Inertial focusing significantly enhances miniature flow-cytometry throughput

Flow cytometry is the gold standard for counting and identifying cells within heterogeneous samples because of the quantifiable data and significant achievable throughput (~50,000 cells per second). These techniques are widely applied to diagnose diseases, particularly through enumeration (counting) of various populations of blood cells. In addition to optical components, an important aspect is the fluid-handling method by which current-flow cytometers prepare cells in continuous flow for optical interrogation. Flow cytometers require large amounts of sheath fluid to hydrodynamically focus or squeeze the sample solution into a narrow stream. This allows alignment with optical excitation to serially examine cells of interest through fluorescence or scatter signatures. Despite current successes, improvements are required to increase throughput for applications in rare-cell analysis, such as identification of circulating tumor cells in blood (on the order of one or fewer positive events per one billion cells). Moreover, current designs that use sheath fluid and complex optical systems are bulky, expensive to operate, and difficult to parallelize.

Several efforts have been made to develop robust, cost-effective, and compact flow cytometers using microfluidic techniques. In most cases, the macroscale cell-focusing mechanism employing sheath fluid is directly translated to the microscale with miniaturized optical detection systems.^1,2 Other groups have explored sheathless miniature cytometers to differentiate cell types without focusing cells into a small interrogated volume, using more sophisticated target-cell enumeration techniques.^3,4 However, further increases in throughput have been challenging because the methods of cell focusing and optical interrogation are not trivially parallelized, while use of sheath fluid lowers the impact of the attempted miniaturization.

To address these challenges, we exploited inertial effects for sheath-free, parallel-flow cytometry with extreme throughput for rapid and accurate cell differentiation.⁵ Inertial effects in microfluidic systems have recently been recognized as a robust way of focusing and ordering microscale particles and cells continuously with only a single input.⁶ A balance of counteracting inertial lift forces (specifically wall-effect and shear-gradient lifts) acting on flowing cells leads to unique lateral and vertical (z) positions in a microchannel with high-aspect ratio (height/width >2) when flow speeds are similar to those used in conventional cytometers.⁷ The ability to localize cells and particles to precise lateral and z positions within a flow and ensure uniform velocities without requiring sheath fluid make inertial focusing an ideal candidate for parallel-cytometry applications. Since particles or cells are focused to one uniform z position, the probability of overlap and out-of-focus blur is negligible, and uniform cell-signature images for accurate detection and analysis are guaranteed. In addition, a stable and uniform particle-flow velocity allows identical residence times for each cell within the given field of view (FOV), yielding identical excitation intensities for laser-based interrogation or the ability to synchronize the frequency of a raster-scanning laser with cell downstream velocity.

We have developed a microfluidic parallel-flow cytometer with a sample rate up to 28 million cells/s by parallelizing 256 high-aspect (width 16μm, height 37μm) microchannels (see Figure 1). The throughput of our device (1 million cells/s) is only limited by the FOV of our high-speed optical interrogation method (a 10-channel FOV). We estimate the overall throughput at 28 million cells/s, upon integration with wide-FOV detection systems.⁸ We confirmed massively parallel inertial focusing abilities, including one uniform z position (standard deviation ± 1.81μm) with uniform downstream velocity (U_ave=0.208±0.004m/s) using monodispersed microparticles (diameter 10μm) as well as dilute whole blood. In addition, the automated red- and white-blood-cell counts in diluted whole blood reveal the high detection sensitivity and specificity (86–97%) of our system while operating at extreme throughput.

Figure 1. Artistic rendering and high-speed microscopic images of our sheathless parallel-flow cytometer with extreme throughput. (Artwork: Mark Lim.)

In summary, we have successfully designed and characterized a high-throughput, sheath-free cell-positioning system toward a compact flow cytometer with extreme throughput. The stable equilibrium positions and particle velocities achieved will enable future parallel fluorescence interrogation, holographic imaging, or dielectric characterization with appropriate integrated detection systems. With future implementation of larger-FOV acquisition technologies, new applications are possible, including fast total complete blood counts with limited logistical footprint and statistically significant identification of rare cells enabled by the extreme throughput.

The authors thank Nicole MacLennan, Karin Chen, and Edward R. B. McCabe for providing de-identified blood samples. We also thank Marc Lim, a University of California at Los Angeles bioengineering undergraduate, for the artistic rendering of a microfluidic-flow cytometer. This work is supported by the National Science Foundation (grant 0930501).