Skip to main content

'Electronic nose' sniffs out condition that can lead to cancer

Wednesday 26 February 2020

"'Electronic nose' could smell breath to warn about higher risk of oesophageal cancer," reports the Guardian.

Oesophageal cancer is cancer of the food pipe. People with a condition called Barrett's oesophagus have a higher risk of oesophageal cancer and are offered regular monitoring to check for cancer signs.

Barrett's oesophagus happens when cells in the oesophagus change as a result of being exposed to digestive acids from the stomach. These cell changes are what are known as precancerous. The cells are abnormal, but they currently do not have the ability to spread. But they could turn into cancer if left untreated. Barrett's oesophagus usually happens in people with a history of experiencing gastro-oesophageal reflux disease (often called GORD).

Barrett's oesophagus can be diagnosed with an endoscopy (a camera on a tube that goes down the throat to take pictures of the oesophagus and stomach). But many people with oesophageal cancer have never previously been diagnosed with Barrett's oesophagus, and their cancer is diagnosed late. This means they have a lower chance of survival.

Researchers in the Netherlands have developed a test where chemicals in the breath (known as volatile organic compounds or VOCs) are identified by sensors in an "electronic nose", then the results fed into a computer. They tested this in people with Barrett's, GORD and a healthy control group, to see if the computer could recognise the typical patterns of VOCs breathed out by people with Barrett's oesophagus. Researchers found the breath test could correctly identify 91% of people with Barrett's oesophagus, and correctly identify 74% who did not have the condition.

This is a promising development of a new screening test that may help to diagnose Barrett's oesophagus earlier. Further study will be needed to validate these findings, and help find out how the test could be used.

Where did the story come from?

The researchers who carried out the study were from Radboud University Medical Centre, Zeikenhuis Bernhoven and Canisius Wilhelmina Hospital in the Netherlands. The research had no specific funding. The study was published in the peer-reviewed medical journal Gut.

The Guardian carried a balanced and accurate report of the study.

What kind of research was this?

This was a proof of concept study using a case control design. This type of study, where you know which people have the disease (the cases) and which do not (the controls), is useful in the early stages of designing a new screening or diagnostic technology, to see whether it is likely to work. Further, larger-scale studies are needed on different populations (for example, a random sample of the general population who are not suspected to have disease) to see whether it truly has potential as a reliable screening test.

What did the research involve?

Researchers recruited 402 people who were due to have an endoscopy because of symptoms. Before their endoscopy, they were asked to take a 5-minute breath test. An "electronic nose" device used sensors to measure the compounds in the breath. The results were then fed to a computer that sorted them using pattern recognition.

The researchers then looked to see how accurate the breath tests were at identifying patients later diagnosed by endoscopy to have:

  • Barrett's oesophagus (129 people)
  • gastro-oesophageal reflux disease (GORD, 141 people)
  • no signs of Barrett's oesophagus or GORD (132 people, described as the control group)

They also looked to see if the test still worked if people had taken proton pump inhibitors (PPIs), a common acid reflux medicine, or if they had a hiatus hernia – where part of the stomach moves into the chest (some of the control group had this).

What were the basic results?

The artificial nose test correctly identified 91% of the people who had Barrett's oesophagus (sensitivity 91%, 95% confidence interval (CI) 84% to 95%).

The test also correctly identified some people who did not have Barrett's oesophagus. However, a quarter of people who did not have Barrett's oesophagus (that is controls or those with GORD only) were wrongly identified as having Barrett's (specificity 74%, 95% CI 69% to 79%). This is known as a false positive.

Further analysis just comparing people with Barrett's vs controls, or with GORD (rather than non-Barrett's grouped together) found that the accuracy then was not so good. Sensitivity to correctly identify people with Barrett's was reduced to 64% (vs GORD) and 57% (vs controls). So there's still some work to be done on the test.

When looking at only people using PPIs or only those with hernias, the test still showed "fair to good" accuracy (according to the researchers), with sensitivity ranging from 68% to 96% depending on which groups were being compared.

How did the researchers interpret the results?

The researchers said: "Given the high tolerability, high acceptability and low costs, breath testing may be a promising approach to be used for non-invasive screening for Barrett's oesophagus in a primary care setting."


This is an early-stage study that needs further research to confirm that the breath test can successfully identify people who may have Barrett's oesophagus, and who would benefit from an endoscopy.

The results are interesting, but limited. They were based on a small group of people already due to have an endoscopy for clinical reasons. This was a group of people who were more likely to have Barrett's oesophagus, and studies in such samples often give better test results. We do not know whether breath testing a wider group of people in the general population would also identify people with Barrett's oesophagus to the same degree of accuracy.

The researchers say they are in the process of testing the breath test on a larger scale in primary care (general practice), recruiting people with symptoms of GORD. These studies may tell us whether the test could have potential for screening in the wider population, or whether it could wrongly identify too many people as potentially having Barrett's oesophagus who turn out not to have the condition. Such "false positive" results could cause a lot of anxiety and unnecessary endoscopies.

Nevertheless, this is a promising step forward in developing new ways to identify people with Barrett's, who may benefit from monitoring so that if oesophageal cancer develops, it can be caught early.

Oesophageal cancer can be hard to detect. It's a good idea to see a GP if you have:

  • swallowing difficulties
  • heartburn on most days for 3 weeks or more
  • any other unusual or persistent symptoms

These symptoms can be caused by several conditions and many cases will not be caused by cancer. But you should get them checked out in case. Find out more about oesophageal cancer.

Analysis by Bazian
Edited by NHS Website

Links to the headlines

'Electronic nose' could smell breath to warn about higher risk of oesophageal cancer

The Guardian, 25 February 2020

Links to the science

Peters Y, Schrauwen RWM, Tan AC, et al

Detection of Barrett’s oesophagus through exhaled breath using an electronic nose device

Gut. Published online 25 February 2020