AMSTERDAM – A test that samples exhaled breath and looks for the presence of just five volatile organic compounds shows promise as an inexpensive method for screening patients with suspected cancers of the gastroesophageal junction (GEJ), investigators from the United Kingdom reported.
In a multicenter clinical trial testing breath samples from patients with adenocarcinoma of the GEJ and matched controls, the test had an overall sensitivity of 80% and specificity of 81% for adenocarcinoma of the GEJ, said Sheraz R. Markar, MD, PhD, from Imperial College London.
“What we may see in the near future is that if a patient presents with nonspecific upper GI symptoms – whether it’s vague abdominal pain or gastroesophageal reflux – to their primary care physician, they undergo the exhaled breath test, and if it’s positive the patient can then be referred to endoscopy,” he said at an annual congress sponsored by the European Cancer Organisation.If the test can be validated in larger studies, it could increase the number of patients screened and has the potential to save medical costs by reducing the number of unnecessary endoscopies and by catching GEJ cancers at earlier, potentially curable stages, he said.
The test uses selected ion flow tube mass spectrometry, or SIFT-MS to identify the olfactory signatures of specific chemical components among the millions of possible odors in a sample of air.
The investigators previously identified 13 volatile organic compounds (VOCs) associated with GEJ cancers and through additional analysis pared the number down to five: butyric acid, pentanoic acid, hexanoic acid, butanal, and decanal.
In tests of the five-VOC breath model, they found it had an area under the curve (AUC) of the receiver operating characteristic of 0.90, sensitivity of 84%, and a specificity of 88%.
They then sought to validate the model in a multicenter blinded study. They enrolled 163 treatment-naive patients diagnosed with nonmetastatic GEJ cancer (stages I-III), and 172 controls matched on a 1:1 basis.
Breath samples from all participants were collected in steel breath bags and sent to a central lab for SIFT-MS analysis. A statistician blinded to patient diagnosis then determined cancer risk based on previously determined odds ratios for each VOC.
The investigators used quality assurance measures to minimize the risk of errors, including sampling of the ambient air where the samples were collected, training of all researchers in uniform breath collection technique, and calibration to water.
They found that in this validation study, four of the five VOCs were significantly dysregulated in cases, compared with controls; pentanoic acid was the exception. The AUC was 0.85, with a sensitivity of 80% and specificity of 81%.
Looking at the association between VOCs and demographics of the patients as possible confounders, they saw that hexanoic acid levels could be affected by smoking history, and that butanal could be affected by smoking, white race, or history of using an ACE inhibitor.
Dr. Markar said that among the strengths of the study are that is was adequately powered, performed in multiple centers, and had quality assurance measures in place. In addition, the results compared well with results from the use of a cytosponge.
He acknowledged, however, that there were more late- than early-stage cancers among patients in the study, and that the 80% sensitivity level meant that one in five cancers would be missed.
Nonetheless, if the test is refined and can be further validated in an unenriched population, it could serve as an endoscopy triage test, he said.
Ian Banks, MD, president of the European Men’s Health Forum and cochair of ECCO2017 commented that the study provides “a wonderful insight into what we don’t know about things.”He noted that we are just beginning to understand the importance of smell, the “most primitive” of the five senses, in relation to human health and joked that, just as many airports have drug-sniffing dogs, clinical practices could have patient-sniffing dogs that could be used to direct patients to the right specialist.
He was not involved in the study, but commented on it as part of a media briefing.
The study was supported by the UK National Institute for Health Research. The authors reported no competing interests.