ABSTRACT

The Sysmex XT-2000i automated hematology analyzerwas evaluated at Saint Louis Children’s Hospital (SLCH),St. Louis, MO, USA. Complete blood count results fromthe Sysmex XT-2000i were compared to results from theSysmex XE-2100 for 114 pediatric and adult patient sam-ples. Manual differentials were performed on each specimenby 2 experienced medical technologists using guidelinesestablished in the National Committee for Clinical Labora-tory Standards (NCCLS) document H20-A. Carryover,precision, linearity, correlation, stability, and mixing-teststudies were also performed. The XT-2000i results showedexcellent correlation with the results from the XE-2100 forthe following parameters: white blood cells; red blood cells;hemoglobin; hematocrit; mean corpuscular volume; meancorpuscular hemoglobin; mean corpuscular hemoglobinconcentration; red blood cell distribution width by stan-dard deviation; red blood cell distribution width by coeffi-cient of variation; mean platelet volume; platelets; percentneutrophils, lymphocytes, monocytes, eosinophils, andbasophils; and reticulocyte percent and number. In ourevaluation of the XT-2000i the correlation coefficients forall complete blood counts and differential parameters com-pared well with those of the XE-2100. We concluded thatthe XT-2000i demonstrated comparable analytical perfor-mance to its predecessor, the XE-2100. Lab Hematol.2003;9:29-37.

KEY WORDS: Performance evaluation · Automatedhematology analyzer · XT-2000i

INTRODUCTION

The Sysmex XT-2000i is a new fully automated hema-tology analyzer that performs a complete blood count(CBC) with a 16-parameter hemogram, a 5-part whiteblood cell (WBC) differential, and reticulocyte (RET)analysis that includes the immature RET fraction (IRF)and a fluorescent optical platelet count (PLT-O). The ana-lyzer is designed for high-volume output in the clinical lab-oratory with a throughput of 80 samples per hour, but theinstrument’s compact size allows it to fit in a physician’soffice lab. The aspirated sample volume required is 85 µLof whole blood for the manual mode and 150 µL for theclosed mode [1].

The XT-2000i uses the newest in applications, includingan Internet access line, flat-screen monitor, and easy-to-navigate software. This instrument uses the current technologyof fluorescent flow cytometry with a semiconductor laser toprovide the differential, RET, and fluorescent optical PLTinformation.

The purpose of this study was to evaluate the perfor-mance of the XT-2000i and compare its accuracy and preci-sion to those of an already established hematology analyzer,the XE-2100.

The performance of the XT-2000i was evaluated at thelaboratory of Saint Louis Children’s Hospital (SLCH), theoldest pediatric hospital west of the Mississippi River.SLCH has 235 beds, including a 26-bed pediatric inten-sive care unit, a 52-bed neonatal intensive care unit, and a5-bed pediatric bone marrow transplantation unit. Thededicated pediatric laboratories process specimens for hos-

Performance Evaluation of the Sysmex XT-2000iAutomated Hematology Analyzer

K. LANGFORD, L. LUCHTMAN-JONES, R. MILLER, D. WALCK

Saint Louis Children’s Hospital Core Laboratory/Hematology, St. Louis, Missouri, USA

Received October 2, 2002; received in revised form January 13, 2003; accepted January 17, 2003

Laboratory Hematology 9:29-37© 2003 Carden Jennings Publishing Co., Ltd.

29

Correspondence and reprint requests: Keith Langford, Saint LouisChildren’s Hospital Core Laboratory/Hematology, One Children’sPlace, St. Louis, MO 63110-1077, USA; 1-314-454-4268; fax: 1-314-454-4156 (e-mail: kxl6966@bjc.org).

ISLH

Official Publication

pitalized patients, subspecialty clinics, and outpatient clin-ics in the St. Louis area. The hematology laboratory ofBarnes Jewish Christian Hospital (BJC), which providedthe adult samples, processes specimens from patientsadmitted to this tertiary hospital as well as from the sub-specialty clinics and various ambulatory care centers in thecommunity.

MATERIALS AND METHODS

SpecimensResidual peripheral venous blood samples collected in

K2EDTA and K3EDTA tubes were used in the comparisontesting. These blood samples were sent to SLCH and BJClaboratories for routine CBC testing. The specimens wereanalyzed on the XE-2100 (Sysmex Corporation of America,Long Grove, Illinois, USA) and the XT-2000i (Sysmex)within 8 hours of collection. A few exceptions were made forvery unusual CBC results because they were received in theevening and night shifts. These specimens were stored untilthe next morning in a refrigerator at 2°C to 8°C and ana-lyzed within 24 hours.

The specimens used in the study consisted of 50 sampleswith CBC results that fell within normal ranges according toSLCH/BJC parameters and 64 samples that were classified asabnormal. The 64 abnormal specimens were obtained frompatients with various clinical diagnoses including leukemias,sickle cell disease, multiple myeloma, viral infections, idio-pathic thrombocytopenic purpura, sepsis, pancytopenia, andiron deficiency. Samples were obtained from preterm andterm sick and healthy infants and from patients in criticalcare units, including the bone marrow transplantation unit.These specimens were used to determine whether unusualCBC findings could be evaluated appropriately on the XT-2000i. Furthermore, the types of suspect and user-definedflags generated for each abnormality were evaluated. Thevariety of abnormal cells and cell parameters encounteredincluded nucleated red blood cells (NRBCs), sickle cells,blasts, immature granulocytes, atypical lymphocytes, plasmacells, large PLTs, low PLT counts, high WBC counts, PLTclumps, and RBC fragments.

Duplicate peripheral blood smear slides were made andstained with a Wright-Giemsa stain. A slide from each sam-ple was given to each of 2 technologists, who performedblind manual differentials of 200 cells each. These differen-tials were performed according to the National Committeefor Clinical Laboratory Standards (NCCLS) document H20-A[2]. The results from both smears were then averaged. If thedifference in results varied significantly between the technol-ogists, then another technologist read a referee slide. Thelaboratory used the following criteria to determine if the dif-ferential was abnormal: >10 bands; >5 atypical lymphocytes;presence of any metamyelocyte, myelocyte, promyelocyte, orblast; and presence of any NRBCs.

InstrumentsThe XT-2000i is a new hematology analyzer capable of

providing a 16-parameter hemogram, a 5-part differential,and RET analysis including IRF information. The XT-2000i stores 10,000 samples in a Windows 2000 packageusing a flat-screen monitor and user-friendly software. Fig-ure 1 is an example of a NEGATIVE XT-2000i screen printincluding the scattergrams and histograms. Figure 2 is anexample of a POSITIVE XT-2000i screen print. The termPOSITIVE indicates that the CBC was interpreted asabnormal and the XT-2000i alerted the operator of the sus-pect flags, which resulted because of the unusual findings.The following is a list of suspect flags/interpreted messagesgenerated with the presence of unusual measurements:immature granulocytes, left shift, blasts, abnormal lympho-cyte/L-blasts, atypical lymphocytes, NRBCs, RBC frag-ments, RBC lyse resistance, RBC agglutination, dimorphicRBC population, turbidity/HGB interference, iron deficiency,HGB defects, PLT clumps, and WBC, RET, and PLTabnormal scattergrams. User-defined flags are defined byeach institution to identify abnormal numerical results withinthe population being tested. Examples of user-defined flagsare neutrophilia, monocytosis, anisocytosis, macrocytosis,and leukocytopenia.

The WBC count is determined by flow cytometry usingforward-scattered and side-scattered light. The differentialuses a specific nucleic acid dye to measure the cells by side-fluorescent light and side-scattered light. The RBC and PLTimpedance (I) are measured using direct-current detection.The hematocrit (HCT) is simultaneously determined usingthe RBC pulse-height detection method. RETs are analyzedusing flow cytometry and a nucleic acid fluorescent dye. Themeasurement uses both forward-scattered light and side fluo-rescent information to determine the RET count, percent-age, and immature RET information. The PLT-O is alsoperformed in this channel.

The following parameters are calculated from directlymeasured data: mean corpuscular volume (MCV), mean cor-puscular hemoglobin (MCH), mean corpuscular hemoglo-bin concentration (MCHC), RBC distribution width bystandard deviation (RDW-SD), RDW by coefficient of vari-ation (RDW-CV), platelet distribution width (PDW), meanplatelet volume (MPV), platelet large cell ratio (P-LCR),IRF, low fluorescence ratio (LFR), middle fluorescence ratio(MFR), and high fluorescence ratio (HFR).

The Sysmex XE-2100 was used for the comparison study.This analyzer is used daily in the laboratory at SLCH. ACBC with a 5-part differential and an RET count was per-formed on both analyzers.

CalibrationAt the time of its arrival, the XT-2000i was calibrated by

Sysmex Corporation service engineers according to themanufacturer’s guidelines using the Sysmex recommended

30 K. Langford et al

Performance Evaluation of the Sysmex XT-2000i 31

FIGURE 1. Example of negative XT-2000i screen print. WBC indicates white blood cells; RBC, red blood cells; HGB, hemoglobin;HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, MCH concentration; PLT,platelets; RDW-SD, RBC distribution width by standard deviation; RDW-CV, RDW by coefficient of variation; MPV, meanplatelet volume; RET#, reticulocyte count; IRF, immature RET fraction; NEUT#, neutrophil count; LYMPH#, lymphocytecount; MONO#, monocyte count; EO#, eosinophil count; BASO#, basophil count.

FIGURE 2. Example of positive XT-2000i screen print. See Figure 1 legend for abbreviation definitions.

calibrator product. Three levels of quality control materiale-Check were used (levels 1, 2, and 3) throughout the com-parison study. Each day before specimen analysis shutdownand startup were performed with close attention to back-ground counts. All instruments in the study had the sameuser-defined flag settings for flagging evaluation, thusensuring that all conditions such as leukopenia, neutrophilia,and microcytosis would trigger the appropriate flag on bothanalyzers.

ReagentsBy way of comparison, the XT-2000i and the XE-2100

use a common reagent system. The XT-2000i uses 3 fewerreagents than the XE-2100. The footprint of the XT-2000i issmaller because it does not use Cell Sheath, NRBC diluentand dye, or Stromatolyser IM, reagents that are unique to the

XE-2100. Table 1 lists the reagents required for both the XT-2000i and XE-2100.

PrecisionPrecision for both the closed mode (automated aspiration

via cap-piercing) and manual open mode aspiration was eval-uated by the performance of 10 consecutive measurementson fresh donor blood samples.

CarryoverCarryover was performed using the International Com-

mittee for Standardization in Hematology (ICSH) procedurefor the following parameters: WBC, RBC, HGB, HCT, PLT,and numbers of RETs, neutrophils, lymphocytes, monocytes,eosinophils, and basophils. Carryover was assessed by 3 con-secutive analyses of a patient sample with high analyte con-

32 K. Langford et al

TABLE 1. Reagent Comparison*

Reagent Description XT-2100i XE-2100

CellPack Red blood cell/platelet and hemoglobin diluent that is used as the main rinse of the instrument. X XCell Sheath Diluent used as a sheath for cells flowing through the aperture. — XSulfolyser Sodium lauryl sulfate hemoglobin measuring reagent. X XStromatolyser-FB Basophil diluent and lyse. Lyses all cells except basophils. X XStromatolyser-DL/4DS Differential reagents used to lyse red blood cells and dilute and stain white blood cells. X XStromatolyser-IM Used to lyse red blood cells and white blood cell cytoplasm for detection of immature granulocyte — X

information.Stromatolyser-NR Used as a lyse in the detection of nucleated red blood cells. — XRet Search II Dilutes and stains reticulocytes and platelets for analysis X X

*X indicates reagent is used on the system; —, reagent is not used.

TABLE 2A. Precision—Open-Tube Analysis of the XT-2000i *

Open-Tube RBC, HGB, PLT, WBC, NEUT, LYM, MONO, RET,Analysis ×106/µL g/dL HCT, % MCV, fL ×103/µL ×103/µL ×103/µL ×103/µL ×103/µL ×106/µL

Mean 5.15 15.9 47.5 92.3 638 12.22 6.41 4.36 0.89 0.13SD 0.03 0.11 0.31 0.3 5.2 0.19 .10 0.08 0.03 0.01CV% 0.6 0.7 0.7 0.3 0.8 1.6 1.6 1.9 3.5 4.9

*RBC indicates red blood cells; HGB, hemoglobin; HCT, hematocrit; MCV, mean corpuscular volume; PLT, platelets;WBC, white blood cells; NEUT, neu-trophils; LYM, lymphocytes; MONO, monocytes; RET, reticulocytes; CV, coefficient of variation.

TABLE 2B. Precision—Closed-Tube Analysis of the XT-2000i *

Closed-Tube RBC, HGB, PLT, WBC, NEUT, LYM, MONO, RET,Analysis ×106/µL g/dL HCT, % MCV, fL ×103/µL ×103/µL ×103/µL ×103/µL ×103/µL ×106/µL

Mean 4.35 13.4 40.4 92.9 305 8.54 4.46 3.14 0.60 0.08SD 0.02 0.06 0.30 0.28 6.1 0.10 0.07 0.09 0.03 0.01CV% 0.6 0.4 0.7 0.3 2.0 1.1 1.5 2.9 4.8 6.3

*See Table 2A footnote for abbreviation definitions.

centration (H1, H2, H3) followed by 3 consecutive analysesof a patient sample with low analyte concentration (L1, L2,L3). The percentage of carryover for each parameter was cal-culated from the formula [3]:

Carryover (%) = L1 – L3 × 100.H3 – L3

LinearityTo assure the linearity of WBC, RBC, HGB, HCT, PLT,

and RET results, blood samples were serially diluted to chal-lenge the reportable range of the instrument. Data fromthese analyses were compared graphically using the actualvalues obtained plotted against values projected from theundiluted sample. High PLT-I linearity was evaluated using asingle-donor pheresis product that had been volume reduced;the resulting platelet-rich product was added to a venousblood tube. High WBC linearity was established by using aspecimen from a patient with chronic myeloid leukemia whohad presented with a WBC count of 410 × 103/µL. The fol-lowing criteria were used for linear performance testing: thedata must fit a linearity regression line for which the coeffi-cient of determination (R2) must be >0.95; a minimum of5 dilutions distributed throughout the linearity range mustbe used; the dilutions must cover the reportable range for theparameter; and each dilution result must be the mean valueof duplicate or more measurements on the same range.

Mixing StudiesThe XT-2000i mixes each tube 10 times by inversion in

the automode prior to blood aspiration. The effectiveness ofthe 10-inversion mixing was validated by performance of3 analyses of 14 samples in the automode and averaging ofthe RBC value. This value was used as the baseline compari-son. The 14 samples were then allowed to sit undisturbed for4 hours at room temperature. After 4 hours, the sampleswere placed on the analyzer without any manual mixing andanalyzed in the automode. The specimens were then mixedwell by hand and reanalyzed in the automode. Both of the

RBC obtained values were then compared to the baselinevalues and had to agree within 1.5% of the baseline.

Comparison StudiesAs discussed previously, a total of 114 venous blood samples

were selected and analyzed on the Sysmex XT-2000i and theSysmex XE-2100. For all specimens, 2 blood films were pre-pared for a 400-cell manual differential that was performed by2 technologists. The CBC and differential data were evaluated.

RESULTS

PrecisionResults from replicate open-tube and closed-tube analysis

are shown in Tables 2A and 2B, respectively. The precisionattained surpassed manufacturer specifications for all parame-ters. RBC parameters showed exceptional precision, with coef-ficients of variation (CVs) less than 1.0%. The WBC and PLTparameters showed CVs of 2.0% or less. Precision of differen-tials for the primary cell types—neutrophils, lymphocytes, andmonocytes—also was excellent, with CVs less than 5%.

CarryoverCarryover data are presented in Table 3 for total WBC,

RBC, HGB, HCT, PLT, and RET number. Carryover on theXT-2000i for all parameters was of minimal magnitude. Theresults of high-to-low carryover testing were less than 1% forall determinations.

Carryover was also performed on a manipulated single-donor pheresis PLT product and a specimen with a high WBCcount to test the XT-2100i, and the results showed no signifi-cant carryover even with extremely high parameters (Table 4).

Performance Evaluation of the Sysmex XT-2000i 33

TABLE 3. Carryover*

WBC RBC HGB HCT PLT RET#

H1 12.05 5.11 15.8 47.6 645 0.1298H2 12.18 5.07 15.8 47.2 638 0.1278H3 12.07 5.15 15.8 47.8 638 0.1344L1 3.73 1.52 4.7 14.5 131 0.019L2 3.57 1.53 4.7 14.5 134 0.0252L3 3.66 1.52 4.7 14.4 129 0.0228Carryover, % 0.83 0 0 0.30 0.39 –3.41

*See Figure 1 legend for abbreviation definitions.

TABLE 4. Carryover of High Platelet and White Blood Cell Counts

Carryover Platelet Count White Blood Cell Count

Low 1 60 23.24Low 2 57 23.78Low 3 52 23.37High 1 6561 109.50High 2 6558 106.37High 3 6597 105.18Carryover, % 0.12 –0.16

TABLE 5. Linearity

Lowest Highest CorrelationParameter Value Value Coefficient Slope Intercept

White blood cell count 0.39 401.83 1.00 1.00 –1.11Platelet count 2 1474 1.00 1.01 4.23High platelet count 102 6560 1.00 1.00 100

LinearityThe linearity results looked very good on all parameters,

with special attention given to the WBC and PLT counts.Table 5 shows the WBC linearity with an upper reportablerange verified at 410 × 103/µL and the upper reportable PLTvalue of 1474 × 103/µL. A specimen was obtained from asingle-donor–pheresis PLT product, and the volume wasreduced to yield a very high PLT count of 6560 × 103/µL.Table 5 also shows the results of this high linearity.

Mixing StudiesThe data from the mixing studies indicated that even after

4 hours of sitting undisturbed, the samples were adequatelymixed on the XT-2000i and showed no significant differencefrom those specimens that were mixed thoroughly before

being analyzed. All changes in the RBC count were less than1.5%. The data are displayed in Table 6.

Comparison StudiesAs shown in Table 7, overall correlation between the XT-

2000i and the XE-2100 for all measured parameters wasexcellent, with R2 values for all parameters >0.92. The onlyexception was the MCHC, which showed an R2 value of0.8552. Only 64 of the 114 specimens used in the compari-son testing were abnormal. These specimens with unusualmeasurements were selectively picked to challenge the linear-ity of the analyzer and its capability to categorize the cells inthe automated differential. See Figure 3 for graphic displayof key parameters.

Comparison of XT-2000i with the Manual Differential Reference Method

The correlation coefficients, slope, intercept point, andnumber of comparison results are shown in Table 8. Experi-enced technologists performed the two 200-cell referencemanual differentials. The autodifferential results obtained onthe XT-2000i compared very well with the manual differen-tial results. More than half of the specimens used in thecomparison had abnormal differential findings, and excellentcorrelation between the 2 instruments was demonstrated. SeeFigure 4 for graphic display.

Morphologic Abnormality Flagging EfficiencyThe results of the reference 400-cell differential indicated

that immature or morphologically abnormal cells werepresent in 48 of the 114 samples. The XT-2000i generateda morphologic flag on 46 of these samples. Two samples,both showing NRBCs present at very low levels (less than1/100 WBCs) were not flagged. Both of these samples hadWBC counts less than 1.8 × 103/µL. No immature or mor-phologically abnormal cells were found by the manual differ-ential on the remaining 66 samples, although distributionalabnormalities were present in some cases. The XT-2000i cor-rectly identified 60 of these samples. The remaining 6 sam-ples had false-positive flags; 1 had a left shift flag, 1 had an

34 K. Langford et al

TABLE 6. Mixing Studies*

RBC 1 RBC 2 RBC 3 RBC 4 RBC 5 RBC 6 RBC 7 RBC 8 RBC 9 RBC 10 RBC 11 RBC 12 RBC 13 RBC 14

Run 1 4.1 4.45 2.23 4.88 5.7 5.49 2.97 4.49 2.79 4.86 4.13 4.63 3.92 4.45Run 2 4.12 4.47 2.26 4.87 5.69 5.49 2.98 4.51 2.82 4.86 4.11 4.61 3.95 4.45Run 3 4.11 4.46 2.25 4.88 5.67 5.45 2.98 4.49 2.8 4.83 4.09 4.62 3.96 4.44Average 4.11 4.46 2.25 4.88 5.69 5.48 2.98 4.50 2.80 4.85 4.11 4.62 3.94 4.454 h unmixed 4.09 4.44 2.25 4.89 5.73 5.48 2.97 4.49 2.84 4.79 4.12 4.62 3.97 4.484 h premixed 4.13 4.41 2.28 4.9 5.72 5.46 2.94 4.51 2.78 4.84 4.11 4.62 3.97 4.43Change unmixed, % 0.49 0.45 –0.15 –0.27 –0.76 –0.06 0.22 0.15 –1.31 1.24 –0.24 0.00 –0.68 –0.75Change premixed, % –0.49 1.12 –1.48 –0.48 –0.59 0.30 1.23 –0.30 0.83 0.21 0.00 0.00 –0.68 0.37≤1.5% OK OK OK OK OK OK OK OK OK OK OK OK OK OK

*RBC indicates red blood cell.

TABLE 7. Comparison of XT-2000i and XE-2100 Complete BloodCount and Differential*

CorrelationParameter Coefficient Slope Intercept N

WBC 1.00 1.12 –0.63 114RBC 1.00 1.03 –0.12 114HGB 1.00 1.02 –0.16 114HCT 0.99 1.00 –0.43 114MCV 0.98 0.99 0.51 114MCH 0.99 1.03 –0.33 114MCHC 0.86 0.84 5.93 114RDW-SD 0.99 0.92 2.72 114RDW-CV 1.00 0.99 0.22 114PLT 0.98 1.04 4.55 114NEUT# 1.00 1.01 0.01 114LYMPH# 0.98 0.96 0.04 114MONO# 0.98 1.10 0.05 114EO# 0.99 1.02 –0.01 114BASO# 0.99 1.09 –0.01 114RET% 0.99 1.22 –0.22 114RET# 0.99 1.20 –0.01 114IRF% 0.97 1.19 –1.22 114

*See Figure 1 legend for abbreviation definitions.

Performance Evaluation of the Sysmex XT-2000i 35

FIGURE 3. Comparison graphs of results for white blood cells (WBC), red blood cells (RBC), and hemoglobin (HGB) obtained onthe XT-2000i and those obtained on the XE-2100.

immature granulocyte flag, 3 had abnormal lymphocyteflags, and 1 had multiple abnormal/immature flags. The XT-2000i flagging sensitivity and specificity are shown in Table 9.The XT-2000i showed very good flagging efficiency for bothadult and pediatric samples.

Separate analysis of the NRBC and blast flags also showedgood efficiency. Blasts were identified on 9 of the manual dif-

ferentials. The XT-2000i flagged 6 of these samples as havingblasts present. The 3 samples not flagged by the XT-2000i allhad a manual blast count of <1.0%. All 3 of these sampleshad other flags that would have prompted a manual review.On these 3 samples, a rare blast was reported by only 1 ofthe 2 technologists doing a 200-cell differential. On furtherreview, the referee could not positively determine whetherthe cells were blasts without performing cytochemical stains.According to the referee, the cells appeared to be abnormalimmature lymphocytes with only a rare blast. Detection of ablast cell during a manual review of 100 cells is an extremelyrare occurrence. An additional 3 samples had generated afalse-positive blast flag; however, all 3 samples had significantnumbers of pathologically abnormal lymphocytes. The blastflag alone had a negative clinical accuracy of 97% and an effi-ciency of 94.7%. NRBCs were identified on 24 of the sam-ples in a range of 0.25 to 84/100 WBCs. The XT-2000igenerated an NRBC flag on 20 of these samples. The 4 samplesnot flagged all had NRBCs present at less than 1/100 WBCs.

36 K. Langford et al

TABLE 8. Comparison Results of XT-2000i to the Manual DifferentialReference Method

CorrelationParameter Coefficient (r) Slope Intercept N

Neutrophils 0.95 0.95 3.38 114Lymphocytes 0.96 0.85 1.67 114Monocytes 0.90 1.37 1.89 114Eosinophils 0.94 0.87 0.04 114Basophils 0.76 0.48 0.24 114

FIGURE 4. Comparison graphs of results obtained on the XT-2000i and those obtained with the manual differential reference method.

A false-positive NRBC flag was generated on 6 of the 114samples. The NRBC flag alone showed a negative clinicalaccuracy of 95.5% and an efficiency of 91.2%.

DISCUSSION AND CONCLUSION

Our evaluation has shown that the XE-2100 and the XT-2000i have excellent correlation statistics for all CBC, RET,and differential values. Because the technology is very similarin both analyzers, this result is not surprising. The use of fluo-rescent flow cytometry for the CBC differential gives verygood correlation with the manual reference method.

By challenging the linearity of the instrument withextremely high- and low-count samples, we validated thelinearity of the XT-2000i for WBC counts up to 410 ×103/µL and PLT counts up to 6560 × 103/µL. This extendedlinearity could eliminate the labor-intensive and error-pronemanual dilutions required when dealing with specimenswith drastically increased values. The confidence that comesfrom knowing that PLT counts can be measured accurately

by the fluorescent optical method also reduces the need forthe technologist to perform a manual microscopic PLTcount to verify results on a specimen that contains enlargedPLTs or RBC fragments, both of which interfere with theimpedance PLT methodology. PLT testing is of great impor-tance in the pediatric population because the majority ofcapillary specimens received have a higher incidence of PLTclotting or clumping.

The technical comparison performed resulted in out-standing statistical results indicating that the XT-2000i is avery precise and accurate hematology analyzer that wouldperform well in any clinical setting. Our laboratory person-nel were impressed that the XT-2000i uses the alreadyproven Sysmex methodologies from the XE-2100 [4] andimproves on them, resulting in an instrument that is moreuser friendly and a little more compact without sacrificingany technical reliability.

REFERENCES

1. XT-2000i Product Brochure. Long Grove, Ill: Sysmex Corpora-tion of America; 2002.

2. National Committee for Clinical Laboratory Standards. Referenceleukocyte differential count (proportional) and evaluation ofinstrument method. Villanova, Pa: NCCLS; 1992. Approved stan-dard, NCCLS Document H20-A.

3. International Committee for Standardization in Hematology. Pro-tocol for evaluation of automated hematology analyzer. ICSH.1984:;6:69.

4. Gould N, Connell B, Dyer K, Richmond T. Performance evaluationof the Sysmex XE-2100, automated hematology analyzer. Sysmex JInt. 1999;9:120-125.

Performance Evaluation of the Sysmex XT-2000i 37

TABLE 9. Flagging Efficiency

Positive Negative

Manual differential 48 66XT-2000i 52 62Sensitivity 95.8%Specificity 90.9%Positive clinical accuracy 88.5%Negative clinical accuracy 96.8%Efficiency 92.9%