A trained dog can detect the odour of bowel cancer in breath and stool samples to a high level of accuracy, even in the early stages, according to researchers in Japan who hope this will lead to new ways of screening for the disease, not necessarily using dogs, but based on the principle that cancer tumors secrete specific volatile compounds that can be detected with equipment as sensitive as a dog’s nose.

You can read how Hideto Sonoda, assistant professor at Kyushu University at Fukuoka, and colleagues, carried out their study, aided by Marine, a 9-year-old female Labrador retriever, specially trained in scent detection of cancer, in a paper published online first on 31 January in the journal Gut.

Previous studies have already reported that trained dogs can sniff out melanoma, bladder, lung, breast and ovarian cancer.

For example, using just breath samples in the case of lung and breast cancer, and tissue samples in the case of ovarian cancer, trained dogs can tell the difference between samples from cancer patients and healthy volunteers to a high level of accuracy.

Also, using gas chromatography and mass spectroscopy, chemists have identified several volatile organic compounds isolated from the exhaled breath of lung and breast cancer patients as potential candidates for early detection of cancer; however this data is still fairly speculative and should be treated with caution.

For this study, which took place between November 2008 and June 2009, the researchers obtained exhaled breath and watery stool samples from patients with colorectal cancer (CRC) and healthy controls before they underwent a diagnostic test with a colonoscopy.

There were 33 groups of breath samples and 37 groups of stool samples, each group comprising one sample from a patient with CRC, and four control samples from volunteers who were free of cancer.

The samples were randomly separated and put into five boxes. Marine, who was trained at the St Sugar Cancer Sniffing Dog Training Center at Minamibousou, Chiba, in Japan, and her handler were then given the task of picking out a box.

Marine first smelled a standard breath sample from a different patient who definitely had CRC (not included in the test samples), and then sniffed each of the test boxes, and sat down in front of the box that smelled like the standard sample.

According to the Japan Times Online, the chief of the centre where Marine was trained, Yuji Sato, said the dog’s training included “getting her to recognize ingested food from the breath of a person in the belief she might be able to sniff out a disease based on odors inside the body”.

In this study the researchers expressed accuracy of detection in two ways: sensitivity (eg the proportion of people with a disease who have a positive test result) and specificity (eg the proportion of people without the disease who have a negative test result).

The results showed that:

  • For both the breath and the stool samples, Marine showed high accuracy in picking out the CRC sample, compared to diagnosis by colonoscopy (the dog was correct more than nine times out of ten).

  • In the case of the breath samples, the sensitivity of “canine scent detection” was 0.91 and the specificity was 0.99.


  • In the case of the stool samples, the accuracy was even greater: sensitivity was 0.97 and specificity was 0.99.

  • The accuracy was high even for early cancer, wrote the researchers, and these results were unchanged when they took into account current smoking, benign bowel disease or inflammatory disease, or the presence of human haemoglobin or transferrin.

The researchers concluded that their findings show that there is a specific scent for cancer, and that cancer-specific chemicals could be circulating in the bodies of people with cancer.

“These odour materials may become effective tools in CRC screening,” they wrote.

Studies that identify cancer-specific volatile organic compounds will play an important role in opening up opportunities to develop new ways of detecting bowel cancer early, they added.

“Colorectal cancer screening with odour material by canine scent detection.”
Hideto Sonoda, Shunji Kohnoe, Tetsuro Yamazato, Yuji Satoh, Gouki Morizono, Kentaro Shikata, Makoto Morita, Akihiro Watanabe, Masaru Morita, Yoshihiro Kakeji, Fumio Inoue, and Yoshihiko Maehara.
Gut, Published Online First 31 January 2011
DOI:10.1136/gut.2010.218305

Additional source: The Japan Times Online.

Written by: Catharine Paddock, PhD