How Dogs' Sensitive Noses Could Change Cancer Diagnosis
4/25/2017, Kathryn Bowers, Slate
This article is part of Future Tense, a collaboration among Arizona State University,New America, and Slate. On Thursday, April 27, at noon, Future Tense will host an event in Washington, D.C., called “Do We Need to Stop Talking About ‘Curing’ Cancer?” For more information and to RSVP, visit the New America website.
If you could “see” a smell, it would look something like a drop of food coloring dispersing through a cup of water with intense knots in some places and misty tendrils around the borders. Olfaction scientists call these “odor landscapes”—it’s why sometimes you get a whiff of a nasty smell while your cubiclemate sits blissfully unaware. Moving through these odor landscapes with a human nose is relatively unremarkable. Unless you’re near a bakery or a dirty diaper, a perfume counter or a freshly varnished bookcase, you’re unlikely to notice the many scent fields continuously curling and colliding around you.
Equipped with a dog's nose, however, your experience of the world would be completely different. Dogs' senses of smell -- tens of thousands of times more acute than ours -- give them an sensual experience as vibrant and full of information and nuance as what we get through our eyes.
We humans have harnessed this canine aptitude by training our best friends to find hidden drugs and bombs, injured people and stowaways, and contraband fruit at airports. A few years ago, scientists and veterinarians started investigating another application: using dogs to sniff out cancer.
Canine cancer detection began, as science often does, with anecdote. Veterinarians and oncologists semi-routinely heard stories of dogs who licked or sniffed a little too intently at some part of their owners’ bodies. A thigh, a breast, the back of the ear: It would turn out that the site of this overnuzzling contained a malignancy. The first published case of the behavior was in a 1989 Lancet letter describing a border collie–Doberman mix who sniffed and bit so aggressively at a mole on her owner’s leg (even through her trousers) that the owner was moved to check with her doctor, who diagnosed a malignant melanoma.
The oncologist who wrote the letter hypothesized that tumors, “with their aberrant protein synthesis,” “emit unique odors” which dogs can detect. In the decades since, studies have supported the claim that dogs can identify cancers by their smells, which are caused by the volatile organic chemicals in the waste products the cells release.
According to the research, dogs have been able to detect a range of cancers, including bladder, breast, colorectal, lung, ovarian, prostate, and skin. (They can’t distinguish between them—a “common cancer odor” hasn’t yet been pinpointed—but they can indicate that cancer is present.)
So how does canine screening work in a clinical environment? You would not be alone if you’re picturing a scene like this: A gowned patient lies on an exam table when suddenly a golden retriever bounds in, sniffs the patient scalp-to-toenail, and barks if he discovers a suspicious odor along the way.
What actually happens, according to Cynthia Otto, a DVM-Ph.D. and the founder of the Penn Vet Working Dog Center, is that the dogs sniff samples, not people. Blood, tissue, urine, and breath can all contain traces of the molecules emitted by cancerous cells. At Penn Vet, the dogs are trained using what they call a scent wheel—a round table hung with bread loaf–sized, perforated steel boxes around the perimeter. Here’s a demonstration:
These boxes are loaded with odor samples. The dog enters the room, beelines for the wheel, and one by one sniffs the boxes. If she detects the smell of cancer, she sits in front of the box. As part of the training, researchers put decoy and distraction smells into some of the boxes, like coffee grounds, rubber gloves, and other blood samples that are either noncancerous or cancerous but benign. They’re meant to confuse and then refine the dog’s sense of smell.
After decades of polite but skeptical stares from the medical establishment, research by Otto and others has indicated that dogs do have a capacity to pick up malignancies by their scent—although more work is needed to determine the precise specificity and sensitivity of the approach. One goal for the Penn Vet group is to use dog olfaction data to screen for cancers for which there are no reliable early tests, like ovarian cancer.
A major challenge is that dog accuracy varies—by individual, by breed, by time of day, by age, and level of training. Dogs are so attuned to human communication that, as with drug-sniffing canines, handler bias can influence whether a dog finds what he’s supposed to be looking for. Dogs can get tired, grouchy, and bored, which can skew their precision.
So the best use of their incredible noses probably isn’t to have them waiting in labs to sniff out cancer. Instead, their potential is in bio-inspiring the engineering of electronic and robotic chemical-detection machines. Around the world, multidisciplinary teams are working to invent technologies powered by the evolutionary wisdom of the dog’s nose but with the predictability and scalability (and low emotional affect) of a robot.
Taking e-nose technology one step further is a National Science Foundation–funded, six-university project led by John Crimaldi at the University of Colorado at Boulder. The scientists are studying not simply how animals detect odor, but how they navigate through odor plumes in natural environments to find food, evade predators, and connect with mates. Understanding the whole-body connection across scent, brain, and behavior in many different species would allow scientists to create an algorithm to mimic that behavior, which could then instruct a machine to automatically navigate itself toward a cancerous region.
Crimaldi says that one reason olfaction is so interesting from an evolutionary perspective is that smell is the very first sense to evolve in life, making it a “primal portal into brain evolution.” Even single cells “smell” in a sense: They chemo-detect.
None of this will be easy. Engineering a sniffing tool as powerful as a dog’s nose is as daunting—and at the moment as out of reach—as creating baby formula that comes even close to the complexity of breast milk. But as cancer research focuses on the cutting-edge and new, it’s wonderful to think that something as humble as a canine snout could someday make a difference.An ideal screening tool would diagnose cancer accurately and early with a high degree of sensitivity and specificity (meaning not too many false positives or negatives). A really ideal tool would additionally detect metastases and indicate how a patient is responding to treatment. E-noses affixed to odor-navigating robots may someday be a powerful way to do this. Teamed with visual imaging like mammograms, CT scans, and MRIs, these devices could add accuracy and speed—with less human error. They also could give useful information about the molecular composition of the cancer.
Click here for the original article.