Last updated¬†07:11, June 26 2018
It may be an overstatement to describe him as a prodigy, but young Onyx’s¬†nose for¬†academic research¬†is undeniable.
At the age of 3, he¬†already shares¬†company with university lecturers and graduate students.
But at the heart of Onyx’s¬†success¬†is less a hunger for knowledge¬†than a hankering for doggy snacks.
Inside a quiet, nondescript room at Waikato University, the burly¬†black labrador¬†is being presented with a series of¬†choices.
*¬†Waikato woman has a soft spot for the giant rats detecting landmines and tuberculosis
*¬†Waikato health projects get funding boost
*¬†Waikato University lecturer awarded $238,000 for transgender health survey
To the casual observer, the exercise seems rather puzzling.
Six months of training has taught Onyx to stick his muzzle inside¬†a hole in a white screen.
Behind the screen, out of sight, is a purpose-built machine dubbed an automated canine scent-detection apparatus.
The machine contains¬†17 sealed¬†chambers that can be accessed¬†via a flap behind¬†the screen hole.
Inside five of the chambers is a scent ‚Äď¬†in this case, amyl acetate, an organic compound that gives off a smell¬†similar to bananas.
If Onyx¬†holds his nose inside the chambers containing¬†the scent¬†for a set period of time¬†‚Äď¬†about five seconds ‚Äď¬†it’s considered a positive indication.
That then triggers a doggy treat¬†to be dispensed from a nearby pet feeder.
The machine’s¬†chambers are rotated by the dog hitting a lever with his muzzle.
A chamber without a scent ‚Äď¬†termed¬†a negative sample ‚Äď¬†carries no food reward.
Onyx, eager to get to the treats, doesn’t waste time on the negative samples and quickly pulls his nose out of the chambers that lack the desired scent.
After five¬†minutes, the exercise is over, and Onyx has correctly identified all five scented chambers.
The labrador happily consumes his rewards¬†while, in the next¬†room, researchers¬†Tim Edwards¬†and Margaret Crawford¬†share a knowing¬†smile.
The pair have been watching¬†Onyx via a camera and say¬†the labrador’s¬†impressive¬†showing isn’t¬†a surprise.
“It’s now normal for him to be completing this¬†task with 100 per cent accuracy,” Edwards says.
But as entertaining as it is to see Onyx’s¬†amazing sense¬†of smell in action, the exercise has an important scientific purpose.
Earlier this year, Edwards, a senior lecturer in psychology, was awarded a $233,607 grant from the Health Research Council of New Zealand to study how dogs can be used for lung-cancer screening.
The 36-month study will involve Onyx and six other¬†dogs sniffing¬†saliva and breath samples collected from patients attending Waikato Hospital’s respiratory clinic.
Many of the patients¬†have been diagnosed with lung cancer.
To date, Edwards and his team’s focus has been on¬†training the¬†dogs¬†to¬†use the automated canine scent-detection apparatus.
Last year, Edwards received a $30,000 grant from the Waikato Medical Research Foundation which kick-started his research.
In the next few months, the dogs will be introduced to the medical¬†samples.
While Onyx breezed¬†through the five-minute exercise, it¬†represented months of training¬†‚Äď¬†and patience.
A newer recruit to the¬†project¬†is three-year-old Bramble.
The english springer spaniel is regarded as a “work in progress”, but¬†already she’s demonstrating two key attributes Edwards is looking for:¬†an excellent temperament and a willingness to work. A range of breeds are involved¬†in the study.
Edwards’ optimistic assessment of Bramble is shared by her owner, Richard¬†Shephard,‚Äč¬†who describes her¬†as a “big sniffer”.
“She loves doing anything and is always excited to come here,”¬†Shephard¬†says.
“When we heard about Tim’s research using dogs, we thought it would give Bramble something interesting to do during the day. And also it’s a chance to help other people. It just seems silly not to use the skills that dogs have got, to be honest.”¬†
Several studies around the world have¬†shown dogs’ ability to identify cancer, but few have¬†involved methods that can¬†be used in a clinical setting.
Edwards estimates it will take another six months to fully train all the research dogs.
The Waikato University study will also be¬†the first of its kind to use saliva samples in its research.
“One of the reasons we’re keen to use saliva samples is the collection process is so much easier,” Edwards says.
“Even though patients don’t have to breathe forcibly into our collection¬†tubes, it’s still asking someone with a lung condition to breathe into something. And when it comes to analysis, it’s much easier to work with a liquid, rather than something that’s trapped on an absorbent material.”
Although reports¬†exist¬†that suggest¬†dogs naturally act differently around cancer-related stimuli, Edwards says¬†he and his team aim to¬†train the dogs to learn the difference between lung cancer and non-cancer samples.
The dogs will be rewarded with a food treat for detecting¬†lung cancer-positive samples.
If a dog singles out a non-cancer sample, there is no reward.
A common question Edwards gets asked is what will Onyx, Bramble and the other dogs be smelling when they detect a lung cancer sample?
The answer is it’s likely to be a “whole bouquet of compounds”, he says.
A chemical analysis of breath or salvia samples from a lung cancer patient could show 20 to 30 different chemicals. Most¬†of these chemicals won’t be exclusive to lung cancer patients.¬†
“It’s really an exercise in concept formation,” Edwards says.
“It’s like the way some people¬†can identify a specific make and model of a car without seeing its logo. Because they’ve seen the car from several different angles before, they¬†can tell what it is without necessarily being able to say why.
“If you give a dog a whole different bunch of lung cancer examples and a whole bunch of non-examples, and teach them to distinguish between the two, they’ll eventually get the concept of lung cancer.¬†
“There may be no one or two or three identifying features. Instead the dogs will have to be able to figure it out intuitively.”
Crawford compares teaching a dog to identify lung cancer to teaching someone to appreciate a fine piece of craftsmanship.
Eventually, after enough exposure, a¬†person can identify a fine object without necessarily being able to point out what they’re responding to.
Edwards, who grew up on a ranch¬†in¬†Utah, in the United States,¬†has been based at Waikato University for the past three years.
Previous to that, he worked for a humanitarian organisation¬†in Tanzania, studying the ability of¬†giant African pouched rats to detect¬†tuberculosis.
After arriving at Waikato University, Edwards set up a scent-detection research facility at the campus. The automated canine scent-detection apparatus is his own creation.
He’s currently writing a full description of the device, and intends to publish the article¬†in a peer review journal. That way, the technology will be available to anyone to use. The move will also prevent anyone from putting a patent on it.
Although working with pet dogs in a research facility in Hamilton seems far removed from studying giant rats in Africa, Edwards says his previous research work will benefit¬†his current study.
“One of the main things I learnt from my time in Tanzania is you just can’t walk into an industry, show them some flash data, and expect them to all fall over themselves to work with you,” he¬†says.
“With cancer detection, we need to work with the people who are doing the diagnosis services and find out what the challenges are, how a product could potential help them, and find out how we can make their job easier.”
The hope is that Edwards and his team’s research can help the development of machine-based sensor technology.
“Everyone’s goal is that we can get to a stage where you can breathe into a small device and find out fairly quickly and accurately whether or not you have condition A, B or C.
“In the meantime, if our testing works well enough, and the technology is just not there yet, then using dogs could potentially work as a stopgap measure.”
Each year about 1600 Kiwis¬†die from¬†lung cancer, making it the leading cause of cancer-related death in New Zealand.
By comparison, about 600 people die from¬†breast cancer each year and 300¬†from melanoma.
Waikato Hospital respiratory physician Cat Chang says¬†there is a great need to develop an affordable and non-intrusive method of screening for lung cancer.
Many people in¬†the early stages of lung cancer don’t exhibit any symptoms.
Common¬†symptoms of lung cancer¬†include a cough (often with blood), weight loss, wheezing and chest pain.
“Lung cancer is our biggest cancer killer for the reason that most people present late, when the disease is advanced, and we can’t do as much to treat these people,” Chang says.
“To improve treatment for people with lung cancer what we need is early detection. With cervical cancer, women can have a smear test, and with breast cancer we have mammograms. But with lung cancer, we don’t have a good screening tool.”
Ideally, such a¬†tool would be non-invasive, accurate and cheap to perform.
That’s where Tim Edwards and his¬†dogs come in.
Chang says that, throughout the history of medicine, doctors have¬†used¬†breath to help diagnose medical conditions.
“What we hope is these¬†dogs can detect tiny amounts of chemicals in people’s breath that might be associated with lung cancer and hopefully be able to do this when the cancer is at an early stage. And of course, breath is easy to get and it’s non-invasive.”¬†
To date, about 200 patients attending¬†Waikato Hospital’s respiratory clinic have given¬†breath and saliva samples for Edwards’ research project.
Despite smoking rates in New Zealand declining in recent decades, Chang says¬†it had yet to translate into a drop in lung cancer rates.
Currently, 16 per cent of adult New Zealanders smoke. And although smoking is a leading cause of lung cancer, the disease also impacts non-smokers.
The growing popularity of electronic cigarettes also has some health professionals concerned.
“We don’t know the long-term effects of vaping,” Chang says.
“The short-term effect from animal studies and biopsies from healthy volunteers that have vaped show definite cell damage in the lungs, similar to what you get if you were smoking a cigarette.
“We don’t have long-term data, but the very short-term data we have on vaping is that it’s not benign.”¬†
If Edwards’ study shows dogs can reliably sniff out lung cancer, Chang says there’s no reason why the screening tool can’t be applied to other cancers, such as prostate cancer.