stroke imaging, focus point, CT techniques, image processing.File | Dimensione | Formato | |
---|---|---|---|
Barzan_Luana.pdf
accesso aperto
Dimensione
4.06 MB
Formato
Adobe PDF
|
4.06 MB | Adobe PDF | Visualizza/Apri |
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/20.500.12608/56924