Introduction

The analysis of body fluids (BF)—pleural, ascitic, cerebrospinal fluid (CSF)—is an important diagnostic modality that gives evidence of malignancy12 and therefore it has vital role in the management of cancer patients. The gold standard of BF examination is manual hemocytometer followed by cytomorphological examinations of BF with Papanicolaou3 and other stain like hematoxylin and eosin. The manual methods have high turnaround time and large interobserver variability; therefore, automated BF analysis is useful as it enables laboratories to cut short the turnaround time affording rapid diagnostic information to clinicians.456 The BF module on the XN-1000 (Sysmex Corporation) represents a convenient platform for accurate and fast quantification of total nucleated cells.7 The BF mode (Sysmex-1000) differentiates cells as neutrophils and eosinophils, lymphocytes, and monocytes. This channel records high fluorescence body fluid (HF-BF) cells, which are not included in the white blood cell (WBC) count, but are included in the total nucleated cell count.4 HFCs have higher fluorescence signal because of large amount of nucleic acids. The HF-BF cells are segregated by their differential fluorescence intensity, and they include malignant cell, mesothelioma cells, and macrophages.

The cutoff percentage of HF cells is not defined in literature, and it therefore is a limitation in the use of hematology (XN-1000) as a screening tool. This study was aimed to reliably estimate the cutoff for HF cells. More such studies are needed to validate the findings of this study.

Materials and Methods

A descriptive type of observational (validation) cross-sectional study was conducted over a period of 1 year, extending from January 2019 to May 2020, in the department of pathology in a tertiary care hospital in west India. This study group comprised of patients who submitted BFs sample such as CSF, pleural fluid, and ascitic fluid to the laboratory for examination for malignant cells. This study was approved by institutional ethics committee and subjects were included in study after written informed consent was obtained from the participants.

The following exclusion criteria were applied:

• Cases with inadequate quantity of sample for both manual count and automated analysis

• Samples were excluded if more than 10% degenerated cells were identified in the counting chamber

• If WBC count using the manual method was less than 5 cells/µL in non-CSF samples.

A total of 120 specimens of BFs, including 53 (44.17%) ascitic fluid, 46 (38.3%) pleural fluid, and 21 (17.50%) CSF, were analyzed on BF mode on Sysmex XN-1000 (Sysmex, Kobe, Japan). Sample collection, preservation, transport, and analysis were performed in accordance with the Clinical & Laboratory Standards Institute (CLSI) H56-A guideline.7 The cell identification was done by both manual and automated methods. All fluid samples were collected in K2-ethylenediaminetetraacetic acid anticoagulants tubes or sterile recipients without anticoagulant. Sample were received by laboratory immediately after collection and automated analysis and slide preparation was completed within 2 hours. Cytology section received separate sample and manual microscopy was done on slide stained by special stains.

The samples were first run in the BF mode in XN-1000automated hematology analyzer. Manual microscopic counting was performed using a Neubauer hemocytometer using samples diluted 1:10 (50 μL CSF + 450 μL of solution) in Turk's solution (crystal violet and glacial acetic acid dissolved in distilled water). Cell counting was performed at 200× magnification using a standard light microscope. The average number was converted into cells/ μL according to the following formula provided in the CLSI H56-A guidelines.7 The differential WBC counts were performed after sample cytocentrifugation (7 minutes at 1250 rpm) and later smear preparation stained with Wright staining. At least 100 cells were counted at 400× magnification.

The results of automated XN 1000 analyzer were compared with manual method for BF examination. The statistical analysis was done using online free statistical calculator (Medicals).

Results

This study comprised of 160 samples of BF received in the department of pathology between January 2019 and May 2020. The majority of samples were from the age group 41 to 50 years (20.83%) followed by 51 to 60 years (20.0 %). The mean age group for this study population was 52.48 years (Fig. 1). Out of 120 patients, 70 patients were male and 50 were female with male female ratio of 1.4:1 (Fig. 2).

Close

Fig. 1: Distribution of the cases according to age groups.

Export to PPT

Close

Fig. 2: Distribution of the cases according to sex.

Export to PPT

There were three types of BF samples: 53 ascitic fluid (44.17 %), 46 pleural fluid (38.33%), and 21 CSF (17.50%; Fig. 3). Fifty (41.67%) fluids had malignant cells on cytopathological examination (Fig. 4).

Close

Fig. 3: Distribution of the cases according to type of fluids. CSF, cerebrospinal fluid.

Export to PPT

Close

Fig. 4: Distribution of the cases according to malignant status.

Export to PPT

Of the 53 ascitic fluids, 24 (45.3%) samples were reported as malignant, and 29 (54.7%) samples were reported as nonmalignant cells. Out of 46 pleural fluids, 24 (52.2%) samples were reported as malignant and 22 (47.8%) as nonmalignant. For 21 CSF only, 2 (9.5%) were reported as malignant and 19 (70.5%) CSF as nonmalignant (Fig. 5). Mean value of total WBC count for malignant sample was higher than nonmalignant samples (statistically nonsignificant as p-value more than 0.05; Table 1).

Close

Fig. 5: Distribution of the cases according to malignant status with type of fluid. CSF, cerebrospinal fluid.

Export to PPT

In our study, the mean polymorphonuclear cells (PMN)% for malignant and nonmalignant samples were 34.80 and 36.03%, respectively, with p-value 0.78. There is no significant correlation found in between PMN% and malignant status of fluid (statistically in significant, p-value 0.96; Table 2).

In this study, mean mononuclear cells (MN)% for malignant and nonmalignant samples were 58.40 and 62.57%, respectively. MN% was slightly higher in nonmalignant sample in our study; however, the results (statistically in significant, p-value >.05; Table 3). Table 4 shows mean WBC count for malignant and nonmalignant BFs—ascetic, pleural, and CSF.

The mean WBC count for malignant fluid was 995.83 and for nonmalignant fluid it was 254.48. For pleural fluid, mean WBC count for malignant fluids was 3433.33 and for nonmalignant fluids, it was 802.27. Similarly, for CSF mean count was 645 and 21.16 for malignant and nonmalignant fluids, respectively (Table 4).

The mean absolute HFC count for malignant fluids was 450 and 16.40/µL for nonmalignant fluids in all BFs. For ascitic fluid, mean HFC count for malignant/nonmalignant cell was 458.05 and 12.37/µL, for pleural fluid mean HFC count for malignant/nonmalignant cell was 467.58 and 29.49/µL, and for CSF it was 164.50 and 7.41/ µL, respectively (Table 5).

Mean absolute HF% for all malignant fluids was 21.1 and 9.9% for nonmalignant fluids. Mean HF% for ascitic fluid was 18.6 and 7.0% for malignant and nonmalignant fluids, respectively.

Mean HF% for pleural fluid was 20.7 and 16.9% for malignant and nonmalignant fluids, respectively. Mean HF% for CSF was 55.4 and 6.1% for malignant and nonmalignant sample, respectively (Table 6).

The receiver operating characteristic (ROC) curve analysis was performed to determine the optimal cutoff values of significant variables (Hf#) detected between the two groups ROC for Hf# showing (1-specificty) on the X axis and sensitivity on Y Axis exercise different cut off value to provide the greatest sum of sensitivity and specificity (Fig. 6).

Close

Fig. 6: Receiver operating characteristic (ROC) curve for the optimal cutoff value of absolute high fluorescence cells count.

Export to PPT

The optimum cutoff value was obtained by points of test values that grants the highest Youden Index comes out to be 15.70% with a sensitivity of 82% and a specificity of 81% (Table 7). Similarly, for HF%, the optimal cutoff was 5.05%. The sensitivity and specificity of 72% were determined with SE 0.039 (Table 8; Fig. 7).

Close

Fig. 7: Receiver operating characteristic (ROC) curve for the optimal cutoff value of relative high fluorescence cells count.

Export to PPT

Discussion

The past two decades have seen rapid advancement in automation in the field of laboratory medicine, from high output automatic analyzers to the use of machine learning algorithms of artificial intelligence.891011121314151617 The scarcity of trained manpower, need for rapid turnaround time, and minimal subjective bias are some of the factors that have driven manufacturers to bring new equipment for automated analysis. This study included three types of BFs similar to studies by Rastogi et al18 and Cho et al19 for performance evaluation of automated BF analyzer using HF-BF parameters as discriminators, whereas others like Wu et al,20 Wrong et al,21 and Zimmerman et al22 used single type of BF for analysis (Figs. 8,9,10).

Close

Fig. 8: Ascitic fluid—(A) Benign scatter gram, (B) malignant scatter gram, (C) benign microscopy, and (D) malignant microscopy.

Export to PPT

Close

Fig. 9: Pleural fluid—(A) Benign scatter gram, (B) malignant scatter gram, (C) benign microscopy, and (D) malignant microscopy.

Export to PPT

Close

Fig. 10: Cerebrospinal fluid—(A) Benign scatter gram, (B) malignant scatter gram, (C) benign microscopy, and (D) malignant microscopy.

Export to PPT

The total WBC count in this study was higher for malignant BFs than nonmalignant ones. The finding is similar to studies by Huang et al,23 Xu et al,24 Labaere et al,25 and Sun et al.26

As per ROC analysis, the HF% cutoff of 5.5% equivalent to HF cell count of 15.70/µL should yield good sensitivity and specificity for malignant cell detection in this study that could be used as discriminator. The comparison with other literature studies is shown in Table 9. The findings of the study are in line with those by Rastogi et al,18 Buoro et al,27 Huang et al,23 Xu et al,24 and Larruzia et al28 had higher cutoff value determination than this study. The use of automated analyzer is mainly as a screening tool, whereas the cytological examination remains the gold standard. The cutoff selected in this study is justified as the automated analysis being used as screening tool should not compromise on sensitivity. The results are comparable to studies by Labaere et al,25 Cho et al,19 and Zimmerman et al22 who found false positive similar to this study—11 malignant cases out of 70 nonmalignant cases. They also reported interference by histiocytes, mesothelial cells, and plasma cells. In terms of false negatives, there were 7 out of 50 cases; majority of them were CSF samples that have low cell yield. There was no carry over of WBC in this study as reported by Caroline et al.9 The advantage of automated BF analyzer could not be undermined as the analysis is quick and reliable as stressed by Caroline et al29 and Aguadero et al.30

We acknowledge that the limited sample size and inability to capture time interval between sample collection and analysis were limitations of the study. It mandates that in the absence of clear-cut discriminatory value for HF cells to detect malignant cells, each laboratory should establish its own cutoff in BF analysis on automated analyzer.

Conclusion

Automated BF analysis using Sysmex XN-100031 could be a sensitive tool in predicting the nature of BF offering quick screening as adjuvant to routine cytological examination.