Advertisement

Good boy! Robot ‘sniffer dog’ trained to detect cancers

Dogs have detected diseases that doctors have missed. Can a gadget replicate their success?

Dogs have detected diseases that doctors have missed. Can a gadget replicate their success? Photo: Getty

Could robot sniffer dogs really be the way of the future?

US scientists are trying the replicate the proven cancer-detecting skills of furry dogs, with artificial intelligence.

In experiments, dogs have successfully identified lung, bladder, ovarian, breast and prostate cancers – and they’re being trained to detect COVID-19 at Australian airports.

How successful are they?

Dogs have detected prostate cancer with 99 per cent accuracy. They tend to do a better job than standard tests.

In some cases, dogs have detected cancer in a control group – where the participants were thought to be healthy. And they’ve identified multiple cancers in the one patient from a single urine sample.

Which is great. But training dogs to root out disease takes time and money – and having a canine unit in a hospital setting is thought to be impractical.

Hence the robot project.

Robot sniffer dogs sure do have a sensitive nose

Dr Andreas Mershin is a research scientist at the Massachusetts Institute of Technology’s Centre for Bits and Atoms. He’s a co-author of a new machine learning study that sought to mimic a dog’s cancer-detecting abilities.

Dr Andreas Mershin and Florin, a cancer-detecting dog used in the study. Photo: MIT

“Dogs, for 15 years or so, have been shown to be the earliest, most accurate disease detectors for anything that we’ve ever tried,” Dr Mershin said in a prepared statement.

“So far, many different types of cancer have been detected earlier by dogs than any other technology.”

For a number of years, Dr Mershin’s team has worked on developing a system that can detect “the chemical and microbial content of an air sample with even greater sensitivity than a dog’s nose”.

The system incorporates mammalian smell receptors that have been “stabilised” to act as sensors

The system is actually “far more sensitive” than a dog’s nose in terms of detecting the presence of tiny traces of different molecules – but in terms of interpreting those molecules, “it’s 100 per cent dumber,” Dr Mershin said.

In other words, the system can smell out the molecules, but isn’t capable of identifying the disease.

They don’t recognise in the molecules the patterns that dogs can infer from a scent.

“The dogs don’t know any chemistry,” Dr Mershin said.

“They don’t see a list of molecules appear in their head.

“When you smell a cup of coffee, you don’t see a list of names and concentrations – you feel an integrated sensation. That sensation of scent character is what the dogs can mine.”

AI to the rescue

You’ve probably read about skin cancers being identified by artificial intelligence more accurately than doctors.

For this to happen, millions of photographs of different cancers were analysed by machine learning algorithm.

The artificial neural network essentially learnt what different skin cancers “look like”.

In its study, the MIT team has done something similar with urine samples taken from men with confirmed cases of prostate cancer, as well as control samples taken from men known to be free of the disease.

Throw in some high-tech chemical analysis and microbial profiling … and woof!

The machine-­learning program teased out the similarities and differences between the samples that in turn helped the sensor-based system identify the disease.

The performance of two dogs trained in cancer detection was compared to the performance of Dr Mershin’s system. They both achieved accuracy rates above 70 per cent.

“We knew that the sensors are already better than what the dogs can do in terms of the limit of detection, but what we haven’t shown before is that we can train an artificial intelligence to mimic the dogs,” he said.

“And now we’ve shown that we can do this. We’ve shown that what the dog does can be replicated to a certain extent.”

Dr Mershin is working to shrink down the technology, until it fits inside a smart phone.

He envisions a day when phones “are routinely equipped with scent detectors that could pick up early signs of disease far sooner than typical screens – and could even warn of smoke or a gas leak as well”.

Stay informed, daily
A FREE subscription to The New Daily arrives every morning and evening.
The New Daily is a trusted source of national news and information and is provided free for all Australians. Read our editorial charter
Copyright © 2024 The New Daily.
All rights reserved.