February 06, 2011
Belgian Malinois Detects Cancer Too

A female black lab named "Marine" who excelled at using her nose to detect bowel cancer is not alone. A Belgian Malinois in Paris shows a knack for detecting prostate cancer by sniffing urine. Given that dogs are going to sniff urine anyway might as well as make this instinctive desire useful.

Arnhem, The Netherlands, 7 February 2011 -- In the February 2011 issue of European Urology, Jean-Nicolas Cornu and colleagues reported the evaluation of the efficacy of prostate cancer (PCa) detection by trained dogs on human urine samples.

A reminder on why this matters: Dogs show the potential to detect cancers at earlier stages. If cancer can be caught before metastasis then the odds of death go way down.

In their article, the researchers affirm that volatiles organic compounds (VOCs) in urine have been proposed as cancer biomarkers. In the study, a Belgian Malinois shepherd was trained by the clicker training method (operant conditioning) to scent and recognize urine of people having PCa. All urine samples were frozen for preservation and heated to the same temperature for all tests. After a learning phase and a training period of 24 months, the dog's ability to discriminate PCa and control urine was tested in a double-blind procedure.

The dog turned out to be right that one of the controls really had undetected cancer. Good doggy!

Urine was obtained from 66 patients referred to an urologist for elevated prostate-specific antigen or abnormal digital rectal examination. All patients underwent prostate biopsy and two groups were considered: 33 patients with cancer and 33 controls presenting negative biopsies. The dog completed all the runs and correctly designated the cancer samples in 30 of 33 cases. Of the three cases wrongly classified as cancer, one patient was re-biopsied and a PCa was diagnosed. The sensitivity and specificity were both 91%.

This study shows that dogs can be trained to detect PCa by smelling urine with a significant success rate. It also suggests that PCa gives an odor signature to urine. Identification of the VOCs involved could lead to a potentially useful screening tool for PCa.

This is the journal published version of the preliminary report.

What's needed: a heavily automated training program for a large number of dogs.

Share |      Randall Parker, 2011 February 06 10:44 PM  Biotech Assay Services

philw1776 said at February 8, 2011 8:15 AM:

Wicked Pissah as we say in MA!

Next step, train pussycats to sniff for uterine cancer!

PacRim Jim said at February 8, 2011 9:49 AM:

Why stop at dog-level sensitivity?
If it were possible to build hand-held sensors orders of magnitude more sensitive, imagine the possibilities (and the spurious readings).
Imagine scanning millions of people. It would be an irresistible tool for product marketers. Perhaps Apple would put them in iPhones.

Mike Anderson said at February 10, 2011 2:34 AM:

Arrgh! This all sounds great until you factor in all the false positives. With colorectal cancer having a prevalence of 46.1 per 100,000, a test with sensitivity and specificity of only 0.91 gives a posterior predictive value (probability of CRC, given a positive test) of only 0.46%. While that's a ten-fold improvement over random chance, it's nothing to adopt any time soon. This needs more discerning dogs before it needs more dogs.

Radford Neal said at February 10, 2011 11:51 AM:

--- This needs more discerning dogs before it needs more dogs.

These may be the same. If one dog doesn't give accurate enough results, take a vote of three dogs. That will at least get rid of random effects such as a dog being momentarily distracted. There might still be errors due to whatever the dogs are smelling not being 100% indicative of cancer.

Randall Parker said at February 10, 2011 7:11 PM:

Mike Anderson,

If the false negatives are low then dogs could be great screeners to tell who should get a colonoscopy. Getting sniffed by a dog is a lot less invasive than getting knocked out and then probed up your intestines.

Radford Neal,

We really need to see several dogs tested in parallel so we can see if there are any correlations with the false positives.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright