Gray-Level Guided Image-Activated Droplet Sorter for Label-Free, High-Accuracy Screening of Single-Cell on Demand

Single-cell encapsulation in droplet microfluidics has become a powerful tool in precision medicine, single-cell analysis, and immunotherapy. However, droplet generation with a single-cell encapsulation is a random process, which also results in a large number of empty and multi-cell droplets. Current microfluidics sorting technologies suffer from drawbacks such as fluorescent labeling, inability to remove multi-cell droplets, or low throughput. This paper presents a gray-level guided image-activated droplet sorter (GL-IADS), which enables label-free, high-accuracy screening of single-cell droplets by rejecting empty and multi-cell droplets. The gray-level based recognition method can accurately classify droplet images (empty, single-cell, and multi-cell droplets), especially in differentiating empty and cell-laden droplets (accuracy of 100%). Crucially, this method reduces the image processing time to ≈300 µs, which makes the GL-IADS possible to reach an ultra-high sorting throughput up to hundreds or even KHz. The GL-IADS integrates the novel recognition method with a detachable acoustofluidic system, achieving sorting purity of 97.9%, 97.4%, and >99% for single-cell, multi-cell, and cell-laden droplets, respectively, with a throughput of 43 Hz. The GL-IADS holds promise for numerous biological applications that are previously difficult with fluorescence-based technologies.