The first time I heard about dogs being able to “sniff” cancer in a sample of urine was many months ago. I have to admit that at the time I was very, very skeptical.
The research concept that was reported in an article in the BMJ (also known to many of us as the British Medical Journal) in September of 2004 was fairly straightforward: dogs have a very well developed sense of smell, and maybe they can smell something in a body fluid that may be able to detect whether or not the patient had cancer.
As the researchers reported, up to that time there had been several anecdotal reports of pet dogs alerting their owners to the fact that something was amiss, and that something turned out to be cancer.
What the researchers found was that in a group six ordinary dogs they were able to train the dogs to detect something through their sense of smell in the urine of cancer patients that was different from the urine of people without bladder cancer.
The test wasn’t perfect. The dogs were right 41% of the time, compared to 14% which would have been the expectation if the dogs were just guessing. Despite that caution, the story received a lot of media attention for the obvious reasons: it was offbeat, and it suggested a possible simple solution to a complex problem. Even 60 Minutes weighed in.
I didn’t hear anything more about this concept until this week, when a paper published in a journal called Integrative Cancer Therapies and made available online generated a considerable amount of media interest, probably for the same reasons noted above.
In this more recent article, the researchers reported that they were able to train several local dogs to determine which breath samples came from patients with lung or breast cancer. Not only could the dogs pick out the right sample almost all of the time, but they could also detect samples from folks with even early stage disease.
Before everyone goes out and starts creating cancer screening programs based on the unique ability of dogs to differentiate body fluid or breath samples from cancer patients compared to people without cancer, there a couple of points that have to be made.
First, let’s talk about what may be going on here.
Our bodies in fact are, among other things, pretty neat protein and chemical factories when you stop and think about it. We are made up of cells that create proteins, and create waste products and other compounds that are part of our daily existence.
For many years, there has been considerable interest in identifying chemicals, proteins and other markers in our body chemistry that may give a clue to the presence of cancer. The goal, ultimately, is to be able to measure something through a lab test that is produced specifically by a cancer, and that would allow us to become aware a cancer is present before it can be felt by examination, or found by an x-ray or other imaging test. Then, maybe the cancer could be treated before it ever becomes a problem.
Research into these types of detection possibilities has accelerated over the past several years, as our knowledge of cancer and the quality of our research tools have improved rapidly. A recent editorial in the Journal of Clinical Oncology nicely summarizes these efforts and some of the problems that researchers have faced, including some of the false starts that resulted in considerable excitement but did not pan out.
As part of our “body factory,” there are going to be products produced that are “volatile” in nature. That means they are excreted through the body in a gaseous form and may be associated with certain odors.
For example, when a person has poorly controlled diabetes, they produce a substance called ketones which can be measured in their blood and urine. However, the body also excretes this material through the lungs, much like it gets rid of carbon dioxide while it takes in oxygen. Those ketones smell, and that is what gives the characteristic odor associated with diabetic coma. We doctors usually are able to smell the ketones when they get to a high enough level in a patient’s breath.
Similarly, as a physician, you can also smell alcohol on the breath of a chronic abuser. That is, in fact, one of the ways the body gets rid of the breakdown products of alcohol consumption. Most of us are very familiar with the breath test that police use to determine the alcohol level in a person’s body.
There are many other examples, but I suspect you get the idea: the body creates chemicals that may be produced in small amounts, but in fact in larger concentrations may create a smell that is characteristic of the process. We humans can smell these things when they are present in large amounts, but we can’t smell these chemicals when the concentration in the air is very small. Dogs, with a very highly developed sense of smell, may be able to detect these odors.
If we could prove the association between cancer and the production of some of these chemical products in small amounts that could be smelled by a dog, we could then concentrate on trying to find that compound. If we could do that, then the compound could be measured. We would, in essence, use the skills of the dogs to guide us to find the compound.
It is possible that the dog could be better than our fancy machines. We would be able to short circuit the process of sifting through thousands and thousands of possible candidates to zero in on the one that counts. The compound, whatever it is, may be in the body in amounts too small to actually isolate it without getting a clue from our canine friends.
That is really where, in my opinion, the action is for this particular approach.
I do not envision long lines of lab specimens lined up, with a trained pooch walking down the aisle and sitting down in front of a urine sample or a breath sample when the dog smells the smell.
The experiments themselves and how they were constructed do not point in that direction (pardon the pun regarding pointing). What the researchers did was demonstrate that the dogs were able to pick the cancer sample out from four normal specimens in one test, and from six samples in the other test. The dogs did not have to go down a row of 100 samples, and pick out all the cancer samples from all the normal samples, nor did they go through thousands of samples of healthy patients and find those who had a lung, breast or bladder cancer.
So what have we ended up with after all of this excitement?
We know that your neighborhood pooch can be trained to pick out a cancer sample from samples which came from a group of folks not known to have cancer. We don’t know what it is they smell, and we don’t know how much of whatever compound they smell is present in the body.
We don’t know if this is a real breakthrough. We do know that it has not been demonstrated yet—based on these studies—that this is going to have significant impact on the early detection of cancer.
But you are hearing from someone who was a skeptic that there may be something here. Stranger things have happened in science, and in the treatment of cancer. Far be it from me to be an absolute doubter about this possibility.
Skeptical? Yes, but that’s my nature about a lot of new discoveries and claims. An absolute skeptic? No longer. My mind is open, and I look forward to continued research reports with interest.
