Integrating Cell-Based Numerical Analysis and Machine Learning for Assessing the Binding Behavior of Circulating Tumor Cells in Microfluidic Devices and Achieving Phenotype Classification

Circulating tumor cells (CTCs) are malignant cells that detach from the primary tumor and enter the bloodstream. Early detection of CTCs is vital for diagnosis, yet it poses a challenge due to their low frequency in blood samples. Microfluidic devices emerge as a promising detection method, actively enriching CTCs through external fields or passively separating them based on physical properties. Our collaborator at Texas Tech University has proposed a microfluidic device to isolate CTCs, experimenting with various micro-post sizes and layouts for optimal capture efficiency. However, the intricate transport and adhesion behaviors of CTCs in blood cell suspensions remain not fully understood. In this study, we introduce a cell-based numerical approach, employing the Lattice Boltzmann method, to assess the binding behavior and trajectories of CTCs under diverse flow conditions. This includes variations in cell size, coating density, microfluidic design, and cell collisions. Validated results from our approach contribute to the enhancement of the microfluidic device design. Furthermore, we plan to integrate the Long Short-Term Memory (LSTM) approach to analyze the trajectory and capture location of each CTC, aiming to achieve subphenotype classification.

Please report video problems to itsupport@slmath.org.

See more of our Streaming videos on our main VMath Videos page.