Dogs have an Extraordinary Sense of Smell, Capable of Detecting Medical Conditions in Humans
If you are a dog person, you know your canine friends sometimes seem to know you better than yourself.
Scientific research appears to bear this out.
For example, researchers in the UK conducted a study (doi.org/10.1038/s41598-024-66147-1) that confirms that dogs have strong reactions to the isolated odor of stressed individuals but remain calm when presented with the odor of more relaxed people.
A follow-up double-blind study (doi. org/10.1371/journal.pone.0274143) sought to quantify these findings and found that dogs in the study were able to correctly identify the odor of stressed people 90 to 96% percent of the time.
What Other Human Diseases or Medical Conditions Can Dogs Detect?
Intriguingly, dogs can do more than assess our stress levels – they also appear capable of identifying cases of clinical depression.
Researchers have also shown they can also detect the presence of serious medical conditions, including chronic diseases such as cancer. Among different types of cancers, dogs appear able to detect melanomas (skin cancers), breast cancer, lung cancer, ovarian cancer, and prostate cancer with varying degrees of accuracy.
In some cases, the dogs were able to successfully identify a condition by sniffing the patient directly while in other instances they were able to identify the presence of a medical condition by smelling bodily fluids produced by the patient, such as urine.
Other research has found that dogs can identify other serious diseases as well, including cases of Parkinson’s, tuberculosis, malaria, or the presence of certain viruses, such as Covid 19.
Dogs’ Unique Physiology Allows Them to Identify Smells Millions of Times More Accurately than Humans
It’s common knowledge that dogs have an incredible sense of smell (with some breeds, such as bloodhounds, up to 100 million times more sensitive to smells than humans), making them ideal for assisting in sniffing out the presence of drugs and explosives as well as finding humans (or human remains) during disaster rescue and recovery operations.
What makes a dog’s sense of smell so accurate compared to that of humans?
First of all, dogs are very interested in detecting novel odors, a phenomenon known as neophilia.
Another key factor is the way dogs perform rapid breathing when they are trying to smell – they inhale and exhale air through their nose up to 5 times a second! (Unlike humans, dogs have two pathways in the nose, allowing them to pull samples of air in and out of one pathway without breathing through the lungs, which is much slower.) Dog noses also have vents to direct the exhaled air backward, away from their nose, allowing new fresh air samples to enter the nostrils.
Once the air samples enter the dog’s nose, they pass to a large number of canine olfactory receptors (ORs), e.g. smell sensors.
Dogs have humans beat in this department as well.
While humans have around 5 million olfactory receptors crammed into an area roughly the size of a quarter, dogs can have between 125 to 300 smell sensors, depending on the breed, comfortably spread out across several square inches.
Once the olfactory receptors pass the information to the brain, dogs have the advantage again. About 1/3 of a dog’s brain is dedicated to decoding smells, compared to only around 5% of the human brain.
Why Animals Aren’t Used to Detect Human Diseases in the Clinic
So why aren’t animals on the frontline of human disease detection?
After all, one of the first instances (doi.org/10.1016/S0140-6736(89)92257-5) of a dog detecting a human disease (melanoma) to be published in a scientific journal dates back to 1985.
The problem with using live animals at scale (as the Silicon Valley bros like to say) is it becomes a little problematic. In the case of dogs, each sniffer dog must be trained, and the results are not always consistent nor accurate enough to be medically useful in some cases.
For example, in a double-blind study funded by the Bill & Melinda Gates Foundation, researchers found that dogs were able to identify children infected with malaria parasites only 70 percent of the time. Useful to be sure, but what about the 30% of kids whose condition went undetected?
Could other animal species be useful instead?
Researchers have also been investigating rats, which also have a powerful sense of smell, to see if they could be a better fit.
In one study published in Nature (doi.org /10.1038/s41598-021-81086-x), researchers found that African giant pouched rats (Cricetomys gambianus) – exposed to human sputum taken from patients with pulmonary tuberculosis (TB) – were quick to identify the presence of TB, but the accuracy was disappointing – the rats were only correct 81% of the time.
Researchers at Michigan State University have also investigated (doi.org/10.1016/j.bios.2024.116466) if honeybees can accurately identify whether patients had a case of either non-small cell lung cancer (NSCLC) or small cell lung cancer (SCLC) based exposure to odorants (e.g. biomarkers) in human breath.
The honeybees, held in place with 3D-printed work jackets, were able to identify the presence of lung cancer with an accuracy rate of 93% or greater and could correctly distinguish between the two different types of cancer 82% of the time.
Unfortunately, while the honeybees can produce results in near real-time, they have a limited “work span” and can only process about 100 samples before losing their abilities.
Can We Replicate a Dog’s Sense of Smell in the Laboratory with Sensor Hardware Connected to AI?
In response to the problems of using live animals to detect diseases in the clinic, researchers have been investigating the possibility of creating artificial noses, e.g. bioelectronic noses, based on the mechanisms of the highly sensitive olfactory receptors (ORs) found in species such as dogs, rats, and honeybees.
Early iterations of this idea have resulted in experimental electronic bio-detection devices that could detect a single disease, such as tuberculosis (doi 10.1016/j.jinf.2017.08.003).
However, recent advances in AI, such as machine learning (ML) and Large Language Models (LLMs) are encouraging researchers to marry the bio-detection hardware with dedicated AI systems that can analyze the results from the sensors, hopefully creating much more accurate devices that can be reliably deployed in a clinical testing environment.
One recent example is work undertaken by researchers at MIT, John Hopkins, and the Medical Detection Dogs organization in the UK to identify cases of prostate cancer.
In a paper published in PLOS One, the team developed a miniaturized detection device to detect volatile organic compounds (VOCs) in patient urine samples in a manner that mimics how dogs process these smells. After the readings were taken, the researchers set up an artificial neural network to train the device – based on data gathered by two dogs participating in the experiment – to identify Gleason 9 prostate cancer in urine samples from patients who had biopsies confirming their cancer diagnosis.
This combination of olfactory hardware sensors combined with machine learning holds great promise.
We will likely see devices come on the market that can perform a variety of clinical tests in the near future.
Formaspace is Your Laboratory Research Partner
Evolving Workspaces. It’s in our DNA.
Talk to your Formaspace Sales Representative or Strategic Dealer Partner today to learn more about how we can work together to make your next construction project or remodel a success.
