Integrated machine learning approaches for flow cytometric quantification of myeloid-derived suppressor cells in acute sepsis.

Highly heterogeneous cell populations require multiple flow cytometric markers for appropriate phenotypic characterization. This exponentially increases the complexity of 2D scatter plot analyses and exacerbates human errors due to variations in manual gating of flow data. We describe a semi-automated workflow, based entirely on the Flowjo Graphical User Interface (GUI), that involves the stepwise integration of several, newly available machine learning tools for the analysis of myeloid-derived suppressor cells (MDSCs) in septic and non-septic critical illness. Supervised clustering of flow cytometric data showed correlation with, but significantly different numbers of, MDSCs as compared with the cell numbers obtained by manual gating. Neither quantification method predicted 30-day clinical outcomes in a cohort of 16 critically ill and septic patients and 5 critically ill and non-septic patients. Machine learning identified a significant decrease in the proportion of PMN-MDSC in critically ill and septic patients as compared with healthy controls. There was no difference between the proportion of these MDSCs in septic and non-septic critical illness.