Tedlar bag free: Accurate volatolomics of ⅠA stage non-small cell lung cancer come out in wash

Bohao Liu,Xue Jiang,Ruizhi Ning,Heng Zhao,Yanpeng Zhang,Junnan Zhang,Tianqing Liu,Danyao Qu,Yinhui Bao,Zhanchen Guo,Xiaoyan Zeng,Shan Gao,Kun Fan,Runyi Tao,Jian Ji,Guangjian Zhang,Weiwei Wu
DOI: https://doi.org/10.1016/j.cclet.2024.110301
IF: 9.1
2024-01-01
Chinese Chemical Letters
Abstract:Breath analysis diagnose diseases non-invasively. Accurate measurement of volatolomics is critical for breath analysis to be a gold standard. Tedlar bags (TB) are often used to collect breath samples, but they emit contaminants that affect accuracy. This issue was overlooked in previous studies. We found contamination issues with TB (Inc., siloxanes and aromatic impurities, etc.) that affect the identification of volatile organic compounds (VOCs) due to impurities. Then, home-designed equipment (HD) made with poly-tetrafluoride (PTFE) and quartz glass for breath collection was developed and employed in clinical trials. 15 healthy individuals and 32 non-small cell lung cancer (NSCLC) patients at IA stage participated in this study. 610 VOCs can be collected through TB, which is less than HD (1109 VOCs), demonstrating that the inner wall of the TB easily adsorbs VOCs, leading to decreased detection concentrations. Otherwise, utilizing orthogonal partial least squares discriminant analysis (OPLS-DA), we identified chemical markers with significant discriminatory power (VIP > 1.5, P < 0.05). The HD method identified 12 target VOCs, surpassing the 3 target VOCs discerned by the TB method. A model combined with a machine learning algorithm for distinguishing early-stage lung cancer patients was established based on biomarkers, which were selected based on OPLS-DA. The results showed strong predictive capabilities for the HD-based model. It indicated that 12 biomarkers derived from the HD model were more effective in distinguishing NSCLC patients, with an AUC value of 0.92, compared to the AUC value of 0.5 from 3 markers obtained from the TB model. The sensitivity and specificity in the confusion matrix reached 100% and 80% for the HD test, but TB test reached only 40% and 60%. This work demonstrated that optimizing and standardizing VOCs collection methodology from breath of lung cancer patients is essential to identify actual volatiles, which could promote disease volatolomics worldwide.
What problem does this paper attempt to address?