|
By detecting the exhaled breath for the study of the diagnosis of the disease has a nearly 40-year history, and in recent years, many researchers will exhaled volatile organic compounds (VOCs) in the gas detection applied to the diagnosis of lung cancer, but characteristic for lung cancer VOCs component of its mechanism is not yet formed a unified conclusion. In addition, there are many researchers to detect exhaled breath condensate (EBC) in lung cancer markers, such as carcinoembryonic antigen (CEA). The breath test is a rapid, non-invasive, novel detection methods have broad application prospects. This paper analyzes the volatile markers of lung cancer in exhaled breath and exhaled breath condensate of nonvolatile markers of lung cancer, with self-developed electronic nose to detect the characteristic of the exhaled breath of lung cancer VOCs. The main contents include the following aspects: the thesis further analysis early in Kam et al from the acquisition of 85 patients with lung cancer, 70 cases of benign lung disease and 88 healthy people the breathing gas sample mass spectrometry data to extract 41 kinds endogenous VOCs. ROC curve, and the area under their ROC curve and significant difference p-value selected 25 kinds of VOCs statistically significant difference in lung cancer and control groups, as lung cancer markers characteristic of each of VOCs . Then using linear discriminant analysis, to establish the best lung cancer diagnostic model, the sensitivity and specificity of the best models reached 95.29% and 96.20% respectively. Our laboratory developed two electronic nose for detection of VOCs in exhaled breath, a metal oxide semiconductor (MOS) sensor, a sensor based on surface acoustic wave (SAW). This paper developed two sets of software, and is set with the MOS sensor-based CN e-nose II supporting breath test electronic nose detection analysis software, the software to complete control of the instrument and sensor data processing and analysis. Another set of software MOS-SAW the composite sensor lung diagnostic software, the software implementation of the MOS sensor and SAW sensor data analysis and diagnosis of lung cancer model. In this thesis, two electronic nose analysis of exhaled breath samples from 42 healthy subjects and 47 patients with lung cancer, 138 eigenvalues ??response curve extracted from sensor data processing. Eigenvalues ??of the ROC curve analysis, to 53 significant distinguishing between lung cancer and healthy groups eigenvalue extracted as a model of argument based on the area under the ROC curve. Finally using principal component analysis (PCA), linear discriminant analysis (LDA), Artificial Neural Network (ANN) and partial least squares regression analysis (PLS) four pattern recognition algorithm to establish the following six models: LDA model ANN model, PLS model, PCA-LDA model, PCA-ANN model, and PCA-PLS model. The final PCA-ANN model has the highest specificity and sensitivity of 90.48% and 93.62%, respectively, and higher modeling efficiency. The paper also collected EBC samples of patients with lung cancer, EBC carcinoembryonic antigen (CEA), neuron-specific enolase (NSE) and squamous cell carcinoma antigen (SCC) the three protein content were Detection and analysis. Although EBC markers of lung cancer in recent years has been many researchers, but these three markers are still very rare, although these markers are common serum markers of lung cancer, but few studies in EBC concentration. This thesis examined three markers in EBC, CEA and SCC detection rate of about 30%, the paper also analyzes the relationship between the three markers and pathological type of lung cancer.
|