Before they underwent an endoscopy, each patient was asked to breath into an “electronic nose” — a device that can detect different volatile molecules. In the breath, such molecules result from processes in the body, however while many of these occur in a healthy individual, some may be linked to particular diseases, either reflecting changes in cells or changes in the local community of microbes caused by a disease. As a result, a particular composition within a breath sample may act as a hallmark of a condition. The team’s portable electronic nose incorporated a type of artificial intelligence to look for these patterns. […] Overall the results reveal that the nose correctly identified patients with Barrett’s oesophagus 91% of the time, while it correctly identified those without the condition 74% of the time. When the test was restricted to only those with either gastro-oesophageal reflux disease or Barrett’s oesophagus, the system was still able to distinguish patients, albeit less accurately.
Read more of this story at Slashdot.