Microfluidic Channels for Separation of Circulating Tumor Cells from Blood cells Using Dielectrophoresis and Performance Analysis Using ANFIS

Abstract

The traditional isolation process of Circulating Tumor Cells (CTCs) is a huge technical hurdle. In this work, simple electrode arrangements are proposed that utilizes Dielectrophoresis and Fluid Dynamics to separate CTCs from blood cells that can be used effectively in microfluidic channels. Dielectrophoresis mechanism aided the microfluidic channel that has been made considering the Clausius-Mossotti (CM) factor, electrical and other mechanical properties of RBC and CTC particles to accumulate the rare CTCs being isolated from blood cells to a specified outlet. Single to multi-phase separation-based microfluidic channels have been proposed one of them can separate CTCs from RBCs in a comparatively low voltage of 8 V peak-to-peak in the first phase separation region and 4 V peak-to-peak in the second phase separation region with 100 kHz Alternating Current (AC) for the inlet sample stream speed of 420 μm/s. The other microfluidic channel can separate CTCs from WBCs and RBCs in a comparatively low voltage of 6-8 V peak-to-peak with 100 kHz Alternating Current (AC) for the inlet sample stream speed of 150 μm/s. A comparative analysis with microfluidic channels with single and multi-phase separation and different electrode arrangements by computer-assisted multi-physics simulations using Finite Element Method (FEM) with various governing parameters using COMSOL, MATLAB, and MyDEP software has been done in this study to validate the performance of the proposed microfluidic channels. The proposed microfluidic channels have achieved 100% separation efficiency (SE) and 100% separation purity (SP) while separating CTCs from different blood cells. Analysis of the inputs and outputs from the simulation models have been done to suggest specific values of inputs for the most efficient separation of the channels through Adaptive Neuro-Fuzzy Inference System (ANFIS) where two machine learning algorithms were used to give an overview of the microfluidic channel’s input-output relationship.