The nose knows: Canine olfaction helps detect human ovarian cancer

Ovarian cancer is the fifth-leading cause of cancer-related deaths in women in the United States, with about 22,000 cases occurring annually, the American Cancer Society reports. Worldwide, there are more than 204,000 new cases of ovarian cancer annually, accounting for about 4 percent of all cancers diagnosed in women. The highest rates are in the U.S. and Northern Europe.

If diagnosed early, ovarian cancer has a five-year survival rate of more than 90 percent. However, early cancer screening is not currently available—yet. A group of dogs at the University of Pennsylvania may change all that.

In an interdisciplinary collaboration, the University of Pennsylvania School of Veterinary Medicine’s Working Dog Center (photo above courtesy of University of Pennsylvania), the School of Arts and Science’s Department of Physics and Astronomy, Penn Medicine’s Division of Gynecologic Oncology, and the Monell Chemical Senses Center have joined together for a research investigation using canine olfaction, along with chemical and nanotechnology analysis, to detect early-stage human ovarian cancer. A grant of $80,000 from Kaleidoscope of Hope Ovarian Cancer Foundation will fund the project.

The olfactory capability of dogs is unparalleled. According to research from Auburn University, dogs have more than 220 million olfactory receptors in their noses, while humans have only 5 million. A dog’s basic scenting ability, from search and rescue of persons to detecting explosives, is a tribute to its olfactory sensitivity.

“I like to describe it as the dog’s ability to ‘smell in color,’” says Cynthia Otto, DVM, PhD, associate professor of critical care and director of the Penn Vet Working Dog Center.

The research program

The Working Dog Center currently has 16 dogs in foundation training. Of those, three have been identified for cancer detection work. The veterinary research group is waiting for final regulatory approvals to receive the human samples. “Since the human patients consented to give their samples without knowledge of the dogs involved, we want to be perfectly clear that those patients won’t have any concern with that aspect of our work,” Otto says. The collaborating oncologic surgeon, Janos Tanyi, MD, PhD, assistant professor of obstetrics and gynecology, has already collected samples from 31 cancer patients and 30 healthy individuals for a control group.

Once the research regulations are in place, the studies will begin. “We will then imprint the dogs on the tissue samples, and then we’ll test them on the plasma samples,” Otto says. Her team’s work builds on that of Swedish researcher György Horvath, who’s shown that dogs have a 100 percent sensitivity and 98 percent specificity to detect ovarian cancer in plasma and tissue samples.

With the dogs exposed to both tissue and plasma samples, Penn Vet hopes to give the research the best chance for success. The ovarian tissue samples have the most concentrated odor, with fewer conflicting, confounding odors versus plasma. “If we can document that the dogs are good at picking up the odor in plasma after being exposed to the tissue, then we can try moving forward to see if the dogs could be imprinted solely on the plasma samples,” Otto says.

According to Annemarie DeAngelo, the Working Dog Center’s training manager, it will take about a week to imprint the dogs on the odor, then additional training will begin. The team is budgeting about two months to initiate the training, then they’ll test the dogs with fresh samples that are different from those the dogs were trained on.

Are there breeds that might be better suited to this type of work? “It’s hard to tell,” Otto replies. “We have two Labradors and a springer spaniel, both hunting breeds we think will be best-suited for the work.”

While the canine ability to detect the presence of ovarian cancer is well-established, researchers aren’t sure exactly which chemicals the dogs are detecting in the tissue and plasma samples. But they’re getting closer.

“Monell Chemical Senses Center has some preliminary evidence that the chemicals are volatile organic compounds, odorants that are altered in the earliest stages of ovarian cancer, even before the cancer can be detected by current laboratory equipment methods,” Otto says. And although the researchers don’t know what the chemicals are, they can identify a signature from cancer cells grown in culture that are different from natural ovarian tissue cells.

“As cancer cells’ entire metabolic machinery is shifted when they become malignant, their unique odor is evidently a signature the dogs are able to identify, with different cancer types having different signatures,” Otto says. “Some of [Hovarth’s] studies noted that the non-ovarian cell types were not picked up by the dogs. It is really fascinating, and once we know more about these chemicals, it will hopefully allow us to better understand the physiology of how these cancers grow.”

With its pilot grant, Penn Vet is looking at four to six months to complete this line of research (with more work to follow if funded), which they’ll spend training dogs with high sensitivity and specificity to the human ovarian cancer tissues and plasma samples.

The research outcome

Of course, Penn Vet’s ultimate goal is not to send dogs out to every laboratory in the country to perform broad-scale cancer screening. “Though I think a lab full of Labradors would be absolutely fabulous, it’s probably not practical,” Otto says. Rather, the researchers will take what they learn from the dogs and use it to develop a laboratory test for early ovarian cancer screening.

“Though the studies by Horvath have shown promise in dogs being able to be so trained, we want to document that our dogs can do it and be precise,” Otto says. “Once we establish that, the next step is to develop a high-throughput screening method with our colleagues at Monell doing gas chromatography-mass spectrometry, our colleagues in the Physics Department doing nano-sensor technology work, and our dogs.”

The first stage is to prove the principle, establishing that the dogs can be trained to scent the target odors, and to utilize the machinery to do likewise. It will take about a year to 18 months to characterize the components of the chemicals and to refine the analysis to enable the machines and canines to accomplish that goal.

“Once the first step is established, we will determine whether there may be an assessment of tissues that will be less invasive, for example from breath odor samples or urine samples,” Otto says. “Those may be more difficult, because those samples may have components that are more contaminating, that may contain confounding odors. We hope to answer the questions: Can we use the dogs to help make the machines better? Will we be able to use the less invasive breath odor or urine samples via a high-throughput system?”

If the team is successful in its goal is to enable the dogs and machinery to detect early-stage human ovarian cancer, they’ll use both the dogs and the equipment to identify other cancers. But that stop may be a ways off.

“At this point we don’t want to get too diffuse,” Otto says. “We want to confirm with this multidisciplinary approach that the dogs are indeed doing what we think they are, so that there is a real sound understanding of their abilities to detect early-stage ovarian cancer.”

Saving women’s lives

Ideally, the team would like to have a high-throughput, easy methodology to successfully screen women for ovarian cancer. The ideal would be to screen women every six months, even beginning at 25 years of age, especially for those patients with a family history of ovarian cancer. To determine the most appropriate screening time is up to the oncologists, Otto says.

When Otto has spoken with experts at the Fox Chase Cancer Center in Philadelphia, they’ve noted that millions of samples per year need to be screened if the test is going to make a difference in the lives of women. “We know how much of an impact an early diagnosis of ovarian cancer will have to save women’s lives,” Otto says.

Suggested reading

1. Johnson JM. Canine detection capabilities: Operational implications of recent R&D findings. Institute for Biological Detection Systems (now Canine Detection Research Institute), Auburn University College of Veterinary Medicine 1999.

2. Horvath G, Anderson H, Paulsson G. 2010. Characteristic odour in the blood reveals carcinoma. BMC Cancer 2010;10:643.

3. Horvath G, Jarverud GK, Jarverud S, et al. Human ovarian carcinomas detected by specific odor. Integr Cancer Ther 2008;7(2):76-80.