Evaluating cell viability heterogeneity based on information fusion of multiple adhesion strengths

• Published in: Biotechnology and bioengineering Vol. 118; no. 6; pp. 2360 - 2367
• Main Authors: Wei, Mingji, Zhang, Rongbiao, Zhang, Fei, Zhang, Yecheng
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01.06.2021
• Subjects: Adhesion; adhesion strengths; Antineoplastic drugs; Antitumor agents; Assaying; Biological Assay; Cell Adhesion; Cell fusion; Cell Line, Tumor; Cell Survival; Cell viability; cell viability heterogeneity; Data integration; Dependent variables; Heterogeneity; Humans; Independent variables; microfluidic chip; Microfluidics; Models, Theoretical; polynomial regression; Polynomials; Regression models; Shear stress; Stress, Mechanical; microfluidic chip; shear stress; adhesion strengths; cell viability heterogeneity; polynomial regression
• ISSN: 0006-3592; 1097-0290
• DOI: 10.1002/bit.27749

Cell viability evaluation is significantly meaningful for cellular assays. Some cells with weak viability are easily killed in the detection of anticancer drugs, while others with strong viability survive and proliferate, ultimately leading to the treatment failure or the inaccuracy of biological assays. Accurately evaluating cell viability heterogeneity still remains difficult. This article proposed a multiphysical property information fusion method for evaluating cell viability heterogeneity based on polynomial regression in a single‐channel integrated microfluidic chip. In this method, adhesion strengths τN, that are defined as the magnitude of shear stress needed to detach (100 − N) % of cell population, were extracted as the independent variables of polynomial regression model by calculating the nonlinear fitting of the impedance–response curves for shear stress (cell detachment assay). Besides, by calculating the nonlinear fitting of the drug dose–response curves for cancer cells (IC50 assay), the half‐maximal inhibitory concentration (IC50) was extracted as the dependent variables of polynomial regression model. The results show that the mean relative error of our fusion method averagely reduces by 6.04% and 62.79% compared with the multiple linear regression method and the cell counting method. Moreover, a simplified theoretical model used to describe the quantitative relationship between cell viability and its adhesion strengths was built to provide a theoretical basis for our fusion method. Polynomial fitting multiple cell adhesion strengths can realize accurate prediction of cell viability heterogeneity. Multiple cell adhesion strengths can be extracted by using cell detachment assay, and cell viability is calculated by IC50 assay.