Several studies have indicated that volatile organic compounds (VOCs) in exhaled breath can be altered by lung cancer and serve as identifiable biomarkers. A limitation of using these VOCs as clinical biomarkers has been the fact that hundreds of such molecules are present in exhaled breath and it is experimentally challenging to monitor the molecular concentration changes of all the VOCs and further use them in lung cancer detection.

In a study published on Nov. 30, 2022, in PLOS One, researchers at the University of Louisville, Kentucky, and Indian Council of Agricultural Research (ICAR), New Delhi, analyzed the metabolic carbonyl compounds present in exhaled breath of the patients and developed a machine learning approach involving relevant VOC selection and use in cancer patient classification model training.