Single-cell-level and population-level biophysical analysis of spiked tumor cells

The detection of circulating tumor cells (CTCs) from whole blood is critical for cancer diagnosis and prognosis. Cellular impedance serves as a key biophysical parameter for characterization at both the single-cell and population levels. In this study, we present a label-free method for CTC detection based on establishing impedance models for CTCs, white blood cells (WBCs), and red blood cells (RBCs). A microfluidic device featuring four parallel constriction channels sharing a pair of electrodes was employed for impedance measurement. As cells traverse the constrictions, CTCs undergo deformation, whereas WBCs and RBCs remain largely undeformed, enabling differentiation based on impedance signatures. The acquired impedance–time data were processed using machine learning algorithms to identify signals corresponding to individual CTCs as well as clusters of WBCs and RBCs. Specifically, a segmentation algorithm combining k-means clustering and relevance analysis was developed to extract single-cell and cell-cluster events. At the population level, impedance distributions were characterized using histogram-based models, enabling accurate classification of different cell groups. At the single-cell level, receiver operating characteristic curve analysis was applied to evaluate the performance of a regression-based classification model. The results demonstrate high accuracy in distinguishing CTCs from blood cells, achieving a prediction rate of 98% at the single-cell level. At the population level, overlap between cell types was minimal, with most overlap rates below 5% at selected frequencies. Overall, this method provides proof of concept for biophysical analysis of CTCs using impedance-based, label-free techniques.