How Do Dogs Sniff Out Cancer?

Published Apr. 1, 2015

A curious headline popped up along my Twitter feed the other day: “Could Dogs Sniff Out Thyroid Cancer?” I read the words and paused a few seconds, contemplating taking the bait before opening the link.

Convinced I would be disappointed in what I was about to read, I contemplated how a dog could be able to detect cancer given the complex nature of the disease and how troubling it is to uncover even under the best of circumstances. I figured the title was just a catchy way to drive readers to an advertisement for something completely banal like air fresheners.

One the other hand, what if it were true? What if dogs were really able to pick up on the subtlest of changes in our biochemistry, leading them to distinguish those of us with disease from those without? What if doctors could somehow capitalize on a dog’s powerful sense of smell and bypass the need for invasive diagnostics? How remarkable would that be?

I clicked the link.

To my surprise, the sensational headline was completely legitimate. In early March 2015, during the 98th annual meeting of the Endocrine Society, a group from the medical school of the University of Arkansas presented a research abstract entitled “Scent-Trained Canine Prospectively Detects Thyroid Cancer in Human Urine Samples.”

As if that weren’t fascinating enough, this presentation was actually a follow up to a previous study by the same group that demonstrated that dogs could reliably discriminate between urine samples obtained from patients already diagnosed with either metastatic thyroid cancer or benign thyroid disease.

What I wouldn’t have given to be a fly on the wall during that conference just to listen to this fascinating topic!

In the study, a single dog (which an unconfirmed source reports is a German shepherd mix named “Frankie”) was trained to either lie down when he detected the presence of papillary thyroid cancer (PTC) in the urine sample, or turn away or do nothing if the sample was ‘clear.’

Urine was collected from 59 human subjects who presented for evaluation of one or more thyroid nodule(s) suspected to cancerous. In the meantime, Frankie was “imprinted with urine, blood, and thyroid tissue obtained from multiple patients with PTC, and trained over 6 months to discriminate between PTC and benign urine samples.”

During the experiments, a gloved handler, lacking information about the diagnosis of the person providing the sample, presented Frankie with the urine samples. Frankie sniffed the samples and responded with the cues above. The handler verbally communicated Frankie’s response to a blinded study coordinator. Control samples (both cancerous and benign) were interspersed with the unknown samples and Frankie was rewarded with positive reinforcement when his answer was correct.

Frankie’s diagnosis matched the final surgical pathology diagnosis in 24 out of 27 cases (92.3% correct, 2 false negatives and 1 indeterminate), yielding a sensitivity of 83.0% (10/12) and specificity of 100% (14/14). Not too shabby for a four-legged ball of fur that never graduated more than a basic puppy training class!

In all seriousness, the most fascinating aspect to me is that the researchers have no idea what the dog actually smells to trigger the response. There clearly must be a chemical present that is excreted by affected individuals. However, research thus far has failed to identify this particular biomarker.

Much energy and effort in medicine is spent on early disease detection and veterinary oncology is gaining much ground in this aspect of medical care. We routinely recommend preventive screening diagnostics in order to discover illness at an earlier stage. We model our testing algorithms off those presented to our human counterparts.

But what if the reality is that we simply need to learn how to listen to our animals in a different way to understand their capacity for communication about their health?

Veterinarians lament the lack of ability to communicate with our patients and their inability to tell us where it hurts. It seems maybe we just need to heed their warnings a little harder. 

The old wives tale of a cold, wet nose indicating a healthy pet may not be as farfetched as we presume. How wonderful would it be if man’s best friend were also the best advocate for not only their health, but for that of their owner? 

I suppose maybe Frankie's nose knows the best answer to that question.

Dr. Joanne Intile

Image: Rita Kochmarjova / Shutterstock



Help us make PetMD better

Was this article helpful?


Get Instant Vet Help Via Chat or Video. Connect with a Vet. Chewy Health