Haemodynamics of the Fontan connection : investigating patient specific model creation and validation from MRI data

Show simple item record

dc.contributor.author Moyle, Keri R. en
dc.date.accessioned 2020-07-08T04:50:41Z en
dc.date.available 2020-07-08T04:50:41Z en
dc.date.issued 2003 en
dc.identifier.uri http://hdl.handle.net/2292/52065 en
dc.description Full text is available to authenticated members of The University of Auckland only. en
dc.description.abstract The haemodynamics of the Fontan connection have been investigated using patient-specific, anatomically realistic, computational models. The models were constructed to extrapolate beyond the measurable data obtained from Magnetic Resonance Imaging (MRI), to simulations of sufficient complexity to allow meaningful interrogation of the fluid mechanics. The aims of this project are three-fold: • To investigate the utility of MRI data in the construction of transient, three- dimensional patient-specific computational simulations, and to determine the degree to which this process may be automated. • To substantiate the solutions created, by both internal (mesh refinement) verification and external (comparing to independent MRI data) validation; completing the Ioop from measured data to patient specific model creation and back to the real world.• To extrapolate the known data to regions where measured data are not available, and to use the models to investigate the haemodynamics of the Fontan connection. Methodology, Processing techniques and methodologies were developed in conjunction with a set of data from coarctation patients Iate after repair. An automatic image segmentation method based on the ratio of intensity to velocity in MRI data was developed. Application to the coarctation data set yielded accurate contours, where parameter selection by a medically inexperienced and blinded operator compared well with manual contouring by an experienced medical practitioner. The efficacy of the automatic segmentation method applied to the Fontan data set was limited by an inability to distinguish moving blood from moving heart tissue. The amount of a priori knowledge required motivated manual contouring of all Fontan images. The interpolation from a boundary description on slices to a three-dimensional surface was undertaken using a stepwise method of fitting cubic splines in radial coordinates to the vessel cross-section, and successfully automated for the coarctation data set. This method was extended to process junctions and large chambers such as atria, and realised its goal in test problems. Employment in processing the Fontan data sets was accomplished in one patient. The small length of vessels involved meant automatic reconstruction was not practicable in the other four data sets, whose surfaces were constructed manually. Simulations were created using an existing Voronoi-based mesh generator, and a transient, second-order correct, 3D Navier-Stokes solution algorithm. Solution stability was dependent on mesh refinement, geometry and advection approximations. The presence of fluctuating, diverging flow in the anastomotic region tended to unsettle Iess robust solutions. Transient MRI data were used as velocity inlet boundary conditions, following the establishment of initial conditions from a steady-state solution to the first time-step. Solution error estimation, validation and verification are presented for test cases, as well as being applied to selected patient models. The effect of inlet conditions on the clinical validity of the solution is assessed, and while not negligible, its effect on clinically relevant parameters such as work Ioss and pressure drop is within the margin of error derived from mesh convergence studies, and comparison to measured data. It is suggested that the degree of model complexity required to provide clinically relevant solutions is over-estimated. Results, A method of precisely identifying and quantifying the flow structures responsible for energy Ioss is formulated based on entropy generation in the fluid domain. Areas of complex colliding flow do cause an increase in dissipated work, but the greatest irreversible work Ioss occurs through wall shear stress, and flow impinging on the walls. Contrary to conclusions implied elsewhere, the contribution of complicated flow to the overall in efficiency of the connection is minor, but represents the only improvable flow region. It is suggested that future connection design be based on wall shear stress management, due to its dual physiologic and energetic significance. Quantitative comparison of the effect of specific flow structures on the overall connection haemodynamics is investigated, revealing that the slow recirculation regions and tight vortices present in the atriopulmonary connections are of detriment to both the connection efficiency and the physiological conditions generated. The transition in surgical opinion towards the total cavopulmonary connection is justified in these results. Established techniques for the extraction of pressure field information from a velocity data set were implemented and extended to three dimensions. Results showed a comparable pressure drop between modelled pressure, pressure fields reconstructed from sampled modelled velocity fields, and MRI data. Conclusions, This study has fulfilled the aims of creating and validating patient specific models of the Fontan connection. Five models have been created and their fluid dynamics investigated in detail. Computational solutions have been validated by comparison to measured data, and the error bound of the most detailed model is estimated to be within the measurement errors of the velocity MRI data set. The simulations have been used to examine the detailed haemodynamics involved, highlighting the irreversible work dissipation at the walls as being a more significant design parameter than the inefficiencies present in colliding caval flows. en
dc.publisher ResearchSpace@Auckland en
dc.relation.ispartof PhD Thesis - University of Auckland en
dc.relation.isreferencedby UoA99119544014002091 en
dc.rights Items in ResearchSpace are protected by copyright, with all rights reserved, unless otherwise indicated. en
dc.rights Restricted Item. Full text is available to authenticated members of The University of Auckland only. en
dc.rights.uri https://researchspace.auckland.ac.nz/docs/uoa-docs/rights.htm en
dc.title Haemodynamics of the Fontan connection : investigating patient specific model creation and validation from MRI data en
dc.type Thesis en
thesis.degree.discipline Mechanical Engineering en
thesis.degree.grantor The University of Auckland en
thesis.degree.level Doctoral en
thesis.degree.name PhD en
dc.rights.holder Copyright: The author en

Files in this item

Find Full text

This item appears in the following Collection(s)

Show simple item record


Search ResearchSpace