Application of advanced, digital engineering techniques in a user-guided vascular diagnostic framework: FlowAI

Ugolini, Edoardo (2026) Application of advanced, digital engineering techniques in a user-guided vascular diagnostic framework: FlowAI, [Dissertation thesis], Alma Mater Studiorum Università di Bologna. Dottorato di ricerca in Meccanica e scienze avanzate dell'ingegneria, 38 Ciclo.
Documenti full-text disponibili:
[thumbnail of PhD_thesis_Edoardo_Ugolini.pdf] Documento PDF (English) - Richiede un lettore di PDF come Xpdf o Adobe Acrobat Reader
Disponibile con Licenza: Creative Commons: Attribuzione 4.0 (CC BY 4.0) .
Download (4MB)

Abstract

This PhD project focuses in developing a computational framework for patient-specific analysis of aortic aneurysms, combining image segmentation and computational fluid dynamics. The objective is to develop a semi-automated pipeline capable of processing patient-specific anatomical and flow data and providing quantitative morphological and hemodynamic descriptors of complex aortic diseases, including aneurysms and congenital variants such as ASA. The framework is designed to reduce the technical expertise required to operate advanced computational tools such as image analysis and CFD within an integrated, user-oriented environment. The platform supports both morphological and haemodynamic analyses. Users can perform three-dimensional measurements relevant to preoperative planning of endovascular procedures such as EVAR. In parallel, the system provides patient-specific CFD simulations for the quantitative characterization of local hemodynamics. The workflow integrates deep-learning-based segmentation of lumen and thrombus from contrast-enhanced CT images, enabling the reconstruction of three-dimensional thoracic and abdominal aortic geometries. These geometries, together with patient-specific boundary conditions, are processed within a containerized, HPC-compatible CFD environment. A centreline-based parametrization enables the extraction of cross-sectional slices orthogonal to the vessel axis and the region-wise computation of hemodynamic descriptors, including WSS, and OSI. The geometric reconstruction module was validated on patient-specific abdominal aortic datasets. The pipeline demonstrated sub-voxel accuracy in lumen surface reconstruction and good agreement with manual measurements across different acquisition protocols and clinical settings. Results support reproducibility of the geometry extraction workflow, basis for subsequent morphometric and haemodynamic analyses. CFD engine was evaluated through simulations of transient, non-Newtonian flow in patient-specific thoracic aortic and abdominal geometries, demonstrating physically consistent hemodynamic behaviour and applicability to larger statistical studies. The resulting framework provides an integrated computational environment linking medical-image-derived geometry, automated geometric processing and patient-specific CFD analysis. FlowAI and its digital transition potential may represent a step toward accessible patient-specific hemodynamic diagnostics and more complex risk-assessment tools.

Abstract
Tipologia del documento
Tesi di dottorato
Autore
Ugolini, Edoardo
Supervisore
Co-supervisore
Dottorato di ricerca
Ciclo
38
Coordinatore
Settore disciplinare
Settore concorsuale
Parole chiave
Aortic Aneurysm, Vascular CFD, Patient-Specific Diagnostics, FlowAI
Data di discussione
15 Ottobre 2026
URI

Altri metadati

Statistica sui download

Gestione del documento: Visualizza la tesi

^