Label-free Cell Sorting Based On Image Recognition And Surface Acoustic Wave Microfluidic Technology

With the development of cell therapy and gene therapy technology,cell sequencing and cell culture have higher and higher requirements on the purity of cell acquisition,and cell sorting and enrichment technology has become a key research topic in the medical field.Although the traditional Fluorescence Activated Cell Sorting based on fluorescent labeling has high purity and recovery rate for cell separation,it needs to carry out complex labeling treatment on cells in the early stage,which affects cell viability.Moreover,the equipment is expensive and complex.And it is easy to produce photobleaching and phototoxicity in the experimental process.In order to solve these problems,a cell sorting device called Image Acoustofluidics Cell Sorter(IACS)based on Image recognition technology,microfluidic technology and acoustic separation technology was proposed in this paper.Real-time intelligent label-free cell classification was performed in combination with CNN.In order to reduce the cost of equipment,a cheap and repeatable microfluidic device is selected as the core of the separation system.In cell separation,Acoustic Radiation Force(ARF)is used to cause cell displacement.However,the separation of ARF is based on the size,density and other physical characteristics of cells,and the sorting effect for cells with similar size is poor.Combined with CNN,cell image information can be precisely extracted,providing another sorting basis for cells with similar physical properties.The IACS proposed in this paper adopts a labeling free method to automatically classify cells,so as to achieve the purification and collection of target cells.First,most of the waste cells were separated from the cell mix sample by ARF into the waste liquid pool.Then,cell images of purified samples were collected under a bright field microscope.Morphological characteristics of cells were analyzed using CNN,and cells were classified in real time by effective classification criteria.In this paper,the YOLOv3 target detection model based on CNN is improved,and residual structure and small target detection layer are added on the basis of the model to effectively prevent the degradation of deep network,so as to further improve the feature extraction capability of the network and the detection capability of the small target.Finally,each detected target cell in the purified sample was catapulted from the microfluidic channel into the collection area through ARF and flowed into the collection tank for collection.The results showed that the highest recovery and purity of IACS were 90.7% and 98.2%respectively.IACS provides a low-cost,repeatable,automated method for real-time,highpurity,high-activity,label-free sorting and collection of target cells compared to traditional FACS(purity >95%,recovery >90%).This technique is expected to achieve high purity classification and enrichment of different cells in the future,and further culture and analysis of the recycled cells offers hope for earlier detection of primary cancer metastasis.