“Dogs are quick to pick up specific odors,” he says.“The difficulty is applying that in the field and isolating that odor from all others.We never wean them off [the food rewards]. For our dogs, it’s the best way to maintain high accuracy.” At the University of Florida’s entomology department, Peruyero’s dogs achieved a 98 percent accuracy rate and a false-positive of less than 4 percent in clinical trials. To maintain that rate in the field, Scott Umphenour,who handles Kirby, a Beagle trained by Peruyero for Falcon Termite and Pest Control, makes sure Kirby finds these bugs every day, whether he’s on the job or at home.
Because the ability of dogs to precisely distinguish scents is being tested and confirmed, people are thinking of new targets. In Florence, Texas, the Southern Star Ranch has begun training dogs to sniff minute quantities of peanuts in any form: raw, cooked, oil, butter, even dust. And recently, Peruyero has begun working with the University of Florida to train dogs to detect melanomas. If dogs can sniff out mouse poop the size of a sesame seed in a forest, and peanuts in dust, couldn’t they differentiate between normal and abnormal chemical changes in our bodies? The answer is yes, and this probably explains the reactions of alerting dogs.
At the Pine Street Foundation in San Anselmo, Calif., teams of highly trained dogs inhale breath captured and frozen in vials to detect molecules of ovarian cancer. In the UK, anecdotal studies published in The Lancet, a medical journal, led to the first clinical trial involving dogs sniffing for bladder cancer. That 2002 trial resulted in a scientific study published in the British Medical Journal in 2004, and the founding of a charity, Cancer and Bio-detection Dogs, in 2008. The charity works with researchers supported by another charity, the Amerderm Research Trust through the Buckinghamshire Hospitals Trust.
Claire Guest, a training and behavior consultant at the organization, is currently involved with a double-blind, yearlong clinical study of the cancerdetecting ability of the dogs. During training and practice, the dogs are achieving an impressive success rate; she rewards the correct answers with clicks and treats. “We’re talking serious detection,” she says. “These are not fluffy dog stories.”
During the double-blind study, the trainers don’t know which samples, if any, are from people with cancer, so the dogs receive no affirmation or reinforcement for correct responses. Still, the success rate is good. “We’re using very bright dogs,” Guest says.“No matter what we try, they know which times they are not going to get rewards.”
In addition to dogs who screen urine samples for bladder cancer, the organization trains dogs to alert diabetes patients to impending hyper- and hypoglycemia. The overriding goal for this group and others like it is to help scientists learn which odors—which complicated patterns of molecules—dogs use to detect disease, so that someday, for example, doctors might have mechanical cancersniffing devices in their offices and people with conditions such as diabetes might wear alert bracelets.
“I think the dogs’ role, really, is to accelerate the research,”Guest says. Recently, one of Guest’s dogs began behaving anxiously around her. “The dog was almost neurotic,” she says. When Guest found a painful spot on her chest, she decided to investigate and discovered that she had breast cancer. “I’m in my 40s, and we don’t get routine mammograms here until we’re in our 50s,” she says.“The consultant said they would never have felt the tumor—it was so far in.”
Fortunately, because of her work, Guest was canny enough to know something might be wrong rather than to think her dog had a behavior problem.