International Journal of Laboratory Hematology

The Ofcial journal of the International Society for Laboratory Hematology

ORIGINAL ARTICLE

INTERNATIONAL JOURNAL OF LABORATO RY HEMATO LOGY

Optimization of laboratory workow in clinical hematology laboratory with reduced manual slide review: comparison between Sysmex XE-2100 and ABX Pentra DX120
M. HUR*, J.-H. CHO*, H. KIM*, M.-H. HONG*, H.-W. MOON*, Y.-M. YUN*, J. Q. KIM

*Department of Laboratory Medicine, Konkuk University School of Medicine, Seoul, Korea Konkuk University, Seoul, Korea Correspondence: Mina Hur, Department of Laboratory Medicine, Konkuk University School of Medicine, Konkuk University Hospital, 4-12, Hwayang-dong, Kwangjin-gu, Seoul 143-729, Korea. Tel.: +82 2 2030 5581; Fax: +82 2636 6764; E-mail: dearmina@hanmail.net This work was supported by Konkuk University in 2010.
doi:10.1111/j.1751-553X.2011.01306.x

SUMMARY

Received 24 September 2010; accepted for publication 21 December 2010 Keywords Slide, review, Sysmex XE-2100, ABX Pentra DX120, hematology

Introduction: The validation of automated hematology analyzer results by manual slide review (MSR) is currently an inevitable work process in clinical hematology laboratories. The laboratory workload would be optimized if the requirement for MSR could be reduced without compromising patient care. We investigated whether slide-making rates would be different between two hematology analyzers, which were paired with their own automated slide makers/stainers: Sysmex XE-2100 with SP-1000i (Sysmex, Kobe, Japan) and ABX Pentra DX120 with SPS evolution (ABX-Horiba, Montpellier, France). Methods: A total of 943 samples were run in parallel on the Sysmex XE-2100 and ABX Pentra DX120. Reex slides were automatically made in each analyzer according to its own criteria, which reected the criteria of MSR in our laboratory. The slide-making rates were compared, and the results were further conrmed using the criteria of MSR. Results: The slide-making rates in Sysmex XE-2100, ABX Pentra DX120, and manual review were 22.5% (212/943), 15.91% (150/ 943), and 11.5% (108/943), respectively. In 774 (82.1%) samples, the three methods showed concordant results, and all made slides in 82 samples. Using the manual method as a standard, the sensitivity and specicity were 86.1% and 85.8% in Sysmex XE-2100 and 89.8% and 93.7% in ABX Pentra DX120. Conclusion: Our data show that the slide-making rates are variable in different hematology analyzers. It also implies that although MSR cannot be fully substituted by modern hematology analyzers, it can be effectively reduced to optimize laboratory workload.

INTRODUCTION
The complete blood count (CBC) with leukocyte differential counts (LDC) is one of the most frequently
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

requested tests in clinical laboratories. Technical evolutions in automated hematology analyzers have improved the analytic performance greatly and have broadened the range of information provided (Buttarello
1

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY

& Plebani, 2008). Among the traditional CBC parameters, the red blood cell count (RBC), hemoglobin concentration (Hb), white blood cell count (WBC), and mean cell volume generally show excellent analytical function, while the results for reticulocytes, platelet counts, or certain components of LDC are less satisfactory (Buttarello, 2004; Segal et al., 2005). The clinical applications and analytical function of the newly introduced parameters, such as reticulocyte indices, immature reticulocyte fraction, reticulated platelets, or immature granulocytes, have not yet been fully established or standardized (Buttarello et al., 2002; Sandhaus & Meyer, 2002; Briggs et al., 2004; Thomas et al., 2005). Despite the high-quality performances and expanded capabilities of automated analyzers, the manual examination of blood smears still plays an important role in hematology laboratories. Each laboratory has its own decision-making criteria or rules for the validation of quantitative abnormalities or qualitative alterations that are highlighted (agged) and which trigger slide preparations and reviews as reex tests (Lantis et al., 2003; Barnes et al., 2005). The laboratory workload and efciency of generating the nal CBC report are widely affected by the number of manual slide review (MSR) performed and are variable between laboratories (Novis et al., 2006). Laboratory work processes would be optimized if the number of MSR could be reduced without compromising patient care. Automated slide makers and strainers, together with their hematology analyzers, are widely used in high-volume hematology laboratories (Simson, Gascon-Lema & Brown, 2009). In such situations, reex slides are automatically prepared after the generation of abnormal CBC data, with no manual intervention by experienced laboratory staff. Although each laboratory has criteria that trigger reex slides, the implementation of these criteria in the different hematology analyzers could vary according to the characteristics of each analyzer. Consequently, the rates of slide preparation and review might be highly affected. This study investigated whether the slidemaking rates were different between two hematology analyzers that were paired with their own automated slide makers/stainers: Sysmex XE-2100 with SP-1000i (Sysmex, Kobe, Japan) and ABX Pentra DX120 with SPS evolution (ABX-Horiba, Montpellier, France).

MATERIALS AND METHODS

A total of 943 patient samples were randomly selected from the daily routine workload of our hematology laboratory and were run in parallel on the Sysmex XE-2100 and ABX Pentra DX120. Reex slides were automatically prepared by each analyzer according to its own criteria, which were programmed to reect the criteria of MSR in our laboratory. To increase the clinical sensitivity, the criteria in each analyzer were developed not to miss any pathological samples. The slide-making rates were compared between the two analyzers, and the results were further conrmed with those of manual review. One expert senior technologist checked all the CBC data and reex slides that were automatically prepared from each analyzer and slide maker/stainer combination to choose the slides to be reviewed. Using the manual review as a standard, the sensitivity and specicity of each analyzer were obtained. The decision-making criteria for MSR in our laboratory are presented in Figure 1. The Sysmex XE-2100 is used as the main hematology analyzer in our laboratory, and the workow of slide review was developed according to the use of this analyzer and its automated slide makers/stainer, the SP-1000i (Figure 2). One of the criteria for slide review was to review all the samples from the hematology department regardless of CBC abnormalities. Accordingly, in addition to the main workow of slide preparation through SP-1000i, we needed another process to identify samples from the hematology department, for which slides were not prepared. For such samples, slide preparation and staining were performed manually. The rules for slide making in the Sysmex XE2100 are listed in Table 1, and the workow of the rules in ABX Pentra DX120 is summarized in Figure 3. The main differences in the rules between the Sysmex XE-2100 and ABX Pentra DX120 were that department information and delta check could be included in the criteria for the ABX Pentra DX120. To compare the diagnostic performances of Sysmex XE-2100 and ABX Pentra DX120, their sensitivities and specicities were analyzed using MedCalc Software (version 11.2.1; MedCalc Software, Mariakerke, Belgium). The statistical differences in their sensitivities and specicities were obtained with chi-square test.
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY 3

Figure 1. The decision-making criteria for manual slide review.

Figure 2. Workow of slide review with Sysmex XE-2100 and SP-1000i.

RESULTS
The comparison data of slide preparation by Sysmex XE-2100, ABX Pentra DX120, and manual review are presented in Table 2. The slide-making rates in Sysmex XE-2100, ABX Pentra DX120, and MSR were
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

22.5% (212/943), 15.91% (150/943), and 11.5% (108/943), respectively. All the three methods showed concordant results in 774 (82.1%) samples: all positive in 82 samples and all negative in 692 samples. Discrepant results were observed in 169 samples. No case showed a positive result by manual review but

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY

Table 1. The rules for slide making in Sysmex XE-2100 Quantitative abnormalities Leukocytopenia/leukocytosis: WBC < 2.0 109/l or WBC > 20 109/l Neutropenia/neutrophilia: neutrophils < 30% or neutrophils > 85.5% Lymphocytopenia/lymphocytosis: <10% or >70% Monocytosis: monocytes > 15.5% Eosinophilia: eosinophils > 20% Basophilia: basophils > 2% Nucleated RBCs > 2/100 WBCs Thrombocytopenia/thrombocytosis: PLT < 100 109/l or PLT > 600 109/l Flags PLT clumps or abnormal distribution Fragments Blasts Immature granulocytes Left shift Atypical lymphocytes Abnormal lymphocytes/lymphoblasts WBC, white blood cell count; RBC, red blood cell count; PLT, platelet count.

Regarding false-negative results, the causes were department (n = 11) and absence of result (n = 4) in Sysmex XE-2100, and absence of ag (n = 11) in ABX Pentra DX120. The quantitative values or ags in 11 samples with false-negative results by ABX Pentra DX120 are presented in Table 5. The Sysmex XE-2100 also showed ags and/or quantitative abnormalities that triggered MSR in these 11 samples. However, these did not correspond to our criteria for the manual conrmation, except for one sample with 3% basophils (sample 8).

DISCUSSION
The manual examination of blood smears is time-consuming and expensive and may not be always necessary. To increase the clinical sensitivity, most laboratories tend to develop less strict criteria so as not to miss potentially important abnormalities. This would be more conspicuous especially when a hospital has a large pool of hemato-oncological patients. According to the College of American Pathologists Q-Probes Study with 263 participating institutions, the rates of MSR varied considerably among participants (26.7% in the median, 9.9% in the 10th percentile, and 50.0% in the 90th percentile institutions) and were elevated with increased numbers of hospital beds (Novis et al., 2006). That study showed that the rates of MSR were directly related to the efciency in generating CBC results. Most of the MSR were triggered by hematology analyzer ags, and these threshold limits also varied widely among participants. Recently, the use of automated slide makers and stainers has increased in large-sized clinical laboratories. They are used in combination with their multiparameter hematology analyzers, and their performances are reported to be comparable to well-prepared manual processes (Simson, Gascon-Lema & Brown, 2009). Compared with manual procedures, the introduction of automated slide makers and stainers has signicantly reduced the workload of slide preparation as well as the turn-around-time of the nal CBC report. On the other hand, not all the automatically prepared slides are reviewed manually. If slide-making rules of the instrument do not perfectly match the slide-review criteria of the laboratory, there may be a discrepancy between the prepared slides and reviewed slides. Considering the general policy of laboratories not to miss
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

negative results by both Sysmex XE-2100 and ABX Pentra DX120. However, there were discrepant results between the Sysmex XE-2100 and ABX Pentra DX120 in 26 samples with positive MSR results: 15 were Sysmex negative and Pentra positive and the other 11 were Sysmex positive and Pentra negative. Using MSR as a reference method, the sensitivity and specicity were 86.1% (95% condence interval (CI), 78.192.0%) and 85.8% (95% CI, 83.288.0%) in Sysmex XE-2100 and 89.8% (95% CI, 82.594.8%) and 93.7% (95% CI, 91.895.2%) in ABX Pentra DX120, respectively (Table 3). Signicant differences in the specicity and specicity of the analyzers were observed, with a difference of 3.7% (95% CI, 0.69 6.71; P = 0.0165) for sensitivity and 8.0% (95% CI, 5.210.8; P < 0.0001) for specicity (chi-square test). The causes of false-positive and false-negative results were analyzed (Table 4). In the 119 samples with false-positive results by Sysmex XE-2100, slide ag was the most frequent cause followed by delta check and department; slide ag and/or delta check comprised three-quarters (75.7%) of the total falsepositive causes. In ABX Pentra DX120, the two causes of false-positive results were slide ag (64.2%) and monocytosis with large immature cells (35.8%).

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY 5

Figure 3. Workow of the rules for slide making in Pentra DX120. PDX, ABX Pentra DX120; SPS, SPS evolution; PML, Pentra multilink data management system.

Table 2. Slide preparation by Sysmex XE-2100, ABX Pentra DX120, and manual review Sysmex XE-2100 Positive Positive Negative Negative Negative Positive Negative Positive ABX Pentra DX120 Positive Negative Positive Negative Negative Negative Positive Positive Manual review Positive Positive Positive Positive Negative Negative Negative Negative Number (%) 82 11 15 0 692 90 24 29 (8.7) (1.2) (1.6) (0) (73.4) (9.5) (2.5) (3.1)

Table 3. Comparison of the results between Sysmex XE-2100, ABX Pentra DX120, and manual review Sysmex XE-2100 Positive Manual review Positive 93 (n = 108) Negative 119 (n = 835) Negative ABX Pentra DX120 Positive Negative

15 716

97 53

11 782

Positive means the slide preparation, and negative vice versa.

any possible pathologic samples, the presence of these unnecessarily prepared slides might be regarded as inevitable. However, construction of the rules to reect the laboratories own criteria may be variable in
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

each hematology analyzer, and consequently, the gap between the prepared and reviewed slides may be different. To the best of our knowledge, no study has focused on this practical issue in routine hematology laboratory so far. This study investigated whether the slide-making rates would be different between two hematology analyzers when paired with their own automated slide

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY

Table 4. Causes of false-positive and false-negative results by Sysmex XE-2100 and ABX Pentra DX120 Sysmex XE-2100 False-positive Total (n = 119, 100%) Slide ag (n = 46, 38.7%) Delta check (n = 34, 28.6%) Department (n = 21, 17.6%) Slide ag and delta check (n = 10, 8.4%) Department and delta check (n = 8, 6.7%) Total (n = 15, 100%) Department (n = 11, 73.3%) No result (n = 4, 26.7%) ABX Pentra DX120 Total (n = 53, 100%) Slide ag (n = 34, 64.2%) Mono + Lic (n = 19, 35.8%)

False-negative

Total (n = 11, 100%) No ag (n = 11, 100%)

Mono + Lic, monocytosis and large immature cells.

Table 5. Quantitative values or ags in samples with false-negative results by ABX Pentra DX120 Quantitative values (%) or ags Dep 1 2 3 4 5 6 7 IM IM IM IM IM IM IM ABX Pentra DX120 Lic: 1.0 E: 16.9 No ags Lic: 1.0 Lic: 2.2 M:13.6, Lic: 0.4, Aty L: 1.5 nRBC, platelet aggregates, monocytosis, (M: 11.4, Lic: 0.6, Aty L: 1.4) Lic: 0.8, Aty L: 0.9 Neutropenia (M: 10.2, Lic: 0.8, Aty L: 2.2) Lic: 2.0, Aty L: 1.7 M: 12.8, Lic: 0.9, Aty L: 1.4 Sysmex XE-2100 Abn L/L-Blasts Eosinophilia, Blasts Abn L/L-Blasts Aty L Immature Granulocytes Abn L/L-Blasts Basophilia Manual differential count (%) N: 67, Nb: 2, L: 25, M: 6 N: 65, L: 8, M: 8, E: 19 N:73, L:16, M:6, E:3, B:2 N: 73, Nb: 2, L: 17, M: 4, E: 1, B: 2, Aty L: 1 N: 78, Nb: 2, L: 12, M: 4, E: 2, B: 1, Mm: 1 N: 43, Nb: 1, L: 39, M: 13, E: 3, B: 1 N: 65, Nb: 1, L: 22, M: 7, E: 4, B: 1

8 9 10 11

IM PED IM IM

Blasts Neutropenia, Abn L/L-Blasts Aty L Aty L

N: 46, Nb: 2, L: 43, M: 6, B: 3 N: 15, Nb: 4, L: 64, M: 11, E: 3, B: 2, Aty L: 1 N: 79, Nb: 4, L: 10, M: 5, B: 1, Aty L: 1 N: 61, L: 25, M: 9, E: 2, Aty L: 3

Dep, department; IM, internal medicine; PED, pediatrics; Lic, large immature cells; Aty L, atypical lymphocytes; Abn L, abnormal lymphocytes; nRBC, nucleated RBCs; L-Blasts, lymphoblasts; N, neutrophils; Nb, band-form neutrophils; L, lymphocytes; M, monocytes; E, eosinophils; B, basophils; Mm, metamyelocytes; My, myelocytes.

makers/stainers: Sysmex XE-2100 with SP-1000i and ABX Pentra DX120 with SPS evolution. Our data showed that the sensitivity and specicity of slide preparation were different between these two systems and that the performance of ABX Pentra DX120 with SPS evolution was superior to that of Sysmex XE-2100 with SP-1000i (Table 3). In particular, the number of false-positive samples was decreased in ABX Pentra DX120, and the main causes of difference were attributable to delta check and department, which could not be included in the rules for Sysmex

XE-2100 (Table 4). Department was also the main cause of false-negative results in Sysmex XE-2100, and in the routine practice of our laboratory, this was one of the main causes of extra workload. Eleven samples showed false-negative results by ABX Pentra DX120 with no ags. In contrast, they mostly showed ags of blasts, atypical lymphocytes, or immature granulocytes and/or quantitative abnormalities by Sysmex XE-2100 (Table 5). According to our decision-making criteria for slide review, MSR was triggered by the results of Sysmex XE-2100, and the
2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.

M. HUR ET AL.

REDUCED SLIDE REVIEW IN HEMATOLOGY LABORATORY 7

results by ABX Pentra DX120 were considered falsenegative. The manual counts on these 11 samples, however, showed that the review process was not necessary for most of the samples, except for one sample with 3% basophils. Accordingly, the actual false-negative rate of ABX Pentra DX120 would be lower and the specicity would be higher than the obtained data. Our study has several limitations. It was conducted to compare the slide-making rates between the two analyzers, but the rules of ABX Pentra DX120 were constructed in accordance with our decision-making criteria, in which the pre-existent Sysmex XE-2100 played a main part. Accordingly, MSR, a reference method, might have itself affected the comparison data possibly toward increased false-negative results by ABX Pentra DX120. Another limitation is that the manual review rate in this study (108/943, 11.45%) was relatively lower than the usual review rate in our laboratory. The average manual review rate was 26.9% in our laboratory for the period August 2009 July 2010, ranging from 20.0% to 39.0%. It can be explained by the fact that, for this study, we avoided the busiest day of the week and the busiest time of the day, although we tried to conduct it as a routine practice. This bias might have decreased the difference in slide-making rates and affected the causes of discrepant results.

Manual slide review to validate the results from automated hematology analyzers is currently an inevitable work process in clinical hematology laboratories, and the efciency of generating CBC results may be affected by how to set both instrument threshold triggers and laboratory policies (Novis et al., 2006). Despite suggested criteria, the criteria for MSR seem to be still variable across laboratories (Barnes et al., 2005). The use of automated slide makers and stainers has decreased manual workload and increased laboratory efciency. Considering that automated slide preparation triggers MSR, if the slide-making rates are variable among different hematology analyzers, it would directly affect the slide-review rate in the laboratory. Such an impact would be greater in high-volume laboratories, especially if there is a large pool of hematooncological patients. In summary, our data showed that, even with the same laboratory criteria for slide review, the slide-making rules could not be constructed equally in different analyzers and the slidemaking rates therefore varied. This study implies that in addition to the automation of slide preparation, there is a room for further optimization of laboratory workload and efciency by adjusting the slide-making rate to each laboratorys situation. More comprehensive effort is required to streamline the workow and improve the productivity in hematology laboratories.

REFERENCES
Barnes P.W., McFadden S.L., Machin S.J. & Simson E. (2005) The international consensus group for hematology review: suggested criteria for action following automated CBC and WBC differential analysis. Laboratory Hematology 11, 8390. Briggs C., Kunka S., Hart D., Oguni S. & Machin S.J. (2004) Assessment of an immature platelet fraction (IPF) in peripheral thrombocytopenia. British Journal of Haematology 126, 9399. Buttarello M. (2004) Quality specication in haematology: the automated blood cell count. Clinica Chimica Acta 346, 4554. Buttarello M. & Plebani M. (2008) Automated blood cell counts. State of the art. American Journal of Clinical Pathology 130, 104116. Buttarello M., Bulian P., Farina G., Petris M.G., Temporin V. & Toffolo L. (2002) Five fully

automated methods for performing immature reticulocyte fraction: comparison in diagnosis of bone marrow aplasia. American Journal of Clinical Pathology 117, 871879. Lantis K.L., Harris R.J., Davis G., Renner N. & Finn W.G. (2003) Elimination of instrumentdriven reex manual differential leukocyte counts. Optimization of manual blood smear review criteria in a high-volume automated hematology laboratory. American Journal of Clinical Pathology 119, 656662. Novis D.A., Walsh M., Wilkinson D., St. Louis M. & Ben-Ezra J. (2006) Laboratory productivity and the rate of manual peripheral blood smear review. A College of American Pathologists Q-Probes Study of 95141 complete blood count determinations performed in 263 institutions. Archives of Pathology and Laboratory Medicine 130, 596601. Sandhaus L.M. & Meyer P. (2002) How useful are CBC and reticulocyte reports to clini-

cians? American Journal of Clinical Pathology 118, 787793. Segal H.C., Briggs C., Kunka S., Casbard A., Harrison P., Machin S.J. & Murphy M.F. (2005) Accuracy of platelet counting haematology analysers in severe thrombocytopenia and potential impact on platelet transfusion. British Journal of Haematology 128, 520 525. Simson E., Gascon-Lema M.G. & Brown D.L. (2009) Performance of automated slidemakers and stainers in a working laboratory environment routine operation and quality control. International Journal of Laboratory Hematology 32, e64e76. Thomas L., Franck S., Messinger M., Linssen J., M. & Thomas C. (2005) Reticulocyte Thome hemoglobin measurement: comparison of two methods in the diagnosis of iron restricted erythropoiesis. Clinical Chemistry and Laboratory Medicine 43, 11931202.

2011 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem.