|
Imaging flow cytometry (IFC) has become an established tool for cell analysis across diverse biomedical fields. However, the performance of IFC is severely limited by the fundamental trade-off among multi-color, flow speed and exposure time. Here we develop multiplex imaging flow cytometry (mIFC) that overcomes this trade-off by utilizing unique single-source single-detector technology for sensitive detection of ovarian cancer cells with the content-aware image restoration method. Our mIFC achieves efficient, non-interfering 4-channel excitation and 3-channel emission based on a metal halide lamp. The spatial wavelength division multiplexing technology with a knife-edge right-angle prism is the key optical design to simultaneously obtain brightfield and multi-color fluorescence images of individual cells in flow on a single detector. A U-net variant deep learning network based on a 3-layer encoder-decoder structure is employed to perform content-aware image restoration on captured multiplex ovarian cell images. The blurred multiplex images are converted into enhanced images, which helps to balance the trade-off between flow speed and exposure time. Our multiplex imaging flow cytometry (mIFC) with content-aware image restoration deep learning method enables automatic, high-quality detection of ovarian cancer cells and has the potential broad applications in biomedical fields.
|
