Performance evaluation of Mindray CAL 8000(BC‐6800 and SC‐120) hematology analyzer and slidemaker/stainer

Abstract

Introduction

The Mindray CAL 8000 is a cellular analysis line that consists of the BC‐6800, an automated hematology analyzer, and the SC‐120, an automated slidemaker/stainer. We evaluated the performances of the BC‐6800 and the SC‐120.

Methods

Four hundred and eight normal and abnormal samples were analyzed. The performance of the BC‐6800 and Sysmex XE‐2100 were compared, and blood films by the SC‐120 and manual method were compared according to the CLSI guideline H26‐A2 and H20‐A2.

Results

Most parameters measured by the BC‐6800 matched well with the XE‐2100 and manual differential. The flag efficiency of the BC‐6800 for blasts (95.3%) and atypical lymphocytes (92.6%) were higher while immature granulocytes (89.7%) and NRBCs (94.1%) were lower than that of the XE‐2100. Additionally, the BC‐6800 detected four of five samples infected with plasmodium parasites. The SC‐120 showed no carry‐over and expected repeatability. There was good agreement on the five‐part differential including abnormal cells between blood films by the SC‐120 and manually prepared blood films. The shape of the RBC was also comparable between blood films.

Conclusion

The CAL‐8000 analysis line is beneficial for precise, fast hematology work, and even more useful in malaria endemic areas.

1. INTRODUCTION

The complete blood count and leukocyte differential count are the basic, most common tests in hematology laboratories. More tests can be performed in the same time, labor reduced, and results made more precise by an advanced automated hematology analyzer and slidemaker/stainer. However, manual examination of blood films still is required to identify morphologic abnormalities as determined by flags or laboratories' own criteria. The blood film review rates vary from approximately 10%–50% depend on laboratories' review criteria.1 Reducing review of unnecessary blood film without missing true abnormal samples is the goal of automated hematology analyzers, because manual review increases laboratory cost, labor, and turnaround time. For the same reason, automated slidemaker/strainers demand quick, high‐quality blood films.

The Mindray CAL 8000 (Mindray, Shenzhen, China) is a cellular analysis line consisting of the Mindray BC‐6800 and the Mindray SC‐120. The BC‐6800 is a fully automated hematology analyzer that adopts a flow cytometric technique based on laser light scattering at two angles and fluorescence signals to perform complete blood cell count (CBC) and white blood cell (WBC) differentials. WBCs are classified by size, cellular complexity, and DNA/RNA content. The total WBC count is based on WBC/BASO channel, and the presence of abnormal cells in the sample is flagged.2 Another advantage of the BC‐6800 is the detection of plasmodium parasites in routine tests. The BC‐6800 flags and gives the number of the RBCs infected with malaria parasites in the blood samples. The SC‐120 is an automatic slidemaker/stainer that is a standalone or integrated into the CAL 8000 cellular analysis line. Only 40 μL of blood, the smallest blood requirement in the hematology industry, is required for blood smear preparation. The blood spreader angle is automatically adjusted according to the consistency of blood samples. All patient information is thermally printed on the slide, and blood films are dried before and after staining for optimizing quality and reducing contamination.

This study evaluates the performance of the BC‐6800 compared with the Sysmex XE‐2100 (Sysmex, Kobe, Japan), and the performance of SC‐120.

2. MATERIALS AND METHODS

2.1. Samples

Four hundred and eight samples were selected to evaluate the CAL 8000. Samples were collected from patients who needed a peripheral blood smear test by physicians' recommendation and from patients who had regular checkups. Two hundred of 408 samples showed flags, and 208 samples did not show any flags by the XE‐2100 used in our laboratory. This study was approved by the Institutional Review Board (IRB) of Gil Medical Center, Gachon University (GDIRB2015‐167).

2.2. Evaluation of the BC‐6800

Parameters measured by the BC‐6800 were compared with data from the Sysmex XE‐2100 and manual examination. Two hundred and eight samples without flags by the Sysmex XE‐2100 were selected for comparison study. This study was performed by the CLSI guideline H26‐A2.3 Two qualified examiners verified a 200‐cell differential on blood films according to the CLSI guideline H20‐A2 and produced 400‐cell reference differential criteria.4

The flag accuracy of the BC‐6800 for blasts, immature granulocytes, atypical lymphocytes, and nucleated red blood cells (NRBC) was evaluated by microscopic examination of blood films and compared with the Sysmex XE‐2100 on the manufacturers' default settings. Positive smear findings for blasts, immature granulocytes, and NRBCs were set at more than one abnormal cell out of 100 WBCs, and positive smear findings for atypical lymphocyte was defined as more than five abnormal cells found, following the consensus guidelines given by the International Consensus Group for Hematology Review.5 The BC‐6800 also generates flags of malaria parasite infection as “infected RBCs.” Any finding of malaria parasite infected RBC was considered a positive finding on blood films.

2.3. Evaluation of the SC‐120

The quality of blood films prepared by the SC‐120 was macroscopically and microscopically scrutinized by two technologists following the CLSI guideline H20‐A2.4 Blood films were judged as acceptable or unacceptable by two technologists.

High WBC counts with many atypical lymphocytes and smudge cells (51.36 × 10⁹/L, a case of chronic lymphocytic leukemia) and extreme low WBC (<0.05 × 10⁹/L) counts samples were selected to perform carry‐over testing. The high WBC specimen was run first, followed by the low WBC specimen. This combination was run two more times. Five fields of view in the working area were randomly selected and microscopically assessed at 100× magnification. In this case, carry‐over was defined if an atypical lymphocyte or smudge cell was found in the working area of the low WBC count sample.6

Five samples without abnormal cells were selected to evaluate the repeatability of the SC‐120. Ten consecutive blood films were prepared per sample, and two technologists performed 200‐cell WBC differential counts on every blood films. The mean, standard deviation (SD), and coefficient of variation (CV) were calculated for WBC differential parameters.6, 7 Results were defined as acceptable if the calculated SD of the neutrophil population was <6, the lymphocyte population <5 and the monocyte population <3.5

A WBC differential count including abnormal cells and RBC morphological assessment of the SC‐120 was analyzed. Two experienced observers performed morphological film review and 200‐cell WBC counts on the manually prepared films and automated films by the SC‐120. A 400‐cell WBC differential of two blood films was compared using 95% confidence intervals of 400‐cell counts were calculated according to the CLSI guideline H20‐A2.4 If 95% confidence interval for intercept includes value zero or 95% confidence interval for slope includes value one, there is no constant difference or proportional difference between two methods. A 400‐cell WBC differential count included abnormal cells such as blasts, immature granulocytes, atypical lymphocytes, and NRBCs. RBC morphological assessment was performed in size, shape, and pallor based on the ICSH recommendations.8 Abnormalities of RBC shape were noted as “present,” and the results of two blood films were compared.

2.4. Statistical analysis

For evaluation of the BC‐6800, correlations of the Sysmex XE‐2100 and manual differentials with the BC‐6800 were calculated using Microsoft Excel (Microsoft, Redmond, WA, USA). Sensitivity, specificity, and efficiency of flag performance were determined based on microscopic examination of blood films according to the CLSI recommendations.3, 4

Passing‐Bablok regression and Bland‐Altman plot were calculated and drawn using MedCalc version 16.4.3 (MedCalc software, Ostend, Belgium) for evaluation of the SC‐120. Sensitivity, specificity, and efficiency of RBC shape between films were determined based on microscopic examination. The area of under the curve (ROC) was determined from the receiver operator curve using SPSS version 13.0 (SPSS, Inc, Chicago, IL, USA).

3. RESULTS

3.1. Evaluation of the BC‐6800

The results of the comparison of the BC‐6800, XE‐2100, and manual differential are presented in Table 1. The CBC with leukocyte differential counts of the BC‐6800 is strongly associated with the Sysmex XE‐2100. The BC‐6800 and manual differential are also strongly associated with leukocyte differential counts, except basophil% (r ²=.25).

Table 1.

Correlation statistics among the BC‐6800, XE‐2100, and manual differentials

Parameter	BC‐6800 vs XE‐2100			BC‐6800 vs manual
	r 2	Slope	Intercept	r 2	Slope	Intercept
WBC	.99	1.00	0.04	N/A	N/A	N/A
RBC	.99	0.96	0.05	N/A	N/A	N/A
HGB	.99	1.01	−0.24	N/A	N/A	N/A
HCT	.98	1.01	−0.08	N/A	N/A	N/A
MCV	.98	1.08	−4.94	N/A	N/A	N/A
MCH	.96	1.00	0.13	N/A	N/A	N/A
MCHC	.69	0.74	7.81	N/A	N/A	N/A
PLT	.99	1.02	−0.71	N/A	N/A	N/A
Neutrophil	.98	0.96	4.59	.86	087	7.73
Lymphocyte	.98	0.98	−0.98	.89	0.88	4.10
Monocyte	.89	0.79	0.31	.69	0.77	1.41
Eosinophil	.95	0.98	0.08	.80	0.95	0.33
Basophil	.72	0.87	0.16	.25	0.35	0.28

WBC, white blood cell; RBC, red blood cell; HGB, hemoglobin; HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; N/A, not applicable.

By microscopic examination of blood films, 22 samples showed more than 1% blasts. Fifty‐five samples with more than 1% immature granulocytes, eight samples with more than 5% atypical lymphocytes, and 21 samples with more than 1 NRBC/100 WBCs were identified. The flagging performance of the BC‐6800 and the XE‐2100 were comparable. The BC‐6800 was more efficient for blast and atypical lymphocyte flags than the XE‐2100 while XE‐2100 was more high efficient for immature granulocyte and NRBC flags (Table 2).

Table 2.

Comparison of flagging performance between the BC‐6800 and the XE‐2100

	TP	FP	TN	FN	Sensitivity	Specificity	Efficiency
Blast (22a, >1/100WBCsb)
BC‐6800	15	12	374	7	68.9	96.9	95.3
XE‐2100	20	27	359	2	90.9	93	92.9
Immature granulocyte (54a, >1/100WBCsb)
BC‐6800	45	33	321	9	83.3	90.7	89.7
XE‐2100	32	13	341	22	59.3	96.3	91.4
Aty. Lym. (8a, >5/100WBCsb)
BC‐6800	5	27	373	3	62.5	93.3	92.6
XE‐2100	7	51	349	1	87.5	87.3	87.3
NRBC (21a, >1/100WBCsb)
BC‐6800	8	11	376	13	38.1	97.2	94.1
XE‐2100	9	11	376	12	42.9	97.2	94.4

TP, true positive; FP, false positive; TN; true negative; FN, false negative; Aty. Lym., atypical lymphocyte; NRBC, nucleated red blood cell.

^aNumber of positive sample determined by manual review.

^bStandard of positive smear finding.

Five samples were diagnosed with malaria parasite infection by microscopic examination of blood films, the gold standard diagnostic method. The BC‐6800 flagged infected RBCs among four of five samples, while the XE‐2100 flagged only eosinophils in four samples.

3.2. Evaluation of the SC‐120

The macroscopic examination for length, transition, and margin yielded a low (<0.5%) unacceptable rate. All 408 blood films secured an acceptable working area. Eight blood films (2%) were deemed unacceptable due to drying artifacts. Twenty (4.9%) blood films had streaks, and stain precipitate was found in 20 blood films (4.9%) through microscopic examination (Table 3).

Table 3.

Quality assessment of blood film made by the SC‐120

Desirable quality	Acceptable	Unacceptable	% Unacceptable
Sufficient working area	408	0	0.0
Minimum 2.5 cm in length terminating at least 1 cm from the end of the slide	407	1	0.2
Gradual transition in thickness from the thick to thin areas, ending in a feather edge	408	0	0.0
Acceptable morphology with the working area	408	0	0.0
Narrower than the slide on which the film is spread, with smooth continuous side margins that are accessible for oil immersion examination	406	2	0.5
No artifact introduced by the technique	400	8	2.0
Minimum distributional distortion	407	1	0.2
A far end that becomes gradually thinner, without grainy streaks, troughs, or ridges, all of which indicate an increased number of WBC carried into this area	388	20	4.9
Stain precipitate	388	20	4.9

There was no carry‐over on the SC‐120. No atypical lymphocytes or smudge cells were seen in the low WBC blood films, while about 150 atypical lymphocytes and 300 smudge cells were found in the high WBC blood films. Neutrophils and monocytes were also found in the high WBC blood films, not in the low WBC blood films.

The mean %, SD, and CV of three‐part differential in each of the five samples were calculated for repeatability. The calculated SDs of all parameters were <2.

For comparability of WBC differential, scatter diagrams with regression lines (solid line) and Bland‐Altman plots were drawn on a graph of manually prepared films versus automated films by the SC‐120 (Fig. 1). Regression lines were almost fitted to identity lines (dotted line), and intercepts and slopes were closed to zero and one. Confidence intervals for the regression lines were dashed. On Bland‐Altman plot, mean of differences were all closed to zero and there were no definite bias although outliers were identified. The RBC shape was comparable between the manually prepared films and automated films by the SC‐120, with high sensitivity (94.1%), specificity (97.4%), efficiency (96.3%), and AUC (95.8%).

Figure 1.

Graphs of qualifying counts for WBCs with 95% confidence intervals and Bland‐Altman anaysis between manually prepared blood films and blood films made by the SC‐120. (a) Neutrophil. (b) Lymphocyte. (c) Monocyte. (d) Eosinophil. (e) Basophil. (f) Blast. (g) Immature granulocyte. (h) Atypical lymphocyte. (i) NRBC

4. DISCUSSION

Manufacturers have developed automated hematology analyzers to measure parameters using electrical impedance, light scatter, or both. They also produce automatic slidemaker/stainers that can rapidly make good quality blood films. Hematology laboratories apply their own criteria to increase clinical sensitivity by reviewing blood films although manual determinations are time‐consuming and labor intensive. All of these set a goal of accurate and quick diagnosis without any omission. Achieving this goal starts with accurate automated hematology and good blood films.

We studied 408 samples using an objective protocol for evaluation of the Mindray CAL 8000. The CAL 8000 is a cellular analysis line that consists of the BC‐6800, an automated hematology analyzer, and the SC‐120, an automatic slidemaker/stainer. Our study showed that the CAL‐8000 performed well for goal of hematology laboratories.

The performance of BC‐6800 was comparable with the XE‐2100. The association between the BC‐6800 and XE‐2100 was good. Lippi et al.9 reported similar results among the BC‐6800, Siemens Advia 2120 (Diagnostic Solutions, Milan, Italy) and manual differentials. Recently Ciepiela et al.10 reported a poor level of agreement among the BC‐6800, Sysmex XN‐2000 (Sysmex, Kobe, Japan), and Beckman Coulter LH750 (Beckman Coulter, Miami, FL, USA) with pediatric blood samples. The BC‐6800 and XE‐2100 was weakly associated with manual differential for basophil% (r ²=.25 and r ²=.17) because basophils are a very low proportion of total WBCs. For MCHC, the association between the BC‐6800 and XE‐2100 was low (r ²=.69). MCHC are calculated using MCV, which is directly measured through the electrical impedance method. Therefore, MCHC values are more variable because the MCHC value has a narrower range than MCV.11 For the flagging performance on the manufacturers' default setting, flags of blast and atypical lymphocyte were more efficient with the BC‐6800 compared with the XE‐2100, while left shift flags were less efficient (Table 2). Although the BC‐6800 was more specific than the XE‐2100, it was less sensitive than the XE‐2100 on blast and atypical lymphocyte. These results should be helpful for laboratory criteria when the hematology laboratories first set the BC‐6800 up. We did not separate benign and malignant atypical lymphocytes in this study, and atypical lymphocyte samples were few. Further studies with large samples with atypical lymphocytes, and separated analysis should be helpful for benign and malignant atypical lymphocytes. The BC‐6800 performed comparably with the XE‐2100 for NRBCs.

Another advantage of the BC‐6800 is the malaria parasite infection flag. This study included five samples with malaria parasite infection confirmed by manual examination, the gold standard method. The BC‐6800 showed infected RBC flags on four of five samples, while the XE‐2100 showed eosinophil flags. One sample not flagged was a follow up sample after treatment with less than one infected RBC/1000 RBCs. The infected RBC flag function is only in the BC‐6800. Therefore, the BC‐6800 should be very useful for screening malaria infection in endemic areas.

This study was the first performance evaluation of the SC‐120. Accurate assessment of microscopic examination starts with good quality blood film preparation, and the quality of the staining reaction is dependent on the appropriated proportion of stain and buffer and the precise timing sequence of each step.6 Blood films by the SC‐120 were excellent, except for streaks and stain precipitate (Table 3). There was no definite interference for morphological assessment although stain precipitate was inevitable during process and streaks were placed by stain precipitate. However, for future studies, the stain should be more delicately adjusted before evaluation.12 There was good agreement between the SC‐120 and the manual differentials for WBCs including abnormal cells (Fig. 1). RBC morphology was strongly associated between the manually prepared films and automated films by the SC‐120, but blood films made by the SC‐120 showed more poikilocytes. This may be because blood films by the SC‐120 were made after blood films by Sysmex and manually prepared blood films.

5. CONCLUSION

The performance of BC‐6800 was comparable with the XE‐2100, and has a further advantage of detecting infection of malaria parasite. The SC‐120 agreed well with manual differential and morphology assessment although the SC‐120 needed the adjustment for stain precipitate and steaks. Therefore, the CAL‐8000 is effective for precise, fast hematology work, and has useful application for malaria endemic areas.