[1874] Flow Cytometry Underestimates the Bone Marrow CD34-Positive Blast Population Compared to Immunohistochemical and Morphologic Assessments

EA Morgan, KY Oh, CG Waghorne, D Neuberg, K Stevenson, JL Kutok. Brigham and Women's Hospital, Boston, MA; Dana-Farber Cancer Institute, Boston, MA

Background: Accurate quantification of bone marrow (BM) blast percentages is critical for the diagnosis and classification of hematological malignancies. We evaluated the validity of using flow cytometry (FC) or immunohistochemistry (IHC) as a method of quantifying CD34+ blasts compared to BM aspirate morphologic (morph) blast counts.
Design: We identified 112 samples from 90 patients having concurrent BM aspirate counts, BM biopsies with CD34 IHC, and FC analysis with CD34+ blasts performed at our institution between 1/1/2006 and 1/1/2007. IHC estimates of CD34+ blasts were made by dividing positively staining blasts by the total nucleated marrow cells. FC detection and quantification of the blast percentages was performed by defining a blast gate using low to intermediate forward and low side scatter, and dividing the CD34+ cells in that gate by the total number of cells analyzed. Cases of subset expression of CD34 by the blast population were excluded from this study. Paired comparison of the difference (Δ) in blast % between the three methods was assessed by the Wilcoxon signed-rank test and correlation analysis was assessed by Spearman's rank correlation test. A locally weighted scatter plot smoother fit was used to investigate trends between these methods.
Results: While there was a positive correlation between all three counting methods (correlation coefficients ranging from 0.75-0.79; p<0.001) (Fig 1A-C), further analysis revealed FC consistently showed lower values compared to IHC and morphology counts (median Δ=FC-Morph=-13.0, range (-68, 56); median Δ=FC-IHC=-17.0, range (-85, 45); both p<0.001) (Fig 1D).
Conclusions: Simultaneous comparison of 3 methods for blast determination (morphology, IHC, FC) in a large series revealed that FC underestimates the BM blast population compared to the other methodologies. Given the importance of blast determination in the diagnosis of hematological malignancies, use of the blast count as determined by FC may lead to the misclassification of these entities and, therefore, alter clinical decision making.

Category: Quality Assurance

Monday, March 22, 2010 1:00 PM

Poster Session II # 216, Monday Afternoon


Close Window