Flow cytometry is one of the most widely used techniques in cell sorting, but its lack of spatial resolution limits its utility for quantifying complex cellular phenotypes. Goda et al. present an imaging flow analyzer that can capture microscopic images of single cells at a rate of up to 100,000 cells/s. To achieve this high throughput, the authors combine a technology known as 'serial time-encoded amplified microscopy' with a microfluidic device, which provides uniform flow velocities and cell positioning in the front of an objective lens, and an 'optoelectronic time-stretch image processor' that can analyze the raw data in real time. By screening a heterogeneous budding yeast population, the technique is able to sort cells into different stages of the budding process solely based on cell morphology. The system is also able to detect rare mammalian cells in a blood sample. In a 'spike in' experiment, MCF7 cells were labeled with 1-μm metal beads attached to specific antibodies. The authors show that image-based cell identification is sensitive enough to detect one labeled MCF7 cell in a million white blood cells with a false-discovery rate 100× smaller than conventional fluorescence flow cytometry. (Proc. Natl. Acad. Sci. USA. 109, 11630–11635, 2012)