Baby Boo was an unassuming medium-size dog with ears that flopped out sideways and rotated forward like small triangular satellite dishes. When she trotted down the sidewalk of her leafy London neighborhood, they flapped like the wings of a happy cartoon bird. Her saucer eyes conveyed more emotion than the average dog’s. Every time she looked slightly to the side, the whites showed, giving her the appearance of intelligent concern.
She had the sleek black-and-brown coat of a Doberman pinscher. But the eyes were definitely and intensely border collie. One darting glance would deflate a rebel sheep’s fantasies of breaking from the flock.
Baby Boo was the daughter of a border collie–ish dog named Frisky Fru. The Whitfield family had adopted Frisky Fru when their neighbors moved and didn’t want to take her. The Whitfields didn’t realize Frisky Fru was pregnant. They were able to find homes for all but one pup—the runt of the litter. One of the children, smitten with her from the moment she was born, called her Baby Boo. It stuck.
A caring soul from the start, Baby Boo was always tending to Frisky Fru and the other family dog, Lucy Lou, who happened to be a kleptomaniac. Lucy Lou, a retriever-Dalmatian mix, lived to escape from the house or yard and dash off to nearby shops, where she would pilfer whatever she could get her mouth on. Usually it was small items, like cookies or little cakes, but once she brought home a can of baked beans.
“Oh no! Why do you steal things, Lucy Lou?” her family asked. But Baby Boo didn’t care that her housemate was a thief. She groomed her and watched over her almost as much as she looked after her own mother.
Then an odd thing happened when she was about two. Baby Boo took an interest in a spot on the leg of Bonita Whitfield, who was forty-four. As Bonita washed dishes at the sink, Baby Boo would walk up behind her and sniff the back of her left thigh for several seconds until Bonita shooed her away. This went on for a few months, with Baby Boo snortling into the back of Bonita’s thigh whenever opportunity allowed, even when Bonita was wearing slacks.
Bonita found it annoying, but what could she do? Dogs will be dogs.
One warm summer afternoon, on a day off from her banking job, Bonita decided the outside of the house’s windows needed cleaning. She changed into shorts and brought her supplies and a ladder to the yard. The dogs followed. Partway through her task, she came down from the ladder, and sweet, gentle, caring Baby Boo jumped up and bit at the back of her leg. She did it again and again. It was the same spot she’d been sniffing all along. The nips didn’t draw blood, but they hurt. Bonita couldn’t figure out what had gotten into her.
Later that day, Bonita was changing into a skirt when she touched the spot that had so interested Baby Boo. It was raised. She twisted around and saw it was a fairly large black bump. She had never seen it before.
She brought it up to the bank’s secretary, who had once worked at a skin clinic in Canada. The secretary took a look and insisted Bonita call her doctor immediately. The doctor saw it and sent her to King’s College Hospital in London. A doctor there excised the lesion and sent it for a biopsy.
It turned out to be a melanoma, the deadliest form of skin cancer. They’d caught it just in time. The doctor told her that in another year, the cancer would have spread throughout her body.
Baby Boo became the family hero, getting extra treats and hugs all the way round. With the mole gone, she no longer examined Bonita’s leg. Her discovery would have just been another interesting family story if not for a young researcher named Hywel Williams.
Dr. Williams was training to become a dermatologist at King’s College Hospital. While reviewing more than two hundred cases of malignant melanoma from previous decades, he came to Bonita’s file. He noticed that someone had written, “Dog sniffed at mole!!!”
“I laughed it off, but that night as I drove home, I kept thinking about it. I called her the next day and learned her fascinating story,” Dr. Williams, MD, DSc, told me when I got in touch twenty-eight years after it happened.
“One could say this dog literally saved her life.”
He wrote about the case with dermatologist Andres Pembroke in a letter to the editor of the prestigious general medical journal the Lancet. It appeared in the April 1, 1989, issue. The doctors told the patient’s story (not referring to Bonita or her dog by name) in a couple of paragraphs and suggested that this might be worth exploring scientifically.
“Perhaps malignant tumors such as melanoma, with their aberrant protein synthesis, emit unique odors which, though undetectable to man, are easily detected by dogs with their well-developed rhinencephalon.”*
The letter concluded, “We have not as yet proceeded to a trial of sniffer dogs in our melanoma clinic but the adjunctive use of animals with highly developed sensory modalities in cancer diagnosis is worth considering . . .”
The letter was one of forty letters published that week. It livened up the space between “Warning against Use of Intrathecal Mitoxantrone” and “Immunosuppressive Properties of Cyclosporin Metabolites.”
“Sniffer Dogs in the Melanoma Clinic?” was the least scientific-sounding title in the journal that week, unless you include “Impregnated via a Bullet?,” which you probably should.* But it’s likely that it got far more publicity than any of the other letters or articles.
The Associated Press picked up the story, and reporters from around the world sought interviews with Dr. Williams, and sometimes with Bonita. The story appeared in different versions in media outlets around the globe, from respected newspapers to the Weekly World News, which added dramatic flair with the headline “PET DOG SNIFFS OUT GAL’S CANCER and saves her life!” The half-page piece appeared after a full-page article entitled “SPACE SHUTTLE ATTACKED BY 200-FT. UFO!” with the subtitle “Bug-eyed aliens invaded Discovery & terrified crew.” (If you had read about Baby Boo in that article, you could not be blamed if you didn’t believe it.)
There had been anecdotes of people’s dogs detecting their cancer before, but nothing had appeared in a scientific journal. This gave legitimacy to the idea. Still, the science world took much longer than mainstream media to warm up to it.
Dr. Williams tried to get funding for cancer-detection dog research after the Lancet letter, but couldn’t. “I think there was a lot of skepticism in the scientific world, and we were just in the wrong place at the wrong time,” he said.
For years, there was no news of any research in the field.
Then, in the mid-1990s, a Tallahassee, Florida, dermatologist took an innovative approach to dogs helping with cancer detection. Armand Cognetta, MD, who specialized in skin cancer, was frustrated that one in five melanomas was not discovered in time for the patient to survive.
“I literally go to sleep sometimes thinking, ‘What’s a better way?’” Dr. Cognetta said in an Associated Press article.
He had heard a radio report of a dog who sniffed out a body in a lake and wondered if it would be possible for a dog’s great nose to be used for cancer detection. A search of the scientific literature led him to the Lancet letter. He asked local dog expert Duane Pickel if he thought a dog could be trained to detect cancer. Duane, who had handled and trained dogs in the military and was retired after twenty-two years in the Tallahassee Police Department’s K-9 unit, told him what he’d hoped to hear:
“A dog can be trained to find anything you need it to find.”
Duane volunteered his dog George for the cause. George was a standard schnauzer he’d already trained as a bomb-detection dog. Duane developed a fun way for George to want to sniff out cancer: He hid test tubes containing melanoma samples around a room and asked George to find them. When he did, Duane rewarded him. George was always happy to find a test tube with melanoma. The next phase of training saw George alerting to a test tube with the melanoma in a rack with other test tubes that didn’t contain a malignancy. He did extremely well over repeated trials, with a reported accuracy of 99 percent.
Then George went from detecting melanoma in vitro to doing it at least somewhat in vivo. A nurse with a family history of skin cancer volunteered to wear dozens of bandages, one of which would have a melanoma sample sealed under it. Would George know melanoma without the test tube? They waited as he circled her and sniffed.
He seemed attracted to one of the bandages. He sniffed it extra hard. Then he sat and looked at it. Nailed it!
Later he would be trained to gently touch the spot with his paw. They did more than forty trials in a year and reported that George again reached nearly 100 percent accuracy. When it came to real patients with melanoma, George met with seven over a couple of years and correctly identified a malignancy in four of them—maybe five, depending on how you interpret it. It wasn’t anywhere near 100 percent, but Dr. Cognetta and Duane thought the misses could have happened because George had been trained on the nurse, who always showered first, washed her clothes with scent-free detergent, and didn’t smoke. Patients probably had many olfactory distractions.
Dr. Cognetta’s study was understandably rudimentary, but it was a brave foray into the idea proposed by Dr. Williams and Dr. Pembroke. He had no question that dogs can detect cancer. But how accurately?
“That’s for the next phase, done elsewhere,” he said in the AP article.
It would be several more years before there was serious research on cancer-detection dogs. This new phase may have been kicked off, yet again, by another letter to the editor of the Lancet. It was coauthored by none other than Dr. Williams, and entitled “Another Sniffer Dog for the Clinic?”
Dr. Williams and Dr. John Church, a retired orthopedic surgeon, reported on a sixty-six-year-old man whose Labrador retriever, Parker, kept pushing his nose into his man’s thigh. It was a spot the man had been treating as eczema for eighteen years. But after two years of his dog hounding him, he gave in and went to the doctor. In September 2000, the lesion was removed and turned out to be basal cell carcinoma. After it was gone, Parker showed no more interest in the spot.
This letter to the Lancet concluded with a call to arms:
Although all these data are anecdotal, we believe that the phenomenon of some dogs seeming able to detect unique odours of certain skin cancers worthy of investigation in rigorously controlled experiments. Whether they can detect odours associated with other specific diseases such as tuberculosis or Ebola virus should also be investigated to aid early detection.
The letter was published in the September 15, 2001, issue—a mere four days after 9/11. It could easily have gotten lost, buried under the rubble of the tragedy that gripped the world. Mainstream media didn’t give the letter much attention. Reporters had their hands full.
But in a switch, this time it was science that took notice. The seed had been planted by the first letter, and after a long period of dormancy while conditions for its survival were improving, the idea was ready to germinate. In 2003, Dr. Church held a conference in the United Kingdom about ways to use dogs for studies on cancer detection. Among the speakers at the packed event were Tadeusz Jezierski, PhD, DSc, a revered canine-olfaction expert from Poland, and Michael McCulloch, PhD, who had begun his own cancer-detection dog study in California.
The conference helped launch a new field at the intersection of science and dogs. It was the beginning of the golden age of research on cancer-detection dogs.
Funding and studies would take time. One study here, one study there, a year or two or three in between—like a promise of rain after a drought. Even now, there’s not a deluge of studies, but the number of researchers using dogs to sniff out cancer continues to increase.
Most of the published studies mention the 1989 Lancet letter as the starting point for science acknowledging the possibility that dogs can detect cancer. So do the mainstream articles that report on these studies.
Even though Dr. Williams—with the help of Bonita and Baby Boo—had kicked off this new field, he would be involved in only one study of dogs detecting cancer. He has moved on to other areas of dermatology research and practice.
He’s now a professor of dermato-epidemiology at the Centre of Evidence Based Dermatology, University of Nottingham. He is also the director of the National Institute for Health Research’s Health Technology Assessment Programme. It’s a big-deal national job. As he wrote me in an email: “Basically I am in charge of most of the major publicly funded trials happening in the UK NHS today (around 400 studies worth around £1/2 billion). I am the first dermatologist to hold such a post and it is a great honour. I remain director until Oct 2020, then I shall fall in a heap.” He still does research (he has published more than five hundred papers) and has a weekly clinic in pediatric dermatology.*
He fancies whimsical bow ties and speaks in a charming, singsongy South Wales accent. His report on Baby Boo’s feat is still a highlight of his storied career.
“I look at it with fondness, especially to see some good research eventually developing that might help cancer patients in the future,” he told me. “It was quite a risky thing for me to report at such an early stage of my training, but we felt it was important to suggest the scientific idea so that others could make a judgment for themselves.
“It showed me that research isn’t linear. Sometimes it’s enough to have an idea and let others test it out.”
He said of all his patients, in all his years of practice, Bonita was the one who had the most effect on his work.
“It taught me the importance of listening to patients. They are often right about things even though we might not have the technical and scientific knowledge to explain everything when we first encounter them.”
I wondered what had become of Bonita. I looked for her and didn’t find any contact information. I feared the worst. Maybe another melanoma had eluded Baby Boo.
But when I poked around social media, I found her on Facebook. There hadn’t been any activity for a long time. I reached out to her and, just in case, to her daughter, Jane, hoping they wouldn’t think I was a creepy stalker. A couple of weeks later, I heard from Jane. She told me that Bonita is alive and well and would love to speak with me.
Bonita is seventy-five, a great-grandmother, and an animated talker. She has no more dogs. Frisky Fru died at eighteen, tended to more tenderly than ever in her dotage by daughter Baby Boo. Lucy Lou also passed on at an old age after a life of debauched thievery. Baby Boo lived to be nineteen years old. They are all buried in Bonita’s yard. The only pet she has now is a cat named Bert who likes to drink water from the tap.
“I know I wouldn’t be here today if it weren’t for Baby Boo,” she said. “Since that time, I always listen to my pets. They talk to you. You just have to be able to know what they’re saying.”
Bonita didn’t seem to realize what Baby Boo’s actions had started. I told her about the many studies and articles that have referred to the Lancet letter, and the current state of research involving dogs detecting cancer.
“Oh, that’s lovely, isn’t it?” she said. “It’s her little legacy. Baby Boo would like that.”
Claire Guest bursts into her small office at the headquarters of Medical Detection Dogs with four dogs in tow. “I’m terribly sorry for keeping you,” she says, wiping a streak of dog slobber from her leather riding boots with something she pulled from her vest pocket. “They needed a walk and it always takes longer than you think it will.”
Medical Detection Dogs is involved in several important national and international studies. Camilla, Duchess of Cornwall, and her husband, Prince Charles, are patrons of the nonprofit. British press reported that the duchess got emotional at a demonstration in which she and the prince watched as three Labrador retrievers detected cancer samples with ease.
Claire has even met with Queen Elizabeth II on a couple of occasions.
But today Claire is sitting on the floor of a small office that now smells of dog breath, and brushing burrs out of the fur of two panting English cocker spaniels.
Medical Detection Dogs is located in the midst of rich green swaths of farmland in the village of Great Horwood, about ninety minutes northwest of London. It’s glorious countryside, but during certain times of year, the burrs seem to Velcro themselves onto the dogs, especially the ones with the more resplendent coats.
As we talk, and Claire picks at the more stubborn burrs, people dash into her office every few minutes to ask questions, or take dogs for training, or tell her about upcoming appointments. She’s used to multitasking. She’s busy nonstop these days.
“It’s never-ending,” she says. “I love the energy and the excitement because the world is really finally taking it seriously that dogs do seem to be able to detect cancer.”
The organization officially started in 2008, but its roots go back to 2003, when Claire and Dr. Church—the coauthor of the second Lancet letter—collaborated with several others on a study to find out whether dogs can detect cancer. (This is known as a proof-of-principle, or proof-of-concept, study.) Six dogs—including a wee papillon—were trained to detect the scent of bladder cancer in urine in a lineup of six healthy subjects. They were successful 41 percent of the time—not stellar compared with today’s standards, but a statistically significant success rate compared with the 14 percent expected by chance. It was an exciting result because it was a scientifically rigorous study.
The study was published in the BMJ in 2004. It proposed the idea that tumors release volatile organic compounds that have distinctive odors dogs can detect even in small quantities thanks to their “exceptional olfactory acuity.”
“The results we achieved should provide a benchmark against which future studies can be compared,” the authors wrote, “and we hope that our approach to training may assist others engaged in similar work.”*
Even more intriguing than the dogs’ overall success was when all six dogs appeared to fail. They all “unequivocally” alerted on what had been deemed a control—a urine specimen declared negative for bladder cancer based on cystoscopy and ultrasonography. The paper explained what happened next: “The consultant responsible for the patient was sufficiently concerned to bring forward further tests, and a transitional cell carcinoma of the right kidney was discovered.”
In other words, the dogs beat the doctors and their sophisticated equipment. And they detected a cancer they had not been trained to detect.
The study, as predicted by the researchers, helped inspire more canine cancer-detection work around the world. “For me,” says Claire, “the paper was the moment when I knew that this work must continue.” The research led to the founding of Medical Detection Dogs. Claire had worked for Hearing Dogs for Deaf People for twenty years when she left to cofound Medical Detection Dogs with Dr. Church. It was harder to get backing—financial and otherwise—than they’d anticipated.
“There was huge skepticism to overcome and it took us considerable time to build support and then register the charity,” she says.
In 2009, Claire’s fox red Labrador retriever, Daisy, was among the dogs being trained on cancer samples. She had been acting strangely around Claire. One day, while Claire was letting her dogs out of the back of the car at a park, Daisy just sat and stared at Claire. Claire thought she looked anxious. Then Daisy smashed her nose into Claire’s chest a couple of times, prodding her hard. That evening she felt that area of her chest. There was a lump.
Claire knew many stories of dogs detecting cancer in their own people. She figured it was nothing, but just in case, she made an appointment with her doctor. The lump she felt was a harmless cyst, but deep behind it was something the doctors thought needed further investigation. After a mammogram and core biopsy, she was diagnosed with breast cancer.
She had a lumpectomy, several lymph nodes removed, and five weeks of radiation, and has been cancer-free since. “Because the tumor was so deep,” she says, “the doctor told me by the time I discovered it on my own it may have been too late. Daisy saved my life.”
As Claire tells me the story in more detail than I had already read in her book Daisy’s Gift, the star of the book is lying in a cushy dog bed. She is thirteen and has arthritis in most of her joints. She looks frail, but she is alert as ever, with her large brown eyes trained on Claire. Someone walks in the office.
“Time for PT, Daisy! You ready?” Daisy lumbers out of her bed, stretches, and wags down the hall for physical therapy.
“Daisy is everything to me. She saved my life, she’s my best friend and confidante, and she helped give me hope to keep going with detection work despite all the skeptics,” she says. “I dread the day she won’t be with me anymore.”* She tries not to think about it. Focusing on the organization helps keep her mind from the inevitable.
Medical Detection Dogs has about thirty biodetection dogs on its team these days. Most work in cancer detection, and some others are in pilot projects for detecting other diseases. All of the dogs live in homes and come in a few days a week for training or working.
With just a couple of exceptions, that’s the setup for almost all cancer-detection dogs in the world. They spend most of their time with their families—often fosters taking care of them during their couple or so years of work. Training is fun and reward-based, and working periods are kept short.
Medical Detection Dogs is involved with two large cancer studies. One is a breast cancer study with rigorous protocols—so rigorous that the study had been paused shortly before my visit. The dogs had already completed work on breath samples from one thousand patients, but the amount of breath in the tubes was not consistent enough. The trial had been kicked back to research and development. It’s not the first breast cancer study to use the noses of dogs, although when it proceeds, it will be by far the most comprehensive.
A groundbreaking study by Dr. McCulloch, Dr. Jezierski, and colleagues published in 2006 found that dogs could distinguish between exhaled breath samples of women with breast cancer and healthy women. The dogs demonstrated 88 percent sensitivity (the ability to detect the disease, a true positive) and 98 percent specificity (the ability to identify samples without the disease, a true negative; in the case of dogs detecting cancer, that means ignoring the negative sample rather than alerting to it). But the results of another study published two years later were disappointing. None of the six dogs performed better than chance for detecting the positive samples, and only two out of the six dogs showed specificity better than chance.
In a promising study in 2017 by the Curie Institute in Paris, 130 women with breast cancer kept a square of gauze, held in place by a clean bra, against their affected breast overnight. They sealed the square in a sterile jar, and the samples were later sniffed by two Belgian shepherds. The dogs had a success rate of 90 percent on the first trial and 100 percent on the second trial. (No figures are available on false positives.) The institute hopes to start a three-year clinical trial to test the feasibility of this technique for cancer screening.
The other cancer work that’s keeping the dogs at MDD busy is a proof-of-principle study on prostate cancer, with dogs looking for cancer in urine samples of three thousand participants—the largest of any such study so far. In initial training trials, the dogs had an impressive 93 percent reliability rate. Four previous prostate cancer studies by others have ranged from a low of 13 percent sensitivity and 71 percent specificity to a high of 99 percent sensitivity and 97 percent specificity.
Claire has no doubt dogs can accurately detect cancer. Not usually, at least.
“Some nights I wake up in a hot sweat,” she says. “What if I made this all up? Then I watch the dogs alerting to cancer, working so well, and all is well.”
The two cocker spaniels Claire had been grooming during our conversation pass her inspection. One lies down on a chair; the other pads off to a dog bed on the floor.
“Excuse me just for a second,” Claire says. She brings over a handheld vacuum, gets on her knees, and hoovers the crimson carpet of all the burrs and fur. I want to grab the vacuum and do it myself. “You shouldn’t be doing this,” I would tell her. “You know the queen of the United Kingdom!”
But before I can get the words out, she’s done.
“So much to do, so little time. The clock is always ticking.”
Lives are on the line. People around the world are dying from cancers they might have been able to survive if detected earlier. Too often, the tests available today don’t find cancers until it’s too late. Or they’re invasive or expensive, or both.
Some researchers would like to figure out a way to have dogs screen large numbers of patient samples quickly and reliably in laboratory settings. Having dogs screen patients in person is not the goal of serious scientists.
“That would be rather like the grim reaper, wouldn’t it?” says Claire. “We don’t want dogs to become that.”
Besides the bad PR of having dogs giving dire news by offering a paw or sitting and staring, most researchers say there are too many confounding factors to ever have dogs examine people in real life. They want this to be science, not educated guessing. It’s challenging enough to get accurate results with dogs in a laboratory setting for one specific cancer, much less on a whole person for many types of cancer.
“The doctor dog will see you now” is not a likely phrase you’ll hear at your physical in the future, despite occasional news headlines about dogs performing “PET scans.”* The dogs who may one day help save lives will likely remain behind the scenes.
So far, dogs have been trained with varying degrees of success to look for breast, cervical, colorectal, lung, stomach, liver, ovarian, prostate, skin, and thyroid cancers. Trained dogs have sniffed for cancers in more than a half dozen sample types, including blood and sweat, urine and feces (no dog would complain about this assignment), and tissue samples and breath.
Dogs seem to be able to detect cancers across all four stages—not just when cancer is advanced.* That’s exciting news, and encouraging. Early detection not only would save lives, but also could make treatment far less debilitating or prolonged.
Good science takes time. Proof-of-principle studies are going on around the world. Ideally, science learns from its mistakes. The authors of papers often write about ways to improve future research. The authors of the breast cancer study where dogs didn’t perform better than chance, for instance, suggested that “better management of urine samples and a more stringent training protocol during our study may have provided new evidence as to the feasibility of using canines for cancer detection.”
Yet even studies that seem to be glowing successes can have flaws that aren’t apparent until later research fine-tunes techniques. What researchers have learned during these studies highlights the intelligence and remarkable senses of canines.
In some of these studies, handlers have been in the same room and visible to the dogs. It may not be a big deal, but it could be a confounding factor—something researchers try to pare away to get to the purest and best results.
The problem is that dogs can be so attuned to our body language—sometimes even the subtlest shift in expression—that they may be taking unintentional cues from the person in the room. It’s a phenomenon known as the Clever Hans effect.
Hans was a horse who lived in the early 1900s and wowed crowds with what seemed to be his mental prowess. It looked like he could tell musical notes from one another, read, spell, do simple math, compute fractions, tell time, and identify dates on a calendar. He answered spoken and written questions by tapping numbers or letters with his hoof. Even biologists, doctors, and psychologists believed he could do these feats.
A commission in his native Germany was set up to figure out what was going on, and determined it wasn’t a hoax. But shortly after, a biologist and psychologist figured out that Hans was very clever, although not in the way everyone—including his human partner—had thought. He concluded that Hans was so observant that he could read the “almost microscopic signals” in his owner’s face.*
It has become more common—almost standard—for handlers to be hidden behind a screen or in another room during studies of dogs sniffing for disease. I watched dogs who sniff for colorectal cancer in a pilot study in the Netherlands burst with wagging tails through a little doggy door into a room with the samples on a scent carousel as researchers observed from a large one-way window. The handler was hidden in the room behind a white box with a little one-way window for observing, and popped out to reward the dog for correct answers.
One of the dogs in the study was a Belgian Malinois. I’ve never seen this hard-core breed doing work other than for the police or military—tracking bad guys, taking them down, or sniffing for drugs or bombs. Seeing the same breed of dog brought by Navy SEAL Team 6 on the Osama bin Laden compound raid now checking for colon cancer in a neat Dutch lab gave me an even greater respect for this breed’s capabilities.
Cancer studies have used many breeds of dogs—including Labrador retrievers, German shepherds, poodles, dachshunds, golden retrievers, Large Munsterländers, rottweilers, schnauzers, cocker spaniels, springer spaniels, Havanese, and Portuguese water dogs—and mixed breeds as well. The breed of dogs doesn’t seem to matter as much as the dog’s drive for a reward.
If a dog isn’t toy- or food-driven, he or she won’t have much motivation to work. As much as we’d like to think otherwise, dogs are not doing this work to benefit mankind. They’re doing this for a ball or a special toy or a tasty snack, mixed with sincere praise for a job well done.
But just because dogs have good noses and are happy to work for rewards doesn’t mean all dogs will be great at detecting the scents we ask them to find. Dr. Angle, the Auburn University researcher we met in the book’s introduction, says this is one reason he and his team use several dogs in each study. Even though the dogs they work with come from the CPS breeding program, which is on its seventeenth generation of Labrador retrievers bred to be top-notch sniffers, not all the dogs are going to be able to do the job.
“We can have ten dogs on a study. Maybe eight can detect something, and the other two can’t detect it at all,” he says. “It’s a physical capability like in humans. Same breeds, same bloodlines, but very different results.”
It’s kind of like how some people, thanks to what may be a genetic variation, can’t smell the odor asparagus gives urine.
If someone runs a cancer study using only one or two dogs, and those dogs happen to be the ones who can’t detect that scent, it could lead to unfortunate outcomes. Along the lines of that soon-to-be-tired analogy, you’d hate to run a study with humans sniffing the urine of people who recently ate asparagus, and inadvertently use humans whose sense of smell can’t pick up the scent.
In general, though, dogs are masters at finding most scents.
“If there’s an odor, we can train them on it,” says Cindy Otto, DVM, PhD, founder and executive director of the Penn Vet Working Dog Center at the University of Pennsylvania. “They’re heroes without knowing how important they are.”
But how do dogs even get to the stage where they can detect cancer? No one knows what the volatile organic compounds of cancers are, so researchers can’t simply whip up a soup of known VOCs to train a dog. The dog has to be a sleuth and put together an olfactory picture from the VOCs in samples presented to him or her. Considering there are hundreds of VOCs in even healthy samples, it’s a mind-boggling feat.
Dr. Otto likens it to people learning to find Waldo on a page of a Where’s Waldo? book—without knowing what Waldo looks like. They can’t even exclude objects while they search. They may point to the woman in the bathing suit, the striped umbrella, the piece of cake, in their attempt to find out what a Waldo is. But eventually they will find Waldo, and boom! From then on, the happy, nerdy fellow with the red-and-white-striped shirt, matching hat, and blue pants is the guy they’ll pursue.
It’s more complicated than that in cancer detection. Dogs may be sniffing for patterns of VOCs. There is no one Waldo.
In most of this research, dogs are looking for the common odor within one type of cancer. It’s believed that each cancer has a scent that’s unique to it and it alone. Lung cancer may have its own scent, and breast cancer may have its own. Researchers are asking dogs to find the fingerprint scent.
But an important question remains: Does cancer itself have an overall common odor—something shared by most or all cancers?
It’s not a question most researchers have tried to answer yet. But thanks to anecdotes about cancer-detecting dogs being able to find more than one kind of cancer, many in the field are willing to conjecture there may be a generalized cancer smell.
“It may be like listening to a piece by Mozart,” says Claire. “It’s complex, but every so often you come across a bar of music that repeats. It could be like that with cancer. There could be that repeating bar, that one small common factor.”
She’s careful when she talks about this idea, though. “There’s no published evidence that dogs can detect different cancers by one scent. Stating this as fact could result in discrediting the work, and there’s some fantastic work being done.”
Claire has had personal experiences with a dog being able to detect multiple types of cancers, but she tries not to let it get in the way of pure science. Daisy, credited with detecting Claire’s breast cancer, has also successfully sniffed a few types of cancer at Medical Detection Dogs. She had originally been trained on bladder cancer for a study. But she later rooted out samples of prostate cancer and kidney cancer.
It could be that Daisy learned to differentiate cancers by their individual, unique scents. Or maybe she discovered a general scent that cancers have in common. Claire sometimes wishes Daisy could talk. “Oh, the things she could tell me!”
Researchers are in awe of the odors dogs can detect, but it’s a double-edged sword. Canine olfaction is so keen that it can end up being frustrating when trying to do real science.
While dogs are figuring out what they’re supposed to be finding in a study, they may choose samples based on the location where the samples were obtained. Dogs have picked up on hospital disinfectant odor and alerted to that. They’ve even selected samples based on the person preparing the samples.
Dr. Angle says since people shed so many skin cells—thirty thousand to forty thousand per minute—dogs have alerted to the person dealing with the samples, even though the person is wearing gloves.
“Dogs are natural biosensors. They track things,” he says. “They’re innately able to pick up on a human odor, so you have to have this in mind when conducting studies.”
Researchers continue to try to improve the ways dogs can detect cancer. There are dozens—sometimes hundreds—of details they need to consider when putting together a strong study that might make it into a peer-reviewed journal to further the science. What’s the best way of training the dogs? Who will collect the samples? How should the samples be stored? How long should detection sessions be?
A more recent question—and a vital one: Can we get enough samples so a dog doesn’t have to sniff the same sample twice?
It turns out that dogs have really good memories for smell. “Their olfactory memories are crazy!” says Dr. Otto. Dr. Angle seconds that. “Dogs can remember samples for years,” he says. “We’ve pulled synthetic odors from old projects years later, and the dog still alerts to it as if they’d been trained on it yesterday.”
That’s good, in a way. That might be part of the reason dogs may be able to zero in on “Waldo” so well. But there’s a flip side to this great memory: Dogs have been known to memorize the sample itself instead of responding to the cancer in a sample.
“They basically learned that odor of that sample, not of ovarian cancer generally,” says Jennifer Essler, PhD, a postdoctoral fellow at the Working Dog Center.
It would be as if you were in an experiment where you have to push a button every time you see an image of someone with earrings in order to get a reward. The images are random and can be shown more than once. Eventually you might take the shortcut, and instead of looking for the earrings, you might recognize the person wearing the earrings. So every time you see, say, Emma, with her curly brown hair and wide-set eyes, you’ll push the buzzer. You don’t have to go to the trouble of looking at her ears to see if there are earrings anymore.
With dogs, it’s much the same. The scent of a sample might be stronger than the scent of the cancer. If dogs have enough repetitions of a sample, they might end up having a hard time generalizing the odor for the cancer in their study.
“Dogs cheat,” says Dr. Otto. “They love to do this, and if you put a shortcut in inadvertently, they will take it. They don’t have any skin in the game.”
Samples are already like gold. As common as cancer is, researchers in the field have to work hard to get adequate numbers of samples. When the dogs’ memorization skills became apparent, it created a new challenge. Scientists need to figure out ways to get more samples, or get creative.
Dr. Otto and her team are trying to do both while training dogs to detect ovarian cancer. Since their approach is unique, and since I have a personal interest in this particular cancer, I thought a visit might be in order.
Dr. Otto, her training manager Pat Kaynaroglu, and I are in stealth mode. We’re hiding behind a one-way mirrored window in a large storage room of the Penn Vet Working Dog Center, waiting quietly. We’re about to see if Osa, a German shepherd, can alert to blood plasma samples of ovarian cancer.
We don’t want her to know we’re there. She’s been training nearly a year for this day, and if we do more than whisper, she’ll hear us and might get distracted. I even have to be careful about flipping the pages of my reporter’s notebook because the sounds travel from our area to hers almost as if we’re in the same room.
A few dogs at the center can already detect ovarian cancer in plasma. But this is Osa’s first time trying. Until now, Osa has been trained on ovarian cancer cell lines. Cell lines, which you may have learned about in Rebecca Skloot’s book and the subsequent movie, The Immortal Life of Henrietta Lacks, if not in Biology 101, keep dividing and growing in the laboratory and are widely used in cancer research. They can be bought online from laboratories around the world.*
If Osa could alert to the plasma of women with ovarian cancer, it would be a big deal. Not only would it mean that the dogs are alerting to the actual cancer, and not the body’s reaction to cancer, but it also would mean the team might not have to “waste” as much of its precious store of ovarian cancer plasma samples during training. Dogs could initially train on the cell lines, then switch to plasma, and more samples could be reserved for studies.
Osa strides into the detection room, and I hold my breath. She walks quickly around the scent carousel, sniffing at samples in each port at the end of eight steel arms. She doesn’t alert anywhere.
But the next time she goes around, she stops and sniffs one of the samples for a couple of seconds. We hear the click of the training clicker, and she runs to the next room for her reward. The handler, with changed gloves, puts a new sample in another port and we wait.
This time, Osa stops and sniffs the port with the new sample. It’s not a full alert, but it’s close enough. Click!
“She’s really telling us something!” Dr. Otto exclaims in a whisper while Osa is out of sight enjoying her reward. Dr. Otto is beaming. Osa goes around one more time, stops at the correct port, and is done.
“She has so found a Waldo!” Dr. Otto says. “She did it!”
I learned that day that scientific discovery is not always glamorous. It can be three people hiding like Peeping Toms, silently thrilling to a dog’s slight change of behavior. It was exciting to be a witness to something that could make a difference in how dogs detect ovarian cancer in the future.
The day before, I’d watched a mostly black shepherd named Bobbie as she got ready to sniff diluted blood plasma from someone with ovarian cancer. She had already been trained to detect a single drop of undiluted plasma. Now she had advanced to a one-to-one ratio, half plasma, half saline. Researchers had mixed one drop of plasma and one drop of saline, then extracted one drop of this mixture as the sample for detection.
Keep in mind this isn’t whole blood. Plasma is just one part of blood. It’s the yellowish clear liquid that carries red blood cells, white blood cells, platelets, and other components of blood. That a dog can detect a single drop of plasma is remarkable. That a dog can detect plasma diluted one to one with saline is extraordinary. Some dogs at the center had even been alerting to a one-to-two ratio.
It’s a scientific limbo dance with cancer scent, seeing how low it can go before it’s too low for the dogs to be able to smell it. The diluted plasma wasn’t riding the carousel by itself that day. The other ports contained controls, including plasma from women with benign ovarian tumors, plasma from women with no known tumors at all, saline, and ordinary items like paper clips and latex gloves.
Before starting, Bobbie trotted over to a large bin of dog toys in the room adjoining the detection room. “Dogs get to choose their own paychecks here,” said Dr. Essler.
Bobbie dipped her nose into the bin and selected a rubber bone-shaped toy and walked off with it. Dr. Essler was surprised. “That’s not her normal choice,” she told me.
Bobbie stopped and seemed to think. She walked back to the bin and returned it. She chose a blue rubber ball instead. “Ha-ha, that’s her usual choice!” Dr. Essler said. Bobbie is a dog who knows what she likes. The hope was that she would also like diluted plasma from ovarian cancer patients.
The first time around the scent wheel, she missed the diluted plasma. A handler switched the positions, but she missed again. I wondered if the scent could have become too low for Bobbie. Was a one-to-one dilution too much?
She took a third stroll around the carousel. This time she alerted to port six.
Yes! Click, cheers, the bounce of a ball. Next time she found a new sample in port eight. After that, she nailed it every time.
If Henry Higgins had been in attendance, he would have had occasion to exclaim, “I think she’s got it! By George, she’s got it!” As it was, the room of researchers and handlers let her know she was a very good girl. After each alert, she chomped the ball and wagged her tail and reveled in the praise.
The only times Bobbie didn’t alert were when there was no plasma from ovarian cancer patients. This wasn’t an error; it’s called a blank. In a blank, all arms of the carousel hold some kind of scent, but the cancer is left out.
When a session is blank, dogs shouldn’t alert. In fact, dogs are rewarded for not alerting. They always have to have positive feedback for the right answer. If there’s nothing in a particular round and dogs have been trained with the expectation of always finding something, they’ll often “find” something anyway just to get a reward.
In other words, they’ll cheat.
So they need to learn that it’s OK if there’s nothing. They should find a cancer sample only if it’s really there. That’s pretty obvious, but until recently, blanks hadn’t been commonly used.
Blanks are important if the science of sniffing cancer is to go beyond studies of known cancer samples and move to actual screening for cancer in the real world. In cancer screening, no one knows if a sample is positive or negative. Dogs have to be able to slog through round after round, and possibly come across no positives for several sessions. If they don’t get rewards, they could start alerting to cancer that isn’t there. That’s a problem. Even worse would be if a dog were bored or distracted or overwhelmed and missed a positive sample.
Researchers at Krems University Hospital in Austria set up part of a lung cancer study as if it were a screening. “To imitate a real-life situation, the dogs were not rewarded during the actual testing,” explains Klaus Hackner, MD, coauthor of the study. “We did not want to reinforce a possible wrong result.” The handlers also didn’t know which—if any—of the five samples of breath during each detection round was from someone with lung cancer.
In traditional studies, someone is in the know, and the dog is rewarded. But in this screening-like study, Dr. Hackner says, “the lack of feedback was very stressful for both the dogs and the handlers.”
Dr. Hackner, a dog-lover himself, notes that the dogs didn’t end the test days stressed. “They were treated with much care, and ‘after work’ they were of course rewarded.” (It’s what he’d want for his own dog, a beloved decade-old beagle mix named Pauli. “He enriches my life every day.”)
At the end of the testing, the dogs recognized 78 percent of the samples that were positive for lung cancer. But they “diagnosed” 66 percent of healthy samples as cancerous. (In the scientific parlance introduced earlier, that’s 78 percent sensitivity, 34 percent specificity.) If this had been a real screening, it would have resulted in quite a few sick people not being diagnosed, and a lot of worried, healthy people being flagged for potential cancer.
In their paper in the Journal of Breath Research, Dr. Hackner and his coauthors concluded that “canine scent detection might not be as powerful as is looked for in real screening situations. One main reason for the rather poor performance in our setting might be the higher stress from the lack of positive responses for dogs and handlers.”
In other words, without a regular paycheck, dogs may not be reliable. The authors suggested “positive feedback mechanisms for future study designs. In fact, it seems to be favourable to confront dogs relatively often with the pattern odours.” Adding positive samples during screenings may help dogs remain accurate and inspired.
Most researchers in the field agree that, so far, dogs seem to be good at detecting cancer in a laboratory setting where someone knows if a sample is cancerous or not and dogs get rewarded when correct. As methodologies improve, so should accuracy.
But will dogs one day be able to reliably screen samples from patients whose cancer status isn’t known? I had heard of only one scientist in the world attempting to screen the public.* I wanted to know more—to see what this would look like in real life. I packed my bags and headed west. Way west. The story of doctor dogs took me to a beautiful community far beyond anything I expected.
It is a challenge to get to the town of Kaneyama, in Japan’s Yamagata Prefecture. First there’s the trip to Tokyo. If you’re leaving from San Francisco, as I did, it’s a ten-hour flight across the Pacific Ocean, over features like the Chinook Trough, the international date line, and the massive Hawaiian-Emperor seamount chain. You can’t actually see most of these except on your seat-back monitor when you get tired of hopelessly cramming essential polite expressions you should have learned weeks ago from your Japanese phrase book.*
Then there’s the Shinkansen ride out of Tokyo. It’s best to spend some time in Tokyo before embarking on this journey. Getting to the high-speed rail for the big trip is an adventure of its own; the Tokyo Metro subway map looks like a plate of spaghetti someone has dropped on the floor. Once finally on the Shinkansen—the best form of public transportation you will ever take if you like cleanliness, pleasant and on-time service, and sweet chiming arrival and departure tunes—you bullet your way north.
Eight stops and 272 kilometers later, you arrive in Fukushima,* where the Shinkansen becomes more of a local train. It slows down a little, which is good, because you’ll soon be streaming through some of the most beautiful scenery in the world. Rice fields flash by, gold, green, then soba fields with their white blossoms. Gold, green, white, gold, green, white, backed by rich green pyramids of hills and mountains, each with its own faithful entourage of bright white clouds.
In ten more stops, you reach Shinjo, the end of the line. You are now 421 kilometers north of Tokyo. But you’re still about fifteen kilometers south of Kaneyama. There’s a nearby bus that can get you there. It takes about thirty-five minutes. Or you can rent a car if you’re confident about driving on what Americans tend to call the “wrong” side of the road.
Luckily for my trip, Akiei Shibata, a Kaneyama town clinic official, was waiting at the station. He would escort me along with Masao Miyashita, MD, PhD. Dr. Miyashita had joined me on the Shinkansen in Koriyama, the stop before Fukushima. He had spent the morning at a hospital there performing endoscopies—something he does once a week on a “day off.” Normally he works at the Nippon Medical School Chiba Hokusoh Hospital, in Inzai, Chiba Prefecture, east of Tokyo. He’s a busy guy—deputy director of the hospital and head of digestive surgery there. He’s also a professor at the medical school, and involved in several research projects.
The doctor’s pet project was the reason I had made this trip. Several years ago, after reading studies from around the world about dogs and their cancer-detection capabilities, he wanted to explore the possibilities himself. Anything that could help with early cancer detection sounded good to him. He’d seen too many deaths, watched too many patients and their families suffer.
He knew there was a dog available for some studies he wanted to do, thanks to an earlier study by Hideto Sonoda, MD, PhD, from the Department of Surgery and Science at Kyushu University in Fukuoka, Japan. Dr. Sonoda’s team reported in the journal Gut that the sole dog in the study was highly accurate at detecting colorectal cancer from breath and stool samples. It seemed the dog could even detect early-stage cancer.
Dr. Miyashita met with Yuji Sato, who runs St. Sugar Japan, the cancer-detection dog training center used by Dr. Sonoda. He was impressed by Marine, the black Labrador retriever who appeared to be a one-dog cancer-detection unit as she sniffed out several types of cancer. Dr. Miyashita and other researchers from Nippon Medical School ended up working with St. Sugar on two studies—one on breast cancer, one on cervical cancer. Marine worked on the cervical cancer study, and a dog trained specifically for breast cancer was the sole dog in the other study.
After Marine died, Sato-san brought on a few other Labrador retrievers he had been training to detect the scent of cancer.* Dr. Miyashita was happy to see that these dogs, like Marine, appeared to be able to detect more than two dozen types of cancer—almost any cancer presented to them.
“There seem to be common VOCs for all or most cancers,” he had told me on the train, “and I think each cancer might also have its own unique VOCs. Of course, much more work will need to be done, but I find it amazing.”
In a way, Dr. Miyashita came to Kaneyama because of horses. A few years back he was looking into offering horse therapy at his hospital for patients and staff. The expert he was working with knew a man who does horse therapy in Kaneyama. In addition to his work with horses, Wataru Inoue produces a lecture series. He invited Dr. Miyashita to see the horses and give a seminar on any topic.
Dr. Miyashita’s topic was “amazing animals”—“the super abilities of whales, dolphins, owls, and butterflies,” he told me, “and of course, cancer-sniffer dogs.” After the talk, Inoue-san brought the doctor into town and introduced him to Mayor Hiroshi Suzuki.
Mayor Suzuki was troubled. A thick issue of a magazine had recently been published listing cancer mortality rates and numbers in Japan. Of Japan’s 344 “secondary medical districts,” Mogami—the rural area that includes Kaneyama—had the highest mortality rate for stomach cancer in women. The ranking for men’s mortality from stomach cancer was 22 out of 344.
Areas of northern Japan are known to have high rates of stomach cancer. Dr. Miyashita and other gastric cancer specialists say there are probably a number of causes, including high intake of salt from the pickled and preserved foods common to the region, genetics, an aging population, and infection with the bacteria Helicobacter pylori.
The mayor was worried about the health of Kaneyama’s citizens, and embarrassed by the ranking. He asked the doctor for advice. Dr. Miyashita went over all the usual causes of stomach cancer. Mayor Suzuki wanted to know if there were easy, noninvasive screening tests. There are not.
Dr. Miyashita told him about his work with cancer-detection dogs. He explained that the dogs haven’t ever screened for cancer, just worked in studies. He couldn’t make any guarantees, but it would be interesting to see if the dogs could succeed. He thought they’d have a good chance.
Mayor Suzuki was excited about the possibility of the residents receiving screenings that could detect cancer early and save their lives. He brought it up at the next town council meeting.
The council voted to spend 11 million yen ($98,000) annually to conduct a screening study of up to 1,000 of its 5,600 residents. The members approved the project for three years, but funding would have to be approved each of the following years. People who signed up would be tested for cancer—cancer in general, not just stomach cancer—by submitting a urine sample, which would be frozen and sent to the Chiba hospital, then brought to St. Sugar. Results would come back within three months, and anyone who tested positive through the dog analysis would receive medical consultation and further testing through traditional methods like endoscopy and CT scan.
It had been four months since the start of the study when I arrived at the Shinjo train station with Dr. Miyashita. He was scheduled to give a talk to Kaneyama’s residents that evening, updating them on the results of the screening so far. The dogs had screened about 250 samples and alerted to five samples. Dr. Miyashita and a local doctor would be meeting with these men and women the following day to discuss options and quell fears.
Shibata-san drove us to Kaneyama as opera music from a Verdi CD filled his Honda. The farther we got from Shinjo, the more spectacular the scenery: cedar forests, streams, rivers, traditional old Japanese houses, colorful rectangular fields of vegetables and grains almost ready for harvest, mountains shrouded in mist.
It was hard to believe that this pristine paradise could be so high up on the cancer-mortality list.
When we arrived in Kaneyama—whose name means “golden mountain”—school was getting out. Dozens of uniformed elementary school students were walking home. They smiled and told us “Konnichiwa!”
Many children seemed surprised to see me. Tall Caucasian women with wavy hair made unruly by the September humidity are not common in these parts. A few, with big grins, carefully practiced their English: “Hello! How. Are. You?” I smiled back. “I’m doing great, thank you! How are you?” They laughed and gave me thumbs-up or peace signs.
Shibata-san told me (through Dr. Miyashita, who would be my translator for the duration) that it is “very, very uncommon” to have Western visitors here. And yet at least one of its brochures is in both English and Japanese—just in case foreign tourism ticks up.
In a little park in town, next to a mountain stream burbling with big and colorful koi, sits a monument commemorating British explorer and writer Isabella Bird. She steamed to the area in 1878 and traveled Japan by train, foot, horse, “man cart,” boat, and, once or twice, a cow. She briefly mentioned Kaneyama in her book Unbeaten Tracks in Japan when describing the area’s stunning mountains, which seemed to block her northward progress: “At their feet lies Kaneyama in a romantic situation, and, though I arrived as early as noon, I am staying for a day or two, for my room at the Transport Office is cheerful and pleasant, the agent is most polite, a very rough region lies before me, and Ito has secured a chicken for the first time since leaving Nikko!”
I wondered how much the area had changed since she passed through. Several large old homes and buildings in the main part of town appear quite old, but I don’t know if they could have been around when she was here.
They look as if a Swiss Alps architect and a Japanese architect had gotten together long ago and designed a new breed of mountain structure. Their outer walls are thick and smooth, and coated with bright white paint. Dark brown cedar roofs with steep slopes jut far past the walls, protecting them from the deep snows of winter, and other wooden “mini roofs” shelter every door and window. Shibata-san said that long ago, these had been miso storage buildings.
One of these buildings now houses a gift store that specializes in cedar products, made from local wood that is a backbone of the economy here. You walk in and are greeted by the fresh woodsy scent of cedar emanating from the cedar walls and from all the items for sale. While I shopped, Enya’s Shepherd Moons album was playing, and a light rain pattered against the thick windows upstairs.
The mayor was expecting us, so I made my purchases, which the shopkeeper wrapped in white paper marked with little koi and green mountains.
We arrived at the office of Mayor Suzuki. A tall, elegant man with light steel-colored hair, he was wearing no-rim glasses and a crisp blue-striped shirt. His secretary served us tea in locally made ceramic cups set on cedar saucers. Dr. Miyashita briefed him on the project so he’d know what to expect at the town meeting. I asked questions here and there, as the doctor translated for me.
Mayor Suzuki stated that he was “embarrassed and concerned that the mortality rate for stomach cancer is so high. We need to help the people in whatever way we can to keep them healthy.
“We hope and believe the dogs can succeed, because people will be happier to have the easy screening of the dog than the usual methods.”
Mayor Suzuki said his mother had died of stomach cancer fifteen years earlier at the age of eighty-two. Her doctor didn’t tell her she had terminal cancer. This wasn’t unusual. In fact, it was the norm. Until the early 1990s—or, in some cases, much more recently—it was a long-standing custom in Japan for doctors not to tell patients the truth about their cancer diagnoses, especially if they had a terminal illness. A 1988 UPI article stated that eight out of ten doctors said they lied to their cancer patients.
The patient’s immediate family members, if there were any, would be told, and it would be up to them to tell the dying loved one. But chances were that the dying family member would not find out. Another survey, reported in 1989, said that only 21 percent of people would pass along a cancer diagnosis to their sick relative.
Until the late 1980s Dr. Miyashita himself often used to withhold the truth because it was viewed as something that would be upsetting to the patient. People with hopeless gastric cancer might be told they had an ulcer.* Even Emperor Hirohito, who was diagnosed with duodenal cancer in 1987, was apparently never told of his diagnosis. He died in January 1989, after a dreadful last few months. It was only after his death that his cancer was announced to the public.
Gradually Japan changed, and patients are now told if they have cancer, no matter how bad it is. Some more traditional rural areas came along more slowly, which is why Mayor Suzuki’s mother, who died in 2002 three months after her diagnosis, did not know.
And now, with this study at hand, people would not only be learning if they had cancer, but the news would be coming, indirectly, through dogs. I found the residents’ ability to not only embrace the times but try to race ahead of the times admirable.
That night about 250 people gathered in a large meeting hall to hear Dr. Miyashita discuss the progress and to ask questions. The faces were serious as he went through a PowerPoint presentation showing dogs and statistics.
For some reason I was seated at a table in the front of the room next to the stage. The mayor had introduced me to the audience. I knew this when I heard my name and something like the words “American journalist.”
Several TV and newspaper reporters, including an NHK news crew, filmed or took notes as Dr. Miyashita spoke. I understood nothing he said except “cancer-sniffing dog.” Those words brought me back to a road trip I’d taken with Dr. Miyashita a couple of days earlier to meet one of the cancer-sniffing dogs on the project. Since I had nothing else to do, I jotted down some extra notes I hadn’t had time to write from that day.
On the two-hour drive from Dr. Miyashita’s hospital to St. Sugar, it became clear that he is someone who thinks outside the box—a quality that goes along with his willingness to screen a remote town for cancer via dogs.
Shortly after we started our drive, he connected his mobile phone to his SUV’s radio and played an album in the background. I didn’t pay much attention to the music since I was busy interviewing him. But after I began to feel carsick from taking notes on winding roads, I took a break from the interview. With a lull in the conversation, I listened to the music.
After a couple of minutes, he asked if I liked it. I did. I told him it reminded me a little of the early Beatles.
He beamed. “Ha-ha, that is me singing! I remastered it last year!”
He explained that he and his friends had recorded the music when he was twenty. One friend had written the lyrics, and young Masao was the music and voice and guitar talent. It was the year 1973, just before he was going to start medical school.
They sent the tapes to several US studios and waited. His dream was to be a singer, or write music, or both. If there had been interest, he would have dropped the idea of going to medical school and focused on his musical career. “I thought maybe I could be the fifth Beatle,” he said, and laughed heartily.
But they never heard back from anyone.
With plan B dashed, he stuck to plan A. It wasn’t a bad plan. I asked him why he had wanted to become a doctor.
“You will laugh, but it’s true. I was inspired by American TV shows I watched when I was young.”
First was The Fugitive, which ran in the US from 1963 to 1967, although he watched it a little later in reruns. It featured David Janssen as the devoted Dr. Richard Kimble, convicted of a murder he didn’t commit and hunting down the real murderer while trying to stay a step ahead of the law. It impressed young Masao that the TV doctor would always stop to help people along the way. He admired the doctor’s devotion to aiding others, even if it meant the doc’s life would be harder because of it.
But his greatest role model came along a couple of years later with Medical Center, a series that aired in the US for seven years starting in 1969. Masao idolized Dr. Joe Gannon, a dashing and sensitive young heart surgeon who worked at a fictional hospital in Southern California. He was so influenced by Dr. Gannon that he visited UCLA’s teaching hospital during a summer break from Nippon Medical School to get a feel for where his hero worked.
Dr. Miyashita pointed out that neither TV doctor used dogs in their practice.
“If someone had told me back in medical school that one day, when I was much older, I’d be trying to find cancer by using dogs, I would have laughed,” he said.
We arrived at St. Sugar, in the coastal city of Tateyama, Chiba Prefecture. The training building is close to Sato-san’s house. Both are powder blue and flanked by palm trees. They look out onto the Pacific Ocean across the road. A bright yellow memorial to Marine stands at the entrance to St. Sugar. Sato-san, a slight man with silver hair, greeted us and ushered us into the training center. We exchanged our shoes for slippers at the entrance, and he brought us to a small training room.
There I met the current star of the training center—a slender yellow Lab named Bea (short for Phoebe). She immediately ran to me for a snuggle. I had been missing Gus and was happy to be with any dog, especially another yellow Lab. After a couple of minutes of hugs and pats, Bea sat and gave me her paw.
Again. And again. And again.
“Um, is she trying to tell me something?” I asked Sato-san.
If Stewie had done this back in Chico, I would have made an appointment with my doctor, stat. But Sato-san assured me that Bea doesn’t alert to cancer on people, and that this isn’t her signal anyway. This was just the universal dog signal for “Please?” or “More?” It made me feel better when she headed over to Dr. Miyashita and gave him her paw several times as well.
Sato-san called her to one end of the room and tossed the ball for her a few times. It was her warm-up, a little thank-you, a down payment before her detection work began. “Let’s go and smell,” he told her quietly in Japanese. Today’s demo was on urine samples. One test tube contained urine from a woman with breast cancer, the rest contained specimens from women without breast cancer.
Rows of test tubes were lined up on the floor in wooden boxes with screened tops. Bea walked from one to the next. She seemed tentative sometimes, not quite sure. It wasn’t the quick confidence I was used to with detection dogs. But judging by her reward—the tossed ball—she had gotten it right. Sato-san changed out samples while she wasn’t in the room and did it again. She alerted accurately again.
This was just a demonstration. During studies, Sato-san is not in the room, although the experimenter handling the dogs is. I wondered if this could somehow influence the dogs, even though the experimenter is blind to which sample is positive.
He says the dogs work fifteen minutes, three times a day at most. He wants to keep them well rested and not overworked. To this end, after the demo on the way back to the car, we walked by a long in-ground swimming pool under construction in front of the training center. Sato-san said he was having it built for the dogs, as an extra-special treat in the hot, humid summers.
“They work hard. They deserve to play hard,” he said.
Dr. Miyashita wrapped up his forty-five-minute talk to the people of Kaneyama just as I ran out of details to write about our trip to St. Sugar. A town official moderated a question-and-answer period. I heard him say “Maria-san” a couple of times toward the end. Heads turned toward me, and I smiled and nodded dumbly. I vowed to myself that if I ever came back, I would know a lot more Japanese.
I spent the night in a cozy Alpine-style hotel in Kaneyama that serves as a ski lodge in winter. Its German name, Hotel Schönes Heim, surprised me, but inside, Japanese hospitality flowed, as did the water in the hot onsen spa bath.
The next day, Dr. Miyashita and Shibata-san picked me up and drove me back to the cedar gift store. Upstairs, we met with a seventy-seven-year-old woman, Eiko Tan. She sat on a bright orange couch next to Dr. Miyashita and told me about her late husband. “Everybody loved him and he loved everybody. He was always so happy. He would play golf in the rain and come home smiling and wet,” she said. A tear came to her eye and she blinked it away.
She said that even the dogs of the town loved him. When he walked around town, it was like a parade behind him. He was the pied piper of dogs in Kaneyama.
They were together for forty-nine years. When doctors discovered his stomach cancer, it was too late to do anything that would save him. His doctor was honest with him and told him he should spend as much time with his family as possible. His widow said she wishes the dogs were sniffing cancer when her husband was alive.
“Maybe the dogs can save other people in the future,” she said, smiling at Dr. Miyashita.
“I hope so, very much,” he said. “We are working so hard on this.”
We spoke with other older residents whose spouses had died of stomach cancer. Some had given urine samples in hope of helping advance science so others wouldn’t have to succumb to the disease. Others didn’t want to. They felt it was too strange.
Our last stop was a traditional old Japanese home, with tatami-mat floors and sliding rice-paper doors called shoji. Four generations reside there. The oldest is eighty-two-year-old Tsuruko Chigahara. She smiled with her whole face and seemed delighted as she welcomed us into the large home she shares with her son, daughter-in-law, granddaughter, grandson-in-law, and great-granddaughter. We took off our shoes, and she ushered us to a nearby room to sit on floor cushions around a low circular black table.
Chigahara-san adjusted her floral bib apron and asked if we’d like some food. She and her son brought in several plates of fresh, homegrown edamame, strawberries, apples, pears, tomatoes, and some fruits I didn’t recognize. They laid them on the table next to our teacups and a pot of green tea.
We wiped our hands with hot, wet washcloths and enjoyed the bounty as she told us about her husband, who was diagnosed with stomach cancer five years earlier. He used to smoke and drink—both risk factors for stomach cancer—but he hadn’t for years. He had had H. pylori, but it was long ago eradicated. He didn’t want surgery or chemo. He was eighty and had lived a full life. She said he didn’t want the pain and problems caused by interventions. When it was his time, he would go.
Chigahara-san had married him when she was twenty-one years old, after relatives set them up. “It was the only way, back then.” She laughed, and her eyes smiled. She said her husband was a group leader in construction, and was a little rough around the edges, but always tried to make her happy.
“Later in life he became kind and gentle naturally, to everyone,” she said.
Her husband lived five years after his diagnosis, and only had noticeable problems at the very end. He went into the clinic because he got a little wobbly the week before he died, but he came home and was talking with his family until the night before he died.
He had passed away in this house a little more than three months before my visit.
Chigahara-san brought me over to a large altar on the other side of the room. It seemed to be made of mahogany, or some other deep, smooth wood, and had multiple levels. On top in the center a golden Buddha statue sat in prayer, under two round lanterns. Golden vases with golden metallic flowers sat on both sides of the altar, and on a shelf below, golden platters held three oranges each. There was a book of some sort, with family names. Maybe a family record of births and deaths, Dr. Miyashita later told me.
On the shelf below was an assortment of wrapped snacks—her husband’s favorites. I noticed that the yellow gift bag containing a box of special chocolates I’d brought from San Francisco had been placed there. It was touching to see how she had automatically put a gift on his altar, as if to let him enjoy it, too. Candles, papers, fresh flowers, a small gong and wooden hammer, and bowls with incense lay at the altar’s base. In front of the altar was a framed photo of her husband of sixty-two years. He was probably in his midsixties in the photo, clad in a formal black robe with a white lining around the neck. His mouth was set in a serious pose, but his raised eyebrows gave the impression that he may have been amused by all the fuss of taking a formal photo. It seemed like an expression he’d picked up from his wife.
As she walked us to her door at the end of the visit, Chigahara-san giggled at the idea of dogs helping people know they have cancer. “Dogs are good animals. Please wish them good luck.”
Dr. Miyashita and a young clinic doctor, Kyoichi Seo, were meeting that afternoon with the five people the St. Sugar dogs had indicated were positive for cancer. They told me the people were understandably concerned. Dr. Miyashita said he would discuss their cases individually, suggesting various medical follow-ups, and letting them know that the dogs are good but aren’t perfect. He would explain that this is the first time for such a screening, and it’s a time of learning and refining, and he’d reassure them that they’d be carefully monitored.
These were private patient meetings, so I decided to go for a walk. Shibata-san joined me because he wisely thought it was best to have a guide, even though he didn’t speak more than a few words of English.
It was drizzling, so we spent a little time enjoying the natural beauty of the town, and more time inside buildings: A children’s library above the post office. A sumptuous old café decorated with antiques and featuring a variety of typewriters. Another converted miso storage building. I couldn’t tell what it was now, but it looked interesting and smelled a little like a bonfire.
The door was open, so we walked in.
A man welcomed us to Miura Pottery. Its main room stood two stories tall, not including the peaked roof of the building. Coffee-colored wood beams contrasted with the clean white walls. It was a perfect backdrop for the traditional pottery and paintings around the gallery. The bonfire smell I’d caught outside had come from a hearth, called an irori. Three women sat around it, drinking tea from a cast-iron teapot kept warm in the embers.
The man, Seitero Miura, didn’t speak English. But thanks to a couple of different mobile apps, and with Shibata-san helping with apps of his own, we were able to communicate.
I told him why I was visiting Kaneyama. He typed something and showed it to me when it came out in English.
“My father died of stomach cancer.”
I typed on my app: “Oh I am so sorry! When did he die?”
“I was 11. I am now 28. 17 years ago.”
“You were so young. Did he live a long time with it?”
He brought over his mother, Sanae Miura. He said something to her, and she spoke into his phone as the app translated.
“We thought it was colon cancer. He had pain in his lower belly. He had stage 4 stomach cancer.”
I learned that she and her son had moved the gallery to this building a few years ago, but had always lived nearby. They are both artists.
I looked around at the pottery and asked Miura-san if his father, Masashi Miura, still had any pottery there. He showed me three large vases. I asked if there was something smaller. He brought me to a table with several miniature cups that looked like large sake cups or miniature teacups.
Of all of them, only two had been made by his father. Each year, there is less of his father here. I bought a single cup. It has a base of bare clay, and the rest of the cup is painted deep brown with a glazed rim the color of milk chocolate.
The cup made it through the rest of my travels in Japan. Every time I drink from this little cup, I think of its creator and his family, and all the people in the Mogami district lost to stomach cancer. I say a toast to them, and I wish the dogs good luck: Ganbattene! A lot of people are counting on them.
It has been a year since my visit. The second year of screening has been approved by the town council, even though the first year ended with disappointing results. Of the 924 samples collected, dogs alerted to nineteen as positive. Of those positive samples, so far only one person has been found to have cancer. Among the 905 “negatives,” five residents have been diagnosed with cancer.
Dr. Miyashita does not blame the dogs. He says the failure is all human—part of the learning curve. “It is a matter of learning how to help the dogs do this consistently and better. We are all a little frustrated by the results but still optimistic to believe the dogs’ ability to sniff out cancer samples. The question is how to design the screening test.”
There could be many explanations for the results. Maybe looking for all cancers at once is too much. If the dogs had been specifically trained on gastrointestinal cancer only—almost all studies have focused on one cancer—maybe they would have been more successful.
Dr. Miyashita thinks one reason for missing the positives was the size of the test tubes (tiny, and tops screwed on tightly). He says that once they moved new samples to larger test tubes with tops screwed on but less tightly, the dogs alerted to each of the five they had missed.
As far as the nineteen people the dogs alerted to, he said it could have something to do with reward methods. He’s aware of Dr. Hackner’s study that showed the problem with dogs who can’t get rewarded in a screening scenario. He isn’t the dog expert, so he hadn’t been working closely with Sato-san on the dog side on detection methods, but they’ve now developed a more consistent reward system.
He and the local doctors will continue following the people whose samples the dogs marked as positive. “CT and endoscopy, blood and urine tests were all done for positive cases, but only one of those was found to have cervical cancer that was missed through the routine cervical checkup,” he explains. “CT scans successfully visualized the tumor. Among other positives, they may have the latent tumors that may include thyroid tumor, prostate tumor. These are less malignant cancer and sometimes are carried lifelong. These are extremely hard to diagnose.”
In the second year, the dogs will be screening fewer people—about six hundred—so Sato-san and his assistant can take more time with the dogs and run them through in a more careful manner, with the new and inspiring rewards system.
“I know that most scientific action begins with the failure,” says Dr. Miyashita. “We are now set to do better, and hope others can learn from our mistakes and, I hope, from our future successes.”
With all the work that’s going into dogs detecting cancer, it may come as a surprise that most researchers don’t envision dogs as the final stop on the road to better cancer detection.
As good as the dogs appear to be at sniffing cancer, they are not predictable machines. That’s part of the joy of working with them, but for something as serious as cancer detection, reliability trumps fur appeal.
Dogs are happy and willing to do the job, but researchers are aware of their limitations. After the poor results of his lung cancer screening-style study, Dr. Hackner wrote in the Journal of Breath Research about how dog personalities play into achieving accurate and reproducible effects. “Dogs can only be trained for a limited set of applications and get tired, thus requiring a high turnover. In contrast to analytical instruments, dogs are subject to boredom, limited attention span, hunger, fatigue and external distractions.”
Even if researchers figure out how to accurately screen for cancer using dogs, there would probably never be as many dogs as would be needed.
One day, dogs may be entirely out of the cancer-sniffing picture—replaced by analytical instruments that will be able to detect cancer easily, painlessly, inexpensively, and reliably. This type of device is commonly known as an electronic nose, or e-nose. It also goes by nanonose, electronic olfactory, or artificial nose. (I think it should be called Dog-e-nose—a little play on words that gives credit where it’s due.)
The work dogs are doing now may help make e-noses for cancer viable. Many of the researchers you’ve read about here, including Claire, Dr. Otto, Dr. Hackner, and Dr. Miyashita, are collaborating with other scientists and cancer-sniffing dogs to create e-noses.
Since e-noses detect patterns in complex mixtures of VOCs, the scientists are trying to isolate and identify the components of cancer a dog might be smelling as that cancer’s olfactory signature. They typically use a chemical-analysis technique called gas chromatography–mass spectrometry, or GC-MS. Cancer-sniffing dogs are given parts of a cancer odorant that have been isolated using GC-MS. If the dogs alert to those components, scientists will try to isolate them further and give them back to the dogs to see what they think. The process goes back and forth, with dogs informing the scientists and scientists checking back in with the dogs.
The amount of odor pulled from the GC-MS is much lower than what the dogs typically detect, Dr. Essler explains. That’s why researchers at the Working Dog Center and some other cancer-sniffing venues are trying to find the lowest threshold for the dogs’ detection abilities, and why dogs like Bobbie, detecting more diluted samples, cause their trainers to cheer when they show they have what it takes. The hope is that dogs will help researchers find out exactly what part of the scent makes a cancer smell like cancer.
There’s much optimism about this partnership of dogs and researchers.
“If nature can do it, we can do it. It’s just a matter of time,” says physicist Andreas Mershin, PhD, research scientist and director of the Label Free Research Group at Massachusetts Institute of Technology. His group is part of MIT’s Center for Bits and Atoms, described on its website as “an interdisciplinary initiative exploring the boundary between computer science and physical science. CBA studies how to turn data into things, and things into data.”
In keeping with that mission, the MIT team has entered a collaboration with Medical Detection Dogs, Johns Hopkins University, and the Prostate Cancer Foundation. They’re working on a pilot project in part to determine whether the e-nose technology MIT is developing can be used to detect urological cancer.
Most researchers considering the e-nose envision it as a relatively small device doctors could use in their offices. This kind of advance in the war on cancer would be revolutionary.
The MIT team’s goal is bigger than some others. Or, really, smaller. An e-nose for specific cancers at the doctor’s office is great, Dr. Mershin says, but he calls that a “conservative” look at the future. He wants this cancer detection available to anyone, anytime, anyplace, for any cancer. How to do that?
“We want this to be part of your cell phone,” he says. “It smells you day in and day out, all the time, it learns what your normal scent is, and at the first instance of something amiss, sends you a notification to get checked out.”*
Dr. Mershin is confident there is an overall cancer “fingerprint” dogs can recognize, and thinks that as more researchers work with dogs and technology, it will be found.
He’s hoping such a device will be intriguing to a company like Apple, Amazon, or Google, and one will want to work with MIT to get the technology into the phones of tens of thousands of people who want to participate in a pilot study. (Fifty thousand people is his dream number.) Users would need to track their activities while having the phone with them at all times. And they’d need to skew older. The younger population may be more willing to do this, but it’s rarer for them to get cancer, and cancer is what’s going to teach this phone-e-nose and the researchers what they need to know.
If someone in the pilot finds out he or she has cancer, the researchers can go back and look for any signs of minute odor changes. As data accumulate, the researchers can fine-tune. The dogs will still be doing their part behind the scenes until this is perfected.
What about the e-noses that could be used in doctors’ offices? Could they be a reality more quickly? Most scientists won’t guess about the timeline of these e-noses either. There are too many variables.
For a brief moment, I had hope that e-noses might be around the corner. I was in an olfaction laboratory at Monell Chemical Senses Center in Philadelphia, speaking with George Preti, PhD, an analytical organic chemist who has been studying human odors at the company for forty-five years. He’s working with Dr. Otto’s group to find the unique odor signature for ovarian cancer—an essential step to developing an e-nose.
He is seventy-two. He told me he hopes there will be an e-nose for ovarian cancer before he retires. That seemed like great news. At his age, retirement can’t be too far down the pike.
But then I took a good look at him. The Brooklyn native stoops slightly, but he also has a wiry runner’s body and seems to be in excellent shape. He told me he can still bench-press 165 pounds if his arm joints aren’t acting up. He wears a Fitbit.
“So, do you know when you might be retiring?” I asked, my hope waning.
“Oh, I have no intention of retiring in the near future.”
It’s likely to be several years before even larger versions of e-noses are introduced for limited clinical use.
Until then, our faithful companions will be at the forefront of the research, sniffing out cancer in exchange for a bounced ball and happy praise.
“When it comes to the least invasive, the earliest, and most precise cancer detection,” Dr. Mershin says, “the ability of the humblest trained dog still far surpasses our best analytical laboratory tests.”