Dynamic range and uncertainty analysis of a serial microfluidic cytometer

M.A. Catterton, M. DiSalvo, E.W. Esch, P.N. Patrone, G.A. Cooksey
John Hopkins University,
United States

Keywords: microfluidics, flow cytometer


Flow cytometers are a robust and versatile tool for many biotech applications, which include immunophenotyping, cell cycle and viability analysis, and drug development. The instruments are capable of make high speed measurements of optical properties of individual cells (such as size and biomarker abundance) in flow. However, limitations make classification and comparison of measurements challenging, thereby reducing their capacity to inform decision making. For example, in conventional flow cytometers, cells are only measured once per laser wavelength, which leads to uncertainty of a fluorescent measurement being convolved with the distribution of the population being measured. NIST has recently developed a serial microcytometer that offers the unique capability to repeat flow cytometry measurements on individual particles, thereby enabling uncertainty quantification on a per-particle basis.[1] The serial microcytometer is a microfluidic device with integrated optical waveguides that uses hydrodynamic focusing to align particles in a single streamline to pass through two interrogation regions. Comparing measurements between regions requires very low velocity variation to facilitate particle tracking. This stability is achieved using a novel hydrodynamic focusing system to focus particles to an inertial equilibrium position in the channel (which, counter to conventional use, is not centered in the channel).[1] The performance of the microcytometer was evaluated using traditional metrics for commercial flow cytometers. First, we determine sensitivity (Q) and background (B)[2] using a 9 peak bead set in the fluorescein channel. The microcytometer Q was very similar to the commercial cytometer (2.54x106 and 2.53x106, respectively). Determination of B was inconclusive since the chosen beads did not meet the criteria for this approach. However, we measured a background value for the unlabeled bead, which was 1877 ± 499 molecules of equivalent fluorescein (MEFL) for the microcytometer and 551 ± 251 MEFL for the commercial cytometer (median ± standard deviation). The 9 intensity peaks could be separated in the fluorescein (488nm excitation) and DAPI (395 nm excitation) channels, but we could only clearly distinguish 8 peaks within the dynamic range of the APC channel (642 nm excitation) (Figure 1). For the fluorescein measurements, the intensity distribution of the peaks in the microcytometer was comparable to the conventional instrument. With the microcytometer, however, we can use the repeated measurements to calculate each particle’s uncertainty and show the dependence of uncertainty on the intensity of the bead (Figure 1). Overall, we demonstrate measurement uncertainty in flow cytometry that relates to the measured object and does not require comparison to a reference material, which in many conditions do not match the optical and mechanical properties of biological samples. Improvement in our understanding of measurement uncertainty will enable better classification and diagnostic decision making in research and clinical use.