Construction Of Colorimetric Sensor Array And Their Applications In The Screening Of Lung Cancer

Investigation from the World Health Organization(WHO)indicates that large-scale screening of lung cancer is one of effective methods to lower the mortality.Among various screening technologies,colorimetric sensor array provides an ideal non-invasive method with the advantages of fast,accurate,inexpensive,portable and visualized for lung cancer screening via breath test.However,most of colorimetric sensor arrays cannot identify the single component of exhaled gas mixture,and usually suffer from poor selectivity and sensitivity,and higher limits of detection(LOD),which largely limits their applications in the screening of lung cancer.Hence,it is of importantly clinical significance to analyze the factors that affect the array's sensitivity and selectivity,to design various receptors,and finally to construct colorimetric sensor arrays with excellent properties via the study of interactions between arrays and analytes for the large-scale screening of lung cancer.Given the limitations of previous arrays,this paper uses the combination of the specific reactions and the cross-reactions and the functionalization of sensitive dyes and matrix with nanomaterials as guidelines,and construct different colorimetric sensor arrays to wholely improve the arrays'sensitivity,selectivity,and accuracy in the identification of specific volatile organic compounds(VOCs)and screening of lung cancer.Through the principal component analysis(PCA)and a correlation model,interactions between arrays and analytes and inherent factors were emphatically studied to provide theoretical basis for the construction of new colorimetric sensor arrays.Using pattern recognition and cross validation models,we deeply investigated the performance of the arrays in the identification of specific VOCs and breath samples.Additionally,we fully evaluated the improvements of specific reactions and functionalization with nanomaterials on the performance of arrays via statistical analysis.Finally,the feasibility that the responses of a colorimetric sensor array to the contents of exhaled breath are capable of identifying and characterizing lung cancer were demonstrated.Specific results are as follows:(1)On the basis of redox between aldehyde group and Ag~+,using NaOH to reduce activation energy and utilizing sensitive dyes to indicate the change in pH respectively,a specific sensing system was designed for selected detection of aldehydes.Pattern recognition shows that the specificities and accuracies of the system are more than 92.9%and 93.3%respectively.A wide linear relationship(0.391~7.82 mM/L)was obtained for acetaldehyde with a lower limit of detection(1.002 mM/L).Pearson correlation model indicates that the reaction rate is determined by the attacking of OH~-on electropositive carbon atom in carbonyl.The difference in electro-positivity of carbon atom in carbonyl dominates the difference in responses of the system to various aldehydes.(2)Based on Br?nsted acid-base and ion interaction,a disposable colorimetric sensor array was constructed with citrate-capped gold nanoparticles-functionalized dyes,graphene oxide monolayer-functionalized dyes,and silica gel-functionalized dyes for rapid identification of amines.The recognition specificities and accuracies for 8common amines are greater than 95.8%and 92.6%respectively,and a wide linear range(5 ppb~50 ppm)and an extremely low LOD(1.3 ppb)were obtained for trimethylamine.The relative standard deviation(RSD)of septuplicate experiments is lower than 5%,implying excellent reproducibility.The principal component analysis(PCA)and Pearson correlation model show that different intensities of Br?nsted acid-base interactions between sensors and amines induce the difference in responses of various amines.Incorporation of nanomaterials significantly enhances the capacity of the array to differentiate analytes.(3)For array sensing methodologies,cross-reactive receptors are indispensable.A colorimetric sensor array was constructed with diverse chemo-responsive dyes to incorporate various chemical interactions for the identification of 20 VOCs that are closely related to lung cancer.The statistical analysis showed that the recognition specificities and accuracies are greater than 89.5%and 85%respectively.A batch of widely cross-reactive dyes are prepared for the construction of composite array.(4)A composite sensor array was constructed with specific receptors for aldehydes,nanomaterials-functionalized dyes,and various cross-reactive pigments.For the identification of 20 VOCs,the specificities and accuracies are more than 94.7%and 95%respectively,and the sensitivities are approximate to 100%.In the dendrogram,20VOCs are clustered into five chemical classes:arenes,alcohols,alkanes,aldehydes,and alkenes.Linear discriminant analysis(LDA)also obtained the results that are highly consistent with hierarchical cluster analysis(HCA).All 20 VOCs can be identified by the first three discriminant factors which only account for 60.7%of total variance.PCA implies an extraordinarily high dispersion of the array that 22 dimensions of principal components are required to explain 95%of the total variance.The results above show that the composite array has stronger physical/chemical forces when compared to the pure chemo-responsive dyes array.The improvement rate of the visualized response,sensitivity,specificity and accuracy is 29.7%,64.7%,2.2%and 5.3%respectively.(5)Using the composite nanomaterials-functionalized dyes array,the screening of breath samples was carried out for individuals with different histologies/stages of lung cancer and the healthy controls.The results demonstrat that if only based on breath molecular predictors,the screening accuracy is greater than 95%for all lung cancer histologies and the healthy controls,84.2%for squamous cell carcinoma vs.adenocarcinoma,and 81.8%for small cell carcinoma vs.non-small cell carcinoma.The recognition accuracy for stage III and stage IV is as low as 54.2%.Pearson correlation model shows that clinical risk factors(e.g.,gender and age)have significant correlation with histology phenotypes and stages of lung cancer(p<0.05).The accuracy of lung cancer staging can achieve an increase of 12.5%by evaluating breath molecular feature and incorporating clinical risk factors,that is,the combination of clinical risk factors and breath predictors shows higher accuracy.This study demonstrates that the colorimetric sensor arrays are capable of characterizing lung cancer via breath analysis.