Thesis
Patient-specific multi-dimensional CFD simulations based on 4D Flow-MRI for the haemodynamic assessment of aortic dissections and perfusion optimisation of vascular grafts
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2024
- Thesis identifier
- T17108
- Person Identifier (Local)
- 201857497
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- Aortic dissection is a vascular pathology which affects 5-30 per million people. Due to regions of high shear stress and weakness in the vessel wall, the intimal layer of the aorta tears, separating it from the media and creating a channel known as a false lumen. AD is a progressive condition due to the cyclical relationship between structural changes and haemodynamic instability. Often, it is fatal in the absence of surgical intervention, with mortality rates up to 90% depending on the dissection type and severity. The preferred treatment for Type A and Type B aortic dissections is open and endovascular surgical intervention, respectively. With both treatment options, there are associated complications including graft migration, branch vessel hypoperfusion, endoleaks, stent strut fracture, thrombosis, and graft limb occlusion. Generally, these failure mechanisms are related to the internal flow regime and post-surgical haemodynamics. At present, it is difficult to predict the internal haemodynamics within these grafts before they are deployed. Therefore, this thesis seeks to understand whether we can use 4D Flow MRI in combination with CFD modelling to build pre-surgical models of aortic dissections to assist in surgical planning. Leveraging CFD in combination with 4D Flow-MRI mitigates the intrinsic limitations of each approach. With 4D Flow-MRI, it is possible to extract in vivo flow rates and wall motion, and elucidate qualitative and quantitative information on the evolution of blood flow throughout the cardiac cycle. However, the spatiotemporal resolution is limited and it is not possible to extract pressure or near-wall haemodynamics. CFD, in contrast, offers a significantly enhanced level of detail, permitting the calculation of clinically relevant parameters such as pressure, TAWSS, and OSI with high spatiotemporal resolution. To generate high-fidelity CFD models would require a methodology to process the quantitative blood flow data to extract anatomical information and calibrate boundary conditions. Commonly, this requires multiple imaging scans and boundary conditions rely on invasive measurements or several assumptions from multiple sources. In this thesis, we seek to extract all relevant information from a single 4D Flow-MRI scan to generate patient specific CFD models. To the best of our knowledge, this has not yet been performed before. We therefore present a methodology to generate high-contrast anatomical images from retrospective 4D flow-MRI data. This permitted successful segmentation and reconstruction of a healthy aorta, along with the true lumen and branch vessels of the dissected aorta. However, it was not possible to generate sufficient contrast within the false lumen due to low flow rates. To do so would require multi-VENC 4D Flow-MRI imaging which was not available during this study. Though it is possible to directly prescribe pressure (from an invasive catheter) and flow (from 4D Flow-MRI) waveforms as BCs to the CFD model, this is inappropriate for several reasons. Primarily, this is because branch flow and pressure waveforms are part of the desired solution for surgical planning. Secondly, the direct prescription of flow waveforms fails to yield correct pressure measurements since the downstream resistance and compliance is not accounted for, unlike in 3EWM BCs. Thirdly, the prescription of pressure waveforms requires invasive catheter measurements and increased patient burden. Therefore, we describe a methodology for the rapid estimation and calibration of patient specific Windkessel boundary conditions based on 4D Flow-MRI data. This yielded a perfusion distribution very similar to in vivo data without the need for requiring invasive pressure or flow measurements. Finally, we evaluated the haemodynamic environment in the aortae of healthy volunteers and Type B aortic dissection cases via coupled 0D-3D numerical modelling. Such simulations may assist in determining regions of vessel wall instability to identify patients who are at the highest risk of false lumen rupture. The present thesis shows that all the essential components required for a patient-specific CFD analysis could be derived from a single 4D Flow-MRI scan, with a view to replace CT imaging and non-specific boundary conditions. The methodologies presented could further be improved in the future, by utilising multi-VENC imaging and prescribing 4D Flow-MRI derived wall motion. This may reduce the burden on patients since it is a non-invasive, non-ionising approach which does not require intravenous contrast agents.
- Advisor / supervisor
- Ritos, Konstantinos
- Kazakidi, Asimina
- Maclean, Craig
- Brodie, Robbie
- Resource Type
- DOI
- Funder
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T17108 | 2024-10-08 | Public | Download |