Biomedical engineering

Haemodynamics of brain aneurysms

Brain aneurysms are localised enlargements of an artery wall inside the skull.  (They are also known as cerebral aneurysms and intracranial aneurysms.)  It has been reported that up to 5% of the general population may have brain aneurysms.  Most people are not affected by the presence of an aneurysm, but some aneurysms are unstable:  when such aneurysms rupture, the patient is likely either to die or to be severely disabled.  Robust diagnosis of aneurysm stability - the risk of rupture - for individual patients has not been possible.

Our approach is to study the flow within a brain aneurysm at a higher resolution than currently possible in any hospital.  We do this mainly by applying computational fluid dynamics (CFD) technology to simulate the flow through real aneurysm geometries obtained by medical imaging.  This work has led us to discover a parameter, "energy loss", that has strong statistical correlation with aneurysm stability.  We also support this research with in-vitro modelling, in which silicone models are constructed using a 3D-printing process, and the internal flow examined by particle image velocimetry (PIV).

We focus on development a realistic FSI simulation model that incorporates accurate individualised vascular wall thickness and elastic mechanical properties. In order to validate the accuracy of the newly developed FSI model by applying four-dimensional computed tomography (4D-CT) image, and particle image velocimetry (PIV) to patient-specific samples and models.

We currently seek PhD candidates with strong backgrounds in analysis of fluid flows and numerical methods.  Qualifications may be in applicable branches of engineering (chemical/mechanical/aerospace/civil/environmental) or physics.

Vascular haemodynamics

There are many pathologies (diseases) and medical procedures affecting the human vasculature.  Some examples include the following.

• Cardiovascular:  heart bypass surgery in adults;  vascular rerouting in children congenital heart diseases (e.g. Fontan procedure); heart fail simulation, and virtual surgical simulation.
• Cerebrovascular: cerebral aneurysms, moyamoya disease; AVM, etc.

Our approach is typically to study a large data set of patients, for whom we have obtained medical images of the relevant arteries.  We can then use computational fluid dynamics (CFD) technology to simulate the blood flow through the affected arteries and veins.

We currently seek PhD candidates with strong backgrounds in analysis of fluid flows and numerical methods.  Qualifications may be in applicable branches of engineering (chemical/mechanical/aerospace/civil/environmental) or physics.

Medical imaging

Medical imaging is often the basis for diagnosis of patient condition, besides surgical planning and other uses.  In modern medicine, three-dimensional (3D) imaging technologies have become standard.  Common modes of 3D medical imaging include x-ray computed tomography (CT) and magnetic resonance imaging (MRI).  Within the machine the raw signal data measured at the machine's sensors are converted to a 3D raster (bitmap) image.  However, to be most useful to clinicians that 3D intensity information should be 'segmented', so that physiological features such are identified.  For example, we may wish to distinguish between the blood-filled lumen of an artery, the wall of the artery, and surrounding brain matter;  or between a tumour and normal tissue.  While this is conceptually simple, in practice the different algorithms can yield quite different estimates of feature dimensions (size, area, volume), reflecting substantial systematic errors.  Such errors can result in misdiagnosis of illness severity, complicate comparison of follow-up observations, and produce consequential errors in computational simulations.

Our approach is to apply the latest knowledge regarding image analysis to this problem, and to provide independent validation of the optimal method, along with the optimal method parameters, and measures of method robustness.

Our recent work has included development and validation of a threshold-based level-set method for intracranial aneurysms.  There are many opportunities to take this work further, such as improvement of the existing method, development of alternative methods, and application for different medical conditions (e.g. cancer, œdema).  Further interests include the combination (or 'registration') of imaging data from multiple modalities.

We currently seek PhD candidates with strong backgrounds in analysis of multidimensional (2D, 3D,...) data, mathematical/numerical methods, and computer programming.  Qualifications may be in applicable branches of engineering (software/materials/electrical/electronic/chemical/mechanical/aerospace/civil/environmental) or science (physics, mathematics, computer science).

Surgical training

Surgical training follows an 'apprenticeship' model that has been essentially unchanged for decades, if not centuries.  Each trainee observes a senior surgeon:  what the trainee learns, how they learn it, and how much practice they get, can vary significantly from one mentor or institution to another.  The ultimate proof of surgical efficacy is in patient outcomes;  however, trainees need feedback on their skill level before reaching this stage.  Due to limits on the opportunities to perform surgeries (or surgical procedures) on live animals, human cadavers, or human patients, we propose an increasing the use of surgical simulation technologies.

Surgical simulation technologies may be cheaper than other options, and do not face the same ethical concerns.  Moreover, there is scope to rigorously measure each small detail of the trainee's performance and provide both an overall 'skill score', along with specific and quantitative feedback on individual elements of the task.  We propose that this objective and quantitative feedback will allow surgical trainees to learn faster, and perhaps reach a higher level of skill.

We have already developed a prototype rig for the study of anastomosis (surgical connection of two vessels), and collected preliminary experimental data for both trainees and experienced surgeons.  The next steps will be (i) to study in further detail the existing data, (ii) thereby to identify hallmarks of surgical skill, (iii) to refine the experimental rig and data collection procedure, (iv) to conduct an observational study on a cohort of trainees, (v) to develop a surgical skill score and provide feedback on the trainees, (vi) to validate any benefits of training on our surgical simulator.

We currently seek PhD candidates with strong backgrounds in design of physical measurement systems, data/signal analysis, and numerical/statistical methods.  Qualifications may be in applicable branches of engineering (electrical/electronics/mechanical/chemical) or physics.

Medical devices

Artificial heart

An artificial heart was used in a human for the first time in the middle of last century.  Since then, technological advances have led to continuously improved designs.  Prof. Qian was principal designer of a pioneering left-ventricular assist device, and the group has an ongoing interest in artificial heart design optimisation.  There are many opportunities to improve design, in which our approach is to take a comprehensive view.  Thus, we consider not only mechanical performance (e.g. pump efficiency) but also other factors such as biological compatibility (e.g. minimising damage to red blood cells), size (e.g. for children), and manufacturing costs (to make available to more patients).

Endovascular stent delivery, and vascular stent optimal design

A second area of focus within the group is to stent design, especially for the treatment of cerebral aneurysms.  In this treatment, a compacted mesh tube is introduced through the vascular network, typically at a remote site in the leg.  This is fed through a catheter to the site of pathology and deployed there.  The effect of the stent depends upon its mechanical design, and also upon the manner of its deployment.  The latest innovation in stents are known as flow diverters:  these have only been on the market for less than a decade, and there remain many opportunities to optimise their design, and moreover to develop guidance as to which types of pathology they are best suited for, in comparison to alternative treatment strategies (e.g. coiling, clipping).

Research projects in the two areas above will largely be based on simulation using computational fluid dynamics (CFD), but may be supported by in vitro experiments.  Research may include fluid-solid interaction, requiring finite element modelling (FEM) of the solid components.

We currently seek PhD candidates with strong backgrounds in analysis of fluid flows, solid deformation, and numerical methods. Qualifications may be in applicable branches of engineering (chemical/mechanical/aerospace/materials/civil/environmental) or physics.