Testing ovarian cancer cell lines to train dogs to detect ovarian cancer from blood plasma: A pilot study

• Published in: Journal of veterinary behavior 2019-07, Vol.32, p.42-48
• Main Authors: Murarka, Monisha, Vesley-Gross, Zoë I., Essler, Jennifer L., Smith, Paige G., Hooda, Jagmohan, Drapkin, Ronny, Otto, Cynthia M.
• Format: Article
• Language: English
• Subjects: cancer odor profile; cell lines; medical detection dogs; ovarian cancer
• ISSN: 1558-7878
• DOI: 10.1016/j.jveb.2019.04.010

Ovarian cancer is known as a “silent killer” because of its nonspecific symptoms and generally late diagnosis due to the lack of reliable early detection tools. Medical detection canines have been shown to recognize the odor profile of malignant tumors of many types, including ovarian cancer, and may be utilized to help produce a much-needed early detection system for ovarian cancer. One significant drawback, however, is that to train detection canines on the odor signature of ovarian cancer, different samples or sets of samples must be used in every training session so that the canines do not begin to recognize an individual patient's odor signature or attune to patient-specific noncancerous olfactory cues. Therefore, numerous patient-derived samples must be obtained and stored for use in the training process. Ovarian cancer cell lines may be able to solve this obstacle, as they have low sample-to-sample variability and do not present patient-specific confounding olfactory cues. In this study, we attempted to train three medical detection dogs to alert to an ovarian cancer cell line (OVKATE) but not the culture media or nonmalignant cultured fallopian tube cells. Only one dog was able to discriminate between OVKATE and culture media and was able to be tested on other ovarian cancer cell lines (OVCAR-4, OVCAR-8, and SK-OV-3). The dog's responses suggest a common or related olfactory signature of malignant ovarian cancer cell lines. We then tested whether an OVKATE-trained dog recognized the blood plasma of patients with confirmed ovarian cancer. We did not find strong evidence that the dog recognized the blood plasma of a patient with confirmed ovarian cancer within three testing trials, with the dog only hesitating at the target odor, suggesting that while training on the cell lines may prepare medical dedication canines to recognize the blood plasma of patient with ovarian cancer, it is not a spontaneous switch to blood plasma. Despite this outcome, the behavior of the dog led us to believe that this method could help further the use of medical detection canines to inform the production of early detection tools for ovarian cancer, which would lead to earlier diagnoses and more favorable outcomes from therapy.