Automated Analysis of Clinical Flow Cytometry Data: A Chronic Lymphocytic Leukemia (CLL) Pilot
Traditional manual gating analysis of cytometry data cannot effectively address the scale and complexity of data generation from modern cytometry instrumentation. Bioinformatics investigators have developed a collection of computational methods for automated identification of cell populations from high-dimensional flow cytometry data. A small subset of these methods has been evaluated for their use in diagnostics applications of leukemia and lymphoma with promising results. By applying computational pipelines to classify CLL samples from healthy controls, the pilot study reported in this paper illustrates the use of these methods to determine that traditional CLL definition based on CD5 and CD19 alone can be improved by also examining the expression levels of CD10 and CD79b in an automated fashion. Clinical validation of these computational approaches is ongoing and essential to realize the true potential of these methods for use in the clinical diagnostic laboratory.
This work is funded by the National Center for Advancing Translational Sciences (NCATS) under grant no. U01TR001801