So the presentation from the AUA yesterday that has already garnered media attention is the idea that we may be able to train dogs to detect prostate cancer based on the aroma of male urine.
Several years ago it was initially suggested that dogs were able to detect patients with bladder cancer based on smell. We are not aware of any compelling, long-term follow-up on that research, so we are a little skeptical about the data presented yesterday, However, for what it is worth …
A Paris-based research team has reported that a Belgian Malinois was trained to recognize and identify volatile organic compounds (VOCs) in urine that are derived from prostate cancer cells. First, the research team used urine samples from 33 patients with biopsy-confirmed prostate cancer to train the dog to recognize the characteristic aromas of prostate cancer-derived VOCs. They then trained the dog to differentiate between the smell of urine from cancer patients and urine from non-cancerous controls. Finally, the dog were challenged to find the one urine sample from a cancer patient among five samples when the other four samples were all from non-cancerous patients.
According to the researchers, the dog was able to correctly identify 63/66 samples, with a sensitivity of 100 percent and a specificity of 91 percent. The negative predictive value was 100 percent. This is an accuracy significantly higher than anything accomplished with PSA or PCA3 or any other test available today. However, The “New” Prostate Cancer InfoLink is of the humble opinion that we are going to need a great deal more evidence before we can start to base the decision to biopsy patients on the quality of Lassie’s nose.
In fairness to the canine member of the research team, the Belgian Malinois breed of dog (which is a type of German shepherd) already has a well-recognized history of being able to sniff out bombs in airports and to detect VOCs in urine from other types of cancer patient.