Operating on the virtual human

By Anna-Marie Lever BBC News health reporter

Advertisement

Join Dr Peter Kohl as he flys through the virtual heart

Dr Peter Kohl has a big idea for the future of heart surgery.

A patient needing an immediate operation will have a scan and in the 45 minutes it takes for blood test results to come back, surgeons will have simulated alternative operations using computer models, and know the best way to proceed.

They will have investigated different surgical scenarios in virtual reality and will understand the effects these have on the individual's heart.

For example, they will know the optimum place to fit stimulation leads that are tailored to improve that patient's heartbeat.

And they will have gained vital insight into how blood flow is affected when a tube is inserted to re-open a blocked vessel and decide the best location for it in that person.

All this, they are confident, will result in a faster procedure, when time is of the essence, and less trauma to the patient's body through prior experimentation on the virtual surgical table.

How long does Dr Kohl, a physiologist at Oxford University, think this idea will become reality?

"50 years - no 10 years - maybe even 5 years. I am getting ahead of myself, but I am very optimistic," he explains excitedly.

Virtual Physiological Human

Dr Kohl is one of the principal investigators of the Virtual Physiological Human (VPH) initiative, an international collaboration, funded by the European Union, to produce biomedical models simulating the human body both structurally and functionally.

Computer models can assess the best surgical options

He explains that over the last few decades biology has focused at stripping the body down to understand how it works and now it is time to build it back up again.

"We have developed better tools to look at smaller parts of the puzzle in terms of structure and function," he said.

"We have drilled down into the detail. At the same time, our knowledge has become a fragmented construction site.

"We now need to understand how the pieces interact with each other and the environment."

Through funding from the Biotechnology and Biological Sciences Research Council (BBSRC) and British Heart Foundation (BHF) Universities in the UK, including Oxford and Leeds, are working towards this vision.

In an international effort research is being shared by engineers, bioscientists, physicians and computer scientists to build systems, apply technologies, assess applications and analyse data.

No perfect model

However, Dr Kohl believes a complete model of the body will never be finished.

Computers can already calculate more steps ahead than a chess master - we need to make use of this to be able to cope in the real world." Dr Kohl, Oxford University

"A model cannot aim to capture every aspect of the original; otherwise it would be just as complex and unwieldy as the real thing," he said.

"A model is a simplified representation of reality. Different models are built for different purposes.

"Models need to be constantly updated with testing, and new models will be needed for different specific reasons.

"Like tools in a tool-box, you need to select your model to fit the question."

How reliable?

Such developments may worry people, who fear that these kinds of methods will be less reliable than traditional approaches.

Dr Kohl argues: "Why should a quantitative computer model be any less reliable than a doctor-based in-brain assessment of a patient?

"A computer model can be used as a tool to access plausibility and help a human researcher or clinician work more efficiently through scenarios.

"Computers can already calculate more steps ahead than a chess master - we need to make use of this to be able to cope in the real world."

However, Dr Kohl stresses the need for computer predictions to be assessed thoroughly.

"If you compare your predication with real life and you get an exact match, this is great for clinicians.

"Within the given framework your model addresses reality. This increases your confidence - but there is no knowledge gained.

"However, if your model and reality are different - this drives insight.

"Either your data is incorrect, implementation is wrong or understanding inaccurate. You have more to learn."

Luckily, Dr Kohl and his colleagues are up for this challenge.