A study of an electronic nose for detection of lung cancer based on a virtual SAW gas sensors array and imaging recognition method

In this paper, we propose an electronic nose for non-invasive detection and diagnosis of lung cancer based on a kind of virtual array of surface acoustic wave (SAW) gas sensors and an imaging recognition method. It includes a gas path constructed from a two-bag system, solid phase micro extraction (SPME) and a capillary column to pre-concentrate and separate volatile organic compounds (VOCs) in patients' exhaled air. A pair of SAW sensors, one coated with a thin polyisobutylene (PIB) film, is used to detect chemical compounds. Eleven VOCs that are validated as the markers of lung cancer according to a pathology study can be detected qualitatively and quantitatively by this electronic nose. Then, an improved artificial neural network (ANN) algorithm combined with an imaging method is proposed for the recognition of patients. In addition, the concept of a virtual sensors array based on SAW sensors using a capillary column separation technique and imaging is also proposed to simulate a large scale of sensor array response. Finally this e-nose is calibrated by these 11 VOCs separated in three concentrations and is used to diagnose lung cancer patients in Run Run Shaw hospital. The experimental results show that this kind of electronic nose is effective in the recognition of lung cancer patients.