Dog Sniffs Out Cancer Based on Patients' Urine Samples, Study Shows

A trained scent dog could be used to detect thyroid cancer in patients, according to a study to be formally presented in San Diego, California, on Friday, March 13.

The leaked results of the study, set for presentation at the Endocrine Society's annual meeting, showed that a trained scent dog could identify 88.1 percent of the time if a patient has malignant thyroid cancer or benign tumor based on urine samples.

"Current diagnostic procedures for thyroid cancer often yield uncertain results, leading to recurrent medical procedures and a large number of thyroid surgeries performed unnecessarily," said Dr. Donald Bodenner, chief of endocrine oncology at the University of Arkansas for Medical Sciences in Little Rock, and the study's senior investigator.

"Scent-trained canines could be used by physicians to detect the presence of thyroid cancer at an early stage and to avoid surgery when unwarranted," he said.

Bodenner said the dog's accuracy is only slightly less than that of a fine-needle aspiration biopsy, which is used first to test thyroid nodules for cancer.

He said canine scent detection has the advantages of being non-invasive and inexpensive.

Bodenner's co-author, Arny Ferrando scent-trained Frankie, a rescued male German Shepherd-mix, to recognize the smell of cancer in thyroid tissues from several patients.

"Frankie is the first dog trained to differentiate benign thyroid disease from thyroid cancer by smelling a person's urine," he said.

In the study, 34 patients gave urine samples during their first visit to the university thyroid clinic before undergoing biopsy.

"The surgical pathology result was diagnosed as cancer in 15 patients and benign thyroid disease in 19," according to the Endocrine Society.

Neither the dog handler nor the study coordinator knew the cancer status of the 34 urine samples.

Frankie would lie down if he detected cancer and would turn away if the sample was benign.

"The dog's alert matched the final surgical pathology diagnosis in 30 of the 34 study samples, the investigators reported," the society said.

The sensitivity, or true-positive rate, was 86.7 percent, meaning Frankie correctly identified nearly 87 percent of the pathology-proven thyroid cancers. The specificity—the true-negative rate—was 89.5 percent, which meant Frankie knew that a benign sample was actually benign almost 9 of every 10 times.

Thus, Frankie's average cancer detection rate was placed at 88.1 percent.