Diagnosis of lung cancer by the analysis of exhaled breath with a colorimetric sensor array

Background: The pattern of volatile organic compounds (VOCs) in the exhaled breath of patients with lung cancer may be unique. New sensor systems that detect patterns of VOCs have been developed. One of these sensor systems, a colorimetric sensor array, has 36 spots composed of different chemically sensitive compounds impregnated on a disposable cartridge. The colours of these spots change based on the chemicals with which they come into contact. In this proof of principle study, the ability of this sensor system to detect a pattern of VOCs unique to lung cancer is assessed. Methods: Individuals with lung cancer, those with other lung diseases and healthy controls performed tidal breathing of room air for 12 min while exhaling into a device designed to draw their breath across a colorimetric sensor array. The colour changes that occurred for each individual were converted into a numerical vector. The vectors were analysed statistically, using a random forests technique, to determine whether lung cancer could be predicted from the responses of the sensor. Results: 143 individuals participated in the study: 49 with non-small cell lung cancer, 18 with chronic obstructive pulmonary disease 15 with idiopathic pulmonary fibrosis 20 with pulmonary arterial hypertension 20 with sarcoidosis and 21 controls. A prediction model was developed using observations from 70% of the subjects. This model was able to predict the presence of lung cancer in the remaining 30% of subjects with a sensitivity of 73.3% and a specificity of 72.4% (p = 0.01). Conclusions: The unique chemical signature of the breath of patients with lung cancer can be detected with moderate accuracy by a colorimetric sensor array.