The precise separation and focusing of circulating tumor cells (CTCs) from blood cells are crucial for advancing cancer diagnosis, optimizing therapeutic strategies, and fostering progress in cellular research. Inertial microfluidics offers an efficient, label-free solution with high throughput and simple design. However, challenges remain in improving separation efficiency, purity, and throughput while minimizing fabrication costs. This study introduces an optimized triangular microchannel design to enhance CTC separation performance. A Gaussian Process Regression (GPR) model was developed to predict the separation efficiency and purity of the microchannel based on key design parameters, significantly reducing computational costs and enabling rapid optimization. The simulation results indicated that a 60° triangular microchannel geometry with a curvature radius of R=200μm emerged as the optimal configuration, achieving 100% separation efficiency and purity at an inlet flow rate of 2mL/min. Experimental validation of the fabricated microchip using cultured MCF-7 (CTC) and white blood cell (WBC) demonstrated its remarkable performance, achieving a separation efficiency of 95.7% and a purity of 93.3% for low concentration (1:30) and efficiency of 93.2% and a purity of 92.5% for high concentration (1:5000) at the same flow rate. The integration of machine learning-based modeling with inertial microfluidics in this work provides a powerful approach for optimizing microchannel designs while maintaining high efficiency and purity. The proposed microchip represents a significant advancement in inertial microfluidics, offering a reliable and scalable solution for biomedical and clinical applications, particularly in CTC isolation and enrichment from complex biological samples.
Keywords: Cell migration; Circulating tumor cells (CTCs); Inertial microchannels; Liquid biopsy; Separation.
Copyright © 2025. Published by Elsevier B.V.