An Artificial System Inspired by Dogs Smells Disease

Researchers at MIT and other institutions have developed a miniaturized detection system that seems to detect diseases with even greater sensitivity than a dog's sense of smell.

Dogs are able to smell many things that humans cannot. They have around 300 million scent receptors, compared to around 5 million in humans. And experts have trained dogs, including well-known breeds such as Labrador Retrievers and German Shepherds, to detect and alert us to the presence of disease.

Some service dogs are trained to sniff skin, breath, or urine samples with a very high success rate. For example, medical detection dogs are smelling breast cancer and prostate cancer at a rate of 97% or higher. They are detecting Parkinson's Disease at a rate of up to 90%. And diabetic alert dogs can detect low blood sugar at a rate of 70% on average.

"Dogs, for now 15 years or so, have been shown to be the earliest, most accurate disease detectors for anything that we've ever tried," said MIT researcher Andreas Mershin. "So far, many different types of cancer have been detected earlier by dogs than any other technology."

The new miniaturized detector system is described in a research paper published in PLOS ONE. It incorporates mammalian olfactory receptors that act as sensors. The data captured by the sensors can be handled in real-time by a typical smartphone.

The researchers tested urine samples from confirmed cases of prostate cancer and controls known to be free of the disease. They used both dogs, trained and handled by Medical Detection Dogs in the U.K., and the miniaturized detection system.

The study confirmed that the sniffer dogs had "enormous potential” to detect cancer, said Clare Guest, founder of Medical Detection Dogs, as reported by BBC News. "The dogs have been able to identify these very aggressive cancers. This could lead to lifesaving work in the future that would enable us to understand the difference between other diseases of the prostate and those that will go on to kill men."

But the new detection system has enormous potential as well. Actually, the detection hardware is 200 times more sensitive than a dog's nose in terms of being able to detect and identify tiny traces of different molecules. This was confirmed through controlled tests mandated by DARPA.

However, in terms of interpreting detection results, the hardware is “100 percent dumber" than a dog, said Mershin.

Emulating the nose of a dog is not enough. One has to also emulate the brain of the working dog, which integrates and interprets the molecular signals sensed by the nose.

To do this, the researchers used a machine-learning program based on an artificial neural network. The combination of olfactory sensing and machine learning (“machine olfaction”) was able to match the success rates of the dogs. Both methods scored more than 70%.

“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,” said Mershin. “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.”

The researchers are persuaded that their work could someday lead to automated disease detection systems for smartphones.

As a dog lover, I’m happy to see that our four-legged best friends inspire breakthrough medical technologies.