New Zealand Engineering 1999 September

Interview

Quiet Acheiver

The CAD body
The finite element heart
Team building
Commercial development

Peter Hunter is probably one of New Zealand’s least known entrepreneurs. He is also one of New Zealand’s brightest academics. Quietly, and without much fuss, he has made a career of tackling some of the toughest mathematics in engineering and is reaching the stage where he may become midwife to a range of next century’s high tech export industries.

Softly spoken, with serious glasses and a lean figure, the Professor of Engineering Science at Auckland University owes his life’s work to an almost throwaway theoretical problem proposed to him when a masters student by Mike O’Sullivan, then a lecturer in the same department. How to mathematically model the flow of blood in an artery. Starting from standard fluid mechanics equations the problem required students to take into account both the elasticity of the pipe and the viscosity of the fluid.

"I loved it right from the start. It opened my eyes to a whole new world of non-linear mathematics."

Nearly twenty years later Professor Hunter is the chairman of the Physiome Commission of the International Union of Physiological Sciences, and in the process of helping to organise the 34th IUPS Conference to be held in New Zealand in 2001 - the largest scientific conference ever to be held in this country. He is also one of the leading lights in the global Physiome Project, a collaborative project which will model the human body as completely as computer aided engineering modelled the Boeing 777.

The CAD body
The aim of the project is to develop databases and models which facilitate the understanding of the integrative function of cells, organs and organisms building down through subcellular modelling to the molecular level and the database generated by the Genome Project, and to build up through whole organ and whole body modelling to clinical knowledge and applications. Almost like "unfolding"a Boeing 777 from the metallurgy of every screw to the total performance of the airframe.

In many ways it is a bigger task than the Genome Project and the initial goals include both organ specific modelling (eg. the Cardiome Project, see below) and distributed systems (eg. the Microcirculation Physiome Project).

On the business side Professor Hunter has already begun commercialising his team’s research with software licensing arrangements with Physiome Sciences Inc of New York for the use of heart models to develop drugs, and to Pacific Title Mirage Studio (PTM) of California to model faces for special effects in the movie industry. His long-term aim is to build on New Zealand’s sporting heritage and try to develop an industry based on sports applications of his work.

The finite elemet heart
Physiome modelling is not simply a matter of digitising a lot of magnetic resonance tomography scans and pretending something has been achieved. It is about developing software which can duplicate the characteristics and behaviour of corporal features. For example, the lungs modelled by researcher Merryn Tawhai on a screen are not the result of a scan of any individual’s lungs, they are grown like a fractal into a virtual chest cavity following the rules all lungs seem to follow. The software models the gas mixing that occurs in the asinii as well as the heat and water transfer of the whole organ. The model must be verified against reality and also be useful for predicting reality. This is where the engineering expertise can expand on the physiological science, and where Professor Hunter realised there was a gap more than twenty years ago.

"Medics are typically not well grounded in mathematics and physics. They know little of the conservation of momentum for example which means they find it more difficult to model the whole system in terms of its overall function."

This is of course not to disparage physiologists. Indeed Professor Hunter’s PhD supervisor, after he won a Commonwealth Scholarship to study at Oxford University in 1972, was a physiologist and part of Oxford’s bioengineering collaboration. It was at Oxford that Professor Hunter developed three themes in his career that persist to this day. The first was a fascination with the physics of the heart.

"The complexity of it is marvellous because you are never dealing with just one thing. You have the mechanics of the heart muscle, its strength, thickness and its fibrous structure and orientation, the flow of blood within the organ, the electrical wavefront through the organ, the biochemical aspects," he says.

The second theme was a fascination with the application of finite element analysis in non-linear materials.

"I discovered this book by Tinsley Oden, Finite Element Analysis of Non-linear Continua. It was the only book written about the subject I could find at the time and it became my Bible which I carried around with me constantly."

Professor Hunter adapted Oden’s work to apply to the mechanical properties of tissue which is both grainy, varies in thickness and is elastic.

But none of the above could have come to anything were it not that Professor Hunter was fortunate enough to have access to the Rutherford Laboratories ICL Atlas supercomputer.

"Twenty-five years ago we were ambitious about the sort of things we thought we could achieve. Stupidly ambitious. But we gradually learned to define the problems to meet the available computing power, which has expanded a million-fold since then."

Team building
In 1979, after seven years in the UK, Peter Hunter and his first wife Marnie had reached a critical decision point about their future lifestyle. Their first child was on the point of going to school and it was a case of either staying on in Britain or coming home. They chose to come home. So Professor Hunter returned to Auckland University for a lectureship in Engineering Science.

But one hair-on-fire bioengineer in an engineering science department does not a movement make. That takes a team, both on the staff and among the senior research students. For the last twenty years Professor Hunter and his Medical School colleague Associate Professor Bruce Smaill have been developing such a team, both within Engineering Science and in the Department of Physiology. Today it numbers six faculty staff, six research staff and twelve graduate students. He pays special tribute to the work of Dr David Bullivant, the publicity shy, but vitally important developer of the graphics systems used by the researchers.

"I can’t overstate the importance of good graphics in this sort of research. When equations become visible, when you can see what is going on, it makes the task of mathematical modelling so much easier. Not just for those doing the research but also for those who are being asked to fund it as well," he says.

Senior PhD researcher Nick Smith, who will shortly be following in Professor Hunter’s footsteps to join the Oxford University bioengineering group, pays tribute to Professor Hunter’s skills in orchestrating the research effort. He says keeping the team employed on meaningful research which is at once separate and an integral part of the whole effort, is the key to maintaining the momentum and morale of the team.

Most of the graduate engineering science students are involved with modelling the heart and lungs, sometimes with direct practical results in mind. Merryn Tawhai’s work on water and heat transfer in the lung has demonstrated the clinical importance of humidifiers in respirators to prevent the lung drying out - particularly important when dealing with premature babies. Leo Cheng, under the supervision of Dr Andrew Pullan in the Engineering Science Dept, is working with Green Lane Hospital on non-invasive imaging of the heart for diagnostic purposes using an array of electrodes embedded in a jacket. By comparing the actual recordings of the electric pacing wave spreading through the heart muscle against the model it becomes possible to determine a more precise location of dead heart tissue (infarction) which is causing fibrillation. This can be used to guide surgeons using catheter ablation techniques on the heart wall. Tim Kirk is studying mechanical deformation of lung tissue and the effects of fluid and gas pressure. The object is to find methods for mechanical ventilation which reduce stress and strain on the lung and assist the design of asthma drugs. Carey Stevens is working on finite element analysis of the heart walls (120 elements with curves captured to the second derivative) to determine points of structural weakness in the heart and the effect of surgery on strength, pacing from different locations and how tissue responds to infarction.

But behind the immediately practical research there is a lot of intense pure research also going on. Others are looking at the modelling of the individual cells within the organ. David Nickerson looks at the chemical "channels" and signalling pathways within the cell, which end up in differential equations with up to thirty variables. Martin Buist looks at electrocardiograph patterns and the transfer of microvoltages from cell to cell, providing input for Mr Cheng’s work.

Whether they are involved in cellular structure or mapping Hollywood actor Jim Carey’s mobile face, they are all part of Professor Hunter’s overall plan.

"I am trying to develop new employment opportunities for these students and ultimately an industry for New Zealand. These students are very good. Some of them are among the best in the world because for some reason New Zealanders excel at taking hard problems and making them solvable. Yes they will need to go overseas, just as I did, and build up their networks and experience but eventually many of them will want to come home. Physiome Sciences in New York is a $100 million a year business which shows the level of growth this area is going to see. I want New Zealand to capture some of that business. There’s no reason we shouldn’t. We have the ability and the product is information so it travels as fast as a file transfer over the net. We just need to get the venture capital and the kind of business model that has allowed the Americans to develop their high technology industries so quickly."

Commercial development
Professor Hunter has never been the kind of academic scared of rubbing shoulders with the commercial world either. Sponsorships, such as that from Physiome Sciences or Fisher & Paykel Healthcare, are actively sought out. He sees the ability to generate revenue streams from research as vital to keeping the programme alive and relevant to the world. However, the University does not actually sell anything outright, it sells licences. This means that the same research can be sold for different purposes for a specific time period either as an exclusive or non-exclusive right. The benefit: ownership still remains with the University.

Having something to sell has also justified the group’s capacity to buy. And of course what every bioengineer wants to buy is an SGI Origin 2000 supercomputer, together with a whole lot of high speed SGI workstations. For although the application this research group works on is bioengineering, the product that they produce is essentially software and for that they have developed a most remarkable software development system called CMISS (Continuum Mechanics, Imaging, System Identification and Signal Processing).

The first notable feature of CMISS is that its code management system is web based (see http://www.esc.auckland.ac.nz/ Groups/Bioengineering/CMISS). Although executing CMISS functions is restricted to the university supercomputer anyone in the world can monitor the status of the code on the system via the website. At its core are two libraries of executables: Fortran 77 functions to translate the algorithms and mathematical equations into output numbers; and C++ functions to paint the graphics depicting those outputs. Putting an experiment together is done using a Unix interface where functions are added on a pick and mix basis. Moreover, the code is managed such that every night functions are tested against a set of specific problems. If someone changes a function such that it departs from the performance criteria of the function, the test picks up the variation and alerts the system administrators.

The group (in this case led by Dr Poul Nielsen, another Engineering Science colleague) is also working with XML, the extensible mark-up language. XML is a subset of the Standard Generic Mark-Up Language just as HTML is but it also allows developers to start including content specific tags for different applications. The problem with XML is developing conventions within any particular body of knowledge such that they can be handled consistently across the entire knowledge domain. By starting now the Auckland group hopes to have an active voice in the application of XML to the world Physiome Project development into the next century.

Bioengineering has only become possible with the fantastic increases in computing power that have occurred this century. No 19th century engineer would have dared to take on the mind-numbing complexity of integrated biological systems. But simply because no 19th century engineer could conceive such an enterprise does not mean 21st century engineers can fail to appreciate the importance of such an endeavour in their time. Indeed the degree of success that Professor Hunter can achieve will be a harbinger for the future of New Zealand’s 21st century economy.

Peter King, editor