OBJECTIVE:The present study evaluates the efficiency of software-generated white blood cell (WBC) "suspect flags" of the hematology analyzers Sysmex SE-9000, Sysmex NE-8000, and Coulter STKS.
DESIGN:Automated WBC differential counts were considered positive if they contained any suspect WBC flag indicating the presence of blasts, myeloid precursor cells, or abnormal lymphocytes. Reference differential counts were performed by microscopic examination of 400 WBCs per sample. After comparison to the reference method, automated differential counts were classified as true-positive, true-negative, false-positive, and false-negative. The flagging efficiency of analyzers was expressed as a percentage of subjects correctly classified.
SPECIMENS:Four hundred sixty-seven blood samples were randomly chosen for comparison analysis from various inpatient and outpatient departments of the Vienna university hospital, Austria.
RESULTS:The efficiency rates of flagging for the presence of > or = 1% abnormal WBCs were 78% (SE-9000), 77% (NE-8000), and 72% (Coulter STKS). The flagging efficiencies were best for samples with normal WBC counts. With regard to the specific suspect flags, the flagging of blast cells was most efficient on all analyzers.
CONCLUSIONS:Our results demonstrate the comparable overall performance of three analyzers, SE-9000, NE-8000, and Coulter STKS. They further underscore the importance of critical interpretation of automated differential counts, because at a detection limit of > or = 1% abnormal WBCs > 20% of samples were not correctly flagged by either analyzer.
DESIGN:Automated WBC differential counts were considered positive if they contained any suspect WBC flag indicating the presence of blasts, myeloid precursor cells, or abnormal lymphocytes. Reference differential counts were performed by microscopic examination of 400 WBCs per sample. After comparison to the reference method, automated differential counts were classified as true-positive, true-negative, false-positive, and false-negative. The flagging efficiency of analyzers was expressed as a percentage of subjects correctly classified.
SPECIMENS:Four hundred sixty-seven blood samples were randomly chosen for comparison analysis from various inpatient and outpatient departments of the Vienna university hospital, Austria.
RESULTS:The efficiency rates of flagging for the presence of > or = 1% abnormal WBCs were 78% (SE-9000), 77% (NE-8000), and 72% (Coulter STKS). The flagging efficiencies were best for samples with normal WBC counts. With regard to the specific suspect flags, the flagging of blast cells was most efficient on all analyzers.
CONCLUSIONS:Our results demonstrate the comparable overall performance of three analyzers, SE-9000, NE-8000, and Coulter STKS. They further underscore the importance of critical interpretation of automated differential counts, because at a detection limit of > or = 1% abnormal WBCs > 20% of samples were not correctly flagged by either analyzer.
