Research On Online Detection Approach Of Cancer Cell Viability In Microenvironment Based On Biophysical And Electrochemical Characteristics

Cell viability detection technology is widely used in many fields,such as cell culture study,tumor immunotherapy,cytotoxicity test,anti-cancer drug test,drugresistance research and cell invasion study.Traditionally,cell viability is defined as the percentage of living cells in total cells.However,due to the heterogeneity of individual cells,the drug responses of the same number of cell groups are different,resulting in inaccuracy and unstability of cell viability evaluation.The existing cell viability detection methods,such as MTT,XTT and CCK-8 method,all belong to the end-point detection method with labels and harms to cells,which damages the continuity and accuracy of cell viability research to a certain extent.In addition,the existing label-free non-invasive testing methods,such as electrochemical impedance method,optical microscopic imaging method and lensfree imaging method,only reflect the unilateral information of cancer cell viability.Therefore,improving the accuracy and stability of cell viability evaluation methods is of great significance to the field of biochemical experimental research and medical treatment.In this paper,a cell viability detection method based on physicochemical characteristics is proposed.This method realizes accurate evulation of cell viability by fusing cell electrochemical characteristics,such as the number of living cells,adhesion state,membrane capacitive character and electrical wound-healing ability of BGC-823 and SGC-7901 cells,as well as the mechanical physical characteristics,such as adhesion strengths of cell population,single cell volume,deformability and membrane surface friction characteristics.Three kinds of artificial neural networks are used to compare their advantages and disadvantages.The specific research contents of this paper are as follows:(1)Aiming at the "drift" phenomenon of the most sensitive frequency in the electrochemical impedance spectroscopy found in this paper,and the impedance error caused by noise such as cell fretting and instrument interference,an area-normalized method based on impedance spectroscopy is proposed in this paper.The results show that the maximum error of the area-normalized impedance method is reduced by 4.4%compared with that of single frequency impedance method.This method effectively reduces the error caused by the "drift" of the most sensitive frequency of impedance spectroscopy and noise,and provides a basis for stably and accurately extracting the characteristics related to cell viability by impedance spectroscopy.(2)In view of the one-sidedness of the existing cell electrochemical detection methods,based on the in-depth study of the principle of cell electrochemical detection,this paper integrates the ECIS biosensor system with the microfluidic system to effectively extract the electrochemical characteristics such as the number of living cells,adhesion state,membrane capacitance and electrical wound-healing ability.This lays a foundation for the evaluation of cancer cell viability based on information fusion of cell physicochemical feature.(3)In order to improve the comprehensiveness of cancer cell viability evaluation,the mechanical physical characterization methods related to cell viability were studied in this paper.In this paper,a single-channel integrated microfluidic system is designed and manufactured.Through cell detachment assay,the adhesion cell detachment curve is obtained,and the adhesion strength gradient of cell population is extracted.At the same time,the cancer cells washed away by the cell detachment assay are guided into the narrowed microfluidic chip.The statistical characterization of the mechanical and physical properties such as the volume,deformability and membrane surface friction characteristics of single cancer cells were studied,which lays a foundation for the evaluation of cell viability based on the information fusion of physicochemical characteristics.(4)Aiming at the difficulties in the normalization process of physicochemical characteristics related to cancer cell viability,based on the in-depth analysis of the correlation and complementarity between various physicochemical characteristics,this paper studies the normalization algorithm of electrochemical characteristics such as cell adhesion state,membrane capacitance and electrical wound-healing ability,as well as biophysical characteristics such as cell population adhesion strength,single cancer cell deformability and membrane surface friction characteristics.Three artificial neural network algorithms are used to fuse the physicochemical characteristics together to evaluate cell viability.The results show that the prediction results of BP neural network is poor and there are occasionally very large deviation prediction values.Besides,Elman neural network takes the second place and RBF neural network has the best prediction results.(5)In order to study the accuracy of cell viability evaluation methods under different FBS concentrations,the cancer cell viability evaluation method based on physicochemical information fusion proposed in this paper was compared with the cell counting method,single adhesion strength method and multiple adhesion strength method.The results show that the total average error of physicochemical feature fusion method was increased by 6.01% compared with multiple adhesion strength method and by 62.72% compared with cell counting method.