, , ,

Just analyze someone’s breath, mucus, saliva, sweat or urine and diagnose whether they have diabetes, cancer, COPD, IBD, Clostridium difficile infection (CDI), Tuberculosis (TB), or any other disease for that matter. Now that would be a seismic revolution in medicine. For one, fully non-invasive in complete contrast to the present-day staples, needles and blood draws, and the pain and fear they entail. Could also be done as often as possible, even when asleep or anesthetized during surgery, even in real-time, as point-of-care, i.e., truly portable and thus truly mobile. Underlying idea is the body’s physiological emanations reliably communicate unique signatures of underlying diseases in the form of singular mixes of volatile organic compounds (VOCs), i.e., the human ‘volatilome‘. The reality, OTOH, is a sharp, painful thud since the ground reality is one where most of these possibilities remain not even remotely feasible in the near future.

Volatilome historians point to the French chemist, Antoine Lavoisier, as the modern inspiration for diagnosis using exhaled breath (1). He showed the body produces and exhales carbon dioxide. In turn this became the basis for Capnography, monitoring the concentration pressure or partial pressure of carbon dioxide in respiratory gases, the most common breath test. Antiquated roots notwithstanding, using unique signatures from breath and other emanations for disease diagnosis still remains deep in research mode and far from much practical utility. A 2014 review lists a total of only 7 US FDA-approved breath-related tests (see below from 2).

Obstacles To Widespread Non-invasive Sampling of Body Emanations For Disease Diagnosis

I. Unlike Animal Olfaction, Human Technology’s Remained Too Constrained In Choice Of What To Analyze

Default, standard analytical approach to sampling and analyzing compounds present in emanations is to rely on ‘headspace’ analysis (3, see below from 4), jargon that means sampling what is already in the gas phase, i.e., already volatile in the material, rather than attempt to extract compounds of interest from it.

Problem is this can cut off too thin a slice of the pie. This shows up in the results since technical approaches continue to fail to mimic what animals do so effortlessly when they use smell to communicate, forage and assess the health of those around them. This brings us to animals and their remarkable capacity to sniff out disease so much so that anecdotal reports suggest they can even be far more accurate in diagnosing human diseases compared to human technologies.

Two diseases with substantial research on animals successfully sniffing them out in humans are Skin Cancers by dogs and TB by Giant African pouched rats.

IA. Dogs Can Spontaneously Detect Human Skin Cancer & Can Also Be Trained To Detect Clostridium difficile infection (CDI)

In 1989, the Lancet published what is perhaps the first modern report of a dog sniffing out its owner’s melanoma (See below from 5)

In 2001, a follow-up anecdotal study (6) added two other case reports of dogs accurately sniffing out skin cancer lesions,

  • One, a London, UK report on Parker, a pet labrador who sniffed it out on a 66 year old man’s left thigh. The patient had developed an eczema patch there. Treated unsuccessfully by topical steroids and antifungals, it grew slowly over 18 years. In 1994, Parker became a family member. Around 1999, Parker started to persistently push his nose against the patient’s trouser leg and sniff the lesion beneath it, i.e., could smell something about the lesion even through clothing. This induced the patient to re-consult his family physician. The lesion was excised in September 2000 and histology showed it to be a fully excised basal cell carcinoma. Once lesion was fully removed, Parker no longer showed interest in that area.
  • The other, by George, a Florida, USA, K-9 unit schnauzer trained by his retired handler to recognize in vitro malignant melanoma samples. A local dermatologist had read the original 1989 Lancet case report and teamed up with the handler to see if such a result was repeatable with another dog. When George was introduced to a patient with several moles considered cancer-free, he went ‘crazy’ over one particular mole, which when excised confirmed ‘early malignant disease’.

Authors of this second case report speculated dogs might also be able to detect odors associated with specific diseases such as TB and Ebola.

In 2012, The BMJ published a proof of principle study on Cliff, a 2 year old beagle trained to sniff out human Clostridium difficile infection (CDI) remarkably accurately (7).

Such reports have provoked more systematic studies, which conclude that dogs could detect unique odors emanating from human melanoma and other cancers (8, 9, 10).

Problem is clinical scope for using dogs to diagnose diseases is limited given the costs, effort, space and time required to train sniffer dogs to detect various diseases (11).

IB. Giant African Pouched Rats Can Be Trained To Reliably Detect Human Tuberculosis (TB)

Used to diagnose lung TB in resource-poor settings, the antiquated microscopic Ziehl–Neelsen stain is a standard method for detecting Mycobacterium tuberculosis in Sputum (coughed up mucus). A few studies suggest trained Giant pouched rat are not just as sensitive and accurate but also able to process >50 times more samples per day compared to a lab technician, i.e., much more economical (12). The WHO recommends microscopists not analyze more than an average of 20 samples per day to minimize misdiagnoses (13) while two trained giant African pouched rats could reach a total of 70 consensus results in 32 minutes over 2 sessions each. This means trained rats could screen larger populations in a much shorter time meaning faster TB diagnoses and hence potential for reduced TB transmission, i.e., a potentially enormous public health benefit.

The video below shows how the Belgian social enterprise APOPO trains these rats in Tanzania to accurately diagnose TB from human samples.

II. Inadequate Research Efforts To Deconstruct The Human Volatilome

Clearly animals are able to smell broader, more complex mix of volatile chemicals that the most sophisticated chemical extraction techniques used in volatilome analysis miss (3). In order for technologies to be able to replicate what animals seem to do so effortlessly, research needs to systematically unravel the human volatilome and establish a reference base of what that looks like in health in order to be able to discern and diagnose cause of ill-health simply from analyzing someone’s emanations.

What compounds are present in normal breath, urine, skin emanations, saliva, blood and feces? A compendium of the healthy human volatilome was first described only in 2014 (14), meaning a foundational study has come along only in the very recent past. This study is foundational for the following reasons,

  • It identified compounds in breath (874), urine (279), skin emanations (504), saliva (353), blood (130), feces (381).
  • It classified these compounds by their CAS Registry Number (CAS), unique numerical identifier assigned to every chemical substance described in the published scientific literature.
    • Hundreds of peer-reviewed scientific papers are routinely published on the human volatilome. Problem is there is as yet no standardization of procedures or data reporting. As a result, the literature is awash in duplicates.
    • Does exhaled breath really have ~3000 different compounds? Umm, looks like it’s less than a fourth of that.
  • This 2014 paper (14) is thus the first step in the right direction, namely to consolidate, synthesize and whittle down published information into a potentially universal ‘megatable’ of compounds present in healthy human emanations.

III. Too Much Technical Sensitivity Can Sometimes Be Too Much Of A Good Thing

  • Iterations over decades have vastly improved sensitivity of state-of-the-art volatilome analysis methods like Gas chromatography–mass spectrometry (GC-MS) and Proton-transfer-reaction mass spectrometry (PTR-TOF-MS), Selected-ion flow-tube mass spectrometry (SIFT-MS) and other techniques such that they can easily measure ~1000 compounds.
    • Problem is most volatilome studies investigate a handful of subjects, not the thousands necessary to validate variables that are different between those with or without diseases.
    • One review (1) suggests number of subjects should be >5X the number of analytes measured, clearly something that adds prohibitive cost to such studies but not doing so increases chances for what they call ‘voodoo correlations’, an issue compounded by dividing test populations further into sub-groups.
  • Volatilome data are also not standardized, neither are the procedures (1, 3, 14, 15, 16, 17, 18). This makes meta-analyses, i.e., comparison of data across multiple studies, well-nigh impossible.
  • As with so many topics in biomedical research, human volatilome studies have heretofore paid scant attention to human Microbiota (19), how it shapes the human volatilome and how that process not only differs between health and disease but also yields different outcomes, i.e., different volatile signatures (1, 3).

That said, there are several diseases with candidate volatile biomarkers that await final validation (see tables below from 20). The sky’s very much the limit for diagnosing diseases through their distinctive odors.


1. Amann, Anton, et al. “The human volatilome: volatile organic compounds (VOCs) in exhaled breath, skin emanations, urine, feces and saliva.” Journal of breath research 8.3 (2014): 034001.

2. Amann, Anton, et al. “Analysis of exhaled breath for disease detection.” Annual Review of Analytical Chemistry 7 (2014): 455-482.

3. Kimball, Bruce A. “Volatile metabolome: problems and prospects.” (2016). http://www.future-science.com/do…

4. Restek, A. “Technical Guide for Static Headspace Analysis Using GC.” Restek Corp (2000): 11-12. http://www.restek.com/pdfs/59895…

5. Williams, Hywel, and Andres Pembroke. “Sniffer dogs in the melanoma clinic?.” The Lancet 333.8640 (1989): 734.

6. Church, John, and Hywel Williams. “Another sniffer dog for the clinic?.” The Lancet 358.9285 (2001): 930. http://www.thelancet.com/pdfs/jo…

7. Bomers, Marije K., et al. “Using a dog’s superior olfactory sensitivity to identify Clostridium difficile in stools and patients: proof of principle study.” (2012): e7396. http://www.bmj.com/content/bmj/3…

8. Pickel, Duane, et al. “Evidence for canine olfactory detection of melanoma.” Applied Animal Behaviour Science 89.1 (2004): 107-116. http://sniffoutcancer.org/images…

9. Moser, Emily, and Michael McCulloch. “Canine scent detection of human cancers: a review of methods and accuracy.” Journal of Veterinary Behavior: Clinical Applications and Research 5.3 (2010): 145-152. https://www.researchgate.net/pro…

10. Jezierski, Tadeusz, et al. “Study of the art: canine olfaction used for cancer detection on the basis of breath odour. Perspectives and limitations.” Journal of breath research 9.2 (2015): 027001. https://www.researchgate.net/pro…

11. Buljubasic, Fanis, and Gerhard Buchbauer. “The scent of human diseases: a review on specific volatile organic compounds as diagnostic biomarkers.” Flavour and Fragrance Journal 30.1 (2015): 5-25.

12. Mgode, Georgies F., et al. “Diagnosis of tuberculosis by trained African giant pouched rats and confounding impact of pathogens and microflora of the respiratory tract.” Journal of clinical microbiology 50.2 (2012): 274-280. Diagnosis of Tuberculosis by Trained African Giant Pouched Rats and Confounding Impact of Pathogens and Microflora of the Respiratory Tract

13. World Health Organization, et al. “Management of tuberculosis: training for district TB coordinators.” (2005). http://apps.who.int/iris/bitstre…

14. de Lacy Costello, Ben, et al. “A review of the volatiles from the healthy human body.” Journal of breath research 8.1 (2014): 014001. https://www.researchgate.net/pro…

15. Pereira, Jorge, et al. “Breath analysis as a potential and non-invasive frontier in disease diagnosis: an overview.” Metabolites 5.1 (2015): 3-55. Breath Analysis as a Potential and Non-Invasive Frontier in Disease Diagnosis: An Overview

16. Boots, Agnes W., et al. “Exhaled molecular fingerprinting in diagnosis and monitoring: validating volatile promises.” Trends in molecular medicine 21.10 (2015): 633-644. https://www.breathcloud.org/wp-c…

17. Bikov, Andras, Zsófia Lázár, and Ildiko Horvath. “Established methodological issues in electronic nose research: how far are we from using these instruments in clinical settings of breath analysis?.” Journal of breath research 9.3 (2015): 034001.

18. Scarlata, Simone, et al. “Exhaled breath analysis by electronic nose in respiratory diseases.” Expert review of molecular diagnostics 15.7 (2015): 933-956.

19. Dietert, Rodney Reynolds, and Ellen Kovner Silbergeld. “Biomarkers for the 21st century: listening to the microbiome.” Toxicological Sciences (2015): kfv013. Listening to the Microbiome

20. Kataoka, Hiroyuki, et al. “Noninvasive analysis of volatile biomarkers in human emanations for health and early disease diagnosis.” Bioanalysis 5.11 (2013): 1443-1459. https://www.researchgate.net/pro…

Thanks for the R2A, Jonathan Brill.