Quantifying microcyte and macrocyte percentages in archived red blood cell volume histogram images.

Abstract

We sought to quantify percentages of microcytes and macrocytes in archived automated hematology analyzer (AHA) red blood cell (RBC) volume histogram images. In preliminary studies, we demonstrated that an on-line application of Gauss' area formula (SketchAndCalc™) measured percentage areas of 20 segments under a computer-generated normal distribution curve (-3.0 standard deviations [SD] to +3.0 SD) with accuracy and precision (Pearson's correlation of measured areas with corresponding theoretical areas r [20] = 0.9962 [p < 0.0001]). Thus, we used SketchAndCalc™ to quantify percentages of microcytes (50-80 fL) and macrocytes (110-200 fL) in archived AHA histogram images in women with previously untreated iron-deficiency anemia (IDA) and previously untreated hemochromatosis. Median microcyte percentages in 13 women with IDA and 13 women with hemochromatosis were 63.6% (range 13.5-76.8) and 6.7% (range 3.4-24.8), respectively (p < 0.0001). Mean macrocyte percentages in women with IDA and hemochromatosis were 8.8% ± 6.1 SD and 33.8% ± 11.7 SD, respectively (p < 0.0001). Spearman's correlations of microcyte percentages with macrocyte percentages, mean corpuscular volume, and mean corpuscular hemoglobin in 26 women were rs [26] = -0.9485, rs [26] = -0.9641, and rs [26] = -0.9036, respectively (each p < 0.0001). This method of quantifying microcyte and macrocyte percentages could enable other studies of RBC volume subpopulations in archived AHA histogram images.