| Abstract|| |
Introduction: Automated body fluid (BF) analysis is gradually replacing the traditional methods of cell counting in all BFs. This study was done to analyze the high-fluorescence (HF)-BF parameter generated on Sysmex XN-1000 and study its correlation with the presence of malignant cells in the body fluids. A correlation between manual and automated differential counts was also done. Materials and Methods: A total of 1985 samples including 797 ascitic fluids (AF), 532 pleural fluids (PF), and 656 cerebrospinal fluids (CSF) were run on Sysmex XN-1000 in BF mode and cytopathology was available for 924 BFs including 389 AF, 379 PF, and 156 CSF. Both manual and automated methods were used for cell differential and cell morphology. Results: Of the 924 samples with corresponding cytopathology, malignancy was found in 59 samples. The HF-BF%/100 WBCs (24.8 ± 72.5) and HF-BF#/μL (329.86 ± 932.35) for malignant BF samples were found to be significantly higher than the nonmalignant samples (4.41 ± 8.1) and (19.57 ± 61.91), respectively. Receiver–operator-characteristic curve cutoffs for all BF for percentage and absolute HF-BF were 2.85%/100 WBCs and >12/μL. A good correlation was found between the manual and automated WBC differential counts in all fluids except CSF with total count <5/μL. Conclusions: BFs can be reliably analyzed on automated analyzers. HF-BF parameter is helpful in identifying malignant samples but cannot be totally relied upon. If HF-BF%/# are above the lab-generated cutoffs, microscopy should be done. A complete validation study on HF-BF parameter in BF mode is desired to set the standards for the analysis of serious effusions.
Keywords: Automated body fluid analysis, body fluid, HF-BF, high fluorescent
|How to cite this article:|
Rastogi L, Dass J, Arya V, Kotwal J. Evaluation of high-fluorescence body fluid (HF-BF) parameter as a screening tool of malignancy in body fluids. Indian J Pathol Microbiol 2019;62:572-7
|How to cite this URL:|
Rastogi L, Dass J, Arya V, Kotwal J. Evaluation of high-fluorescence body fluid (HF-BF) parameter as a screening tool of malignancy in body fluids. Indian J Pathol Microbiol [serial online] 2019 [cited 2020 Nov 27];62:572-7. Available from: https://www.ijpmonline.org/text.asp?2019/62/4/572/269086
| Introduction|| |
Body fluid (BF) analysis is an essential tool in the diagnosis, management, and prognosis of several diseases. Though exfoliative cytology should be done in all effusions, the importance of total cell count and differential cell counts cannot be overlooked. The latter carry an enormous diagnostic significance to know the degree and type of inflammatory response, presence of neoplastic cells, and hemorrhage. Neutrophilic predominance in a body fluid sample suggests acute stage of inflammation while lymphocytic effusion signifies a chronic process., For this reason, the cytological study of the BFs must be precise and accurate to avoid any error while reaching a diagnosis. However, this carries a high degree of imprecision, delayed results, and requires skilled personnel. Shifting to automated analysis reduces the inter-observer variability and improves the turnaround time and precision of the results. Many automated analyzers have a specific BF module explicitly dedicated to the evaluation of cell counts in BF.,,
BFs are frequently sent to the laboratories when malignant effusions are suspected. Traditionally, the conventional microscopic analysis of the BF for both cell count (in Neubauer chamber) and neoplastic cells is regarded as the gold standard. Normally, effusions and BF where the suspicion of a malignancy is high are evaluated in cytopathology laboratories with special stains and adjunctive tools like immunostaining. However, this process has a higher turnaround time than the basic evaluation of the BF for cell counts. The modern cell counters have shown preliminary capability to detect malignant cells during the process of generation of the differential counts with the advantage that these equipments are available in most hematology laboratories 24 × 7 and results can be obtained within an hour of sample collection.
The performance evaluation of BF mode and malignant cell screening have been reported in a few studies.,, However, very few reports on the validation of malignant cell screening of BF have been published till date., The Sysmex XN-1000 is an advanced hematology analyzer, which is equipped with a dedicated BF module (XN-BF) for BF analysis. It utilizes semiconductor laser flow cytometry and nucleic acid fluorescence techniques to study BFs. The various parameters measured for BF analysis are total nucleated cell count (TC-BF), white blood cell (WBC) count, differential cell count for polymorphonuclear cells (PMNs), and mononuclear cells (MNs). Other research parameters are high-fluorescence-body fluid (HF-BF) cells, eosinophils (EO-BF), neutrophils (NE-BF), lymphocytes (LY-BF), and monocyte (MO-BF). The values for all these parameters are generated in both absolute (#) and percentage (%) manner. The primary objective of the present study is to analyze the HF-BF parameter (both HF-BF# and HF-BF%) given by Sysmex XN-1000 and study its correlation with the presence of malignant cells in the BFs. HF-BF includes mesothelial cells, macrophages, and malignant cells. Though these HF cells are obtained separately on the graph, still it is difficult to distinguish benign cells from malignant ones. For this a cutoff should be developed and validated for the identification of these malignant cells for triaging the BFs.
We have also attempted to find the correlation between manual and automated differential counts for neutrophils and PMNs and lymphocytes and MNs.
| Materials and Methods|| |
The present study was carried out in the Hematology and Clinical Pathology department of a tertiary care center in North India. A total of 1985 specimens of BFs, including 797 (40.1%) ascitic fluid (AF), 532 (26.8%) pleural fluid (PF), and 656 (33.0%) cerebrospinal fluid (CSF), were analyzed on BF mode on Sysmex XN-1000 [Sysmex, Kobe, Japan]. All the samples were sent in EDTA tube and were processed within 2 hours of receipt in the laboratories. The cell identification was done by both manual and automated methods.
For manual counts, the samples were centrifuged for 5 minutes at 220 ×g and sediments were smeared on the slide and stained with Giemsa stain. Manual cell count and nucleated cell classification were done by microscopy. The differential counts were expressed in percentage while the mesothelial cells, malignant cells, and macrophages were reported qualitatively.
Automated fluid processing
The samples were run in the BF mode in XN-1000 automated hematology analyzer. The MNs and PMNs were determined by flow cytometry in the differential channel. Fluorescence flow cytometry uses the forward scatter and side scatter properties of cells along with the fluorescence intensity to identify and cluster each cell population. The highly fluorescent cells are identified just above the MN cluster and are given as the HF-BF% in the research mode of the equipment. They are included in the TC-BF section rather than in the differential count. The HF-BF% parameter is reported as %/100 WBCs and HF-BF# as /μL. If the number of HF-BF cells present exceeds a preset cutoff value, a “WBC abnormal scattergram” flag is generated. These HF-BF% and HF-BF# values were noted and compared with the microscopic findings obtained by cytopathological examination. HF-BF% and HF-BF# were also studied as surrogate markers for malignant effusions. We have studied the PMN% and MN% parameters and were then compared with the manual counts. High and low BF XN-check controls were measured each day for quality control and L–J charts were plotted.
Of all 1985 samples, a corresponding cytological examination was sent for 924 samples only. The diagnoses rendered on the cytopathological examination of cytospin slides and immunocytochemistry done wherever applicable were reviewed from the medical records. HF-BF%/100 WBCs and HF-BF#/μL were seen in benign versus malignant samples based on the original cytopathological diagnosis.
SPSS v21.0 software (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. Independent t-test was done to calculate difference between continuous variables. Receiver operating characteristic curve analysis (ROC) was done to analyze the cutoff value, area under curve (AUC), specificity, and sensitivity for HF-BF%/100 WBCs and HF-BF#/μL to predict malignant BF. Correlation between the percentages of PMNs and MNs by manual and automated methods was evaluated by Spearman correlation coefficient. A P value of <0.05 was taken as significant.
| Results|| |
All 1985 samples were subjected to the automated analysis and manual differential cell counts. The details of all the samples with their types are presented in [Table 1]. Among CSF samples, 333/656 samples (50.76%) were from children. Of these, 178 patients were ≤1 year of age. The median age for CSF was 9 years (1 day to 82 years). In contrast, for AF and PF samples, the median age was 52 years (1 month to 90 years) and 53 years (7 months to 93 years), respectively. Children constituted only 4.14% (33/797) and 5.8% (31/532) of all AF and PF samples, respectively.
|Table 1: Demographic data from all body fluid samples included in this study|
Click here to view
The 924 samples (46.5%), which had a corresponding cytopathological examination, included 389 (19.5%) AF, 379 (19.0%) PF, and 156 (7.8%) CSF samples. Malignant effusions comprised 59 samples (2.9%) and included 29 AF (1.4%), 24 PF (1.2%), and 6 (0.3%) CSF samples. The remaining 865 (43.5%) cases were nonmalignant and were constituted by 360 AF (18.1%), 355 PF (17.9%), and 150 (7.5%) CSF samples. There were 1061 (53.4%) cases for which cytological examination was not done. The mean value for HF-BF%/100 WBCs for malignant BF samples was found to be 24.8 ± 72.5 and for nonmalignant samples was 4.41 ± 8.1. The HF-BF %/100 WBCs was significantly higher in malignant effusions (P = 0.001) when compared to benign effusions. The corresponding values for HF-BF#/μL for malignant samples were 329.86 ± 932.35 and for non-malignant samples were 19.57 ± 61.91 (P = 0.0001).
ROC analysis of HF-BF% and HF-BF# parameter for malignant and benign fluids
ROC curve analysis was done to evaluate the ability of HF-BF parameter in differentiating benign from malignant fluids. The cutoff values for HF-BF%/100 WBCs and HF-BF#/μL, positive predictive value (PPV) and NPV, obtained for all BFs together as well as for individual BFs are being presented in [Table 2] and [Table 3] respectively. The corresponding ROC curves are given in [Figure 1] and [Figure 2]. The negative predictive value (NPV) in HF-BF%/100 WBCs and HF-BF#/μL was 96.1% and 97.3% respectively.
|Table 2: ROC determined cutoffs for HF-BF%/100 WBCs to differentiate malignant effusions from benign effusions|
Click here to view
|Table 3: ROC determined cutoffs for HF-BF #/μL to differentiate malignant effusions from benign effusions|
Click here to view
|Figure 1: (a-d) ROC graphs for HF-BF%/100 WBCs for malignant effusions in all fluids (a), AF only (b), PF only (c), CSF (d)|
Click here to view
|Figure 2: (a-d) ROC graphs for HF-BF# for malignant effusions in all fluids (a), AF only (b), PF only (c), CSF (d)|
Click here to view
Correlations between differential counts by manual and automated methods
The correlation coefficients (R 2) were calculated between PMNs% determined by automated method and the manual neutrophil percentage and for MN cell % by cell counter method and lymphocytes by manual method. A significant positive correlation was obtained for all categories including AF, PF, and combined AF and PF. For CSF, since there was a poor correlation between the manual and automated differential counts, further analysis was performed by restricting to CSF samples with count >5/μL. Using this strategy of analysis, a higher correlation coefficient was obtained for both PMNs versus neutrophils and MNs versus lymphocytes. The detailed data and the corresponding P values are represented in [Table 4]. The correlation graphs between manual and automated differential for all fluid subtypes are shown in [Figure 3].
|Table 4: Values of correlation coefficients in all fluids between manual and automated differential counts|
Click here to view
|Figure 3: (a-d) Correlation of mononuclear cell count (MNC) and polymorphonuclear count (PMN) between manual and automated methods in AF only (a), PF only (b), CSF all samples (c), CSF with TLC >5/μL (d)|
Click here to view
| Discussion|| |
BF analysis is an important diagnostic tool for many disorders and is also required for the management and monitoring of some diseases. Since past few years, the use of automated hematology analyzers for the analysis of BFs has become quite popular. Using the fluorescence flow cytometry technology, the BF mode in Sysmex XN-1000 is exclusively used for identifying and quantifying the cells. In addition to giving the TLC and DLC, the Sysmex XN-1000 also gives a HF-BF parameter (HF-BF%/100 WBCs and HF-BF#) that corresponds to cells with a HF. The HF-BF is measured in the area of high side fluorescence. This area has been found to contain large cells like mesothelial cells, malignant cells, macrophages, and plasma cells.
The primary aim of our study was to evaluate the HF-BF parameter and to study its correlation with the presence of malignant cells in the BFs. In our study, both HF-BF%/100 WBCs and HF-BF#/μL were found to be significantly higher in malignant fluids. This is similar to the findings reported in earlier papers.,,, The cutoff value obtained for HF-BF% in all BFs excluding CSF was 3.95/100 WBCs, which was marginally lower than the value of 4.4/100 WBCs reported by Xu et al. using the Sysmex XN-1000 cell counter on pleural and ascitic fluids and also lower than the value of 6.9/100 WBCs reported by Cho et al. in their series comprising 405 CSF, AF, and PF samples. The patient selection in the latter paper was biased toward collection of a higher number of malignant fluids and benign fluids in the same duration was excluded. However, our study includes all samples obtained in the accrual period without any selection bias. The sensitivity obtained in our study was 62.3% which is lower than 79.2% reported by Xu et al., but the specificity was 64.6% which is higher than 55.8% as observed in their study.
For HF-BF#, we obtained a cutoff of >17/μL at a sensitivity and specificity of 71.7% and 74.1% respectively, to differentiate malignant from benign effusions for AF and PF samples which is similar to a previously published Belgian study on ascitic, pleural, pericardial effusions, and continuous ambulatory peritoneal dialysis fluids. They reported a cutoff of ~17.1/μL to detect malignant effusions at a sensitivity and specificity of 88% and 61% respectively. However, in their analysis, the cutoff for HF-BF%/100 WBCs obtained was 2.1%/100 WBCs with a sensitivity of 86% with a markedly low sensitivity of 46%. Although the HF-BF% and HF-BF# parameters were found to be higher in the malignant samples, there was still some overlap in the cutoffs for the benign and malignant BF. Therefore, microscopy becomes mandatory to confirm the presence of malignant cells in samples with high HF-BF% or ones with abnormal WBC scattergram. We obtained a very high NPV for malignant effusions using the ROC generated cutoffs for both HF-BF%/100 WBCs and HF-BF#/μL. All NPV values were >95% and were slightly better when HF-BF#/μL was used. This indicates that if HF-BF is not elevated beyond the specific cutoffs, such fluids may be exempted from the microscopy to screen for malignant cells after developing a laboratory policy.
We also found that the HF-BF%/100 WBCs in CSF was very low when compared to all other BFs at 0.75/100 WBCs and a sensitivity and specificity of 66.7% and 79.3%. Most studies have either not included CSF samples , when analyzing HF-BF or not mentioned the data for CSF samples separately  to draw comparisons from.
In a recent study from Japan, an attempt was made to develop a XN-BF gating algorithm by analyzing the FCS files generated on the Sysmex XN cell counters. They used two rules to identify malignant cells. The first rule was centered on identification of aggregates and clumps of malignant cells. For this the forward scatter versus forward scatter-width plot was used to discriminate aggregates from the cells flowing in a single file. The second rule was used to detect the malignant cells smaller than macrophages but with middle range of fluorescence between WBCs and HF-BF cells. All samples were not positive by both rules with ~18.5% patients detected by both rules, ~33.3% by the first rule, ~7.4% by rule 2 alone, and ~33.3% were missed despite applying the XN-BF gating algorithm. However, even with the added steps of flow cytometric analysis, the authors observed a sensitivity of only 63.0%, specificity of 87.7%, PPV of 68%, and NPV of 85.1%. They reported that true positive samples had a higher HF-BF cell number than false-negative samples and attributed the low sensitivity partially to the low number of malignant cells in false-negative samples. It is apparent that this mode of analysis will possibly improve the specificity of the identifying malignant effusions but will not impact the sensitivity when compared to the cutoffs of instrument generated HF-BF determined by ROC analysis.
A highly significant correlation was found between the automated and manual methods in PMN% and MN% in AF and PF. The correlation coefficient for MN and lymphocytes was 0.907 and for PMN and neutrophils was 0.914. However, we observed poor correlations in CSF when WBC count was <5/μL. Earlier studies have also reported good correlations between MNs and lymphocytes and PMNs and neutrophils in AF and PF samples.,, However, in another series including 32 CSF samples, a poor correlation was found between the manual counts and automated counts for PMNs and MNs in CSF samples (ρ = 0.581 and 0.734, respectively). In contrast, Cognialli et al. using the XE 5000 cell counter found good correlations of PMNs and MNs with manual values in PF (ρ = 0.83, 0.9 respectively) but in their series, PMNs in AF samples showed relatively poor correlation (0.76) while MNs had better correlation (0.81). On the contrary, Buoro et al. found PMNs to have better correlation in the serous fluids (0.90) than MNs (0.71). Cho et al. reported a good correlation between the manual and automated methods for WBC differentials in all non-CSF BFs. Even in CSF samples, the correlation was 0.80 for both PMNs and MNs. This is different from the poor correlations observed in our study as we included all CSF samples even those with a WBC count of <5/μL which were excluded in the study by Cho et al. Our correlations were ~0.742 and 0.734 for PMNs and MNs when the analysis was restricted to CSF samples with a TLC of >5/μL. In addition, Cho et al. found that the nonmalignant CSF samples with WBC count <85/μL show poor correlations (R 2 = 0.55 each) while the correlation improved to 0.96 and 0.91 respectively when the WBC count was >85/μL. This is interesting to note as even when the analysis is restricted to CSF samples with WBC count of >5/μL, the correlation observed was lesser in nonmalignant samples than what we found in our study.
| Conclusions|| |
Sysmex XN-1000 hematology analyzer BF method is capable of rapid and reliable differential count in the BFs especially AF and PF but the results tend to be suboptimal for CSF especially with cell count <5/μL. HF-BF parameter is significantly higher in malignant BFs and a value of HF-BF# and HF-BF% lower than lab-generated cutoffs had a NPV of >95%. Though helpful, it cannot be totally relied upon in view of overlap between malignant and benign cases and low sensitivity. Also, when the scatter plot is abnormal or HF-BF%/# exceeds the lab-generated cutoffs or in case of strong clinical suspicion, the morphology of the cells should be confirmed to improve the quality of the analysis. A complete validation study on HF-BF% parameter in BF mode is desired to set the standards for the analysis of serous effusions.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Buoro S, Mecca T, Azzarà G, Seghezzi M, Dominoni P, Crippa A, et al.
Cell population data and reflex testing rules of cell analysis in pleural and ascitic fluids using body fluid mode on Sysmex XN-9000. Clin Chim Acta 2016;452:92-8.
Cognialli RCR, Comar SR, Souza AM, Singer GMB. Evaluation of pleural and ascitic fluid analysis on the Sysmex XE-5000 hematology analyzer. J Bras Patol Med Lab 2017:53:150-8.
Bottini PV, Pompeo DB, Souza MI, Garlipp CR. Comparison between automated and microscopic analysis in body fluids cytology. Int J Lab Hematol 2015;37:16-8.
Conner BD, Lee YC, Branca P, Rogers JT, Rodriguez RM, Light RW. Variations in pleural fluid WBC count and differential counts with different sample containers and different methods. Chest 2003;123:1181-7.
Xu W, Yu Q, Xie L, Chen B, Zhang L. Evaluation of Sysmex XN-1000 hematology analyzer for cell count and screening of malignant cells of serous cavity effusion. Medicine (Baltimore) 2017;96:e7433.
Paris A, Nhan T, Cornet E, Perol JP, Malet M, Troussard X. Performance evaluation of the body fluid mode on the platform Sysmex XE-5000 series automated hematology analyzer. Int J Lab Hematol 2010;32:539-47.
Ai T, Tabe Y, Takemura H, Kimura K, Takahashi T, Yang H, et al.
Novel flowcytometry-based approach of malignant cell detection in body fluids using an automated hematology analyzer. PLoS One 2018;13:e0190886.
Fleming C, Russcher H, Brouwer R, Lindemans J, de Jonge R. Evaluation of Sysmex XN-1000 high-sensitive analysis (hsA) research mode for counting and differentiating cells in cerebrospinal fluid. Am J Clin Pathol 2016;145:299-307.
Fleming C, Brouwer R, Lindemans J, de Jonge R. Validation of the body fluid module on the new Sysmex XN-1000 for counting blood cells in cerebrospinal fluid and other body fluids. Clin Chem Lab Med 2012;50:1791-8.
Labaere D, Boeckx N, Geerts I, Moens M, Van den Driessche M. Detection of malignant cells in serous body fluids by counting high-fluorescent cells on the Sysmex XN-2000 hematology analyzer. Int J Lab Hematol 2015;37:715-22.
Cho YU, Chi HS, Park SH, Jang S, Kim YJ, Park CJ. Body fluid cellular analysis using the Sysmex XN-2000 automatic hematology analyzer: Focusing on malignant samples. Int J Lab Hematol 2015;37:346-56.
Fleming C, Brouwer R, Van Alphen A, Lindemans J, de Jonge R. UF-1000i: Validation of the body fluid mode for counting cells in body fluids. Clin Chem Lab Med 2014;52:1781-90.
Huang WH, Lu LP, Wu K, Guo FY, Guo J, Yu JL, et al.
Extent of agreement between the body fluid model of Sysmex XN-20 and the manual microscopy method. J Clin Lab Anal 2017;31. doi: 10.1002/jcla. 22101.
Department of Hematology, Sir Ganga Ram Hospital, New Delhi
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]