The newspaper of Imperial College London
 Issue 148, 19 January 2005
Taking Imperial from strength to strength«
UK-Thai scientific collaboration boosted by new agreement«
Cirque du Soleil in the main entrance«
A nose job«
Frizzy hair today, gone tomorrow«
New microscope gives boost to UK nanotechnology«
Lord Sainsbury visits Imperial«
Imperial leads the way in surgical training and innovation«
New programme will train next generation of health leaders«
Tea off to good health«
Success halts trial«
The perfect Formula«
Spotlight on new R&D solutions«
Imperial students are best trainees«
Cash boost for Wye’s top new scholars«
In Brief«
Media mentions«
What’s on«

A nose job

Snorting and sniffing just won’t be the same after Imperial researchers finish their detailed mapping of the airflows in and around the human nose, writes Tom Miller.

A small increase in flow rate disturbs the simple laminar streaming flow and breaks it up into eddies deep in the meatus region (flow enters on the right)

A ride up the nose is not a smooth experience. From a particle’s perspective even the gentlest inhalation makes it a bumpy airway to travel. It sends air up through the nostrils, whirling around narrow anatomical spaces on each side. Some currents take a longer route, eddying and recirculating inside the nose, while the fastest whoosh through in just two hundredths of a second.

In this short time it’s the nose’s job to warm, humidify and filter air, purifying it for the lungs, while also performing an indispensable sensing job.

Two years ago researchers at Imperial started to build highly detailed 3D models of the nose. Their fusion of biological mechanics and aeronautical engineering, funded by a BBSRC grant, has since led to experimental and computational models more accurate and useful than anything attempted previously.

They reveal the complicated geometry of the living nose and its airflows. This information may soon be used to help ear nose and throat surgeons plan their operations, as well as helping pharmaceutical companies determine the best way of administering drugs through the nose-either to unblock the nose itself or deliver them directly into the bloodstream. And because their modelling is so detailed, their work also has implications for animal testing: it goes far beyond what animal models can tell us.

“Our main effort is working out where the airflow goes,” says Dennis Doorly, from the department of aeronautics, and one of the two principal investigators. “From quiet breathing to rapid sniffing, we want to know exactly what’s happening.”

Scientists can’t make direct measurements inside the nasal structure of any animal because it’s too small. A probe up the nose to measure airflow disturbs the local airflow, so they have to work remotely.

Their solution now sits in a lab in South Kensington. Exposed in clear see-through silicone is a perfect copy of one half of one of the most intricate spaces we know in the human body. The scroll-like airspaces around the tiny ‘conchae’-Latin for shells-can just be seen, projecting from the septum in the centre of the nose into the main cavity, forming the air passageways known as meatuses.

The Imperial team began their model with sets of CT scans from anonymous, ‘nasally healthy’ patients at St Mary’s hospital, where their clinical colleagues work.

Compiling together the CT slices, they determined a precisely defined airspace-the inverse of the actual nasal geometry-from which they generated a set of instructions for building a 3D map, using a technique called stereo lithography. These map instructions were then taken to a 3D printer-a rapid prototyping machine-which turned the model out in a soft plaster material. From this they cast the inverse airspace model in optically transparent silicone.

Perched upside down, the transparent double-size nose cast sits with inlet and outflow water pipes at either end. Into it they pump water or a liquid that is matched for refractive index to the silicone. A fluid is used to replicate the flow physics, explains Doorly, because it slows down the speed of the particles ten-fold without losing realistic movement detail.

“It’s probably one of the most complex areas of fluid mechanics in the whole of the body, more complex even than the heart,” says Bob Schroter, department of bioengineering and PI, who has spent a lifetime researching biological systems from an engineer’s perspective.

The complex geometry gives rise to some formidable fluid dynamics. “People are used to the flows around an aeroplane being complicated but that is in some ways simpler than understanding the flows inside the nose,” says Doorly. The nose’s geometry is arbitrarily complex, with no straight lines or curves like a wing, and it has an awkward regime of flow that is neither simply laminar nor simply turbulent, but parts of both.

When the flow is turned on, simulating ‘inhalation’, tiny coloured beads are injected and a high-speed digital camera opposite begins recording. The journey of the beads as they pass into the opening of the nose, through the anatomical landmarks in the meatuses and around the conchae is all captured in glorious detail, and analysed using particle imaging velocimetry.

Our sense of smell relies on a sample of air reaching the olfactory bulb near the top of the nose. For that, one needs the air velocity of a deeper sniff. As air reaches the bulb, the geometry causes the jets to eddy, recirculating air around the olfactory bulb. The model captures this motion-long assumed by scientists but never actually seen-in clear detail.

The experimental model is partnered by a huge computational fluid dynamics model that makes use of over 20 million elements, handled by computers from the London e-Science Centre based at Imperial.

The experiment tells us what is happening with confidence, but for Schroter, the two approaches go hand in hand. “The computational work is highly seductive,” he says, “But you have to ground-truth it. You have to go out there and measure what happens in the real geometry.”

“Provided we keep these two tethered together then we can have faith in what we’re doing.”

Schroter and Doorly’s team have good reason to keep the faith. They recently CT scanned their first model ‘nose’ and found only one per cent difference between that and the patient’s original CT scans.

This article first appeared in the January issue of BBSRC Business magazine.

imperial front page | reporter front page | this issue's front page | feedback
©2003 Imperial College London