Volatile organic compounds (VOCs) are released from our bodies in our breath, sweat, blood and urine as a result of metabolic activity and so provide a window on our metabolic and physiological status. These VOCs change in response to disease, providing the potential for disease-specific VOCs in the breath to be used as non-invasive biomarkers for diagnosis and therapy and relapse monitoring. CLI chatted to Dr John Riches (Centre for Haemato-Oncology, Barts Cancer Institute, Queen Mary University of London, London, UK) to find out more about his work in collaboration with Owlstone Medical (https://www.owlstonemedical.com) looking for VOCs as indicators of blood cancers.

How has the use of breath biopsy come about as a method for detecting disease and how is this done?

The field’s been around for at least 30 years or so, where people have been wondering about the possibility of detecting human disease in breath. For example, you hear these anecdotal stories about people’s pets (usually dogs) realizing that something is wrong with their owner, who when they go to the doctor discovers that they have cancer or some other disease, and the idea is that the dog has detected something by sense of smell, that the person doesn’t smell right. In parallel to this, there’s this whole field of volatile organic compounds (VOCs), compounds that the body releases are obviously volatile at room temperature and could be detected by smell and, of course, we’re releasing them in blood, in urine, in sweat and in our breath. VOCs are produced in as a result of the metabolic processes in cells and are indicators or physiological and metabolic status, and their profiles can be altered in different disease states also, which – when these changes can be detected in the breath – make breath testing (breath biopsy) an ideal non-invasive method for disease diagnosis and monitoring.

The field was kick-started by a publication by a group in the US around 25 years ago who showed that you could discriminate between patients who did or didn’t have lung cancer [Phillips et al., 1999]. The original test was done just as a chest X-ray and biopsy and they were able to discriminate between patients with and without lung cancer on the basis of altered levels of 22 various VOCs in the breath.

There had been some interest before, but this study was one of the biggest and the first ones, as it is technically very challenging to do. By comparison, taking a blood sample, where you take all the sample and you fill up a tube with blood is easy. With breath collection, in the early days, patients were just breathing out through a closed valve system into a big bag. Then, of course, it’s quite challenging to have a big bag of air and how do you deal with that? So over the years there have been a few technical innovations that have helped the field to develop. One was rather than collecting the exhaled gas, it was the actual VOCs themselves that were collected by people breathing onto a metal tube that binds and ‘harvests’ the VOCs from the exhaled breath. A second aspect is that there are obviously a lot of VOCs naturally present in the air in the room, so part of their system is a filter that filters the room air to get remove them. The third component was the way the work within the system. This is necessary because the sample that needs to be collected is the alveolar air from deep in the lungs that has been in contact with the bloodstream for longer and not the air from the so-called anatomical dead space of the mouth and trachea where the concentration of the VOCs produced by the body is relatively low.

So these three technological advances make it really quite an exciting system to work with.

How has breath analysis been developed so far for screening, diagnosis and monitoring disease?

Currently, breath analysis is used for diagnosing digestive disorders, such as lactose intolerance and small intestinal bacterial overgrowth (SIBO), as well as respiratory inflammation where the presence of NO in the breath can help to diagnose inflammatory airway diseases such as asthma.

However, much research is being done into a number of human diseases: lung cancer, obviously, but also respiratory diseases, other cancers (ovarian, prostate, breast), gastrointestinal cancers (pancreatic, bladder), as well as others (hepatic, renal and colorectal disorders) [Moura et al., 2023]. Not forgetting the amazing work that’s been done with Parkinson’s disease looking at VOCs in both sebum and breath [Belluomo et al., 2025].

What is known about the use of VOCs in the exhaled breath of patients with hematological cancers?

When I looked, I realized that nobody had investigated the use of breath analysis for any of the blood cancers, but I thought it would be an interesting idea to test because obviously the whole point of breathing is to facilitate gas exchange between the blood and the air. So I thought that if there are any blood cancer cells in the bloodstream or in close communication with the bloodstream – such as in the case of diseases like lymphoma – then in theory these VOCs will move into the blood and then be exhaled into the breath. Or, of course, they could appear in the urine or sweat.

The plan was to begin the study in late 2019, early 2020, but was hampered because of the Covid-19 pandemic. The idea was to pick out patients with the more rapidly growing, more aggressive types of leukemia and lymphoma. The pandemic meant that everyone involved had to have a recent negative Covid test, and also it meant that most of our healthy controls at that point were staff at the hospital because we were all having regular covid tests and coming to the hospital to do our jobs at the time.

Our study involved a small group of patients (36 in total: 17 with acute myeloid leukemia, 8 with acute lymphoblastic leukemia and 11 with high-grade lymphoma) and 27 healthy controls. The aim was to establish if there was a distinct VOC profile that could differentiated between blood cancer patients and healthy people, and in fact we did see some differences [Stiekema et al., 2025].

In our paper, for the lymphoma patients, we list the top 15 compounds that were differentially present – 6 with increased abundance, and 9 at a reduce level – between patients and healthy controls. Interestingly, four of compounds that we found (4-methyl-decane, decane, 4-methylundecane and 2,3,5-trimethylhexane) were remarkably similar to the ones that had been described in the lung cancer patients 25 years or so before. So, even though it’s a completely different technique, the same compounds are being identified – methylated alkanes. We thought that this was quite striking and so we thought that there is some potential here for developing a rapid and non-invasive diagnostic test for patients with high-grade lymphoma.

The rationale behind why we see these particular VOCs is that, generally, these high-grade lymphomas are usually very meta-bolically active. They are amongst the most prolific types of cancer with 90–99% of the cells dividing and metabolizing very avidly. The idea is, therefore, that the rapid metabolic rate means they’re under some oxidative stress and producing a lot of reactive oxygen species which react with lipids in the cell undergoing a process of lipid peroxidation and that’s where these methylated compounds were generated. We’d like to explore further exactly how that’s happening.

The ideal scenario would be a small handheld point-of-care device that could aid diagnosis, monitor treatment efficacy and warn of disease recurrence (Adobe Stock)

Are there complications and challenges for using VOCs for cancer detection and/or monitoring?

Yes, there are challenges with breath biopsy. The patients with acute leukemia tend to be people who are diagnosed on a blood test, but then they’re rushed into hospital and begin treatment quite quickly. Most of them are very heavily immunosuppressed and so often pick up infections. They have very low numbers of normal white blood cells because their bone marrow is making leukemic cells rather than normal ones, so they pick up infections. Hence, the majority of them were on antibiotics and are in hospital feeling very unwell. This means that they were not eating their normal diets and with the leukemia patients, unfortunately, we picked up changes in the microbiome and changes in the diet in the breath analysis. What we saw was a decrease in levels of compounds found in tea and coffee, garlic and spices and things like that, probably reflecting that all the staff were eating and drinking normally, whereas the patients were not feeling well enough and their diets had changed.

At the moment, the technology is still very much in development. For example, in the LuCID (Lung Cancer Indicator Detection) study, which was Owlstone Medical’s main lung cancer study (https://clinicaltrials.gov/study/NCT02612532), I don’t think there was a particularly strong signal there, and so I think these are the challenges that the field has are sensitivity and the specificity. Specificity was immediately an issue, because while we were very heartened to see exactly the same methylated alkanes (very similar and in some cases identical compounds) that had been seen with lung cancer, this means that these compounds are probably not going to be lymphoma specific and the test may be detecting other kinds of cancers. For example, lymphoma is caused by rapidly proliferating lymphocytes, so the test may detect other non-cancerous inflammatory states.

Additionally, regarding screening, this means that we are not going to suddenly start doing breath tests on everybody across the population for lymphoma then. However, where it could be useful is if you’ve got a patient who at diagnosis has high levels of maybe up to, for example, five of these compounds in their breath we could monitor the levels it and see that they are high at the start and then decrease and/or disappear with treatment and then it would be a rapid test to do when they come back to see me in clinic a few months later just to check that there was no abnormal signal or to monitor for recurrence.

Also, with regards to sensitivity, our study deliberately chose patients with the highest grade lymphomas, at a more advanced stage of their disease, mostly at stage 4 where the disease is most disseminated around the body. I think sensitivity has also been an issue in detecting low levels of lung cancer, too. A good screening test needs to be capable of detecting early stage disease because that’s when you can intervene most effectively. Also, if the sensitivity is low, the false-negative rate will be high and patients with disease will be missed.

What are the potential future developments/applications for this technology?

I think the future of this technology is about learning more about exactly what we’re seeing and working out in what sort of situations it could be most useful. For example, in the UK we have the benefit of all of these fancy scanners and histopathologists looking at biopsies in the lab, but for low-resource settings and low- to middle-income countries there may be some benefit to developing a quick and easy test. Particularly, for example, in West Africa where Burkitt lymphoma is endemic it might be very useful to have a very simple handheld device for disease diagnosis and therapy monitoring in those situations.

Also, in our study, we were prospecting – looking to see if there was a difference in VOC profiles between different patient and control groups. We managed to be around 99% certain that we had identified some of the compounds that we were detecting, but for others we were down in the 80s. Once we are absolutely sure of the VOCs that we are looking to detect in a test, we could potentially make a short list of a handful, a small panel, or one, two, three or up to five, as I said before, then we could develop a completely different way of analysing them rather like the roadside ethanol breath test.

What I imagine happening in the future is that there may be some people with the ‘right kind’ of lymphoma, the genetic changes, the high metabolic rate, the right microbiome so that at the outset they have high levels of a particular methylated alkane is going to be a really useful biomarker for their disease. And we could use such a test to monitor them longitudinally for treatment response, remission, relapse, etc. However, there may be some people who have the same type of lymphoma, but a lower burden, a different microbiome, or are on antibiotics that interfere with the VOCs present and the utility of such a test will be affected. This is similar with a lot of the clinical tests we do – no one test is perfect.

So, potentially, if we could develop the same sort of roadside technology to have a test that you could do in 5 to 10 seconds and you could be given a number that’s high at the start and goes down with effective treatment and maybe if it goes back up again to alert you to the fact that the cancer is coming back then there would be clinical value in that.

Additionally, I think we would like to understand more about where the VOCs are coming from, what interaction causes them, see whether we can model them just using cancer cells in a test tube initially, or whether it suddenly becomes apparent that the cancer cells on their own or is there also an interaction with resident bacteria. Also, we need to do a bigger study and to focus more on the well described, well annotated specific subtypes of lymphoma which might give us more consistent results, because there was quite a lot of heterogeneity in the patient cohort that we had.
So those are the next steps.

The interviewee

Dr John Riches
MA, BM BCh, MRCP, PhD, FRCPath, FHEA
Consultant
Hematologist for lymphoid
malignancies; Clinical Senior Lecturer; Course Director for the MSc
Cancer & Clinical Oncology Programme

Centre for Haemato-Oncology, Barts Cancer Institute,
Queen Mary University of London, London, UK

Email address: j.riches@qmul.ac.uk

Bibliography
1. Stiekema LCA, Chou H, Craster A et al. Analysis of volatile organic compounds in exhaled breath of blood cancer patients identifies products of lipid peroxidation as biomarkers for lymphoma detection. Hemasphere 2025;9(7):e70168 (https://doi.org/10.1002/hem3.70168).
2. Phillips M, Gleeson K, Hughes JMB et al. Volatile organic compounds in breath as markers of lung cancer: a cross-sectional study. Lancet 1999;353(9168): 1930–1933 (https://doi.org/10.1016/s0140-6736(98)07552-7).
3. Markar SR, Brodie B, Chin ST et al. Profile of exhaled-breath volatile organic compounds to diagnose pancreatic cancer. Br J Surg 2018;105(11):1493–1500 (https://doi.org/10.1002/bjs.10909).
4. Markar SR, Wiggins T, Antonowicz S et al. Assessment of a noninvasive exhaled breath test for the diagnosis of oesophagogastric cancer. JAMA Oncol 2018;4(7):970–976 (https://doi.org/10.1001/jamaoncol.2018.0991).
5. Phillips M, Cataneo RN, Ditkoff BA et al. Volatile markers of breast cancer in the breath. Breast J 2003;9(3):184–191 (https://doi.org/10.1046/j.1524-4741.2003.09309.x).
6. Moura PC, Raposo M, Vassilenko V. Breath volatile organic compounds (VOCs) as biomarkers for the diagnosis of pathological conditions: a review. Biomed J 2023;46(4):100623 (https://doi.org/10.1016/j.bj.2023.100623).
7. Belluomo I, Tarazi M, Lao-Kaim NP et al. Detection of volatile organic compounds as an emerging strategy for Parkinson’s disease diagnosis and monitoring. NPJ Parkinsons Dis. 2025;11(1):161 (https://doi.org/10.1038/s41531-025-00993-2).