A novel device based on a fluorescent cross-responsive sensor array for detecting lung cancer related volatile organic compounds

• Published in: Review of scientific instruments 2015-02, Vol.86 (2), p.025106-025106
• Main Authors: Lei, Jin-can, Hou, Chang-jun, Huo, Dan-qun, Luo, Xiao-gang, Bao, Ming-ze, Li, Xian, Yang, Mei, Fa, Huan-bao
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial neural networks; Cancer; Change detection; Chemistry Techniques, Analytical - economics; Chemistry Techniques, Analytical - instrumentation; Cluster Analysis; Computer simulation; Data acquisition; Early Detection of Cancer; Fluorescence; Isoprene; Lung cancer; Lung Neoplasms - chemistry; Lung Neoplasms - diagnosis; Pattern recognition; Principal components analysis; Rotation; Scientific apparatus & instruments; Sensor arrays; Sensors; Spectrometry, Fluorescence; Time Factors; VOCs; Volatile organic compounds; Volatile Organic Compounds - analysis; Volatile Organic Compounds - chemistry; Xylene
• ISSN: 0034-6748; 1089-7623
• DOI: 10.1063/1.4907628

In this paper, a novel, simple, rapid, and low-cost detection device for lung cancer related Volatile Organic Compounds (VOCs) was constructed. For this task, a sensor array based on cross-responsive mechanism was designed. A special gas chamber was made to insure sensor array exposed to VOCs sufficiently and evenly, and FLUENT software was used to simulate the performance of the gas chamber. The data collection and processing system was used to detect fluorescent changes of the sensor arrays before and after reaction, and to extract unique patterns of the tested VOCs. Four selected VOCs, p-xylene, styrene, isoprene, and hexanal, were detected by the proposed device. Unsupervised pattern recognition methods, hierarchical cluster analysis and principal component analysis, were used to analyze data. The results showed that the methods could 100% discriminate the four VOCs. What is more, combined with artificial neural network, the correct rate of quantitative detection was up to 100%, and the device obtained responses at concentrations below 50 ppb. In conclusion, the proposed detection device showed excellent selectivity and discrimination ability for the VOCs related to lung cancer. Furthermore, our preliminary study demonstrated that the proposed detection device has brilliant potential application for early clinical diagnosis of lung cancer.