13 August 2011
This week’s question comes from Mitchell Vamos in La Jolla, CA:
I haven’t heard much lately about the cancer-sniffing dogs. Will we one day be able to skip the biopsy and just get sniffed by a pooch?
Writing this blog often follows a predictable pattern:
1) Someone asks a really off-the-wall question about some new cancer “cure” or radical prevention protocol.
2) I look into the actual papers (if there are any) and comment on the research.
3) The papers (or lack thereof) are often heavily flawed, or the media has completely mis-represented the results.
I initially felt a sense of banality or fatalism when researching cancer-sniffing dogs. I had heard of this in passing a while back, but the data at the time was not all that convincing, and had all been produced by one laboratory. That was in 2006.
To my amazement, there has been better science performed on the topic in the last few years. In November 2010 a group in Sweden released a study on cancer detection using dogs trained to smell tumors. (1) Giant Schnauzers (right) were trained to distinguish the blood of recently diagnosed ovarian cancer patients compared to healthy controls. The group reported an extremely high rate of accuracy for the cancer-sniffing pooches. If taken literally, these results would indicate these dogs as better than most early detection screens currently available for ovarian cancer!
There are other studies that have claimed success training dogs to recognize the breath of breast and lung cancer patients (2) and others in detecting bladder cancer from urine samples (3) but their results were not as accurate as current diagnostics. Granted, current diagnostics are both more expensive and invasive than letting a pooch give your face a sniff. One could imagine the low cost leading to more widespread screening.
Now, this might not apply across the board, as other studies using dogs to detect breast or prostate cancer in urine samples did not hold up (4).
I’d be really curious to see if other groups can independently train dogs to accurately detect ovarian cancer from blood samples. If so, it would imply that there are yet undiscovered factors in the blood of ovarian cancer patients (and possibly other types of cancer as well) that can be used as diagnostics for early cancer screening! yay!
I could envision several problems for scaling up cancer detection by dogs. The first of which is it is not possible to have two identical dogs, identically trained, and the subjective nature of professional dogs diagnosing cancer patients means the vetting process might be tricky. Bomb-sniffing dogs are trained in consistent faux settings with explosives hidden at the scene. It would not be feasible to give people cancer just to test the tumor-sniffing pooches. Though I could imagine ways to get around this… it would be tricky. Could you imagine having resident dogs in hospitals? As a researcher working in a hospital environment, I would personally love to have a canine co-worker. Though I might be biased… I grew up with big dogs around!
There is something else that caught my intrigue about these studies. If the dogs are accurately smelling something in the blood, urine, or breath, it might be possible to isolate whatever it is they are smelling and use that for a less subjective diagnostic that doesn’t have bad breath and need to be trained. But then again, it could be a balance of multiple factors the dogs are smelling.
Regardless, this area of research might actually lead to real impact on cancer patients. I’ll be following it with great interest.