The Usefulness of Blast Flags on the Sysmex XE-5000 Is Questionable

Faculty of Health Sciences, Oslo and Akershus University College of Applied Sciences, PO Box 4 St. Olavs plass, N-0130 Oslo, Norway
American Journal of Clinical Pathology (Impact Factor: 2.51). 05/2013; 139(5):633-40. DOI: 10.1309/AJCPDUZVRN5VY9WZ
Source: PubMed


Hematology analyzers generate suspect flags that involve microscopic reviews to confirm the presence of pathologic cells. This study investigated the reliability of the blast flag in a side-by-side evaluation of 3 Sysmex XE-5000 instruments (Sysmex, Kobe, Japan). The repeatability of the Q values reported by each instrument for 10 replicates of the same blood samples was low (intraclass correlation coefficient [ICC] values, 0.62-0.74). The reproducibility of the Q values obtained by analyzing 408 samples on all 3 instruments was reasonable (ICC value, 0.85). In addition, a systematic difference was observed among the instruments in the level of reported Q values. With cutoff commonly being 100, the observed reproducibility of the blast flagging among the instruments was evaluated as poor (κ = 0.73). Based on the observed low performances, we question the usefulness of the Q value as a predictor of blasts and whether a blast flag reported by the XE-5000 is sufficient as a criterion for performing a microscopic review.

57 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Criteria for peripheral smear review are designed to include those samples with results outside the reference interval and can be more extreme based on what is considered to have clinical utility. However, we are unaware of previous studies that reported the distributions of various complete blood cell count (CBC) parameters in infants. In the following study we reviewed screening CBC results of 692 infants aged 9-15 months in order to determine the proportion of peripheral smear reviews recommended according to consensus criteria and that after adjusting for the observed distributions of the various parameters. According to consensus criteria the recommended reflex peripheral smear review rate was 39.7% (95% CI 36.1-43.4) whereas after adjustment for the observed distributions, the rate fell to 5.6% (95% CI 3.9-7.3) (p < 0.001). The major reasons for the difference in rates were the high proportion of infants with an absolute lymphocyte count > 7 × 10(9)/L (17.5%), the presence of a plus one blast flag (4.3%), and a large unstained cell count of ≥ 5% (26.2%) (equivalent to + 1 atypical flag). We found that international consensus criteria for reflex peripheral smear review results in a very high peripheral smear review rate in well infants, and might be inappropriate.
    Scandinavian journal of clinical and laboratory investigation 03/2014; 74(4). DOI:10.3109/00365513.2014.893011 · 1.90 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The Sysmex XN (XN) modular system (Sysmex, Kobe, Japan) is a new automated hematology analyzer equipped with different principles from its previous version, Sysmex XE-2100. We compared the performances of Sysmex XN and XE-2100 in umbilical cord blood (CB) specimens. In 160 CB specimens, complete blood count (CBC) parameters and white blood cells (WBC) differentials were compared between the two analyzers. Their flagging performances for blasts, abnormal/atypical lymphocytes, immature granulocytes and/or left-shift (IG), and nucleated red blood cells (NRBC) counts were compared with manual counts. For the blast flagging, Q values by Sysmex XN were further compared with manual slide review. Sysmex XN and XE-2100 showed high or very high correlations for most CBC parameters but variable correlations for WBC differentials. Compared with XE-2100, XN showed significantly different flagging performances for blasts, abnormal/atypical lymphocytes, and IG. The flagging efficiency for blasts was significantly better on Sysmex XN than on XE-2100 (85.0% vs. 38.8%): Sysmex XN showed a remarkably increased specificity of blast flag, compromising its sensitivity of blast flag. Among the 24 specimens with blasts (range, 0.5%–1.5%), only one (4.2%) showed a positive Q value. This study highlighted the remarkable differences of flagging performances between Sysmex XN and XE-2100 in CB specimens. The Sysmex XN modular system seems to be a suitable and practical option for the CB specimens used for hematopoietic stem cell transplantation as well as for the specimens from neonates.
    Clinical Chemistry and Laboratory Medicine 06/2014; 52(12). DOI:10.1515/cclm-2014-0392 · 2.71 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The rate of auto-validation is dependent on the ability of the laboratory information system (LIS) to integrate historical data, on the frequency and methods for identifying analyzer errors, and on the criteria for reflex testing, including the need for peripheral smear review. The rate of auto-validation in outpatient laboratories, however, is unclear. We examined 45,925 consecutive complete blood count (CBC) test results (1 January, 2014–31 January, 2014) from patients aged 50±24 years. The LIS auto-validates all samples according to set criteria. Technicians validated test results when previous CBC test results were required to determine: 1) the need for peripheral slide review and/or sample rerun or 2) the need for reflex testing to detect autoimmune hemolytic anemia or β-thalassemia minor. The auto-validation rates were 97.6% after rejecting results requiring validation to determine the need for a peripheral smear review and/or sample rerun. This decreased to 92.9% after including reflex testing to determine the reasons for normocytic and microcytic anemia. We estimated that auto-validation decreased the workload by 7.7–11.6 h per 3000 test results. We conclude that very high auto-validation rates are possible in outpatient general laboratories, leading to conformity in the validation process and a considerable estimated savings in technician time. Further studies are needed in other settings.
    Clinical Chemistry and Laboratory Medicine 08/2014; 53(2). DOI:10.1515/cclm-2014-0572 · 2.71 Impact Factor
Show more