| Lung cancer is one of the most common primary malignant tumors,which has high incidence and leads to serious mortality in global scope.Clinical statistics show that the vast majority of patients are not diagnosed until the late stage of lung cancer,which is one of the main reasons for the high mortality rate.Early detection and early treatment are of great help to improve the survival rate of patients,so it is of great significance for patients with large-scale early lung cancer screening and diagnosis.The study found that the types and contents of volatile organic compounds(VOCs)in exhaled gases of patients are different from normal people,and are closely related to lung cancer lesions.Thus,they have been used as lung cancer related gas markers by many researchers.At the same time,lung cancer related tumor markers in blood and tissues,such as carcinoembryonic antigen(CEA)and cytokeratin 19 fragment(CYFRA21-1),are expected to be used in clinical lung cancer diagnosis.Compared with medical imaging and needle biopsy,the detection of lung cancer related gas-liquid markers has many advantages,such as low cost,non-invasive,rapid,etc.Although it has a certain false positive rate,it can be used for early lung cancer screening in large populations.Nevertheless,the detection of exhaled gas is mostly based on large-scale instruments such as GC-MS at the present.The tumor markers are mostly detected by enzyme-linked immunosorbent assay with the expensive instruments which are complicated to operate.Besides,it is often inseparable from professional laboratories and is inconvenient for Point of Care Testing(POCT)portable detection.Although some specific sensors,such as photoelectric sensors and biological DNA sensors,have the advantages of broad spectrum,high sensitivity,and good specificity,they also make the detection results highly misjudged due to their single signal component(only one of electric signal,visible light signal,or fluorescent signal)and single system(gas or liquid).Therefore,based on the screening of lung cancer patients and the comprehensive analysis of the existing screening techniques,a new lung cancer related marker screening and detection system based on dual-signal and dual system was proposed in this paper,which realized the associated,rapid,qualitative and quantitative detection of lung cancer related exhaled gas and tumor markers.Simultaneously,a set of dual-signal feature extraction algorithm was established to cooperate with the detection system,providing new ideas for early screening of lung cancer.In this paper,the following aspects of work were mainly completed.1)The whole construction of a new dual-signal and gas-liquid lung cancer related marker detection system.Firstly,based on the cross-response and specific target recognition mechanism,the detection system was constructed according to the functional needs analysis,and guided by the modular idea,the five subsystems of gas,liquid,light,machine and electricity were simultaneously constructed to solve the five key technical problems involved in building the system.When establishing the subsystem,the design of the sensing chamber,the gas-liquid sensing array,the gas circuit,the light source,etc.were involved.And through 3D structural modeling,digital model simulation,fluid mechanics analysis,time-multiplexing,mathematical model calculation,error assessment and other mathematical tools and technical means,the reliability of the design was verified and the optimal parameters was configured for completing the system integration.Finally,the subsystems were organically integrated,and the system debugging was completed by programming the upper and lower computer software programs,which provided a platform for the design of the dual signal feature extraction algorithm and the detection of gas-liquid markers.2)The proposal of the feature extraction algorithm for the visible light signal.Staring from the feature extraction theory,the generation mechanism of visible light signal and its signal characteristics were analyzed.Based on this,through signal preprocessing,morphological image filtering,image segmentation based on grid division,color information extraction and coding of sensor point,and finally the color space conversion,the "fingerprint map" of the marker and the HSI feature data matrix were obtained,which provided the basis for subsequent pattern recognition and metrology analysis.3)The proposal of the feature extraction algorithm for the fluorescent signal.For the fluorescent signal,according to the feature extraction idea,the generation mechanism of fluorescent signal and its signal characteristics were analyzed.And the macroscopic and local microscopic features of the fluorescence difference spectrum were extracted by the theory of spectral preprocessing,time-frequency domain feature extraction,wavelet analysis and fractal dimension.The fluorescence difference spectrum feature coding was performed,and the features and data acquired by the algorithm can be used for subsequent pattern recognition and metrology analysis.4)The detection of lung cancer related gas-liquid markers by cross-response and specific recognition mechanism to test the performance of the proposed system and algorithm.For gas markers,according to the characteristics of the dual signal of the sensor and the acquired data feature matrix,the target gas was qualitatively and quantitatively analyzed by pattern recognition theory such as hierarchical clustering analysis(HCA),Back-propagation neural network(BPNN)and principal component analysis(PCA),and the "fingerprint" and "fingerprint spectrum" corresponded to the marker were acquired.Therein,the idea of using the visible light signal for qualitative analysis and the fluorescent signal to quantify was determined by the misjudgment results appearing in the pattern recognition analysis and the comparative analysis of the two signal results.At the same time,the source of misjudgment was analyzed to design a new sensing array to detect aldehydes.For liquid markers,aptamer technology was used.Ag nanoclusters were combined to construct fluorescent probes,which were cooperated with graphene oxide to form "fluorescence cut-off" system to achieve qualitative and quantitative identification of liquid markers.Finally,through a large number of experimental tests,the qualitative recognition rate of lung cancer related gas markers was 100%,and the semi-quantitative interval of low concentration(50ppb,200 ppb,350ppb,500ppb)were obtained with the overall interval division error being less than 10%.Besides,the qualitative recognition rate of was 100%,and the formaldehyde pollution monitoring as a case was also verifed in the concentration ranging from 40 ppb to 3 ppm through a semi-quantitative ftting analysis with a detection limit of 10.2 ppb and a good system recovery of 91.2% to 110.4%.For CEA detection,the probe has specific recognition specificity,and the linear range of 5.3 ng/ml~14.7 ?g/ml and the detection limit of 4.72 ng/ml were obtained.The feasibility of the detection system and algorithm was verified by the above test results comprehensively.It is expected to provide new methods and equipment for early clinical diagnosis of lung cancer. |
PDF Full Text Request