Labradors and Laboratories

Each Friday morning I awake with anticipation of our weekly lab meeting. Because the research interests of our lab range from epigenetic inheritance to colon cancer and obesity, conversations at these lab meetings are pretty versatile, but usually go something like this:

Scheduled start of lab meeting: Is anyone coming to this meeting? Where is everyone?15 minutes after the scheduled start of lab meeting: Lab meeting actually begins.The next 30 minutes: Mostly professional discussion about recently published scientific paper or presentation of new data from the lab.The final 20 minutes: Transition into discussion about poop, mouse colonoscopies, checking the cervical mucous plug, how to perform the best sperm extraction, which unfortunate lab technician is going to have to collect fecal samples...

As you can tell, the end of the lab meeting is the most lively part of our discussion. In today's meeting we got to learn about this little gem of a study published back in 2011 in which dogs were trained to sniff out colon cancer. Do you have a funny image of a dog smelling someone's rectum yet? How about now?

Turns out the idea of having dogs "diagnose" cancer by smelling a patient's breath, urine or stool samples is not all that new, but somehow has gotten by me for the past decade and I'm just starting to look at the published research. In the case of the study I'm going to share with you the researchers worked with trained dogs that would smell either a breath or watery stool sample (collected during a colonoscopy to make the diagnosis of cancer or cancer-free) of a study participant. Samples of the patients' breath were used to train the dogs. The researchers explain it as follows:

Each cancer detection training session was considered complete when the dog could correctly distinguish between breath samples from a cancer patient and four controls consecutively in dozens of trials. The fundamental training method was a reward-based approach in which the correct behaviour is rewarded by simultaneous play with a tennis ball.

What I think is important to note here is that (1) I've definitely chosen the wrong subject for my thesis because it did not include playing fetch with a dog, and (2) the dogs were trained using samples from other types of cancer than colon cancer. In the end of the training period the dogs could sniff out the patients with esophageal, breast, lung, gastric, pancreatic, hepatocellular, cholangiocarcinoma (bile duct), prostate, uterine, ovarian, bladder or colorectal cancer simply using samples of the patient's breath. This means that some volatile compound (or compounds) exists in both the breath and bodily excretions of cancer patients that is missing or different in cancer-free patients. Moreover, these volatile compounds are seemingly similar between cancer types.

After putting the dogs to the test 74 times (38 watery stool tests + 36 breath tests) the dogs picked out the samples from the cancer patients with amazing accuracy. Remember how I mentioned above that each study participant underwent a colonoscopy to determine whether they had colon cancer or were cancer-free? The dog's cancer diagnosis correlated with the doctor's diagnosis 91% of the time for the breath test, and 97% of the time for the watery stool test. 97%! Maybe one day we'll see Labrador retrievers roaming the halls of cancer clinics, smelling out the patients needing assistance. Or, in a more boring, practical world, we'll discover what it is in the breath and stools of cancer patients that is being detected by the dogs and come up with a lab technique to measure it. Personally, I'm keeping my fingers crossed for doggy doctors.

Sonoda, H. et al. Colorectal cancer screening with odour material by canine scent detection. Gut (2011)