Detection Of Lung Cancer Biomarkers And Early Diagnosis Of Lung Cancer Based On SERS Technology

Lung cancer has the highest morbidity and mortality,which makes a serious threat to human health.Though the incidence of lung cancer has gradually declined,the five-year survival rate for patients remains low.Besides,imaging and histopathologic characteristics are considered to be gold standard for diagnosing lung cancer,but these methods are likely to lead to the risk of false positive and unnecessary treatments or diagnostic interventions for patients.Therefore,it is urgent to seek more convenient diagnostic methods to carry out early diagnosis and treatment of lung cancer.SERS has the unique advantages of high sensitivity,good selectivity,fast analysis speed and small water interference,making it highly promising for clinical examination.In this paper,the detection of lung cancer was studied from two aspects based on SERS technology:On the one hand,the blood was studied.On the other hand,it was studied at the molecular level（plasma miRNA-21 and plasma CEA）The main contents are listed as follows:1.Based on silver nanoparticles（Ag NPs）,we obtained two groups of whole blood SERS spectra:one group from 26 patients with pathologically confirmed lung cancer and another group from 45 healthy volunteers.The logistic regression（LR）,k-nearest neighbor（KNN）,decision tree（DT）,and random forest（RF）algorithms were employed to develop diagnostic model using the same spectral data.The results show that the diagnostic accuracy of LR,KNN,DT and RF models were 87%,66%,77%and 83%,respectively.We computed that the sensitivity and specificity of LR model were 86%and92%respectively as well as KNN model were 33%and 73%.DT model exhibited 66%sensitivity and 81%specificity,while RF model showed 83%and 72%.The receiver operating characteristic curve（ROC）which can accurately reflect the relationship between sensitivity and specificity is drawn,and the area under the ROC curve（AUC）value of machine learning algorithm for binary classification problem is calculated.AUC of LR,KNN,DT and RF models were 0.93,0.79,0.63 and 0.81,respectively.The diagnostic performance of each diagnostic model for lung cancer was compared through sensitivity,specificity,accuracy and AUC.Our results indicate that LR model has a better performance.Furthermore,SERS technology combined with machine learning algorithm to establish a model has great application potential in the diagnosis and screening of lung cancer.2.A biointerference-free,target-triggered core-satellite nanocomposite was developed for the detection of miRNA-21 in plasma.Using seed mediated method,Au NRs with different aspect ratios were successfully synthesized by adding different amounts of the seed liquid,and then the Au NRs with the best SERS enhancement effect were selected and functionalized with DNA1 probe.Au@4MBN@Ag NPs was functionalized with DNA2 probes.These probes are each half-complementary to the entire mir NA-21 target sequence.In the presence of miRNA-21,the miRNA-21 triggered the assembly of core-satellite nanocomposites comprising Au NRs-DNA1 as the core and Au@4MBN@Ag NPs-DNA2 as the satellite.The nanogap between core and satellite NPs generated enormous electromagnetic fields,further intensifying SERS signals of satellite NPs.This enables the sensitive detection of miRNA-21 down to the 0.1 fM-level.We measure the level of plasma miRNA-21 in 20 lung cancer patients and 10 healthy participants.Significantly higher levels of miRNA-21 are determined in lung cancer patients than in healthy participants,with clear lower expression in stage Ⅰ（n=10）than in stage Ⅲ-Ⅳ（n=10）lung cancer patients.We,therefore,believe that this proposed strategy will have high clinical potential for sensitive quantification of miRNAs markers in liquid biopsy samples and act as a complementary method for early detection of lung cancer.3.A"sandwich"sensor with a two-dimensional silver substrate and a silver coated gold nanosphere structure（Au@Ag NPs）with silent regional standard signal was constructed for the detection of carcinoembryonic antigen（CEA）in plasma.Silane coupling agent（KH550）was used to modify the glass sheet,and silver nanoparticles were captured by amino groups on the surface of the glass sheet,and silver nanoparticles were self-assembled to form silver films.Then,4-mercaptophenylboric acid（4-MPBA）was modified to adsorb glycoproteins in the sample,which reached saturation after 24 hours of modification.Au@4MBN@Ag NPs modified with CEA specific antibody using4-mercaptobenzonitrile（4MBN）as internal standard was added.The time and volume of preparation of the mercapto-pegyl-carboxyl group（SH-PEG-COOH）modified in Au@MBN@Au NPs probe were optimized.Finally,CEA in glycoprotein can be captured and detected specifically.Dual signal calibration was carried out by using the internal standard molecule MBN of the probe and the Raman characteristic peak of the signal molecule 4-MPBA of the substrate,which significantly improved the detection sensitivity of CEA,and the detection range was 10^-14-10^-8M.Meanwhile,For verifying the feasibility and accuracy of the suggested method,the recoveries of CEA in clinical plasma samples were determined,and the recoveries ranged from 105.33%-127.00%.