Canine olfaction as a disease detection technology: A systematic review

https://doi.org/10.1016/j.applanim.2022.105664Get rights and content

Highlights

  • Most studies of disease detection use a small number of trained dogs.

  • Lung cancer and prostate cancer have been included in the most studies of canine disease detection.

  • Median sensitivity, specificity, and accuracy of canine disease detection are high.

  • Study design may impact measures of detection success.

Abstract

Capitalizing on canine olfactory capacity is a promising strategy for detecting and diagnosing human, animal, and plant diseases. The purpose of this review was to assess the extent of current research in canine disease detection and to identify factors impacting detection success. In this systematic review and meta-analysis, multiple databases were searched for studies in which dogs were trained to detect diseases or health conditions for both plants and animals. Following PRISMA guidelines, 2109 non-duplicate studies were screened and 58 relevant studies identified. Most studies (n = 33, 57%) took place in Europe. Across all studies, 192 unique detection dogs were tested. The most numerous breed was Labrador Retrievers (n = 27, 14%). The median number of dogs per study was 2 (range: 1–20). To analyze experimental design and results, studies including multiple test paradigms were divided into sub-studies (n = 105). Lung cancer (n = 11, 22%) and prostate cancer (n = 14, 13%) were the most frequently studied conditions. Urine (n = 27, 26%) and breath (n = 15, 14%) were the most common sample materials. In 86% of sub-studies (n = 90), dogs were presented with sets of samples and 72% (n = 76) reported a constant number of samples per trial. The median number of samples per trial was 6 (range: 2–100). Of the sub-studies reporting a fixed number of positive samples (range: 1–10; n = 65), 91% (n = 59) presented one positive sample per trial. A plurality of sub-studies (n = 47, 45%) presented samples in a lineup. Sensitivity (median: 0.90; range: 0.17–1.0; n = 90) and specificity (median: 0.96; range: 0.08–1.0; n = 81) were the predominant measures of detection success. In some cases, study design may have influenced results. There was a positive relationship between specificity and the likelihood of a true negative response based on the number of samples per trial, and specificity was higher in studies that did not include a double blinded test than those that did. Dogs appear to have the capacity to detect disease via olfaction; yet the nascent nature of this discipline yields inconsistency in methodology and reporting.

Introduction

The ability of dogs to detect health conditions via olfaction has been empirically investigated for the past three decades. The first case report of a dog alerting its owner of cancer, published in the Lancet in 1989, described a dog that persistently sniffed at a lesion on its owner’s leg, which was later diagnosed as melanoma (Williams and Pembroke, 1989). A second case report more than a decade later described a similar phenomenon (Church and Williams, 2001). Since then, numerous studies have examined dogs’ ability to detect disease. These have included cancers (reviewed in Edwards et al., 2017), bacterial infections (Bomers et al., 2012, Maurer et al., 2016), seizures (Catala et al., 2019), and humans infected with COVID-19 (Grandjean et al., 2020, Jendrny et al., 2020). Canine olfaction has also been demonstrated to be effective in identifying conditions or diseases affecting non-human animals, cattle and dogs, (Angle et al., 2016a, Dorman et al., 2017) and plants, avocado and citrus trees (Mendel et al., 2018, Gottwald et al., 2020).

Several reviews have previously summarized portions of this field but are relatively narrow in scope. Jezierski et al. (2015) and ​(Pomerantz et al., 2015​) published reviews focused on the utilization of canine olfaction to detect cancer in humans. These works provided valuable insight into the methodological characteristics of detection studies, but excluded those in which dogs were trained to detect diseases other than cancer, or to detect disease in organisms other than humans. Another systematic review of olfactory disease detection by animals included detector animals other than dogs, but excluded patient species other than humans, and excluded some health conditions, such as seizures (Edwards et al., 2017). Hackner & Pleil (2017) focused primarily on offering recommendations for potential applications of canine olfactory disease detection. Other articles have reviewed selected studies of canine olfactory detection of disease, yet these reviews focused on developing technological alternatives to canine olfaction in detection of disease using volatile organic compounds (Pleil and Giese, 2017, Marzorati et al., 2019).

In the context of human medicine, the risk of canine misbehavior or inconsistency and human fear of, or allergies to, dogs may outweigh the benefits of canine olfaction as a rapid, mobile, low-tech screening technology. However, in both animal and plant agriculture, the capacity for individualized laboratory diagnosis of disease is lower than in human medicine (Gottwald et al., 2019, Larson et al., 2005). Dogs may be highly valuable for screening large numbers of animals and plants for conditions that impact animal welfare and can have economic implications, including bovine viral diarrhea virus (BVDV) in cattle and Liberibacter asiaticus in citrus (Angle et al., 2016a, Gottwald et al., 2019). These studies have been excluded from previous reviews of disease detection by canine olfaction. This review is therefore necessary both to include the several studies that have been published since 2017, and to offer a comprehensive collection of all studies pertaining to the olfactory detection of disease by canines.

In addition to varying in subject matter, prior studies have yielded variable results. Dogs’ ability to detect lung and breast cancers has been reported to have near-perfect sensitivity and specificity (McCulloch et al., 2006), with less accurate performance in others (Amundsen et al., 2014, Gordon et al., 2008), sometimes no better than chance. Because few studies include trials of disease detection dogs outside the laboratory (Bomers et al., 2012, Gottwald et al., 2020), there is a lack of information about the potential for dogs in real-world disease screening applications. The circumstances under which dogs can detect any given health condition by scent remain unclear, and the limits to what conditions dogs are able to detect diseases remain unknown.

In analyzing all available published studies of canine olfactory detection of health conditions, the objectives of this review were to characterize the subjects and methodology of such studies and to conduct a meta-analysis that evaluated to what extent dogs can serve as reliable detectors of disease in humans, other animals, and plants. Comparators were aspects of study design such as sample size, test apparatus, and number of samples presented during testing. Outcomes measured included detection accuracy, sensitivity, and specificity. This will provide a clearer view of the canine olfactory disease detection landscape, a strong foundation for future research in the field, and analysis of what training methods and testing paradigms result in the greatest rates of success in canine disease detection.

Section snippets

Methods

This systematic literature review was conducted via a broad, structured search of scientific databases.

Results

Searches of three databases yielded 3731 search results (Fig. 1). After 1618 duplicates were automatically removed and four additional duplicates were merged, 2109 studies moved to title and abstract screening. Of those, 159 studies moved on to full-text review, and 58 were included for analysis (Table 2). For some measures, because a single study might report testing dogs on multiple diseases, or under multiple sets of conditions, these studies were divided into 105 sub-studies.

Discussion

Most of the studies of canine olfactory detection of disease included in this review reported high levels of success, whether reported as accuracy or as sensitivity and specificity. While there are methodological concerns related to blinding and sample re-use in some studies, only blinding had an impact on detection success. Sub-studies without double-blind tests had greater mean specificity, which suggests that dogs may take cues from experimenters in tests that are not double-blind.

Conclusions

Canine olfaction is a promising method for screening individuals for diseases and health conditions. However, training and testing methods, results, and reporting strategies are highly variable across studies, limiting comparability. Further research assessing detection accuracy under realistic conditions is needed before canines can be widely deployed to screen for disease.

Funding

This work was supported by the Universities Federation for Animal Welfare Small Project, United Kingdom Grant #26–20/21 and the Texas A&M University USA Triads for Transformation Round 4.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We would like to thank the student workers who contributed to this project: Gabriela Diaz, Nicole Fernandez, Lily Grajeda, Hana Henderson, Kenzy Hoffman, and Brooke Vandenbergh.

References (73)

  • E. Maa et al.

    Canine detection of volatile organic compounds unique to human epileptic seizure

    Epilepsy Behav.

    (2021)
  • M. Murarka et al.

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

    J. Vet. Behav.

    (2019)
  • A. Pomerantz et al.

    The possibility of inventing new technologies in the detection of cancer by applying elements of the canine olfactory apparatus

    Med. Hypotheses

    (2015)
  • A.G. Simon et al.

    Chemical and canine analysis as complimentary techniques for the identification of active odors of the invasive fungus, Raffaelea lauricola

    Talanta

    (2017)
  • G. Taverna et al.

    Olfactory system of highly trained dogs detects prostate cancer in urine samples

    J. Urol.

    (2015)
  • H. Williams et al.

    Sniffer dogs in the melanoma clinic?

    Lancet

    (1989)
  • T. Amundsen et al.

    Can dogs smell lung cancer? First study using exhaled breath and urine screening in unselected patients with suspected lung cancer

    Acta Oncol.

    (2014)
  • C. Angle et al.

    Canine detection of the volatilome: a review of implications for pathogen and disease detection

    Front. Vet. Sci.

    (2016)
  • T.C. Angle et al.

    Real-time detection of a virus using detection dogs

    Front. Vet. Sci.

    (2016)
  • M.K. Bomers et al.

    Using a dog’s superior olfactory sensitivity to identify Clostridium difficile in stools and patients: proof of principle study

    BMJ

    (2012)
  • A. Catala et al.

    Dogs demonstrate the existence of an epileptic seizure odour in humans

    Sci. Rep.

    (2019)
  • E. Crocetti et al.

    Cancer prevalence in United States, Nordic Countries, Italy, Australia, and France: an analysis of geographic variability

    Br. J. Cancer

    (2013)
  • J.C. Davies et al.

    Training dogs to differentiate Pseudomonas aeruginosa from other cystic fibrosis bacterial pathogens: not to be sniffed at?

    Eur. Respir. J.

    (2019)
  • K. Dehlinger et al.

    Can trained dogs detect a hypoglycemic scent in patients with type 1 diabetes?

    Diabetes Care

    (2013)
  • D. Dorman et al.

    Canine scent detection of canine cancer: a feasibility study

    VMRR

    (2017)
  • R. Ehmann et al.

    Canine scent detection in the diagnosis of lung cancer: revisiting a puzzling phenomenon

    Eur. Respir. J.

    (2012)
  • K.R. Elliker et al.

    Key considerations for the experimental training and evaluation of cancer odour detection dogs: lessons learnt from a double-blind, controlled trial of prostate cancer detection

    BMC Urol.

    (2014)
  • E. Eskandari et al.

    Sniffer dogs as a screening/diagnostic tool for COVID-19: a proof of concept study

    BMC Infect. Dis.

    (2021)
  • Carola Fischer-Tenhagen et al.

    A proof of concept: are detection dogs a useful tool to verify potential biomarkers for lung cancer?

    Front. Vet. Sci.

    (2018)
  • R.T. Gordon et al.

    The use of canines in the detection of human cancers

    J. Altern. Complement. Med.

    (2008)
  • T. Gottwald et al.

    Canine olfactory detection of a vectored phytobacterial pathogen, Liberibacter asiaticus, and integration with disease control

    Proc. Natl. Acad. Sci. USA

    (2020)
  • D. Grandjean et al.

    Can the detection dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of-concept study

    PLoS ONE

    (2020)
  • H. Guerrero-Flores et al.

    A non-invasive tool for detecting cervical cancer odor by trained scent dogs

    BMC Cancer

    (2017)
  • C. Guest et al.

    Feasibility of integrating canine olfaction with chemical and microbial profiling of urine to detect lethal prostate cancer

    PLoS ONE

    (2021)
  • C.M. Guest et al.

    A lesson in standardization – subtle aspects of the processing of samples can greatly affect dogs’ learning

    Front. Vet. Sci.

    (2020)
  • Á. Guirao Montes et al.

    Lung cancer diagnosis by trained dogs†

    Eur. J. Cardio-Thorac. Surg.

    (2017)
  • Cited by (0)

    View full text