In recent years, deep learning–based MRI reconstruction technologies, typically integrated with conventional techniques such as parallel imaging and compressed sensing, have demonstrated the potential to improve diagnostic performance while reducing acquisition times. This thesis presents a local Health Technology Assessment (HTA) evaluating a vendor-agnostic AI-based MRI reconstruction software from clinical, safety, economic, and regulatory perspectives. The assessment included analysis of clinical effectiveness in terms of diagnostic confidence, based on available clinical evidence on Swift-MR, as well as evaluation of safety considerations related to AI-based reconstruction systems. An economic analysis, including a volume sensitivity analysis, was performed to assess financial sustainability under different productivity scenarios. The regulatory framework was examined within the European context, assessing the positioning of the technology under the AI Act within MDR, and GDPR. Overall, clinical effectiveness and safety were demonstrated, and the economic evaluation suggests potential sustainability in medium-to-high volume settings.
Negli ultimi anni, le tecnologie di ricostruzione RM basate su deep learning, tipicamente integrate con tecniche convenzionali come il parallel imaging e compressed sensing, hanno dimostrato il potenziale di migliorare le performance diagnostiche riducendo i tempi di acquisizione. Questa tesi presenta un Health Technology Assessment (HTA) locale volto a valutare un software vendor-agnostic di ricostruzione RM basato su intelligenza artificiale dal punto di vista clinico, di sicurezza, economico e regolatorio. L’efficacia clinica è stata analizzata in termini di confidenza diagnostica, sulla base delle evidenze cliniche disponibili relative a SWIFT-MR, mentre sono stati esaminati i principali profili di rischio associati ai sistemi di ricostruzione basati su AI. È stata inoltre condotta un’analisi economica comprensiva di volume sensitivity analysis per valutare la sostenibilità finanziaria in diversi scenari produttivi. L’inquadramento regolatorio è stato analizzato nel contesto europeo, considerando l’AI Act all’interno del MDR e il GDPR. Nel complesso, l’efficacia clinica e la sicurezza risultano dimostrate, con una sostenibilità economica potenziale nei contesti a medio-alto volume di attività.
HTA for the adoption of AI-based MRI reconstruction software: enhancement and image quality quantification
MUSUMECI, SIMONE
2024/2025
Abstract
In recent years, deep learning–based MRI reconstruction technologies, typically integrated with conventional techniques such as parallel imaging and compressed sensing, have demonstrated the potential to improve diagnostic performance while reducing acquisition times. This thesis presents a local Health Technology Assessment (HTA) evaluating a vendor-agnostic AI-based MRI reconstruction software from clinical, safety, economic, and regulatory perspectives. The assessment included analysis of clinical effectiveness in terms of diagnostic confidence, based on available clinical evidence on Swift-MR, as well as evaluation of safety considerations related to AI-based reconstruction systems. An economic analysis, including a volume sensitivity analysis, was performed to assess financial sustainability under different productivity scenarios. The regulatory framework was examined within the European context, assessing the positioning of the technology under the AI Act within MDR, and GDPR. Overall, clinical effectiveness and safety were demonstrated, and the economic evaluation suggests potential sustainability in medium-to-high volume settings.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_03_Musumeci_Tesi.pdf
accessibile in internet per tutti
Descrizione: testo tesi
Dimensione
10.11 MB
Formato
Adobe PDF
|
10.11 MB | Adobe PDF | Visualizza/Apri |
|
2026_03_Musumeci_Executive Summary.pdf
accessibile in internet per tutti
Descrizione: executive summary
Dimensione
504.24 kB
Formato
Adobe PDF
|
504.24 kB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/251858