IT HAS long been speculated that dogs have the ability to detect cancer in human beings – or at least cancerous cells, and this ability has been borne out by tests conducted over the years. However, official scientific research has now confirmed what the anecdotal tests had already proved: That dogs CAN sniff out cancer and even detect unconfirmed cases.
A study using six trained dogs, a Labrador, three Cocker Spaniels, a Papillon and a mongrel, found that their ability to detect a sample from a patient with bladder cancer was well above chance.
The dogs correctly selected the sample of bladder cancer urine on 22 out of 54 occasions, including one sample from a patient who had previously been declared healthy. That is a success rate of 41 per cent compared to the 14 per cent expected if the dogs had selected a sample at random.
The research, led by Dr Carolyn Willis, a scientist at the Department of Dermatology, Amersham Hospital, Bucks, grew from anecdotal evidence that dogs could sniff out the skin cancer, melanoma.
The report, published in depth in the British Medical Journal explained the procedure: "Six dogs of varying breeds and ages completed a seven month period of training. All were familiar with obedience commands, but none had been previously trained for search or scent discrimination tasks. We made no attempt to include dogs with a particular suitability for scent discrimination."
The training objective was to enable the dogs to discriminate between urine from patients with bladder cancer and urine from diseased and healthy people, using samples entirely new to them, so as to preclude simple memory recognition of participants' unique odour signatures. Dogs were trained to detect ("alert to" or "indicate") one urine sample from a patient with bladder cancer placed among six control specimens.
Training took place with one operator, using clicker training. The dogs were taught to indicate the appropriate sample by lying beside it. Early recognition of the tumour scent was achieved by using search and find games, which were gradually replaced by discrimination phases of increasing complexity. Urine from patients with bladder cancer was presented sequentially against water, diluted urine from healthy people, undiluted urine from healthy controls, urine (containing blood) from menstruating women, and urine from patients with non-malignant active or recent urological disease or other disease. Samples were not pooled at any stage. Two of the dogs were located 150 miles from the study centre and were trained and tested with dried urine samples. The remaining four Buckinghamshire based dogs were provided with freshly defrosted, liquid specimens throughout.
Quoted in the British Medical Journal Dr Willis said the dogs seemed to have the ability to pick up a "chemical profile" of the cancer probably made up of several volatile chemicals.
"We see great value in our findings and we think that trained dogs will be able to help us to identify the key chemicals. This could help us to develop a medical screening instrument that would help in diagnosis."