Xing Chen1, Kanhar Ghulam Muhammad2, Channa Madeeha3, Wei Fu4, Linxin Xu5, Yanjie Hu6, Jun Liu7, Kejing Ying8, Liying Chen9, Gorlova Olga Yurievna10. 1. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: cnxingchen@zju.edu.cn. 2. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: kanhargm@zju.edu.cn. 3. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: madeehagm@zju.edu.cn. 4. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: 21815055@zju.edu.cn. 5. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: zhixiaz@foxmail.com. 6. Zhejiang Sir Run Run Shaw Hospital, Department of Medicine, Zhejiang University, Hangzhou, China. Electronic address: huyanjie1016@163.com. 7. Department of Biomedical Engineering, Key Laboratory of Biomedical Engineering of Ministry of Education of China, Zhejiang Provincial Key Laboratory of Cardio-Cerebral Vascular Detection Technology and Medicinal Effectiveness Appraisal, Zhejiang University, Hangzhou, Zhejiang, China. Electronic address: liujun@zju.edu.cn. 8. Zhejiang Sir Run Run Shaw Hospital, Department of Medicine, Zhejiang University, Hangzhou, China. Electronic address: yingsrrsh@163.com. 9. Zhejiang Sir Run Run Shaw Hospital, Department of Medicine, Zhejiang University, Hangzhou, China. Electronic address: 3197020@zju.edu.cn. 10. Department of Medicine Epidemiology and Population Sciences, Baylor College of Medicine, Houston, TX, USA. Electronic address: olga.gorlova@bcm.edu.
Abstract
BACKGROUND: Breath analysis is a promising noninvasive technique that offers a wide range of opportunities to facilitate early diagnosis of lung cancer (LC). METHOD: Exhaled breath samples of 352 subjects including 160 with lung cancer (LC), 70 with benign pulmonary nodule (BPN) and 122 healthy controls (HC) were analyzed through thermal desorption coupled with gas chromatography-mass spectrometry (TD-GC-MS) to obtain the metabolic information from volatile organic compounds (VOCs). Statistical classification models were used to find diagnostic clusters of VOCs for the discrimination of HC, BPN and LC patients' early and advanced stages, as well as subtypes of LC. Receiver operator characteristics (ROC) curves with 5-fold validations were used to evaluate the accuracy of these models. RESULTS: The analysis revealed that 20, 19, 19, and 20 VOCs discriminated LC from HC, LC from BPN, histology and LC stages respectively. The calculated diagnostic indices showed a large area under the curve (AUC) to distinguish HC from LC (AUC: 0.987, 95 % confidence interval (CI): 0.976-0.997), BPN from LC (AUC: 0.809, 95 % CI: 0.758-0.860), NSCLC from SCLC (AUC: 0.939, 95 % CI: 0.875-0.995) and Stage III from stage III-IV (AUC: 0.827, 95 % CI: 0.768-0.886). The comparison between the high-risk groups (BPN and HC smokers) and early stages LC resulted in the AUC of 0.756 (95 %CI: 0.681-0.817) for BPN vs. early stage LC and AUC of 0.986 (95 % CI: 0.972-0.994) for HC smoker vs. early stage LC. CONCLUSION: Volatome of breath of the LC patients was significantly different from that of both BPN patients and HC and showed an ability of distinguishing early from advance stage LC and NSCLC from SCLC. We conclude that the volatome has a potential to help improve early diagnosis of LC.
BACKGROUND: Breath analysis is a promising noninvasive technique that offers a wide range of opportunities to facilitate early diagnosis of lung cancer (LC). METHOD: Exhaled breath samples of 352 subjects including 160 with lung cancer (LC), 70 with benign pulmonary nodule (BPN) and 122 healthy controls (HC) were analyzed through thermal desorption coupled with gas chromatography-mass spectrometry (TD-GC-MS) to obtain the metabolic information from volatile organic compounds (VOCs). Statistical classification models were used to find diagnostic clusters of VOCs for the discrimination of HC, BPN and LC patients' early and advanced stages, as well as subtypes of LC. Receiver operator characteristics (ROC) curves with 5-fold validations were used to evaluate the accuracy of these models. RESULTS: The analysis revealed that 20, 19, 19, and 20 VOCs discriminated LC from HC, LC from BPN, histology and LC stages respectively. The calculated diagnostic indices showed a large area under the curve (AUC) to distinguish HC from LC (AUC: 0.987, 95 % confidence interval (CI): 0.976-0.997), BPN from LC (AUC: 0.809, 95 % CI: 0.758-0.860), NSCLC from SCLC (AUC: 0.939, 95 % CI: 0.875-0.995) and Stage III from stage III-IV (AUC: 0.827, 95 % CI: 0.768-0.886). The comparison between the high-risk groups (BPN and HC smokers) and early stages LC resulted in the AUC of 0.756 (95 %CI: 0.681-0.817) for BPN vs. early stage LC and AUC of 0.986 (95 % CI: 0.972-0.994) for HC smoker vs. early stage LC. CONCLUSION: Volatome of breath of the LCpatients was significantly different from that of both BPN patients and HC and showed an ability of distinguishing early from advance stage LC and NSCLC from SCLC. We conclude that the volatome has a potential to help improve early diagnosis of LC.