Background: Hematology analyzer flagging system proves to be a significant tool in screening white blood cell (WBC) morphology abnormalities. This study was conducted to determine the accuracy of WBC morphology flagging system of Beckman Coulter LH500 using slide review as the reference standard.
Method: Four hundred thirteen (413) non-consecutive CBC samples for a period of 15 months were analyzed using Beckman Coulter LH500 for WBC differentiation. The flagging system was compared to slide review using 200 cell counts for each case. The suspect flags used were for blasts, immature granulocytes/band I, immature granulocytes band 2 and variant lymphocytes. The statistics included sensitivity, specificity, positive predictive value, negative predictive value for each of the suspect flags, including overall efficiency.
Results: Flagging system of LH500 for detecting the presence or absence of WBC abnormalities generated a sensitivity of 92.2%, specificity of only 35.7%, PPV of 58.8%, and NPV of 82.2%. The Kappa index of 0.279±0.093 conferred some level of disagreement between automated and the manual WBC differential count. The true negative rate was 18% for abnormalities and the false negative rate was 3.9%. The flagging system has an overall efficiency of 92.3% for variant lymph, 87.7% for blasts, 85.7% for immature granulocytes / bands 2 and 78.7% for immature granulocytes / bands 2. The statistics for each suspect flags were detailed in the study.
Conclusion: The ability of LH500 WBC flagging system showed an acceptable level of accuracy compared that of the manual method of differential count. The overall efficiency is high for variant lymphs, blasts, immature granulocytes / bands 2 and immature granulocytes / bands 1 in decreasing order. The true negative rate of 18% may help in the decision to use the suspect flags in performing the manual slide review. The false negative rate of 3.9% requires a review of these cases.