Separation of exfoliated tumor cells from viscoelastic pleural effusion using a microfluidic sandwich structure

A microfluidic device with a sandwich structure is proposed to achieve label-free and size-selective separation of tumor cells from pleural effusion. The sandwich structure is a co-flow system incorporating an initial sample layer, an isolation layer and the target sample layer. The isolation layer is used to provide a size-selective interface between the initial sample layer and the isolation layer. The relative magnitude of the inertial lift force and the interfacial lift force at the interface only allows exfoliated tumor cells to migrate out of the sample layer. The high interfacial elastic lift force of the isolation layer also enables the device to be used for pleural effusion samples, whose properties usually vary across a wide range. The target sample layer is used for large migration distances of exfoliated tumor cells in the contraction-expansion array (CEA) channel and high separation efficiency. Cell washing is also achieved with the target sample layer, demonstrating the integration of our device. Experimentally, an optimal flow rate ratio of 1:1:6 was obtained to ensure the stability of the sandwich structure, and the collected fluid was all from the target sample layer. A critical polyethylene oxide (PEO) concentration of the isolation layer (500 ppm, η₀ = 1.37 mPa·s) was then obtained by particle tests. Twenty-micrometer particles were efficiently separated from different viscoelastic samples (PEO concentration changes from 0 to 400 ppm) at this concentration. For the cell test, exfoliated tumor cells from different pleural effusion samples were successfully separated and washed. The separation efficiency of exfoliated tumor cells and blood cells was about 100% and over 90%, respectively. Compared with a conventional co-flow system of two fluids, this device has great advantages in 1) wide applicability for pleural effusion samples of various viscoelasticity and 2) focusing performance. It shows potential for use in medical research and clinical diagnosis of cancer.