A GLISTENING slice of human brain rests delicately atop a shining metal pedestal, like a freshly served dessert. Lynne Bilston pulls a lever. A hefty weight suspended above falls and the brain is squashed flat. A “squishometer” records exactly how much it flattens. “Why not have it fall from higher up?” I ask, not expecting a serious answer. “If you do that the brain just flies out,” says Bilston nonchalantly. “Believe me, I’ve tried.” No doubt she has: this entire lab is cluttered with all sorts of sophisticated gadgetry for stretching, twisting and torturing soft, slippery objects.
Bilston, a biomechanical engineer at the Prince of Wales Medical Research Institute in Sydney, is one of a growing number of researchers engaged in the often messy business of squishing brains, jiggling breasts, stretching livers and even plucking fallopian tubes like banjo strings. Believe it or not, there is a reason for such gruesome activities. The information they yield – the exact mechanical properties of the body’s squishy bits – might just save your life one day.
Bilston and other researchers are using their findings to create detailed mathematical models of how organs behave under pressure or tension. Such models will allow engineers to develop a new generation of simulators to enable surgeons to practise operations. The simulator will be so realistic that surgeons will not only see what they are operating on, they’ll be able to feel just how hard to push with the scalpel or pull with the forceps. Accurate models could also give medical imaging a major boost: by measuring tiny changes in the elasticity of cancerous tissue, radiologists will be able to find tumours they would otherwise miss. Models could also improve our understanding of what happens during a car crash or landmine blast, and show how to better protect the body’s soft tissues in a host of activities such as running and skiing.
Advertisement
Labs like this are where medicine meets pure materials engineering. The instruments here are usually used for twisting tyres or stretching dough, which has a surprising amount in common with the brain. “They both have complex microstructures with long-chain molecules and they both have a fluid and a solid component,” says Bilston, who frequently collaborates with dough researchers during her work, and is now applying tried-and-tested principles of modelling dough to brains.
The fluid-solid duality, or viscoelasticity, of soft organs makes their mechanical properties fiendishly difficult to understand, since their suppleness varies according to how quickly they are squeezed or stretched. Press slowly on a liver, for example, and the water inside moves out of the way, allowing the organ to flatten. But apply the pressure quickly – as when a driver hits the steering wheel in a car crash – and that water has no time to flow. Now the liver behaves like a solid: it shatters. Between these two extremes the fluid-solid balance of the liver will vary, which is where things get tricky for the model-maker.
The most obvious solution is to record the behaviour of real samples in a range of situations and feed this data into the model. But even measuring the simple mechanical properties of soft tissue is tricky. Tests are usually done on small pieces of tissue taken from dead animals or cadavers, and extrapolating the results to the living can introduce serious errors. There’s the “fire hose effect”, for instance: the more blood is pumping through an organ, the stiffer it becomes. Clearly, cadavers have no blood flow. Nor do they breathe, which also affects an organ’s stiffness since the moving diaphragm constantly stretches and relaxes the liver and other organs inside the abdomen and thorax. And the mechanical properties of the gut vary depending on how full it is. Even the liver itself is slightly stiffer after meals, possibly due to the rush of glycogen molecules that fills its cells as digestion proceeds.
So researchers must concoct clever ways to probe the properties deep inside living things without cutting them up. Bilston, for instance, measures the elastic properties of healthy human brains by having volunteers bite on a vibrating mouthpiece. The resulting vibrations through the skull are then tracked using magnetic resonance imaging (see “Jiggling your body gives a better picture”).
Mapping the mechanical properties of every single part of the human body will be a truly mammoth task, and we are still in the early stages of assembling our “mechanical genome” – or perhaps that should be “squishome”. Yet scientists are already transforming the knowledge they have gained into palpable applications.
Canberra, Australia
I run my probe over the liver like a paintbrush, feeling its slippery, spongy texture. As I nudge one end, a tiny bulge ripples all the way to its far edge.
Even though I’m wearing virtual-reality goggles, this computer simulation looks and feels surprisingly real. It is based on a “finite element model”, the model that most organ-squishers aim to develop once they have nailed down an organ’s most important properties from experiment. Such models provide the basis for simulations from practising surgery to recreating the trauma of car crashes. Some exist only as heaps of mathematics to be crunched by supercomputers. Others have been translated into virtual but tangible blobs like the one I’m using, which was developed by Kevin Smith at the CSIRO, Australia’s national research organisation, in Canberra. Smith’s simulation divides the liver into 3004 triangles and calculates the forces and shape changes between each triangle and every other triangle in the organ.
In addition to producing a realistic computer image of how the organ moves when poked, the simulation also drives motors that control how easily I can move a physical rod that directs a virtual stylus, and hence how much resistance the virtual liver gives me as I press or brush against it. Such simulations with force feedback could be used to help train surgeons before they are turned loose on real patients. But before that can happen, the simulations will need to expand from a single organ to complete body cavities, and a few kinks will need to be smoothed out en route.
Palo Alto, California
Holding the mechanical claw, I try to control my quarry. But it bucks like a water buffalo, jerks demonically from side to side, nearly yanking the claw from my hands. Pretty strong for a fallopian tube.
This is a computer simulation of tubal ligation – clipping and blocking the fallopian tubes – a procedure usually performed by keyhole surgery through a tiny hole cut in the patient’s abdomen. But this morning something is wrong.
The Stanford researcher who is hovering over me steps in, reboots the computer, and next time it all runs smoothly. But the glitch illustrates a daunting challenge for surgical simulations. Determining the forces between each of 3000 points in an organ requires massive calculations. Doing that in real time while someone is pulling on a virtual fallopian tube is even trickier.
Our sense of touch is harder to fool than our eyes. Seeing just 24 still frames per second in the cinema creates the illusion of continuous motion, but to fool our sense of touch a simulator must update the forces we feel at least 1000 times a second. If the computer falls behind, the force feedback comes in spurts and jerks. The fallopian tube yanks back.
This means you have to compromise in order to make surgical simulations work, says Stanford surgeon and simulations specialist LeRoy Heinrichs. “But you can still reproduce things very well.” Rather than basing his tubal ligation simulator on complex mechanical properties that are measured from real tissues, Heinrichs is working by trial and error to make it simply feel right in the hands of experienced surgeons. And there are other ways to recreate the forces experienced during surgery. Some researchers, for example, record the actual forces using a modified surgical instrument based on an endoscope, and then feed them directly into a simulator.
Buffalo, New York
A team led by David Fineberg, a plastic surgeon who runs the living anatomy programme at the State University of New York in Buffalo, is avoiding computer trouble altogether by creating wet, mushy models rather than dry, mathematical ones. Working with sculptors and engineers, Fineberg’s team is building a prototype abdomen simulator to practise “open” surgery such as spleenectomy, where the surgeon’s hands (rather than instruments) are messing directly with the guts. “It will feel like you’re putting your hands in a living human body,” from blood and body warmth to the texture of each organ, says Fineberg.
The idea has a strong historical pedigree. Nearly 2500 years ago, the ayurvedic surgeon Sushruta recommended that medical students learning to probe patients’ bodies first practise on worm-eaten wood and that those learning to extract teeth first try withdrawing seeds from fruit. Fineberg’s simulation goes way beyond this. His team is fabricating organs out of polymers designed to have exactly the same mechanical properties as the real thing. Some organs will be constructed from several materials: for example, the lungs will have bronchial tubes made from firm, cartilage-like polymer, and air sacs made of spongier stuff. Along with nasty smells – to simulate a perforated intestine, for example – and implanted sensors that respond to touch, Fineberg will use “augmented reality” to make the experience as genuine as possible. This is done by overlaying computer-generated images, seen via special goggles, on top of real objects. Surgeons will feel the fake organs with their hands while what they see may be altered to show bleeding or cancerous growths.
It is a clever way to sidestep the number-crunching demanded by virtual-reality approaches. However, for those who study traumatic injuries, lifelike body parts are not the whole story. They may enhance training, but they cannot yet replace the awesome calculations required for detailed simulation.
Detroit, Michigan
Somewhere in a refrigerated room in Motown, a crash-test sledge shoots forward and a head slams into a windscreen.
But instead of the usual crude aluminium dummy, this head is human. It’s fresh: just 72 hours ago it lived, breathed and possibly even spoke. Since then, a team of scientists at Michigan’s Wayne State University has transformed it into a precise instrument for measuring the split-second acrobatics of a brain during impact. Twenty markers have been implanted inside the head and brain, and during the crash their movement is captured in three dimensions by X-ray cameras.
Head-banging is a tradition here at Wayne State. In 1939, two professors dropped cadavers head first down the medical school’s stairwell to find out how big an impact was needed to fracture the skull. Thus were born the head injury criteria (HIC), which in part remain the standard for safe car and helmet design.
Methods for calculating the HIC are a little more sophisticated now, and the Wayne State project aims to update them further. King Yang and Albert King, who run the project, have devised the world’s most sophisticated model of the brain, which slices and dices it into 300,000 chunks. It requires 20 hours on a supercomputer to simulate an impact, but unlike the surgical training simulations these calculations do not need to happen in real time. The model is mainly based on mechanical properties measured in bits of brain, with the gruesome head-impact experiments used to verify that the predictions hold true for the real thing.
So far Yang and King have used their model to replay head impacts from NFL football games in the US, and they are collaborating with Laurie Sparke, an automotive engineer at Holden Innovation in Melbourne who designs head protection systems, to reconstruct car crashes in Australia. And so far, the results are good. Their model predicts 95 per cent of serious brain injuries, a 15 per cent improvement over the present HIC.
Salisbury, Australia
Alexander Krstic, a ballistician with the Defence Science and Technology Organisation in Salisbury, uses synthetic organs to take injury a step further. While car impacts involve millisecond contortions, the military blasts that Krstic studies occur a thousand times faster. With such time frames, says Krstic, you’re playing in a different ballpark. Complex models reduce to a simple distinction between “organs that bounce and organs that splat”.
Much like Fineberg, Krstic uses physical models of each component – muscles, tendons, lungs and so on – made of materials with the same properties as the real organ. The model even includes fake blood as well as sensors that record shock waves travelling through flesh, strains on ligaments and split-second spikes in blood pressure. And when it’s all over, military surgeons examine the carnage.
An important area of research is landmines. To test the boots given to soldiers to protect them from mines, Krstic has developed an artificial leg. He shows me a video of one such test. Ten millionths of a second after the leg sets the mine off, all looks good. But 30 millionths later, something is amiss: the blast-proof boot sole, still intact, has been driven straight up to where the ankle was. And another 30 millionths later the sole is halfway up to the knee – yet the knee itself hasn’t moved. Expanding gases from the blast have lifted the sole at 5000 metres per second. Some of that force would normally reach the torso, but since leg muscles are soft and bones bendable, forces can only be transmitted along them at a third that speed, so the lower leg absorbs all the damage before the rest of the body even feels it. The boot sole survives unscathed, but the sides of the boot rupture, providing the shattering leg with a place to go.
It is not pleasant viewing, even though I know the limb is artificial. But the work Krstic and others are doing should one day help to improve the boots and minimise the damage caused by landmines. Understanding the squishability, twistability and stretchability of the body’s soft bits may be a messy affair, but is essential to improving our knowledge of what is going on inside us, and keeping it there.
Jiggling your body gives a better picture
One day, radiologists will screen for tumours by jiggling our innards like plates of jelly. The technique, called elastography, is like the centuries-old method of palpating the body to find odd lumps, except it can feel deep inside the body without an incision. Instead, it measures the elasticity of tissue. Sideways vibrations are applied to the area under investigation, and then magnetic resonance imaging is used to watch the waves moving through it. Tumours will be stiffer than the surrounding tissue and so the waves will diminish as they pass through.
“Every day,” says Richard Ehman, a radiologist at the Mayo Clinic in Rochester, Minnesota, “surgeons reach in through incisions and feel tumours that we’ve missed with every one of our existing imaging techniques.”
The key is that stiffness varies so widely between tissue types: tumours are 10 to 100 times as stiff as healthy tissue. But this variation only shows up as a fourfold difference in density using conventional body imaging techniques.
Another advantage of elastography is that since malignant tumours are even stiffer than benign ones, it may be possible to differentiate between them without a biopsy. Ehman is experimenting with elastography to spot breast tumours, while others are using it to find prostate tumours or precarious bottlenecks in arteries that may cause heart attacks. It might even be able to spot diffuse abnormalities like the nearly invisible beta-amyloid plaques of Alzheimer’s disease.