Background: Accurate bone marrow (BM) blast counts (BCs) are essential for diagnosis (dx) of myelodysplasia (MDS), MDS/myeloproliferative (MDS/MPD) disease, or acute myeloid leukemia (AML), and may be difficult in hemodiluted bone marrow aspirates (BMAs). Erythroid precursors (EPs) may be indistinguishable from myeloblasts in BM sections (aspirate clots/cores). We compare the usefulness of immunohistochemistry (IHC) [ie, CD34, CD117, myeloperoxidase (MPO), Hemoglobin A1 (HbA1), and terminal deoxynucleotidyl transferase (TdT)] of BM sections (IHC-BM) with BMA, bone marrow touch preparation (BMTP), and flow cytometry (FC) BCs.
Design: The initial BC (48), percentage (%) of Eps (38) (both based on initial 100 to 600-cell counts), and FC expressions of CD34, CD117, Glycophorin A(GLY A), and TdT (44) were tabulated from 50 BMs (MDS, MDS/MPD, or AML). BMAs (48) and BMTPs (25) subsequently received 500-cell counts. IHC-BM was performed (45:formalin, 5:B5-fixed) [CD34 (46), CD117 (45), HbA1 (45), TdT (42), and MPO (45)].
Results: Retrospective BMA BCs revealed a 31% (15/48) discrepant rate between the original/retrospective BMA BCs; 80% revealed an underestimated initial BC. There was a 28% discordance rate between the retrospective BMA and BMTP reviews; 77% showed a higher BMTP BC. IHC showed significantly higher BCs in 19% (9/47), resulting in a different dx (5). However, CD34 and CD117 IHCS revealed lower BCs in 38% and 48%, respectively. The CD34 IHC results were primarily due to CD34-negative blasts by FC. The CD117 IHC results were largely unexplained. EPs were CD34 and CD117-negative.
Conclusions: (1) Evaluation for MDS/AML requires 500-cell counts of BMAs and/or BMTPs. (2) CD34+ and/or CD117+ blasts by FC indicate IHC-BM may increase BC accuracy. (3) CD34 is more reliable than CD117 by IHC; however, in combination, they are most reliable and should be performed on BM clots/cores due to variable reactivity.