In this study, a single dog was tested for her ability to detect cancer. The tasks were 1) to choose the breath sample that came from a person with cancer when it was randomly placed among four breath samples from people without cancer and 2) to choose the watery stool sample that came from a person with cancer when it was randomly placed among four watery stool samples from people without cancer.
The dog was correct in 37 out of 38 of the stool samples and in 33 out of 36 of the breath samples. The dog was not fooled by samples from people who smoke, or those who had benign colorectal polyps, inflammation or an infection.
Although this sort of detection is promising as a non-invasive means of detecting cancer, interestingly, the dog in this study is reported to lose her concentration in the hot summer months. This is a detail that needs to be attended to because obviously, the need for this sort of detection is not seasonal.