Slashdot Mirror

← Back to Stories (view on slashdot.org)

If Dogs Can Smell Cancer, Why Don't They Screen People? (scientificamerican.com)

Posted by msmash on from the now-you-know dept.

An anonymous reader shares an excerpt from a Scientific American report: Dogs can be trained to be cancer-sniffing wizards, using their sensitive noses to detect cancerous fumes wafting from diseased cells. This sniffing is noninvasive and could help diagnose countless people, which begs the question: If these pups are so olfactorily astute, why aren't they screening people for cancer right now? Here's the short answer: Dogs do well in engaging situations, such as helping law enforcement track scents or guiding search-and-rescue teams in disaster areas. But sniffing thousands of samples in which only a handful may be cancerous is challenging work with little positive reinforcement. Moreover, it takes time and energy to train these pups, who, despite extensive preparation, still might miss a diagnosis if they're having a bad day, experts told Live Science.

56 of 106 comments (clear)

  1. Uhh by Anonymous Coward · · Score: 1

    If this is true, than something is being emanated into the air and can be detected by a man-made sensor.

    1. Re:Uhh by gnick · · Score: 3, Interesting

      ...something is being emanated into the air and can be detected by a man-made sensor.

      The same can be said for drugs and we still use dogs for that.

      --
      He's getting rather old, but he's a good mouse.
    2. Re:Uhh by ShanghaiBill · · Score: 2

      The same can be said for drugs and we still use dogs for that.

      But how much of that is just theater? Many "drug arrests" at airports are staged, with actors playing the criminals. The rationale is both training for the dogs and deterrence for would-be smugglers witnessing the "arrest".

      A handheld scanner would take away the drama.

    3. Re:Uhh by Aighearach · · Score: 2, Interesting

      Blind tests show that the dog mostly alerts on suspicion from the handler, rather than smelling anything.

      And humans of course have almost as good of a nose as a dog. Dogs have a long snout because of the shape of their mouth. Dogs are good at tracking humans because they're close to the ground and don't have a social aversion to sniffing the ground. Humans do well at tracking if you get them to stick their face down there and do it.

    4. Re:Uhh by Applehu+Akbar · · Score: 2

      If this is true, than something is being emanated into the air and can be detected by a man-made sensor.

      This was my immediate reaction too. Note that your point was answered only by conspiratorial flames about drug-snifffing dogs.

      There is no incentive for dogs to falsely detect cancer, so if some of them can smell it, let's find out what they're specifically reacting to and design a chip for it.

    5. Re:Uhh by gnick · · Score: 2

      ...your point was answered only by conspiratorial flames about drug-snifffing dogs.

      "Conspiratorial flames"? I was just pointing out that drug-sniffing dogs are still in use. The point I was trying to make is that just because something is being emanated doesn't mean you can just grab a sensor off the shelf.

      --
      He's getting rather old, but he's a good mouse.
    6. Re:Uhh by rogoshen1 · · Score: 5, Insightful

      And humans of course have almost as good of a nose as a dog

      Citation absolutely needed sir.

    7. Re:Uhh by bill_mcgonigle · · Score: 1

      It's not the nose sensors, it's the number and percentage of neurons in the brain wired to olfactory pattern recognition.

      Humans have good hardware for many senses, but our visual system is so good that we tend to favor it over the others and don't pay as much attention to the others because of it, which leads to a matter of practice. Dogs just rely on scent far more and get better at it, and are probably thus evolved to be fast learners with those neurons.

      --
      My God, it's Full of Source!
      OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
    8. Re:Uhh by rogoshen1 · · Score: 2

      "Olfaction, the act or process of smelling, is a dog's primary special sense. A dog's sense of smell is said to be a thousand times more sensitive than that of humans. In fact, a dog has more than 220 million olfactory receptors in its nose, while humans have only 5 million"

      That's a pretty wide disparity. My money is on a dog being better at the smellings.

    9. Re:Uhh by stabiesoft · · Score: 2

      You are just so wrong I don't know where to start. I've had hounds for close to two decades.
      Have you ever guided a friend so you can grab a couple of french fries 3 blocks away in the curb? I had a beagle who would pull me on a different route to do just that. 3 french fries.
      Have you ever smelled a rat in a tree 20 feet above you? I had a beagle who the first time he did that at 10pm at night I just thought was barking at the moon. Until I got the flashlight out and saw two beady eyes staring back at me I would not have believed it. After that I knew, and would get a pole to chase it out of the tree. That particular beagle I had was probably the best scent hound I'd had or fostered. He usually had a kill once a week be it a squirrel, possum, rabbit or rat.
      Dogs have a long snout for extra receptors and hounds do this thing where they collect scent in their noses. Humans do not do this.
      Hounds ears are extra long to funnel scent to their noses. Those ears are shaped like they are for a reason.

    10. Re:Uhh by quetwo · · Score: 5, Interesting

      Not quite. I have two SAR (Search and Rescue) dogs, and they are able to able to smell things we just plain can't. One dog can cover a 40 acre plot of land and find every human within about a half hour. He's found a dozen people (live and dead), and in many cases, places where ground-pounders and police already cleared. Our other dog is a sent-specific trailing dog and will follow the scent that a human leave along a trail for miles.

      In all cases, during training and live searches, the handler has no idea where the subjects are. There is no indication that they can follow -- they lead us.

      So -- read up. Your tests are bull shit.

    11. Re:Uhh by gnick · · Score: 1

      If it's in the air, it should be possible to build a cancer-sniffing machine.

      Of course it "should be possible." A lot of things "should be possible" that we're nowhere near accomplishing. We're pretty good with dogs.

      --
      He's getting rather old, but he's a good mouse.
    12. Re:Uhh by jittles · · Score: 1

      He kept on going over to storm drains and I had no idea until somebody pointed out raccoons frequently live there.

      Hmmm. Interesting. I thought it was just because my dog like the smell of stinky things (she really does). But perhaps this is the case.

      I'll watch him follow the scent of a rabbit that he scared out of the bushes perfectly (he's old and has cataracts, so he doesn't see so well). This isn't one offs either. It's 4 or 5 times per walk.

      My dog actually found a squirrel that was buried in snow looking for nuts a few weeks ago. She started digging like crazy in some snow and I thought she was just being her ridiculous self until I saw the squirrel. The areas we walk are full of homeless people and are traversed by hundreds of dogs, so she tends to get pretty distracted. When we used to live in a more rural location she would find deer and foxes all of the time. She'll also pull me down what appear to be random paths, only to find out later that a friend of mine that she loves had just walked that exact route a few hours before.

    13. Re:Uhh by AK+Marc · · Score: 1

      The number of rods in the eye doesn't determine detail. It's primarily about sensitivity. Dogs can detect fainter scents, but that doesn't mean they can distinguish similar scents any better. Discrimination is a brain function.

      --
      Learn to love Alaska
    14. Re:Uhh by Aighearach · · Score: 1

      Instead of asking for a citation, you should just look it up and see if it is something with a bunch of citations.

      Demanding citations from ignorance is absurd, and it isn't going to get you a specialized tutoring session. Look it up, say something more intelligent.

    15. Re:Uhh by Aighearach · · Score: 1

      My money

      Exactly, when you don't read enough science news to know that there were important papers on this recently, then all you can do is make a guess. The most natural guess is whatever the old (known incorrect) belief was, in this case, that dogs are better at smelling than humans.

      It goes right along with the misconception about the shape of the dog's mouth, that it would have a better sense of smell. Instead of, it has a long mouth so that it can bite at the legs of running animals, and that actually gets in the way of scent and requires oversized sinuses to compensate. It is easy to be confused just looking and guessing.

      If you want a really good sense of smell, look at a vulture, which has a pretty small nasal opening, or a wood burrowing beetle, which can smell freshly exposed wood from miles away with a tiny nose. Cats have an excellent sense of smell, and a short nose. But they don't need a long mouth, because they don't bite at the legs.

      The reality is that humans have a really good nose, comparable in sensitivity to most predators. We're generalists with lots of sensory processing. We're so good at sensory processing that we have few physical defenses, and can challenge large predators with pointy sticks or rocks.

    16. Re:Uhh by Nestea_Zen · · Score: 1

      Nature is amazing

  2. Real answer by Anonymous Coward · · Score: 1

    Drug sniffing dogs mostly respond to the officers' body language telling them to start barking to establish probable cause. Actually training them to respond to the smells is hard.

  3. A good screening test by sjbe · · Score: 1

    Moreover, it takes time and energy to train these pups, who, despite extensive preparation, still might miss a diagnosis if they're having a bad day, experts told Live Science.

    All screening tests have false positives and false negatives. That's why they are screening tests. A good screening test is cheap, fast, and has few (close to zero) false negatives and a modest number of false positives. Anyone with a positive test gets sent to follow up with more accurate and costly testing.

    That said the article is right in that dogs really aren't sufficiently reliable. Same problems exist with search and rescue dogs. If it isn't fun for them even the best dogs get bored and stop cooperating. A better approach would be to try to figure out what the scent is and to develop a mechanical nose to replicate the functionality. The dogs should be sufficiently reliable to help development of a sensor.

  4. Cynical much? by sjbe · · Score: 3, Insightful

    If dogs became mainstream at detecting cancers, there would be immediate palpable disruption to the present status quo. Schools that train doctors, radiologists and the whole associated ecosystem would be in peril.

    Nice try but no. Under the best of circumstances dogs wouldn't be more than a cheap form of screening which would have to be confirmed by other more reliable methods of detection. Screening tests are useful but don't replace entire ecosystems of medicine. At most dogs can tell us that something is going on in the patient but they cannot provide much in the way of details.

    Who in this industry, would support such a move? I do not see any.

    My wife is a pathologist and she would happily support dogs being used to detect cancers if it were a practical and reliable approach. Nearly all doctors are more than happy to utilize any tool that will get better results for their patients. Your cynicism is misplaced.

    1. Re:Cynical much? by kaatochacha · · Score: 1

      This argument is silly, the whole "Oooh, the cancer industry stops any fix for cancer, they could cure it but won't". You know why? Because the first Illuminati cancer industry exec who actually got cancer would immediately turn around and allow the secret fix.
      Unless, of course, they're already secretly doing it, and magically keeping thousands of cancer cures secret...

  5. Scent detectors by sjbe · · Score: 4, Interesting

    The same can be said for drugs and we still use dogs for that.

    That's because it is a hard problem to solve and dogs are exquisitely evolved to detect scents. Learning to replicate even a fraction of that functionality will take many years of hard work. And yes, top people are working on it.

    1. Re: Scent detectors by Anonymous Coward · · Score: 1, Insightful

      No, it's because dogs can key off of subtle, hard to detect signals from their handlers, and then "detect" the drugs and explosives which are carried by the people with the wrong skin color.

    2. Re:Scent detectors by nelvinboy · · Score: 1

      Yes, it's an obvious jump and it's already being reported by mainstream news outlets: https://www.nytimes.com/2017/0...

  6. Bees by gurps_npc · · Score: 4, Interesting

    Bees are just as good as dogs at sniffing things, including drugs and explosives.

    You train the hive ONCE, and they train each other after that.

    Unlike dogs, they have much longer working rules. They don't need as much rest or reward.

    http://www.seattlepi.com/news/...

    http://mentalfloss.com/article...

    --
    excitingthingstodo.blogspot.com
  7. Re: it's not "cancerous fumes from diseased cells" by Oswald+McWeany · · Score: 1

    Pregnancy is cancer.

    Nah... Fetuses are a parasites. They feed and grow off a host organism leaving the host organism weaker and less fit whilst carrying the parasite.

    --
    "That's the way to do it" - Punch
  8. Re:One other possible reason by Tablizer · · Score: 1

    You can't patent dogs and create a monopoly.

    Don't underestimate the power of deep-pocketed lobbyists. If they can bribe corporations into being people, then dogs can be corporations (or something like that).

    --
    Table-ized A.I.
  9. More Important Question... by avandesande · · Score: 1

    What are dogs smelling in the air when they detect cancer? Obviously not the cancer cells themselves, it must be a mixture of volatile compounds that could be detected with mass spectroscopy or other technique on the spot. Certainly less invasive than a biopsy...

    --
    love is just extroverted narcissism
    1. Re:More Important Question... by Immerman · · Score: 3, Interesting

      Actually no. Skin and lung/throat/mouth cancers are perhaps the easiest to smell, but dogs have been shown capable of relyably detecting a wide range of other cancers as well, along with a wide range of other diseases with no obvious surface symptoms. Presumably anything that alters your body's chemistry has a fair chance of introducing telltale molecules into your sweat, breath, urine, etc. And cancer uses metabolic pathways not normally used by healthy cells (most produce energy in the cytoplasm rather than the mitochondria), which probably means unusual waste products dumped into the blood stream.

      --
      --- Most topics have many sides worth arguing, allow me to take one opposite you.
  10. WTF? by JBMcB · · Score: 2

    You think that, once a dog says that you have cancer, that's it? You don't need oncologists? You don't need radiologists? You don't need nurses?

    The cancer dog sniffs you, wags his tail a certain way that means you have Leukemia, you take a Leukemia pill and that's it?

    Have you had any experience with the medical profession at all?

    --
    My Other Computer Is A Data General Nova III.
  11. The post answers it's own question.. by Vegan+Cyclist · · Score: 1

    "But sniffing thousands of samples in which only a handful may be cancerous is challenging work with little positive reinforcement. Moreover, it takes time and energy to train these pups, who, despite extensive preparation, still might miss a diagnosis if they're having a bad day, experts told Live Science."

    Next mystery: why do I get wet every time I have a shower?

  12. Misplaced cynicism and anger by sjbe · · Score: 1

    If cynicism is truly misplaced here, care to tell me why the fuck we haven't evolved in other areas of detection where we still use dogs?

    Because we haven't got better technology yet. You seem to be under the delusion that scent detection is an easy thing to do. It isn't. Replicating even a fraction of a dog's nose is a problem as difficult as replicating human visual recognition. We've had some results but we're still not very good at it yet. There is lots of money to be made with artificial noses so I have little doubt it will get figured out eventually but until then we have to work with what we've got and what we've got are dogs.

    Using "sniffing thousands of samples in which only a handful...is challenging work with little positive reinforcement" as an excuse to not use dogs to screen cancer seems pretty damn weak when that is essentially exactly what drug and bomb sniffing dogs are doing all day, every day.

    We can do it if you are willing to indemnify the doctors for the results just like the police are indemnified against being sued for false positives and false negatives. Right now if the dog misses the diagnosis the doctor is the one who will be sued. Police use dogs because they work (even if imperfectly) and they are effectively immune to liability for useing them.

    Your wife does not directly speak for the trillions made by the cancer industry.

    She doesn't have to. The claim was that no one in the industry would support using dogs to detect cancer which is a claim that is demonstrably false.

    1. Re:Misplaced cynicism and anger by david_thornley · · Score: 1

      We already have screening tests that aren't real accurate but serve to find people more likely to have some problem. Are there many lawsuits over mistakes in screening tests? My wife went through a screening test that showed (inaccurately) that she had Hepatitis C, and nobody ever mentioned a lawsuit.

      --
      "When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
  13. There is no begging by smooth+wombat · · Score: 4, Informative

    This sniffing is noninvasive and could help diagnose countless people, which begs the question:

    It raises the question. It does not beg the question.

    --
    We will bankrupt ourselves in the vain search for absolute security. -- Dwight D. Eisenhower
    1. Re:There is no begging by rogoshen1 · · Score: 1

      maybe it was a play on words, dogs, begging... c'mon expound your mind! It's not that hard to see outside the box. The forest with trees and all that.

  14. "Hello, ladies..." by PopeRatzo · · Score: 4, Informative

    If dogs are cancer-smelling machines, then every single dog in my local dog park must have cancer of the ass.

    --
    You are welcome on my lawn.
    1. Re:"Hello, ladies..." by PopeRatzo · · Score: 1

      Or maybe my dog's just providing free ass cancer-screening services to every dog in the dog park. Yeah, I like that idea better.

      --
      You are welcome on my lawn.
  15. Re: They [conveniently] miss the biggest reason... by Lanthanide · · Score: 2

    Um, why would they be in peril? Their job is to cure cancer. If a new tool helps them find more cancers earlier, they'll have more cancers to cure (as opposed to cancers that aren't found until it's too late and kill people).

  16. You: Dr, I have a history of cancer in my family by rsilvergun · · Score: 5, Funny

    Dr: Please step right into this room full of adorable puppy dogs.

    vs

    Dr: Please step into this room FULL OF BEES.

    --
    Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
  17. What is the point of this post? by kenh · · Score: 1

    The summary describes the question, then provides the answer!

    --
    Ken
  18. Why not build a machine that smells cancer? by Brannon · · Score: 1

    Why do we treat smell as though it's some sort of metaphysical sixth sense that cannot be replicated by something man-made?

    Dogs can do amazing things, but they have approximately the same attention-span and cognitive function as a 2 year-old human--so maybe let's not trust cancer diagnosis to them.

    1. Re:Why not build a machine that smells cancer? by bws111 · · Score: 1

      In order to make a machine to do the same thing, you would first have to know what it is you are looking for. And we don't know that. Sure, you can make a detector to smell a gas leak, we know what gas smells like. And you can make a detector to smell BAC, we know what that smells like. But we have no idea what the dogs are smelling when they smell cancer.

      You can imitate a human eye pretty easily with a single light sensor and some processing. The human nose is not a single sensor (like an eye), it is thousands of sensors, all detecting different molecules. And until you can replicate that (and we aren't even close), you can't use machine learning to do what a dog is doing when he detects a 'cancer smell'.

  19. Re: They [conveniently] miss the biggest reason... by kenh · · Score: 1

    Schools that train doctors, radiologists and the whole associated ecosystem would be in peril.

    Right, because once a dog 'detects' cancer in a person, there's no need for trained medical professionals (AKA doctors, radiologists, etc) - you just open a book that decodes the dog barks into a treatment plan then have Amazon drop-ship you a complete med kit to combat the cancer...

    Detecting cancer =/= treating cancer.

    --
    Ken
  20. Re:You: Dr, I have a history of cancer in my famil by Immerman · · Score: 1

    It'd certainly cut down on the number of hypochondriacs wasting doctors' time...

    --
    --- Most topics have many sides worth arguing, allow me to take one opposite you.
  21. Re:One other possible reason by rogoshen1 · · Score: 1

    over the internet does anyone know they're actually corporations?

  22. Bees by antdude · · Score: 1

    Basically, http://www.youtube.com/watch?v... ... :P

    --
    Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
  23. False positives by satsuke · · Score: 1

    Because you have the potential for a ton of false positives that cause mental anguish of "do I have cancer" and related testing to rule out an actual positive.

    The role the dog (or bees, or what have you) play currently is probably the right one, used only when needed and validated with other methods with higher precision.

  24. Invert the logic - save the day by burni2 · · Score: 1

    citation: "But sniffing thousands of samples in which only a handful may be cancerous is challenging work with little positive reinforcement."

    changed: "But sniffing thousands of samples in which mostly all are not cancerous is -still- challenging work with much positive reinforcement."

    -> Train the dogs to _not_ react on cancerous people.

    1. Re:Invert the logic - save the day by david_thornley · · Score: 1

      Or throw in a patient with known cancer now and then, to raise the positive rate.

      --
      "When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
  25. It doesn't "beg the question." by Doug+Jensen · · Score: 3, Insightful

    It doesn't "beg the question." Check your dictionary (despite the overwhelmingly incorrect use of that phrase.)

    --
    Doug Jensen
    1. Re:It doesn't "beg the question." by Doug+Jensen · · Score: 1

      Let's see...that's the name of that book which included a theme about how modifying language modifies thinking? Oh, yes, I remember several books about that. And of course we are experiencing that action and effect in our political and social discourse these days too. So what does linguistic and cognitive corruption matter ...

      --
      Doug Jensen
    2. Re:It doesn't "beg the question." by Doug+Jensen · · Score: 1

      Let's see...that's the name of that book which included a theme about how modifying language modifies thinking? Oh, yes, I remember several books about that. And of course we are experiencing that action and effect in our political and social discourse these days too. So what does linguistic and cognitive corruption matter ...

      "What's" not "That's."

      --
      Doug Jensen
  26. Dogs need positive reinforcement. by Hans+Lehmann · · Score: 1
    Search and rescue dogs, or drug sniffing dogs, need immediate positive reinforcement when they've done their job correctly. Fido finds the injured hiker? Good boy Fido, here's a steak. Fido points out that hippie when I gave Fido out secret cop signal? Good Fido, here's a hug and a treat.

    When sniffing for cancer, you can't exactly say "Good boy. This poor man has cancer, here's a treat!". You can't base a diagnosis on the reaction of dog, and certainly not in front of the patient. The cancer would have to be verified by biopsy or other means, which if course takes time. Praising the dog for what they did days ago doesn't teach them anything.

    --
    09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
  27. Cat scan by urdak · · Score: 1

    We don't have dog cancer tests yet, but we do have CAT scans!

  28. Re:Very interesting the dog tried to bite it off by TFAFalcon · · Score: 1

    How would detecting cancer have helped the owner to survive in the past?

  29. Cancer detection by sjbe · · Score: 1

    No, it's because dogs can key off of subtle, hard to detect signals from their handlers, and then "detect" the drugs and explosives which are carried by the people with the wrong skin color.

    Do you have a better scent detecting technology available? If not, shut up until you do. In any case the issues with drug sniffing dogs have nothing to do with cancer detection.