Wednesday March 6 2013
Stomach cancer was detected with 90% accuracy by breath test
“A quick and simple breath test can diagnose stomach cancer,” BBC News accurately reports, saying it ‘could revolutionise and speed up the way this cancer is diagnosed’.
Currently, the only way to confirm or exclude a diagnosis of stomach cancer is to carry out an endoscopy, where a tube is placed down the throat and into the stomach. This is expensive and time-consuming, and can be unpleasant.
A quicker, less-invasive method of diagnosis would be valuable, particularly in countries with limited access to endoscopy.
This news is based on a small Chinese study looking at how well an analysis of the organic compounds in a person’s breath can be used to diagnose cancer. Researchers tested the breath of people with stomach complaints, some of who had confirmed stomach cancer, and some who had confirmed non-cancerous conditions, such as stomach ulcers.
They found that the test was 90% accurate for identifying who did and who did not have stomach cancer. It could also tell how advanced any cancers were.
These are positive results, however, as the researchers state, their study was essentially a pilot. Tests on a large number of people across the globe are needed to confirm whether this is a truly viable routine way to detect stomach cancer.
Where did the story come from?
The study was carried out by researchers from the Israel Institute of Technology, The First Affiliated Hospital of Anhui Medical University in China and the University of Latvia. Funding was provided by a European Research Commission grant.
The study was published in the peer-reviewed British Journal of Cancer.
The media coverage was representative of this study, although The Daily Telegraph's website is slightly misleading in stating that the test could ‘rule out the need for an unpleasant endoscopy’. While it is true that this breathalyzer is a quick and simple, non-invasive test, it is not a replacement for endoscopy.
Assuming it is confirmed as accurate in further research, it is most likely this test would be used as early ‘screening’ for people with stomach complaints, to see whether they were at increased risk of having serious conditions and should be tested further.
If the breath test suggested a person had stomach cancer, they would still need to have an endoscopy to view where the cancer is, how large it is, and to take samples for laboratory testing.
What kind of research was this?
This was a diagnostic study that aimed to determine the accuracy of a breath test in identifying the people with stomach cancer from a larger group of people with various stomach complaints.
Earlier diagnosis and treatment improves the outlook for stomach cancer. However, initial symptoms of stomach cancer are usually fairly unspecific and include:
- heartburn (acid reflux)
These symptoms can also develop due to much more common non-cancerous (benign) digestive complaints.
Other general symptoms of stomach cancer, such as feeling tired or becoming anaemic, are also very common.
So a diagnosis of stomach cancer is often not suspected until more advanced symptoms, such as unexplained weight loss and blood in the stools, develop. And by this time the cancer may be harder to treat.
Stomach cancer is currently diagnosed using an endoscope. This is a flexible camera that is put into the mouth, down the throat and into the stomach. It allows doctors to see any abnormal and possibly cancerous areas of tissue, and enables them to take samples (biopsies) that can be tested in the laboratory to give an accurate diagnosis.
While endoscopy is highly accurate, it is an invasive procedure, and one where access may be limited by local resources, particularly in the developing world. Endoscopy is normally only used if a patient has more dramatic symptoms known to be associated with advanced stomach cancer, such as blood in the stools. However, because these symptoms are only associated with more advanced cancer, they can be much harder to treat.
For this reason, a quick and simple early test that could be used to see which people are at high risk of having stomach cancer, and who should definitely have endoscopy, would represent a significant breakthrough. This would be particularly valuable in the developing world where there may be limited access to endoscopy, and in the developed world it could help cut down on the number of unnecessary endoscopies.
The test that has made the news is based on a method of analysing the gases that a person breathes out. In fact, a breath test is already used to detect whether people have the particular stomach bacteria (H. pylori) that is a common cause of stomach ulcers.
In this study, the researchers developed a highly sensitive gas sensor that is able to differentiate between benign and cancerous stomach ulcers and other less serious stomach complaints.
They use the ‘gold standard’ diagnostic technique of endoscopy to confirm how accurate the breath tests were.
What did the research involve?
The research included 160 people with stomach complaints, recruited through the University Hospital in China. All of these people received an endoscopy examination with biopsy to diagnose their complaint upon entry to the study.
After their diagnosis, all the participants took the breath tests. They were required not to eat any food, smoke or consume alcohol for 12 hours prior to the test. The breath tests of only 130 participants were suitable for use in this study, as the remaining 30 were damaged during storage and transport. These 130 people included:
- 37 with stomach cancer (17 early stage, 18 late stages and two without staging information)
- 32 people with benign (non-cancerous) stomach ulcers
- 61 people with less severe stomach complaints (such as people with digestive symptoms but with no abnormalities seen on endoscopy)
The researchers identified which specific organic compounds were raised in people with the different cancerous and non-cancerous conditions, and looked at the accuracy of different models for distinguishing between people with the different conditions. They also checked whether there was any influence from other factors such as whether the person smoked, drank alcohol, or had H. pylori bacterial infection.
They evaluated the accuracy of the tests by calculating:
- the sensitivity of the test (for example, the proportion of people with stomach cancer who were correctly identified as having stomach cancer)
- the specificity of the test (for example, the proportion of people with non-cancerous conditions who were correctly identified as having non-cancerous conditions)
- false positives (for example, the people with non-cancerous conditions who were wrongly identified as having cancer)
- false negatives (for example, the people with stomach cancer who were wrongly identified as having non-cancerous conditions)
- the overall accuracy of the test (for example, of all the people tested, the proportion that were correctly identified as either having cancer or non-cancer)
What were the basic results?
The main results were:
- Overall, for differentiating between stomach cancer and all non-cancerous conditions, the test had 89% sensitivity and 90% specificity. This means it correctly identified 89% of people who had cancer and correctly identified 90% of people who did not have cancer.
- The overall accuracy of the test for detecting stomach cancer was 90%, meaning 90% of all those tested were identified with their correct diagnosis – either cancer or not.
- For the 35 people with stomach cancer who had staging information available, the test had 89% sensitivity and 94% specificity for differentiating early stage from late stage cancer. That is, it correctly identified 89% of those with early cancer and 94% of those with late stage.
- The overall accuracy of the test for staging the cancer was 91%, meaning 91% of all those tested were staged correctly – either early or late.
- For the 93 people with non-cancerous conditions, in differentiating between the 32 people with stomach ulcers and the 61 with less serious stomach complaints, the test had 84% sensitivity and 87% specificity. Again, this means it correctly identified 84% of the people with ulcers and 87% of those without ulcers.
- The overall accuracy of the test for differentiating the type of non-cancerous condition was 86%, meaning 86% of all those with non-cancerous conditions were identified correctly as having either an ulcer or a less severe stomach condition.
The tests were unaffected by history of smoking or alcohol use, or H. pylori infection.
How did the researchers interpret the results?
The researchers conclude that the results of their pilot study "could open a new and promising avenue to diagnose [stomach cancer] and distinguish it from other [stomach conditions]". They acknowledge that their pilot study does not allow 'far-reaching conclusions' to be drawn. However, the results are encouraging and support the initiation of a large multi-centre trial into the use of a breath test to distinguish between cancerous and non-cancerous stomach conditions.
This is a valuable preliminary study that investigated the accuracy of a quick and simple breath test to aid the diagnosis of someone who presents with a stomach complaint. It found the new test had 90% accuracy for correctly identifying those with and without stomach cancer. It also had similarly high accuracy for correctly identifying the cancer as early or late stage.
As with most cancers, one of the most important things influencing the outlook for people with stomach cancer, is early diagnosis and treatment. Currently, diagnosis of stomach cancer relies upon endoscopy examination to visualise the cancer and take samples. A quick, simple and accurate non-invasive breath test could potentially revolutionise how people with stomach complaints are further diagnosed and treated. Earlier detection could potentially lead to improved survival rates, although this would have to be proved through further research. This is particularly important in the developing world where stomach cancer has the highest mortality and access to endoscopy may be limited.
The results of the study are very promising, but the researchers are appropriately cautious when drawing their conclusions.
- Importantly, the current study only included a relatively small number of people (130) with stomach conditions. The researchers say that there were 37 cases of stomach cancer in these people and so the ratio of cancerous to non-cancerous conditions in this sample is not as would be expected if you took a much larger population sample of people with stomach complaints (such as indigestion and acid reflux). In a larger population sample the proportion of people with non-cancerous conditions would be much higher than those with cancer. Therefore, a multi-centre study in a large number of people is needed.
- Also, further research in large numbers of people would give a better indication of the accuracy of this test. The breath test would most likely have a role as an early test in people presenting with stomach symptoms. The implications of false negatives (a wrong ‘all clear’ for someone with cancer) and false positives (wrongly suggesting cancer) would need to be carefully considered before such a test could be brought into large-scale medical use.
- As suggested by this study, there is also the possibility of some practical implications: 30 of the 160 breath test samples could not be used as they were damaged during storage or transport. This could be a potential problem, particularly in the developing world where there may be a greater distance to travel to diagnostic laboratories.
Overall, this is highly promising research and the results of larger studies are eagerly awaited.
Analysis by Bazian. Edited by NHS Choices. Follow Behind the Headlines on Twitter.