International Journal of Laboratory Hematology

The Official journal of the International Society for Laboratory Hematology

ORIGINAL ARTICLE INTERNAT IONAL JOURNAL OF LABORATO RY HEMATO LOGY

Comparison of abnormal cell flagging of the hematology

analyzers Sysmex XN and Sysmex XE-5000 in oncohematologic
patients
J. R. FURUNDARENA, M. SAINZ, A. URANGA, L. CUEVAS, I. LOPEZ, J. ZUBICARAY, A. BIZJAK,
N. ROBADO, M. ARAIZ

Department of Hematology,
Donostia University Hospital,
Donostia, Spain
Correspondence:
Jose Ramon Furundarena,
Hematology, Donostia Univer-
sity Hospital, Begiristain Doktor-
earen Pasealekua z/g, 20014
Donostia, Spain.
Tel.: +34 943003439;
Fax: +34 943007063;
E-mail: joseramon.furundare
nasalsamendi@osakidetza.net
doi:10.1111/ijlh.12575
Received 17 January 2016;
accepted for publication 29 July
2016
Keywords
Hematology analyzer, Sysmex,
flag, blast, abnormal lymphocyte,
atypical lymphocyte
S U M M A RY
Introduction: Hematology analyzers should optimize flagging while
minimizing false-negative results and unnecessary microscopic
reviews.
Methods: We compared flagging performance of Sysmex XE-5000
and XN analyzers in oncohematologic patients. Differential counts
were performed by Cellavision digital system (100 cells) and a
hematologist (another 100 cells).
Results: First, we included 292 samples (86 with blasts): 28 acute
lymphoblastic leukemia, 88 acute myeloid leukemia, 91 myelodys-
plastic syndromes, 45 chronic myeloproliferative neoplasms, and 40
chronic myelomonocytic leukemia. Sensitivity, specificity and effi-
ciency to detect blasts were 59.3%, 88.3%, and 79.8% for XE-5000
analyzer and 70.9%, 91.3%, and 85.2% for the XN analyzer. Then,
we included 111 lymphoid malignancies. In 55 CLL XE-5000
flagged for Abn Lympho/L_Blasts?, XN flagged for Abn Lympho?.
In one-third of 19 samples with splenic marginal lymphoma, none
of the analyzers flagged. In 5 S
ezary syndrome cases, XE-5000 trig-
gered the Abn Lympho/L_Blasts? flag while the flagging in XN was
less consistent: Abn Lympho? Blasts? and Atypical Lympho?. In 5
hairy cell leukemias, both analyzers only flagged one sample. In 13
myelomas, XE-5000 generated Atypical Lympho? flag; XN triggered
more variable flags. In other lymphoid malignancies, flags were
variable. XN analyzer generates less samples with false basophilia.
Conclusion: XN analyzer has improved blast detection in oncohemato-
logic patients. Operators cannot rely on the blast flag alone but have
to consider other flags and hemogram data. In lymphoproliferative
disorders, XN analyzer yields less samples with pseudobasophilia.
Both analyzers must improve flagging for hairy cell leukemia.

© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem. 1
2 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY
INTRODUCTION
Patients with oncohematologic diseases undergo regu-
lar analytical tests including blood cell counts. Many
diseases such as acute lymphoblastic leukemia (ALL),
acute myeloid leukemia (AML), myelodysplastic syn-
dromes (MDS), chronic myeloproliferative neoplasms
(CMPN), and chronic myelomonocytic leukemia
(CMML) have in common the risk of blasts presence
in peripheral blood (PB), so it is essential to identify
changes in the results of the different parameters of
the hemogram and the presence of abnormal cell flag-
ging in the analyzers that require a microscopic smear
review. Bearing in mind that the microscopic review
process is laborious and time-consuming, it is very
important that analyzers optimize flagging while mini-
mizing false-negative results and unnecessary micro-
scopic reviews.
Another group of disorders, the lymphoproliferative
diseases can frequently express neoplastic lymphoid
cells in PB and it is important to know how the ana-
lyzers flag these cases.
In our core laboratory, we have recently changed
previous Sysmex XE-5000 analyzers for the newest
generation analyzers XN-1000 and XN-2000, and in
the interim, we have carried out a comparative study
of the abnormal cell flagging in these hematology ana-
lyzers in oncohematologic patients. Authors are deal-
ing with hematology analyzers. Good performance of
both analyzers to measure the different parameters of
the hemogram, to count erythroblasts and immature
granulocytes (IG), and to reduce the need for blood
film microscopic review has been proved before [1–4].
M AT E R I A L S A N D M E T H O D S
This study was carried out in the Core Hematology
Laboratory in Donostia University Hospital. We used
XE-5000 and XN-2000 analyzers (Sysmex, Kobe,
Japan). XE-5000 analyzer has channels to count and
to identify the different white blood cell (WBC) popu-
lations: in the WBC/BASO channel, it counts total
WBC and basophils, and in the DIFF channel, it
counts neutrophils, lymphocytes, monocytes, and
eosinophils. In the DIFF and IMI channels, it can
count immature granulocytes. In the IMI channel, a
combination of radio frequency, impedance, and
direct current resistance methods is used in
conjunction with a lysing reagent. Mature WBCs are
completely lysed and shrunken by the reagent. The
effect of the lysing reagent is different on immature
myeloid cells; they are not completely lysed, allowing
for their differentiation from mature cells. In the
NRBC channel, erythroblasts are identified and enu-
merated and total WBC are corrected accordingly. XE-
5000 can generate the following suspect flags: Blasts?
Abn Lympho/L_Blasts? and Atypical Lympho?
The new analyzer XN incorporates new channels
and reagents in two different models: XN-1000 and
XN-2000. These analyzers do not need IMI and NRBC
channels. They have WNR (white cell nucleated cell)
channel to count total WBC, erythroblasts, and baso-
phils, WDF (white cell differential) channel that is
similar to the DIFF channel in the XE-5000 model but
with an optimized separation between lymphocytes
and monocytes. Usually, both models work in line. If
the XN-1000 analyzer generates Blasts/Abn Lympho?
flag, the system makes a reflex testing and the sample
is processed in a XN-2000 analyzer that has WPC
(white cell precursor) channel that can generate more
specific Blasts? or Abn Lympho? flags, or no flag at
all. In the WPC channel, surfactants in Lysercell WPC
cause hemolysis and dissolution of red blood cells and
platelets, and affects the composition of the mem-
brane lipids of white blood cells. Mature white blood
cells containing more lipids than immature or reactive
cells are affected to a greater extent, leaving the cells
in a less native stage. This also makes the membrane
more permeable for the proprietary fluorescence mar-
ker that labels the cells in the second stage of the
reaction. The signals corresponding to cell volume and
fluorescence that are detected are then directly pro-
portional to this lipid content. Due to their low lipid
content, immature cells such as blasts show one of
the lowest fluorescence signals, coupled with one of
the highest signals for cell volume compared to
mature cells, and can hence be detected. On the other
hand, neoplastic lymphocytes are sometimes difficult
to distinguish from normal or reactive lymphocytes in
standard procedures. The sensitive optical detection
together with the fluorescence reaction and a propri-
etary algorithm allow the identification of such abnor-
mal cells [5]. XN analyzer also could generate Atypical
Lympho? flag for non-neoplastic lymphocytes. In our
study, we used a XN-2000 analyzer not connected in
line and samples were processed using all channels.
© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY
All analyzers were run in accordance with the
standard operating procedures and were checked daily
with quality control samples.
Samples were run within 6 h of collection in
K2EDTA tubes. They belonged to patients with differ-
ent hematological neoplastic diseases at the diagnosis
or during clinic controls. To compare the performance
of the morphological flags between the two analyzers,
we selected a group of patients at risk of having circu-
lating blasts with a diagnosis of ALL, AML, MDS,
CMPN, or CMML. We considered samples positive for
blasts when they had ≥1% circulating blasts (we
excluded samples with <1% blasts). In the CMPN, we
selected only patients that had ≥1% blasts and/or
≥2% IG (metamyelocytes, myelocytes, and promyelo-
cytes) in PB to evaluate the possible influence of
withdrawing the IMI channel in the new XN analyzer.
In a second group, we selected patients with lymphoid
malignancies with >5% circulating atypical lymphoid
cells.
Blood films were prepared and stained by the May-
Gr€unwald–Giemsa method in a Sysmex SP-1000 slide
maker. Differential counts were performed by the
Cellavision digital system (100 cells) and by a hema-
tologist (another 100 cells). As it has been proved, the
Cellavision digital system, after reclassification of the
cells, has a good correlation with the reference man-
ual 400-cell differential [6] for normal blood cells,
immature granulocytes, blasts, and abnormal lympho-
cytes [7]. The manual count by a hematologist sup-
ports the results.
Statistical analysis was performed by STATA 14.0
software (Sysmex Corporation, Kobe, Japan). The
rates of sensitivity (SE), specificity (SP), efficiency
(EF), positive predictive value (PPV), and negative
predictive value (NPV) for flagging samples with blasts
in PB were calculated by 2 9 2 truth tables. We con-
sidered valid the following flags: Blasts? Abn Lympho/
L_Blasts? and Atypical Lympho? for the XE-5000 ana-
lyzer, and Blasts? Abn Lympho? and Atypical Lym-
pho? for the XN analyzer. The IG flag did not provide
additional information to detect blasts.
R E S U LT S
Between November 2014 and July 2015, we collected
403 samples from 394 patients (we accepted two sam-
ples from the same patient if they were collected at
© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
3
diagnosis and later in remission). In the first group,
we included 292 samples from patients likely to have
blasts in PB: 28 ALL, 88 AML, 91 MDS, 45 CMPN (32
primary myelofibrosis (PM), eight chronic myeloid
leukemia (CML) and five other types), and 40 CMML.
From this group, 86 had ≥1% blasts in PB: six ALL,
33 AML, 16 MDS, 25 CMPN (19 PM, three CML and
three other types), and six CMML.
In the second group, we included 111 samples
from patients with lymphoproliferative diseases with
expression in PB: 29 typical chronic lymphocytic leu-
kemia (t-CLL), 26 atypical CLL (a-CLL), 19 splenic
marginal lymphoma (SML), five S
ezary syndrome
(SS), five hairy cell leukemia (HCL), three mantle cell
lymphoma (MCL), 13 myeloma (M), and 11 other
types (three marginal lymphoma not otherwise speci-
fied, two CD5+ lymphoproliferative neoplasm, two fol-
licular lymphoma, one diffuse large B-cell lymphoma,
one lymphoplasmacytic lymphoma, one angioim-
munoblastic T-cell lymphoma, and one peripheral T-
cell lymphoma).
We show the flags given by the two analyzers for
the different diseases included in the first group with
≥1% blasts in PB and without blasts in PB (Table 1).
When we examined the percentage of circulating
blasts present in the samples without flags in at least
one of the analyzers (false negatives), we found that
the majority corresponded to samples with only 1–2%
blasts in PB. The XE-5000 analyzer did not flagged 14
samples with 1% blasts, nine samples with 2% blasts,
five samples with 3% blasts, two samples with 4%
blasts, and five samples with ≥5% blasts. The XN ana-
lyzer did not flagged 11 samples with 1% blasts, six
samples with 2% blasts, four samples with 3% blasts,
one sample with 4% blasts, and three samples with
≥5% blasts. The samples with ≥5% blasts and without
flags in one or both analyzers were four AML (one
sample in both analyzers), one MDS and two PM;
three of them were leukopenic.
Globally, XE-5000 analyzer presented a SE of
59.3%, an SP of 88.3%, and an EF of 79.8% to flag
samples with blasts in PB. The XN analyzer did a SE
of 70.9%, an SP of 91.3% with an EF of 85.2%. The
PPV and NPV were 68% and 83.9% for the XE-5000
and 77.2% and 88.3% for the XN analyzer (Table 2).
When we measured the results distributed by
pathologies, there was some variability. For ALL and
AML, the results were good for both analyzers. In the
 4

Table 1. Flags given by the two analyzers for the different diseases included in the first group with ≥ 1% blasts in PB

ALL AML MDS PM CML Others CMPN CMML Total

n 28 88 91 32 8 5 40 292
With ≥1% blasts in PB 6 33 16 19 3 3 6 86
% blasts in PB 52 (1–97) 26 (1–96) 4 (1–15) 5.8 (1–51) 2.3 (2–3) 1.2 (1–3) 2.5 (1–6) 16.1 (1–97)
Without blasts in PB 22 55 75 13 5 2 34 206
Without blasts and with IG in PB 1 0 1 13 5 2 7 29
Flags XE-5000 in samples with/without blasts in PB
Blasts? 3/2 9/0 2/0 5/1 1/0 0/1 1/2 21/6
Abn Lympho/L_Blasts? 2/1 14/1 5/6 2/0 1/0 0/0 4/10 28/18
Atypical Lympho? 0/1 6/0 0/0 1/0 0/0 0/0 0/0 7/1
No flags 1/20 8/62 9/78 12/28 1/6 3/4 1/23 35/221
IG 1/1 7/0 2/0 15/7 3/4 2/2 2/7 32/21
Flags XN in samples with/without blasts in PB:
J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY

Blasts? 5/1 21/1 8/2 12/5 3/1 3/1 4/2 56/13

Abn Lympho? 1/0 1/2 2/2 0/0 0/0 0/0 0/1 4/5
Atypical Lympho? 2/0 8/0 0/0 5/1 0/0 0/0 0/0 15/1
No flags 1/22 9/61 6/77 7/19 0/4 0/1 2/33 25/217
IG 1/1 5/1 3/0 15/8 3/4 2/2 3/7 32/22

ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; MD, myelodysplastic syndromes; PM, primary mielofybrosis; CML, chronic myeloid
leukemia; CMPN, chronic myeloproliferative neoplasms; CMML, chronic myelomonocytic leukemia.

© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY 5

MDS group, the results were better for the XN ana- XE-5000 analyzer at the expense of a lesser SP.
lyzer. In the group of patients with CMPN with ≥2% Finally, for CMML, the SE was better for the XE-5000
IG in PB, the XN analyzer duplicated the SE of the analyzer.
With respect to the different flags, the XE-5000
analyzer flagged more frequently Abn Lympho/
Table 2. Sensitivity (SE), specificity (SP), efficiency L_Blasts? than Blasts? while the XN analyzer triggered
(EF), positive predictive value (PPV) and negative
more times Blasts? specific flag (Table 1). The XN ana-
predictive value (NPV) for both analyzers distributed
by pathology lyzer triggered the Abn Lympho? flag alone in seven
samples, two of them were MDS with blasts and five
SE SP EF PPV NPV of them had no blasts in PB. Regarding the Atypical
ALL+AML Lympho? flag, the XE-5000 analyzer generated this
XE-5000 76.9 94.8 88.8 88.2 89.0 flag alone in two samples and corresponded to two
XN 74.4 94.8 87.9 87.9 88.0 AML with blasts in PB. The XN analyzer triggered the
MDS Atypical Lympho? flag alone in three samples and
XE-5000 43.8 92 83.5 53.8 88.5 corresponded to three AML with circulating blasts.
XN 62.5 94.7 89.0 71.4 92.2
In the second group, we included 111 samples of
CMPN
XE-5000 36.0 90.0 60.0 81.8 52.9 lymphoid neoplasms with expression in PB. In
XN 72.0 65.0 68.8 72.0 65.0 Table 3, we show the suspect flags generated by both
CMML analyzers in the different pathologies.
XE-5000 83.3 64.7 67.5 29.4 95.7 In 29 t-CLL samples, the XE-5000 analyzer flagged
XN 66.7 91.2 87.5 57.1 93.9
Abn Lympho/L_Blasts? while the XN analyzer gener-
Total
XE-5000 59.3 88.3 79.8 68 83.9 ated the more accurate Abn Lympho? flag, but for
XN 70.9 91.3 85.2 77.2 88.3 five samples, this analyzer did not flag for abnormal
lymphocytes.
ALL, acute lymphoblastic leukemia; AML, acute myeloid We evaluated 26 a-CLL, and the flagging was similar
leukemia; MDS, myelodysplastic syndromes; CMPN,
to t-CLL samples, but in the XN analyzer, there were,
chronic myeloproliferative neoplasms; CMML, chronic
myelomonocytic leukemia. in addition to the Abn Lympho? flag, six samples with
Atypical Lympho? flag and four samples with Blasts?

Table 3. Suspect flags in lymphoid neoplasms with expression in PB

t-CLL a-CLL SML SS HCL MCL M Others Total

n 29 26 19 5 5 3 13 11 111
Flags XE-5000:
Blasts? 0 0 0 0 0 0 0 2 2
Abn Lympho/L_Blasts? 27 25 12 4 1 2 1 7 79
Atypical Lympho? 1 1 1 0 0 0 13 1 17
No flags 2 1 7 1 4 1 0 2 18
Pseudobasophilia 0 2 0 2 0 0 0 0 4
Flags XN:
Blasts? 0 4 3 2 0 0 7 5 21
Abn Lympho? 23 22 8 2 1 0 4 1 61
Atypical Lympho? 1 6 2 1 0 1 8 1 20
No flags 5 1 6 1 4 2 1 4 24
Pseudobasophilia 0 0 0 1 0 0 0 0 1

t-CLL, typical chronic lymphocytic leukemia; a-CLL: atypical CLL; SML, splenic marginal lymphoma; SS, Sézary syn-
drome; HCL, hairy cell leukemia; MCL, mantle cell lymphoma; M, myeloma. Others (3 marginal lymphoma not other-
wise specified, 2 CD5+ lymphoproliferative neoplasm, 2 follicular lymphoma, 1 diffuse large B-cell lymphoma, 1
lymphoplasmacytic lymphoma, 1 angioimmunoblastic T-cell lymphoma, 1 peripheral T-cell lymphoma).

© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
6 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY
flag. In three of the four samples with Blasts? flag, we
found >10% prolymphocytes in PB (in another three
samples with >10% prolymphocytes in PB, the ana-
lyzer did not flag for blasts). We detected 2 a-CLL sam-
ples with pseudobasophilia in the XE-5000 analyzer.
In approximately one-third of 19 samples with
SML, none of the analyzers flagged even though they
had between 6% and 47% little villous lymphocytes
in PB.
We included five S
ezary syndrome cases with 20%
(9–37%) S
ezary cells in PB. XE-5000 analyzer trig-
gered the Abn Lympho/L_Blasts? flag while the flag-
ging in XN analyzer was less consistent: Abn
Lympho? Blasts? and Atypical Lympho? In both ana-
lyzers, one sample with 11% circulating S
ezary cells
was not flagged. We detected two pseudobasophilia
cases in XE-5000 analyzer and one case in the XN
analyzer (the number of cells mistaken for basophils
counted in the XN analyzer in this last sample was
much lower than in the XE-5000 analyzer).
We evaluated five hairy cell leukemias with 9.4%
(5–16%) hairy cells in PB. Both analyzers only flagged
one sample: one with 7% circulating hairy cells in the
XE-5000 analyzer and one with 16% hairy cells in PB
in the XN analyzer. Of five hairy cell leukemias, two
samples had leukopenia and three a normal WBC
count, one of them with slight lymphocitosis (this
case was flagged in the XE-5000 analyzer).
We only collected three MCL samples with 18, 19,
and 36% abnormal lymphocytes in PB. XE-5000 did
not flag one sample, and the XN analyzer did not flag
two cases.
Thirteen myelomas with 5.5% (1–28%) plasma
cells in PB were included in the group. The XE-5000
analyzer generated Atypical Lympho? in all samples;
the XN analyzer triggered more variable flags: Abn
Lympho? Atypical Lympho? and Blasts? and did not
flag one sample with 1% circulating plasma cells.
In the subgroup of other lymphoid diseases, we
included isolated cases with different B-cell and T-cell
neoplasms and flags from both analyzers were vari-
able (Table 3).
DISCUSSION
The performance of new hematological analyzers is
usually evaluated including normal samples and a
group with variable pathologies, but in our case, we
decided to do the study exclusively with oncohemato-
logic patients. These patients undergo regular analyti-
cal tests to determine whether they need transfusions,
whether chemotherapy can be administered or to
know whether there is a relapse, progression, or leu-
kemic transformation of the disease. Therefore, it is
essential that the analyzer generate suspect flags for
abnormal cells, minimizing false-negative results with-
out increasing false-positive flags that demand labori-
ous and unnecessary microscopic reviews.
We chose two groups of oncohematologic patients:
on the one hand, pathologies at risk of having blasts
in PB (ALL, AML, MDS, CMPN, and CMML), and on
the other hand, lymphoid neoplasms with abnormal
lymphocytes in PB.
With respect to the capacity of the hematological
analyzers to flag samples with circulating blasts, the
XN analyzer improved globally the results of the XE-
5000 analyzer. The sensitivity increased from 59.3%
up to 70.9%, specificity from 88.3% up to 91.3%,
efficiency from 79.8% up to 85.2, PPV from 68% up
to 77.2, and NPV from 83.9% up to 88.3%. Compared
to the sensitivity results published in some papers for
Sysmex XN, XE-5000, and XE-2100 analyzers [1–4, 8–
12], we obtained a relatively lower result. For Sysmex
XE-5000 analyzer, our results were closer to the ones
found in other papers [13, 14]. The sensitivity results
in our group of patients were similar to some pub-
lished results with other analyzers as Abbott Cell-Dyn
Sapphire, ADVIA 2120, and Beckman Coulter DxH
800 and LH 750 [3, 10–13, 15–17]. The specificity and
efficiency results were good and similar to the ones
obtained in the evaluations mentioned before
(Table 4).
The differences in the results between our study
and the published papers could have more than one
reason. The fact that all our samples belonged to
oncohematologic patients may influence flagging; the
mean number of circulating blasts is not mentioned in
many references, and in our group, there were more
patients with very few blasts in PB, which analyzers
fail to detect. The minimal percent of blasts in PB
required to consider a sample as positive is variable
(>0, 1 or 2%) as well as the flags accepted as correct
—some authors accept only blast or LUC flags, others,
like us, accept flagging for abnormal or atypical lym-
phocytes, and there is a third group who includes also
IG flag.
© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY 7

Table 4. Published performance evaluations of different analyzers. We show sensitivity (SE), specificity (SP) and
efficiency (EF) to detect blasts in PB

Total Samples with % blasts

Ref Analyzers samples blasts in PB SE SP EF Comments

Ruzicka et al. [8] XE-2100 486 23 >1% 65 89 88 Accepted flags: Blasts

Stamminger XE-2100 367 25 >2% 86 83 83 Accepted flags: Blasts
et al. [9]
Briggs et al. [1] XE-2100 1002 16 81.2 90.8 90.7
XE-5000 87.5 97.2 97.1
Briggs et al. [2] XE-2100 390 20 95 99.7
XN 100 82
Depoorter XE-2100 517 45 AML 29 85 79 Results obtained
et al. [10] XN-2000 4 ALL 96 60 63 considering correct
Sapphire 34 78 74 the following flags:
DxH-800 1 77 76 Blasts? Abn
Lympho? Atypical
Lympho? and IG
for the XN and
other flag
combinations for
other analyzers
Seo et al. [4] XE-2100 261 17 13 > 1% 84 84 Accepted flags: Blasts?
XN 4 < 1% 82 97
Hotton et al. [3] XN-2000 4375 7 AML (0.5–94.7) 89 92 92 Results obtained
Sapphire 3 ALL 72 89 88 considering correct the
Dx-800 7 post-CT 100 96 97 following flags: Blasts?
1 infectious Abn Lympho? Atypical
Lympho? and IG for
the XN
Meintker XE-2100 202 17 >0% 94 80 Accepted flags: Blasts?
et al. [11] Sapphire 65 86 (and LUC > 3% for
ADVIA 120 71 67 ADVIA)
Dx-800 82 80
Kang et al. [12] XE-2100 430 22 >1% 90.9 90.4 90.5 Accepted flags: Blasts
Sapphire 77.3 95.8 94.9
ADVIA 2120 59.1 85.8 84.4
LH-750 63.6 97.5 95.8
Bruegel et al. [13] XE-5000 349 34 >0% 65 98
XN-2000 97 96
Sapphire 76 93
DxH-800 74 95
ADVIA 2120 65 97
Eilertsen XE-5000 408 48 ≥0.5% 73 60 Accepted flags: Blasts
et al. [14]
Tan et al. [15] DxH 800 330 22 >0% 77.3 88.6 87.9 Accepted flags: Blasts
LH-780 40.9 95.8 92.1
Sapphire 59.1 93.8 91.5
Shelat LH-750 390 29 ≥2% 62.1 85.6 83.8 Accepted more than
et al. [16] ADVIA 2120 100 49.3 53.1 one flag
Jean et al. [19] DxH 800 293 22 4.75 (1–94) 90 90.5 90.4 Accepted flags: Blasts
LH755 70 94.1 92.5

© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
8 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY
In our evaluation, we proved that in the XN ana-
lyzer, an Abn Lympho? flag alone could indicate the
presence of blasts, and in regard to the Atypical Lym-
pho? flag alone, in the XE-5000 and XN analyzers, we
proved the same. So, in the context of oncohemato-
logic patients, these flags could indicate the presence
of blasts in PB and, therefore, the need to carry out a
microscopic review of a blood film.
With respect to the different pathologies included
in the first group of patients, the best results were
obtained in ALL and AML, which seems logical as the
percentage of circulating blasts was higher than that
present in MDS, CMPN, and CMML.
Among patients with CMPN, we selected those
who had ≥2% immature granulocytes in PB that
could hinder the flagging in the analyzers. We wanted
to know how the IMI channel present in the XE-5000
analyzer and the WPC channel present in the XN ana-
lyzer could influence the analyzers’ flagging perfor-
mance. We met that XN analyzer was more sensitive
to the presence of blasts although with a lesser speci-
ficity.
Finally, the patients with CMML represent a chal-
lenge for the analyzers because they can have atypical
monocytes, and promonocytes must be counted as
blasts. As among the 40 samples only six had circulat-
ing blasts, we could not reach any conclusions.
In the second group of patients, we included
patients with different lymphoid malignancies with
abnormal lymphoid cells in PB.
In t-CLL and a-CLL, the XE-5000 analyzer nearly
always flagged Abn Lympho/L_Blasts? while the XN
analyzer usually flagged for Abn Lympho? but it can
also flag for Atypical Lympho? and some samples for
Blasts? which could alert about the presence of many
prolymphocytes in PB. In S
ezary syndrome, the XE-
5000 analyzer did Abn Lympho/L_Blasts? flag while
flags in the XN analyzer were more variable.
In patients with splenic marginal lymphoma, both
analyzers failed to flag in about one-third. In any
case, these are patients who usually have lymphocyto-
sis and require a slide review. In our opinion,
although these cases present with villous lympho-
cytes, their size is similar to the normal lymphocytes
and the analyzer did not flag some samples.
The fact that contrary to what we expected, none
of the analyzers triggered any flagging in four of five
cases, in patients with hairy cell leukemia, who do
not usually have lymphocytosis, points at the need of
an improvement of both analyzers flagging for this
type of cases. We have not found any plausible expla-
nation for these cases, but we think it would be inter-
esting to study lymphocyte research parameters that
the analyzer can measure (LY-X: the lateral scattered
light intensity of the LYMPH area on the WDF scatter-
gram; LY-Y: the fluorescent light intensity of the
LYMPH area on the WDF scattergram; LY-Z: the for-
ward scattered light intensity of the LYMPH area on
the WDF scattergram; LY-WX: the lateral scattered
light distribution width of the LYMPH area on the
WDF scattergram; LY-WY: the fluorescent light distri-
bution width of the LYMPH area on the WDF scatter-
gram; and LY-WZ: the forward scattered light
distribution width of the LYMPH area on the WDF
scattergram) and perhaps they could be helpful to flag
these cases in the future.
Due to new treatment options and a longer sur-
vival of patients with myeloma, more and more sam-
ples with circulating plasma cells are detected, which
is why flagging in these samples is of prime impor-
tance. Plasma cells appeared as high fluorescence-
stained cells in the Sysmex XE-2100 analyzers flagging
as Atypical Lympho? [18]. We included 13 myeloma
with ≥1% plasma cells in PB and XE-5000 flagged for
Atypical Lympho? in all samples while the XN ana-
lyzer generated more variable flags including Abn
Lympho? Atypical Lympho? and some with Blasts?
flag. In an evaluation cited before [3], nine samples
with 0.5–1.2% plasma cells in PB were included. The
results were similar to ours and XN analyzer flagged
eight samples for abnormal o atypical lymphocytes.
It is known that some analyzers can mistake circu-
lating atypical lymphocytes for basophils (pseudoba-
sophilia). Jacomo [19] demonstrated that the majority
of basophilias >2% in an XE-2100 analyzer, particu-
larly when they co-appeared together with Atypical
Lympho? or Blasts? flag, corresponded to atypical
lymphocytes. Similar cases have been detected in
other analyzers [20, 21]. In our group of 111 lympho-
proliferative diseases, we detected four cases of pseu-
dobasophilia in the XE-5000 analyzer (2 a-CLL and 2
S
ezary syndrome) and only one case in the XN
analyzer.
Briggs et al. [1] evaluated the performance of XE-
5000 and XE-2100 analyzers including 43 samples
with atypical lymphocytes in PB (diagnosis was not
© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY 9

specified). SE, SP, and EF were, respectively, 51.2%,
95.3%, and 93.4% for the XE-5000 analyzer and
53.5%, 88.1%, and 86.6% for the XE-2100 analyzer.
In another paper, the same author compared XE-2100
and XN analyzers including 31 samples with atypical
or abnormal lymphocytes of 390 total samples. SE and
SP for the XN analyzer to detect abnormal lympho-
cytes were 37.5% and 98.7%, respectively, and SE
and SP to detect atypical lymphocytes were 78.3%
and 95.2%, respectively. XN analyzer generated less
false positives than XE-2100.
In an evaluation of four analyzers [10], 29 samples
of 517 with atypical lymphocytes in PB were included
(12 CLL, six mantle cell lymphoma, one low-grade
non-Hodgkin lymphoma y 10 reactive lymphocytosis).
Considering as correct flagging for Blasts? Abn Lym-
pho? and Atypical Lympho? the XN analyzer obtained
a SE of 72% and an SP of 79%.
In conclusion, the XN analyzer’s ability to alert
about the possible presence of blasts in PB in
oncohematologic patients with ALL, AML, MDS,
CMPN, or CMML represents an improvement over
the XE-5000. However, and even though the majority
were flagged as Blasts? operators cannot rely on the
blast flag alone and have to take into account other
morphological flags, hemogram data, and the visual
interpretation of the different scattergrams of the ana-
lyzers. As regards to lymphoproliferative disorders
with abnormal lymphocytes in PB, the XE-5000 usu-
ally flags for Abn Lympho/L_Blasts? while the XN
analyzer flag for Abn Lympho? and Atypical Lympho?
although some samples flag for Blasts? XN analyzer
yielded a lower number of samples with pseudoba-
sophilia. Samples from patients with hairy cell leuke-
mia, usually without lymphocytosis, are a challenge
for these analyzers that must improve flagging; lym-
phocyte research parameters that can be measured in
these analyzers (LY-X, LY-Y, LY-Z, LY-WX, LY-WY
and LY-WX) should be explored in the future.

REFERENCES
1. Briggs C, Linssen J, Longair I, Machin S.
Improved flagging rates on the Sysmex XE-
5000 compared with the XE-2100 reduce
the number of manual film reviews and
increase laboratory productivity. Am J Clin
Pathol 2011;136:309–16.
2. Briggs C, Longair I, Kumar P, Singh D.,
Machin S.J. Performance evaluation of the
Sysmex haematology XN modular system. J
Clin Pathol 2012;65:1024–30.
3. Hotton J, Broothaers J, Swaelens C, Can-
tinieaux B. Performance and abnormal cell
flagging comparisons of three automated
blood cell counters: Cell-Dyn Sapphire,
DxH-800, and XN-2000. Am J Clin Pathol
2013;140:845–52.
4. Seo JY, Lee S-T, Kim S-H. Performance
evaluation of the new hematology analyzer
Sysmex XN-series. Int J Lab Hematol
2015;37:155–64.
5. Kawauchi S, Takagi Y, Kono M, Wada A,
Morikawa T. Comparison of the leukocyte
differentiation scattergrams between the
XN-series and the XE-series of hematology
analyzers. Sysmex J Int 2014;24:1–8.
6. Clinical and Laboratory Standards Institute
(CLSI). Reference Leukocyte (WBC) Differ-
ential Count (proportional) and Evaluation
of Instrumental Methods; Approved
standard. Second edition. Wayne,
Pennsylvania, USA: CLSI document H20-
A2: CLSI; 2007.
7. Briggs C, Longair I, Slavik M, Thwaite K,
Mills R, Thavaraja V, Foster A, Romanin D,
Machin SJ. Can automated blood film anal-
ysis replace the manual differential ? An
evaluation of the Cellavision DM96 auto-
mated image analysis system. Int J Lab
Hematol 2009;31:48–60.
8. Ruzicka K, Veitl M, Thalhammer-Scherrer
R, Schwarzinger I. The new hematology
analyzer Sysmex XE-2100. Arch Pathol Lab
Med 2001;125:391–6.
9. Stamminger G, Auch D, Diem H, Sinha P.
Performance of the XE-2100 leucocyte dif-
ferential. Clin Lab Haematol 2002;4:271–
80.
10. Depoorter M, Goletti S, Latinne D, Defour
JP. Optimal flagging combinations for best
performance of five blood cell analyzers.
Int J Lab Hematol 2015;37:63–70.
11. Meintker L, Ringwald J, Rauh M, Krause
SW. Comparison of automated differential
blood cell counts from Abbott Sapphire,
Siemens Advia 120, Beckman Coulter
DxH800, and Sysmex XE-2100 in normal
and pathologic samples. Am J Clin Pathol
2013;139:641–50.
12. Kang SH, Kim HK, Ham CK, Lee DS, Cho
HI. Comparison of four hematology analyz-
ers, Cell-Dyn Sapphire, Advia 120, Coulter
LH-750 and Sysmex XE-2100, in terms of
clinical usefulness. Int J Lab Hematol
2008;30:480–6.
13. Bruegel M, Nagel D, Funk M, Fuhrmann P,
Zander J, Teupser D. Comparison of five
automated hematology analyzers in a uni-
versity hospital setting: Abbott Cell-Dyn
Sapphire, Beckman Coulter DxH 800, Sie-
mens Advia 2120i, Sysmex XE-5000 and
Sysmex XN-2000. Clin Chem Lab Med
2015;53:1057–71.
14. Eilertsen H, Vøllestad NK, Hagve TA. The
usefulness of blast flags on the Sysmex XE-
5000 is questionable. Am J Clin Pathol
2013;139:633–40.
15. Tan BT, Nava AJ, George TI. Evaluation of
the Beckman Coulter UniCel DxH 800,
Beckman Coulter LH 780, and Abbott Diag-
nostics Cell-Dyn Sapphire hematology ana-
lyzers on adult specimens in a tertiary care
hospital. Am J Clin Pathol 2011;135:939–51.
16. Shelat SG, Canfield W, Shibutani S. Differ-
ences in detecting blasts between Advia
2120 and Beckman-Coulter LH750 hema-
tology analyzers. Int J Lab Hematol
2010;32:113–6.
17. Jean A, Boutet C, Lenormand B, Callat
M-P, Buchonnet G, Barbay V, Basuyau J-P,
Vasse M. The new haematology analyzer
DxH 800: an evaluation of the analytical
performances and leucocyte flags, compar-
ison with the LH755. Int J Lab Hematol.
2011;33:138–45.

© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.
10 J. R. FURUNDARENA ET AL. | FLAGS OF SYSMEX ANALYZERS IN HEMATOLOGY
18. Linssen J, Jennissen V, Hildmann J, Rei-
singer E, Schindler J, Malchau G, Nierhaus
A, Wielckens K. Identification and quantifi-
cation of high fluorescence-stained lym-
phocytes as antibody synthesizing/secreting
cells using the automated routine hematol-
ogy analyzer XE-2100. Cytometry B Clin
Cytom 2007;72B:157–66.
19. Jacomo RH, Lozano VF, da Cunha Neto JG,
Soares Costa S. What’s the meaning of
basophilia in Sysmex XE-2100? Arch
Pathol Lab Med 2011;135:415.
20. Chandrashekar V. Basophil differentials as
a marker for atypical lymphocyte morpho-
logic characteristics. Lab Med 2013;44:133–
5.
21. Hur M, Lee YK, Lee KM, Kim JH, Cho HI.
Pseudobasophilia as an erroneous white
blood cell differential count with a discrep-
ancy between automated cell counters:
report of two cases. Clin Lab Haematol
2004;26:287–90.
© 2016 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem.