|Year : 2020 | Volume
| Issue : 6 | Page : 667-671
Assessment and reliability of suspect flags in automated hematology analyzers for diagnosing white blood cell and platelet disorders
Namrata Patro, Archana C Buch, Mukta D Naik, Shruti Vimal, Shirish S Chandanwale
Department of Pathology, Dr. D. Y. Patil Medical College Hospital and Research Centre, Dr. D. Y. Patil Vidyapeeth, Pune, Maharashtra, India
|Date of Submission||14-Feb-2020|
|Date of Decision||11-Apr-2020|
|Date of Acceptance||24-Jun-2020|
|Date of Web Publication||6-Nov-2020|
Archana C Buch
B-603, Gold Coast, Ivory Estates, Someshwarwadi, Pune - 411 008, Maharashtra
Source of Support: None, Conflict of Interest: None
Background: The automated hematology analyzers have replaced the traditional manual methods for assessing hematological parameters. Most of the automated hematology analyzers are programmed to identify abnormalities in the form of “suspect flags.” Aims: The aim of the study was to correlate white blood cell (WBC) and platelet (PLT) flag messages provided by automated analyzers with their respective peripheral blood smear (PBS) findings. Materials and Methods: A descriptive cross-sectional study was undertaken on 100 patients who showed WBC and PLT suspect flags and/or abnormal peripheral smears. The flag messages were analyzed with their respective PBS findings. Pearson's Chi-square test was used for the statistical analysis. Results: Analyzer and PBS showed WBC defects (n = 46, 32), PLT defects (n = 39, 54), and combined defects (n = 36, 43) respectively. WBC defects included leukocytosis (n = 27, 22), leukopenia (n = 20, 18), immature granulocytes (n = 10, 3), left shift (n = 3, 2), myeloblast blast (n = 2, 1), and eosinophilia (n = 4, 4) on analyzer and PBS. Correlation between the WBC suspect flags and their peripheral smears was found to be statistically significant (P = 0.006). PLT defects included thrombocytopenia (n = 22, 30), thrombocytosis (n = 13, 17), giant PLTs (n = 10, 15), PLT clumps (n = 8, 12), and PLT debris (n = 2, 5) on analyzer and PBS. Correlation between the PLT characteristics in analyzers and in peripheral smears was statistically significant (P = 0.042). Conclusion: Suspect flags should be used as a screening tool to pick up pathological samples. These should then be followed up with a PBS examination for definitive diagnosis.
Keywords: Automated analyzer, flagging, peripheral blood smear, platelet, white blood cell
|How to cite this article:|
Patro N, Buch AC, Naik MD, Vimal S, Chandanwale SS. Assessment and reliability of suspect flags in automated hematology analyzers for diagnosing white blood cell and platelet disorders. Med J DY Patil Vidyapeeth 2020;13:667-71
|How to cite this URL:|
Patro N, Buch AC, Naik MD, Vimal S, Chandanwale SS. Assessment and reliability of suspect flags in automated hematology analyzers for diagnosing white blood cell and platelet disorders. Med J DY Patil Vidyapeeth [serial online] 2020 [cited 2020 Nov 26];13:667-71. Available from: https://www.mjdrdypv.org/text.asp?2020/13/6/667/300147
| Introduction|| |
The automated hematology analyzers have replaced the traditional manual methods for assessing hematological parameters. A complete blood count (CBC) report is now considered the initial screening and detection procedure for hematological abnormalities in modern hospitals and clinics.
During recent years, there has been an increasing demand of hematologic tests to diagnose and assess the prognosis of various disease conditions. Automated blood cell counters offer red blood cell count (RBC) and platelet (PLT) counts along with a 3-Part, 5-part, or a 7-part differential leukocyte count (DLC). Newer hematology analyzers also provide information about the morphologic features of blood cells by generating abnormal “cell flags.” Peripheral blood smear (PBS) examination is often used as the gold standard for the diagnosis of numerous RBC, white blood cell (WBC), and PLT disorders.
The usefulness of instrument-generated flags depends on the sensitivity and specificity of the automated machine. Flagging is defined as a system of signaling or communicating the message with a “flag.” In the hematology laboratory, a flag is the signal to the operator that the analyzed data may have a significant abnormality during analyzing blood samples. It thus necessitates confirmation of the results by microscopic examination of a well-stained PBS.,
Hematology analyzers generate suspect flags in the presence of abnormal cells. It is a known fact that false-positive rates for flags are high on all analyzers. Yet, it is said that hematology analyzers provide quick and accurate results in most situations. However, false results related to either WBC or PLT disorders such as blasts, atypical lymphocytes, immature granulocytes (IGs), giant PLTs, or other parameters from CBC may be observed in several instances. Most of the automated hematology analyzers are programmed in a variety of ways to “flag” or otherwise identify samples that may have abnormalities. The warning messages created by the automatic blood analyzer enable the user (technician, nurse, or physician) to detect positive samples and to react with the follow-up actions because of the warnings.
Various WBC disorders such as leukemia, leukemoid reactions, and other conditions leading to leukocytosis and leukopenias are known to generate signals (flags) in the analyzer report which should prompt a PBS examination. PLT count abnormalities due to giant PLTs or PLT clumps also generate suspect flags. Beckman Coulter (UNICEL D×H 800) used in our laboratory is a 5-part automated analyzer which produces several flag messages such as myeloblast (MO blast), immature granulocytes (IGs), left shift, eosinophilia, and giant PLTs. This study was undertaken to compare both WBC and PLT suspect flags with PBSs. There have been many individual studies on RBC, WBC, and PLT suspect flags, but only a few have studied them in conjunction. We have attempted to correlate and compare the WBC and PLT flag messages provided by the automated analyzer with their respective PBS to determine the accuracy of analyzer findings.
| Materials and Methods|| |
A descriptive cross-sectional study was carried out at a tertiary care hospital and research center in Pune which included 100 patients from medicine, surgery, pediatrics, and obstetrics and gynecology departments. Institutional ethics committee clearance was obtained. Data analysis was done based on records of the investigations obtained in Central Clinical Laboratory over a period of 6 months from June 2019 to December 2019. Patients from all age groups including infants to 70 years of age with abnormalities of WBCs and PLTs in the form of flag messages and/or abnormal peripheral smears were included in the study. All patients with normal report and normal PBS were excluded from the study. The venous blood samples were obtained in an ethylenediaminetetraacetic acid (EDTA) tube and run in automated hematology analyzers which included the 5-part Beckman Coulter (UNICEL D×H 800).
Analyzer reports were studied based on the flag messages generated for WBC (MO blast, IGs, left shift, eosinophilia, leukocytosis, and leukopenia) and PLT (thrombocytopenia, thrombocytosis, giant PLT, and PLT clumps). These were followed with their respective peripheral smear examination prepared according to the standard operating procedures and stained by Leishman stain. The count, morphology, and presence of any immature cells were noted for WBCs. PLTs were examined for the count, giant PLTs, and PLT clumps. Statistical analysis was done using Pearson's Chi-square test. P value (P < 0.05) was considered statistically significant.
| Results|| |
The analyzer reports and peripheral smears of 100 patients were analyzed. The age group of patients ranged from 2 days old infant to 77 years. Majority of patients were below 20 years of age (45%). Sixty-seven percentage of patients were males.
The detailed distribution of WBC, PLT, and combined defects, both by analyzer and PBS, is as shown in [Figure 1]. The distribution of cases based on the WBC suspect flags (n = 66; 46 only WBC and 20 WBC + PLT combination) showed that the majority had leukocytosis (n = 27) followed by leukopenia (n = 20), IGs (n = 10), left shift (n = 3), MO blast (n = 2), and eosinophilia (n = 4) [Figure 2]. Based on the PBS examination, there were 50 abnormal cases which showed leukocytosis (n = 22) followed by leukopenia (n = 18), IGs (n = 3), left shift (n = 2), blast (n = 1), and eosinophilia (n = 4). Correlation between the WBC histograms and their peripheral smears was found to be statistically significant (P = 0.006).
|Figure 1: Comparison of analyzer reports versus peripheral blood smear defects|
Click here to view
|Figure 2: Comparison of white blood cell detection by automated hematology analyzer and peripheral blood smear methods. (a) Analyzer report showing “myeloblast blast” (b) peripheral blood smear showing acute leukemia (c) Analyzer report showing “left shift” (d) peripheral blood smear showing leukemoid reaction (e) Analyzer report showing “leucocytosis” (f) peripheral blood smear showing leukocytosis (g) Coulter report showing“leukopenia” (h) peripheral blood smear showing leukopenia|
Click here to view
The distribution of 55 cases based on the PLT Coulter suspect flags; only PLT (n = 39) and WBC + PLT combined (n = 16) showed that the majority had thrombocytopenia (n = 22), followed by thrombocytosis (n = 13), giant PLTs (n = 10), PLT clumps (n = 8), and PLT debris (n = 2) [Figure 3]. Based on the peripheral smear examination (n = 79), majority had thrombocytopenia (n = 30), followed by thrombocytosis (n = 17), giant PLTs (n = 15), PLT clumps (n = 12), and PLT debris (n = 5). Correlation between the PLT and their peripheral smears was found to be statistically significant (P = 0.042).
|Figure 3: Comparison of platelet detection by automated hematology analyzer and peripheral blood smear methods. (a) Platelet suspect flag showing “giant platelets” (b) peripheral blood smear showing giant platelets (c) platelet suspect flag showing “platelet clumps” (d) peripheral blood smear showing platelet clumps (e) platelet suspect flag showing “thrombocytopenia” (f) platelet suspect flag showing thrombocytopenia (g) platelet suspect flag showing “thrombocytosis” (h) peripheral blood smear showing severe thrombocytosis|
Click here to view
| Discussion|| |
We attempted to correlate and compare the WBC and PLT flag messages provided by the automated analyzer with their respective PBS to determine the accuracy of the analyzer findings. Our study included EDTA blood samples of 100 patients having either an abnormality in the form of “suspect flags” in the analyzer generated CBC report and/or abnormal PBS. Majority of patients were young adults and males. The WBC suspect flags provide clues to the presence of any immature cell or presence of any particular cell population in excess or low amounts. WBC morphology can be studied using a well-stained PBS. It helps us to correctly determine if the WBC count is adequate, increased, or decreased. PBS provides information regarding the size, shape, and number of neutrophils, lymphocytes, eosinophils, basophils, and monocytes along with the presence of any immature precursors or blast cells. A combined study of the suspect flags and PBS can thus aid in the rapid detection of WBC disorders.
We found 66 samples showing WBC suspect flags. Based on the PBS examination, 50 cases were found to have anomalies. Leukocytosis (total leukocyte count>11,000/μl) in 22 of 27 cases and leukopenia (TLC <4000/μl) was observed in 18 of 20 cases as per the Coulter suspect flags. We deduced the fact that the analyzer was able to correctly pick up the cases of increased/decreased WBC counts. The minor differences could have been due to the cases which showed a very mild increased or decreased count. IGs were significantly less on PBS (n = 3) compared to those detected by Coulter (n = 10). In sepsis, one may find an increase in band forms which may be considered as IGs by the automated analyzer. These, if in small numbers, may be missed on PBS. One case showed toxic granules which may also have been considered as immature precursors by the analyzer. The left shift was observed in two of three cases shown by the flags and blasts were seen in only one case of acute leukemia on PBS instead of two as per the report. The two cases showing left shift were that of leukemia and leukemoid reaction. The two discordant cases were due to leukocytosis without the presence of left shift or immature blasts. All the four cases who diagnosed to be having eosinophilia had true eosinophilia (>6% of DLC) on the PBS in the range of 8%–20%. Correlation between the WBC histograms and their peripheral smears was found to be statistically significant (P = 0.006). Thus, the efficiency of automated analyzers in differentiating normal from abnormal specimens was well established. A study done by Avi Nahar in 2017 on Coulter flags and peripheral smear examination concluded that the manual method is still required, as it gives additional information about various abnormal morphological findings of blood cells and helps to validate the results of automated analyzers.
Peripheral smear examination of PLTs showed abnormalities in 79 cases compared to the 55 cases diagnosed by suspect flags. Thrombocytopenia was seen in 30 as compared to 22 cases on Coulter, and thrombocytosis was observed in 17 instead of 13 as per the analyzer reports. Giant PLTs (size >7 μ) were found in 15 instead of 10 cases and PLT clumps were found in 12 instead of 8 cases as per the suspect flags on Coulter. This difference can be explained by the fact that the presence of PLT aggregates, and cases of pseudothrombocytopenia were misinterpreted by the analyzer and detected as normal. These cases are purely diagnosed on PBS. PLT flags are produced rarely and only in extreme cases of thrombocytopenia and thrombocytosis. Therefore, to diagnose diseases of PLTs, PBS examination should be considered gold standard. A study was undertaken by Hawkins et al. for assessing the reliability of the Sysmex XE5000 instrument in detecting PLT clump. This was done using two PLT flags namely abnormal PLT size distribution (PAD) flag and plateletclumps (CLP) flag. They found that the overall reliability of the CLP flag is superior to the PAD flag, but there is room for further improvement.
In a study done by Bameni Moghaddam et al. for the validity of selected WBC differentiation flags in Sysmex XT-1800i on NRBC and blast flags, they found tthat there was a significant difference between their manually counted and automated counted NRBCs and blasts (P < 0.05). There was no statistically significant difference between automated and manual counting of flags for WBC scattergram. They found significant difference between automated and manual counting data in flags of eosinophilia and monocytosis (P < 0.05). Our study revealed an important correlation between the WBC and PLT suspect flags and peripheral smear findings. All WBC suspect flags except IGs correlated well with the PBS findings. PLT flags also proved very useful in detecting pathological cases. Based on our findings, we conclude that even in the era of automation and molecular analysis, PBS study alongside clinical history and examination is an important diagnostic tool while managing the various patients of hematological disorders. In the high-output laboratories, the automated data provided by the automated analyzers ensure that the suspect samples are not unintentionally missed or delayed in analysis.
| Conclusion|| |
We should analyze the suspect flags and use it as a screening tool to pick up pathological samples quickly. These should then be followed up with a thorough PBS examination for definitive diagnosis.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Pizarro M, Cepeda C, Laguer A. Performance evaluation of the Beckman Coulter Dxh 500* when compared to Coulter Hmx for the five-part differential. ISLH 2016;38:85-6.
Stamminger G, Auch D, Diem H, Sinha P. Performance of the XE-2100 leucocyte differential. Clin Lab Haematol 2002;24:271-80.
Depoorter M, Goletti S, Latinne D, Defour J. Optimal flagging combinations for best performance of five blood cell analyzers. Int J Lab Hematol 2015;37:63-70.
Turgeon ML. Clinical Hematology: Theory and Procedures. Lippincott Williams and Wilkins 2004;4:503-6.
Green R, Wachsmann-Hogiu S. Development, history, and future of automated cell counters. Clin Lab Med 2015;35:1-10.
Roy M, Akshay A. M2G1G2 white blood cell flag by three-part automated hematology analyzer: A hint to dengue infection in appropriate clinical context. J Lab Physicians 2019;11:103-6.
] [Full text]
Nahar A, Sudhamani S, Sirmukaddam S, Kiri V. Coulter flags & peripheral smear examination: Man v/s Machine. BAOJ Pathol 2017;1:2.
Hawkins JG, Uppal G, Gong J. Assessment of the reliability of the sysmex XE-5000 analyzer to detect platelet clumps. Lab Med 2016;47:189-94.
Bameni Moghaddam P, Mahjoub F, Emami A, Abdollahi A. Validity of selected WBC differentiation flags in sysmex XT-1800i. Iran J Pathol 2016;11:97-103.
Lokwani DP. The ABC of CBC: Interpretation of Complete Blood Count and Histograms. New Delhi: Jaypee Brothers Medical Publishers (P) Ltd; 2013.
Gulati GL, Hyland LJ, Kocher W, Schwarting R. An assessment of the Coulter GenS automated flagging system. Lab Med 2001;32:310-7.
[Figure 1], [Figure 2], [Figure 3]