Multicenter comparison of CD34+ myeloid cell count by flow cytometry in low‐risk myelodysplastic syndrome. Is it feasible?

• Published in: Cytometry. Part B, Clinical cytometry Vol. 94; no. 3; pp. 527 - 535
• Main Authors: Font, Patricia, Subirá, Dolores, Matarraz, Sergio, Benavente, Celina, Cedena, María Teresa, Morado, Marta, Pérez Corral, Ana, Bellón, José María, Díez‐Martín, José Luis
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01-05-2018
• Subjects: absolute cell count; Bone marrow; CD34 antigen; CD34+ cells; CD45 antigen; Correlation coefficient; Correlation coefficients; Enumeration; Flow cytometry; Gating; Histocompatibility antigen HLA; Laboratories; low‐risk MDS; Mathematical analysis; Myelodysplastic syndrome; myelodysplastic syndromes; Myeloid cells; Parameter identification; Patients; Reproducibility; Risk; Standardization; flow cytometry; absolute cell count; myelodysplastic syndromes; CD34+ cells; low-risk MDS
• ISSN: 1552-4949; 1552-4957
• DOI: 10.1002/cyto.b.21538

Background
Accuracy of bone marrow (BM) blast count in low‐risk myelodysplastic syndromes (MDS) still remains a challenge though it is essential for prognosis. We investigated whether the enumeration of CD34+ myeloid cells by flow cytometry immunophenotyping (FCI) could be used as a consistent parameter for clinical MDS studies.
Methods
Six clinical centers entered the study and information on their FCI protocols was recorded. Sixty‐seven flow cytometry listmodes from BM samples of patients with low‐risk MDS with <5% BM blasts were exchanged among participants in two different rounds. Interlaboratory variations on the quantification of CD34+ myeloid cells were calculated and strategies to solve differences were evaluated.
Results
An overall “very good” agreement on CD34+ cell count among participants (intraclass correlation coefficient = 0.720) was observed, but agreement was “low” in 22 files. No single parameter could fully explain all discrepancies, but 3 technical issues were identified as relevant: the use of the CD34/CD45/CD117/HLA‐DR mAb combination, acquisition of ≥50,000 events and a low percentage of debris/aggregates. The frequency of discordant results increased with the accumulation of pitfalls (none, 16%; 1 pitfall, 40%; 2 pitfalls, 83%; P = 0.006). Finally, the use of a common gating strategy for analysis increased the percentage of files with “very good” agreement to 100%.
Conclusions
Prevention of specific technical pitfalls is mandatory to reach a good reproducibility of CD34+ cell count among centers. These recommendations set the basis for laboratory standardization and enable the use of CD34+ cell enumeration as additional information in low‐risk MDS patients. © 2017 International Clinical Cytometry Society