Background: The enormous potential of complex data files generated by 10-color flow cytometry (FC) is hindered by the requirement for exhaustive manual gating and the complexity of multidimensional data visualization. We propose a model using radar plots (RPs), to improve FC data visualization by capturing multidimensionality and integration of FC findings.
Method: We analysed 12 normal/reactive bone marrow (N/R BM) samples and 12 BM samples from patients with myelodysplasia (MDS) with 10-color FC. All identifiable cell clusters were individually marked, grouped, and visualized on radar plots. RPs were optimized to de-clutter the cell clusters and map BM cell composition and maturation.
Results: A total of 27 immature and mature cell clusters were identified and visualized on 8 multidimensional radar plots. The RPs displayed flow cytometry findings of normal BM in an integrated fashion to maximize overall insight into the data set. The constructed map of bone marrow cell composition was reproducible in all normal BM samples analyzed. Analysis of the pilot cohort of patient samples confirmed the presence of MDS-related changes. These changes are readily identifiable on RPs.
Conclusion: We demonstrated that the cell clusters of normal BM can be mapped on multidimensional radar plots, which provide an inclusive insight into BM cell composition and maturation. These reproducible RPs present a comprehensive and comprehensible visual display of differentiation and maturation of haematopoietic cells in normal BM, and can be used as a reference map to assess abnormal haematopoiesis in MDS. © 2017 International Clinical Cytometry Society.
Keywords: MDS; bone marrow; data visualization; flow cytometry; multidimensional; radar plot.
© 2017 International Clinical Cytometry Society.