Hounds Astound: Prostate Cancer Found 100% (Around)

ORLANDO, Florida — With an accuracy rate of nearly 100%, 2 specially trained dogs were able to detect prostate cancer with their olfactory system — by sniffing urine samples, according to a new study that is the largest of its kind.
Italian researchers tested each dog’s ability to sniff out prostate cancer in urine samples from 362 men with prostate cancer and 540 men with either non-neoplastic prostate disease or nonprostatic tumors.
The dogs, German Shepherds named Liu and Zoey, had their day, time and again.
Liu had an accuracy rate of 99.0%, with a sensitivity of 100.0% and specificity of 97.8%. Zoey had an accuracy rate of 97.0%, with a sensitivity of 98.6% and specificity of 95.9%.
“This is a real clinical opportunity,” said lead researcher Gianluigi Taverna, MD, chief of the prostatic disease unit at Humanitas Research Hospital in Milan, about the results. The opportunity includes the possibility that specially trained dogs could provide a noninvasive method of cancer detection.
He spoke during a press conference here at the American Urological Association (AUA) 2014 Annual Scientific Meeting.
The dogs were equally capable of detecting low-risk and more advanced prostate cancers. “The dog has a quality, not quantity, response,” said Dr. Taverna.
When evaluating the rare wrongly detected cases, the researchers found no differences between epidemiologic, clinical, or histopathologic characteristics.
The frozen urine samples used in the study were obtained from local Milan hospitals and then thawed for the testing.
The smell tests occurred in a small room with a video recorder mounted on one wall. A dog handler walked a single dog in a circle around a series of mesh covered bowls. The dog went around the full circle once, and then on the second go-round, stopped at specific bowls if they contained urine with prostate cancer odors.
The dogs are not any old mutts, said Dr. Taverna. They are very well-trained dogs who are also experts at detecting explosives. “They are Ferraris,” he explained, referring to the Italian high-performance cars.
They are Ferraris.
Two important questions remain.
First, what do the dogs actually smell? The answer to this is not known, Dr. Taverna said.
Second, how can a dog be used in daily clinical practice? He explained that an international center could be developed to train dogs, and then send them to other countries and medical centers.
An American urologist has a different vision. “Someone discovers what these organic compounds are and uses gas chromatography or the “electric nose” as a sensor to pick them up,” said press conference moderator Brian Stork, MD, from West Shore Urology in Muskegon, Wisconsin, who is a member of the AUA social media committee.
Electronic nose technology, first developed for the military to detect chemical warfare, analyzes odors.
In a recent proof-of-concept pilot study, one such device — the ChemPro 100 eNose (Environics) — detected prostate cancer from a urine sample with nearly 80% accuracy.
Electronic noses have also been shown to identify lung cancer with a high degree of accuracy, as reported by Medscape Medical News.
However, in commenting on the Italian study, Dr. Stork, who is a self-proclaimed dog lover, said that it would be “more beautiful” if dogs were used to detect prostate cancer.
Indeed, there is a growing body of literature on the ability of dogs to sniff out cancer. Previous studies have reported on dogs that can smell and detect lung and breast cancer from breath samples and colon cancer from stools, and there has been anecdotal evidence suggesting that dogs can detect melanoma and bladder and ovarian cancer.
There has been one other study of dogs sniffing out prostate cancer in 33 patients, according to AUA press materials.
The Italian researchers are studying the volatile organic compounds that are behind the odorous signature of prostate cancer.
Dr. Taverna and Dr. Stork have disclosed no relevant financial relationships.
American Urological Association (AUA) 2014 Annual Scientific Meeting: Abstract PD19-01. Presented May 19, 2014.