A Profile of Exhaled Volatile Organic Compounds Measured Using PTR-TOF-MS for Diagnosis of Chronic Obstructive Pulmonary Disease
Xu Yifan,Ying Liu,Xi Chen,Xinghua Qiu,Chengli Que,Lina Zhang,Guancong Huang,Qi Wang,Yuan Yao,Wu Chen,Tong Zhu
DOI: https://doi.org/10.1289/isesisee.2018.p01.2820
2018-01-01
ISEE Conference Abstracts
Abstract:Background: Chronic obstructive pulmonary disease (COPD) is one of the leading causes of global death. The diagnosis of COPD clinically relying on pulmonary function tests is invasive for the elders.Method: To develop a non-invasive and efficient diagnosis method for COPD, a panel of 102 subjects, including 44 stable COPD patients and 58 healthy control, were recruited in Beijing, China. The exhaled gas of each subject was collected with Tedlar bags and then analyzed with proton-transfer-reaction time-of-flight mass spectrometry (PTR-TOF-MS). Binary logistic regression model was used to evaluate the associations between exhaled VOCs level and the classification of subjects, modified by smoking status.Results: Among 260 VOCs identified as endogenous VOCs, 16 key VOCs had discriminative values. The performance of classification of COPD patients using the 16 VOCs delineates an area under the curve (AUC) of 0.93 in the Receiver Operating Characteristic (ROC) curve. 10-folds cross validation demonstrated that the discriminative model has a predictive accuracy of 0.909. Among the 16 VOCs, the level of [m/z=34.99, 46.03, 47.05, 69.07, 73.03, 101.10, 169.16, 243.08] were not significantly higher or lower in the subjects with asthma, diabetes mellitus, hyperlipemia, or smoke history, which indicates that the method is specific for COPD diagnosis.Tentative conclusion: In conclusion, several specific exhaled VOCs were screened out for the non-invasive diagnosis of COPD using PTR-TOF-MS, which is promising for clinical uses. The screened VOCs should be further characterized and their metabolic pathways require further research.