Dogs sniffing for cancer? Not so fast, experts say
Prognose 220 Mil, an Israeli startup founded and run by guide-dog trainer Uri Bakeman, says it is using canine smelling abilities to screen for cancer
Cancer experts say that even if dogs have the proven ability to detect cancer by discerning the distinct odor that the cells emit, using them to screen for cancer patients is still far from a reliable method, and cannot be a substitute for regular cancer-screening procedures.
The comments of the experts were made to The Times of Israel amid claims by Israeli startup Prognose 220 Mil, founded and run by guide dog trainer Uri Bakeman, that says it is using the unique olfactory abilities of canines to screen for cancer. The firm is soliciting the public on its website, in Hebrew, to send in samples of their saliva to get tested in the founder’s Negev laboratory at a cost of NIS 400 (approximately $100).
In a phone interview, Bakeman said he trained his dogs using techniques similar to those he had learned studying to be a guide-dog trainer in the US, after his service in the Israeli army.
“It’s the same way you train any dog. You teach him what the objectives are, and you encourage the behavior he shows in that direction,” he said. “It’s the same as training a drug detection dog, or bomb detection dog.”

This is not the first time that man’s best friend has been used to assist in the diagnostic stage of cancer treatment. In 1989, the medical journal The Lancet published a study dedicated to the findings of one dog’s persistent sniffing of a mole on its owner’s thigh, where a malignant tumor was later confirmed. A 2006 study that relied on a double-blind methodology (neither the dogs or their handlers knew which samples were cancerous) found more compelling results that dogs were capable of detecting cancer, but did not necessarily recommend streamlining pups for cancer diagnosis.
Dr. Gad Rennert, chairman of the Department of Community Medicine and Epidemiology at Carmel Medical Center and Technion Faculty of Medicine, as well as director of the CHS National Israeli Cancer Control Center, agrees that the idea of detecting cancer through studying volatile organic compounds (VOCs), the molecules that give off the distinct cancerous odor, is not new, nor is the confirmation that dogs are capable of accuracy in detecting these tiny particles.
He said in an email interview that commercializing the use of dogs for detection never picked up, and the reason, as explained in a 2017 article in Scientific American, may come down to the difficulties researchers encounter with a dog’s personality.
In the article, Dr. Klaus Hackner, the first physician to publish a double-blind study with cancer sniffing dogs, explained how the settings for most of these dog studies did not reflect the reality of mass cancer screening in the real world. For instance, in most of the studies conducted, dogs are given a handful of samples with one always containing a cancerous specimen. In the real world of cancer screening, a dog would likely encounter thousands of samples and find only a handful of cancerous specimens.
And, as Hackner pointed out in the article, if both the dog and handler are unaware of which specimens contain cancer, there would be no immediate positive reinforcement for the dog who has accurately identified a cancerous sample.
“We were not able to provide positive feedback because neither one knew in the screening situation if the dog was right or not. This was stressful for both the dogs and the handlers,” Hackner told Scientific American. His recent work was published in the Journal of Breath Research in 2016, and he conceded that while his was the first to simulate these real world conditions for cancer-sniffing dogs, it was the lack of positive reinforcement that ultimately led to the study “failing.”

“I am very skeptical about this new start-up, especially since there is no validated standardized procedure for canine cancer detection. They charge people, this makes me further suspicious,” Hackner said in an email to The Times of Israel.
The dog trainer Bakeman said he was inspired to start his own cancer-sniffing dog company after learning that the National Health Service (NHS) in the UK had partnered with Medical Detection Dogs (MDD), a UK-based company that has been training dogs for the past 15 years to detect the odor of diseases, to run two approved clinical trials with different hospitals in that country.
“I said to myself, Uri, if you don’t start a company now that will enable everyone to do it without any problems, you’ll never be able to do it because it will [catch on like] fire in the world,” he said, adding that shortly after learning about the MMD trials, he began the six months of training required to get his own dog ready to be cancer-sniffing. As of two months ago, his company has been up and running, soliciting samples from people through his Facebook page, and he says he is now receiving up to 20 samples a day.
“We are now 95 percent accurate,” he said, but that accuracy comes from a rather small sample size — “in the tens.” He added that the sample size will be much larger by the end of the year, as he claimed to be in talks ”with one of the major hospitals in Israel,” but declined to specify which one. He also said he is in talks with a few investors “who are interested in investing” in the venture, but declined to disclose who they are. Initial funding for the company has so far come from his own pocket.
Dogs can have bad days too
Dr. Uri Yoel, a specialist in internal medicine and instructor at Ben Gurion University‘s Faculty of Health Sciences, who conducted research on the ability of dogs to smell cancer, said that cancer screening in real-world conditions has not been tested thoroughly enough to say with full confidence that it can be used as a substitute for the current screening programs recommended by physicians.
“Because cancer in the healthy population is not very common, you would need to take a sample of 1,000 healthy people,” said Yoel in a phone interview. “No one did a screening survey of this volume with dogs, because such a study is very complicated and very expensive.”
Yoel’s research, published in 2013, found that dogs were not only capable of detecting various forms of cancer, but that there was also a common smell that they could detect.
“We concluded that dogs can smell molecules that come from cancerous cells and, second, that there is a pattern of cancerous cells produced that are common between different types of cancer,” he said. But, even Yoel, who said he was not familiar with Bakeman’s work and clarified that he was speculating, would not recommend subbing dogs in for traditional screenings just yet.
“I’m not sure about commercial use of dogs for the detection of cancer,” he said.

Another possible issue with relying on canines for cancer screening is the issue of false positives and negatives. This point was also raised in the same Scientific American article by Dr. Hilary Brodie, a professor in the Department of Otolaryngology at the University of California, Davis.
Brodie explained that while today’s screening tests may produce their own false positives and negatives, physicians at least know at what statistical rate these occur. For dogs, however, the false positive and negative rate could vary for each animal, he said. One dog may be more accurate than another, and even an animal’s own accuracy could vary by the day depending on factors like health, mood, or energy level.
“[Dogs] can have bad days, just like you and I,” Brodie told Scientific American. “There’s lots that the dogs can do, but I don’t think wholesale screening of the population is where it’s heading.”
“I think it’s fascinating,” said Prof. Hezi Bernholz, co-inventor of the cancer drug Doxi, on the concept behind Prognose 220 Mil, conceding that if there were no other options, it would be reassuring to know that dogs are capable of sensing the differences between a sick and a healthy person. “But I think you know it’s a very, very uncertain, I would never see them get[ing] the approval by the FDA,” he said in a phone interview.
Bernholz supposed that if 10 physicians were asked whether they would recommend a dog over more traditional screening methods, nine of them would say no. He did add, however, that with more research and advances in technology, we could come to see our furry friends as a complement to standardized screening methods. “Maybe one of those doctors would say yes to the dog. Maybe. If he really liked dogs.”