WHAT will it be like knowing when and how I might die? The question hits me as I look at images showing the force that flowing blood exerts on a brain aneurysm. I’m at University College London (UCL), but the images are being produced by a simulation running on HECToR, the UK’s fastest supercomputer, based in Edinburgh.
The complex fluid-dynamics simulation is similar to those used by the likes of McLaren and Ferrari to design Formula 1 racing cars. Leaps in mathematics and computing power now mean that increasingly complex biological functions can now be simulated in silico instead of in vivo. Welcome to the future of medicine.
These simulations could soon run on desktop PCs. The biggest challenge will be presenting the data to doctors in a useful way, says at UCL. “Today’s medics are not trained in fluid dynamics,” he says.
Advertisement
Ideally, we will be able to run biological simulations of the entire human body that are tailored to an individual. This would run from the level of molecules and genes, through cellular processes and tissues, to whole organs, blood vessels and the nervous system. It would be like having a body double to experiment on, testing outcomes of drugs, surgical interventions, and lifestyle choices.
Data drawn from millions of medical records already give statistical modelling a lot of predictive power. Throw in data mining and machine learning, and straightforward number crunching can then help to identify people who might be at risk. “You could haul them in before they have heart attacks,” says Coveney. But that would just be a hunch based on statistics. These models do not take the body’s chemistry into account, he says.
The Virtual Physiological Human project, of which Coveney’s blood-flow simulator is an example, aims to build such models. As these become more detailed, they can be combined to mimic larger biological systems, connecting a model of the liver to one of the heart, say. “At some point, we will have a virtual human,” says Coveney.
See the work afoot now to model the human body
The individual models are already useful. Aneurysms are caused by blood vessels that weaken and balloon out. In about 1 per cent of cases, these rupture leading to a stroke. The blood-flow simulator could be used by a surgeon to decide whether a patient needs brain surgery, for example.
The tool works by building a 3D model of a patient’s blood vessel from a static image, such as an angiogram, and then simulating blood flow through it to see where the forces are strongest (Interface Focus, ). Only if it looks as if the forces could rupture the vessel would the surgeon intervene. There are average figures for typical forces in such cases, but that is not good enough, says , also at UCL. “We are moving towards personalised medicine.”
Researchers have already developed models of many of the major organs (see diagram). Lars Küpfer at Bayer Technology Services, a pharmaceutical company in Leverkusen, Germany, and colleagues have created a virtual liver, for example, which they can use to investigate the toxicity of substances such as paracetamol. “The virtual human is the ultimate goal of computational biology,” says Küpfer. “Current models, including the ones we are using, are important milestones along the way.”
The ability to simulate biological functions is becoming important for developing more effective drugs. “Pharma companies are in trouble using their existing routes,” says Coveney. Drugs often don’t work for significant subsets of the population. “It’s not one size fits all.”
One of the goals of the Virtual Physiological Human project is for people to participate in their own medical care, simulating the outcomes of certain choices and even self-diagnosing. But with control comes responsibility. Will we listen to our simulated self?
We already implore people to act in their own interests, says at New York University. He studies the way people make decisions about their future, such as whether to buy an annuity. But appeals that rely on self-interest often don’t work, he says.
However, studies have shown that presenting people with a virtually aged version of themselves can be an effective way to persuade people to change their lifestyle. A digital body double could be even more effective, says Hershfield. “It could show you exactly what’ll happen in future,” he says. But too much information can backfire. “There are plenty of legal and ethical issues raised by all of this which are quite new,” says Coveney. For a start, each of us will probably have to decide how much we want to know.
What if my simulation told me that I had a high risk of dying very soon? How would that affect my rational decision-making? “That’s a tough one,” says Hershfield. “Risk surrounding mortality is tricky. People don’t like to think about that.”
“What if my simulation told me I had a high risk of dying soon? How would that affect my decisions?”
This article appeared in print under the headline “My virtual body double”
