You smell cancerous

It turns out that dogs are even more incredible than I knew. Most people are aware that a great dog can be trained to:

Guide the blind

Assist the handicapped

Comfort the depressed

Protect the vulnerable

Detect the drugs

What many people do not know is that – over the course of the last decade – dogs have been trained to detect cancer. Yes, that’s what I said. The studies have been carefully conducted and I am completely convinced that these dogs are a worthwhile method of diagnosis. Given that lung and breast cancers are leading causes of death worldwide, methods of early, noninvasive detection offer exciting new possibilities.

Before the first full-blown experimental study was set up, prior work had shown that biochemical markers in the exhaled breath of cancer patients were distinguishable from those of controls. Since it is known that cancer cells emit different waste products than normal cells, it is not unreasonable to assume that this waste smells differently. Unfortunately, the chemical analysis of breath samples has never been well enough understood for individual patient diagnosis.

The Pine Street Foundation, based in northern California, brought in Polish Dr. Tadeusz Jezierski to lead the first large study, explaining that “a dog’s nose, considered by both dog trainers and chemists alike to be one of the world’s most powerful olfactory sensor, was the medical device used in this research.” Authors trained five ordinary household dogs to distinguish exhaled breath samples of 55 lung and 31 breast cancer patients from those of 83 healthy controls. Using food rewards, the dogs were trained to sit in front of a sample from a cancer patient and to ignore the samples from controls. After being trained, the dogs were tested using breath samples they had not previously encountered. During this phase of the trial, the researchers blinded the dog handlers and the experimental observers as to the identity of the breath samples.  

Low and behold, canine scent detection was impressive when compared to biopsy confirmed conventional cancer diagnosis. Among lung cancer patients and controls, the sensitivity and specificity of the dogs’ scent detection were both 0.99. Among breast cancer patients and controls, sensitivity was 0.88 and specificity was 0.98. (Quick reminder: sensitivity is the percentage of detected positives – sick people who are correctly identified as sick – while specificity is the percentage of detected negatives – healthy people who are correctly identified as healthy.) The numbers from the study were incredible; when the study came out, Dr. Donald Berry, chairman of biostatistics at M. D. Anderson Cancer Center, commented that the results were “off the charts: there are no laboratory tests as good as this, not Pap tests, not diabetes tests, nothing.” It is particularly interesting to note that sensitivity and specificity were similar in all four stages of cancer, meaning that the dogs were able to accurately detect even the early stage patients. This is important, given that treatment is much more effective in Stage I and Stage II, before tumors have grown and spread.

Further, these were not even special dogs. After mere weeks of training, household dogs were able to accurately “diagnose” the patients (or, well, their breath). Okay, you say, but we can’t have dogs going around and sniffing everyone’s breath! Maybe not. What this might lead to is future work that determines which chemical compounds can most accurately identify the presence of cancer. Tumors’ metabolic waste products include alkane and benzene derivatives that differ from those of healthy cells. With that knowledge, certain types of spectrometry could be used to develop a machine that would read breath samples. Personally, though, I like the idea that a dog would be the one to let me know the news.

A few answers to questions before you ask them:

Yes, the dogs’ performance was consistent. The investigation lasted four months and involved 12,295 separate scent trials.

No, the dogs’ diagnostic performance was not affected by the meal most recently eaten by the patient. Neither was detection affected by age or smoking.

Yes, dogs have long been reported to demonstrate unusual behavior around cancer patients. People have reported that – prior to official diagnosis – their dogs displayed persistent behavior around specific body locations.

No, the trained dogs are not currently available to come screen you for cancer. There is still more to be learned before this method would be recommended and used at the widespread clinical level.

Yes, the researchers are discussing the development of some sort of “electronic nose” device.

No, you should not expect to see either dogs or electronic noses in the hospital this year.

Yes, several of the dogs used were Labradors provided by Guide Dogs for the Blind, the organization based in San Rafael, California that I worked with when my family raised Guide Dog puppies.