Cytological examination of fluid is often helpful to differentiate between benign and malignant effusions (Sujathan et al., 1996). However, there are various major diagnostic issues in the cytologic interpretation of effusions. They commonly contain abundant reactive mesothelial cells, histiocytes, and lymphocytes. If the effusion contains rare malignant cells, those can be often obscured by the relative overabundance of other cellular elements and may not be readily detectable on microscopic examination. Malignant cells are often exfoliated as single cells or minute tissue fragments and thus, a relative lack of cellular architecture may hamper an accurate cytologic assessment (Sujathan et al., 1996). The exfoliated foreign cells and the mesothelial cells in the fluid may mimic one another. Moreover, neoplastic cells may significantly change their appearance after a prolonged time of suspension in the fluid. Cells may appear more rounded and cytoplasm may develop pseudo vacuoles. So, most of their morphologic resemblance to the primary tumor could be lost. Determining primary cancer based on only effusion analysis becomes a challenging task due to inadequate patient history or information about prior malignancy. 

Due to morphologic similarities, the cytomorphologic distinction between reactive mesothelial cells, malignant mesothelioma, and metastatic adenocarcinoma can often be extremely difficult. The majority of patients with suspected cancers or known malignancies routinely undergo cytological evaluation. Cytologic examination of serous cavity fluid is a very useful diagnostic tool and is considered to be a highly accurate diagnostic procedure. The sensitivity of conventional cytology for the detection of malignant cells varies from 50-78%. Several ancillary diagnostic methods have been proposed to increase the diagnostic accuracy for the detection of malignant cells (Karki et al., 2012). If cytomorphology is combined with immunocytochemistry, the sensitivity increases from 84 to 94%, and the specificity increases from 92 to 100% (Ali & Cibas, 2012; Mohammad et al., 2021). 

Ancillary techniques like cell block, image analysis, and flow cytometry have proved to be useful in the detection of benign and malignant effusion. However they are not readily available in most of the laboratories (Fagere, 2016). Interest therefore has focused on identifying dependable methods for supplementing conventional smear methods to differentiate malignant cells from benign ones. One such area under investigation is nucleolar organizer regions (NORs) (Sujathan et al., 1996). NORs are chromosomal loops of DNA that are involved in ribosomal synthesis. A comparatively simpler technique used for this purpose is the silver (Ag) staining of nucleolar organizer regions (NORs). Interphase NORs are the structural and functional units of the nucleolus. It contains all the essential components for the synthesis of ribosomal RNA. NORs are located in each of the short arms of the acrocentric chromosomes 13, 14, 15, 21, and 22.6. NORs are argyrophilic because they are associated with two argyrophilic proteins. Nucleolin and nucleophosmin are argyrophilic proteins and are easily stained by silver stains. After silver-staining, the NORs can be identified as black dots present throughout the nucleolar area. The number and size of NORs reflect cell activity, proliferation, and transformation that help to distinguish benign from malignant cells. Evaluation of the quantitative distribution of AgNORs has been applied in tumor pathology both for diagnostic and prognostic purposes. Several studies carried out in different tumor types demonstrated that malignant cells frequently present a greater AgNOR count than corresponding non-malignant cells (Akhtar et al., 2004).

Now a day, in countries like ours only conventional smears are made in almost all of the laboratories. Cell blocks are particularly used when the cytological abnormality is misleading. Cellblock is a good technique that usually requires some extra work which is not needed in the conventional smear method and AgNOR staining method. The study has been undertaken to assess the utility of AgNOR staining in the cytological diagnosis of suspected malignant effusions and compare the diagnostic efficacy of the conventional cytological method with the AgNOR count method in effusion.

This cross-sectional observational study was conducted in the Department of Pathology, BIRDEM General Hospital, Dhaka from July 2019 to June 2021. In this study 115 samples of effusion were included. All the samples were centrifuged and then smears were prepared from the deposit followed by staining with Hematoxylin and Eosin stain, Papanicolaou stain and AgNOR stain. At first the diagnosis was made on conventional smear method. Then the findings were compared and analyzed by AgNOR staining method. AgNOR count, variation in size and dispersion of AgNOR dots as well as proliferative index (pAgNOR) were graded and compared in malignant and non-malignant effusions. Relevant clinical data including age, sex, site of effusion and known history of benign or malignant neoplasm were collected and recorded in a predesigned data collection sheet. Statistical analysis was carried out as required. Ethical practice was ensured in every step of the study.

Sample selection criteria 

Inclusion criteria

Sample of effusion received in the Department of Pathology during the study period. 

Exclusion criteria

1. Patients unwilling to participate in the study.
2. Scanty fluid samples when cell block preparation cannot be done.
3. Poorly preserved material

Laboratory Methods

After receiving the fluid sample, without agitation it was decanted in a 60 mL beaker and the remaining fluid was saved in another container. Then 60 mL fluid was distributed in 3 tubes. The effusion samples were centrifuged. From the deposited cells smears were prepared and fixed according to the method of staining. 

A. Conventional Smear Preparation 

• After transferring the fluid from to centrifuge tube labeled with the specimen identifier then was centrifuged for 5 minutes at 2000 rpm. 
• The supernatant fluid was discarded and the sediment was taken on the slide with the help of a glass rod and spread by the thick and thin method. 
• Three smears were prepared. 
• Two slides were fixed in 95% ethanol and stained with Papanicolaou and H & E stain. The remaining smear was further processed to stain with AgNOR.

B. AgNOR staining Procedure

a. Materials for AgNOR staining

• Silver nitrate
• Gelatin
• Formic acid
• 3:1 mixture of absolute alcohol: acetic acid
• Deionized water
• Distilled water
• Xylene

b. Making of AgNOR solution

Solution A: 50% aqueous silver nitrate

Solution B: 2 gm gelatin and 1 mL formic acid in 100 mL deionized distilled water.

AgNOR solution

2 parts of solution A+ 1 part of solution B

Method of staining

The smear was post-fixed in 3:1 ethanol: acetic acid mixture. It was brought to deionized distilled water through graded alcohols, and covered with filter paper. Smears were covered with 10 drops of working silver staining solution in a dark humidity chamber at room temperature for 30-40 minutes. Smears were washed by covering them with a layer of distilled water 3 times, 5 minutes for each. Dehydrated through alcohol series, clear in xylene, and mounted in DPX. AgNOR stained smear was examined under the light microscope. The nuclei stained light yellow. The AgNORs were visualized as brown-black dots of variable size within the nuclei. Only nuclei of mesothelial, epithelial, or malignant cells were evaluated. Inflammatory cells (PMNs, lymphocytes, and macrophages) were excluded. AgNOR counting was performed under 100x objective using oil immersion.

Scoring system

1) Mean AgNOR (mAgNOR)

The mean number of AgNORs in 100 cells was calculated.

2) Proliferative index (pAgNOR)

The percentage of nuclei exhibiting 5 or more AgNOR dots/ nucleus/ 100 cells called the proliferative index (pAgNOR). The size variation and distribution of AgNORs was performed by the following criteria

Size variation grading 

0 = More or less uniform in size. 

1+ = Two different sizes. 

2+ = More than two different sizes (but not those of 3+). 

3+ = All grades and sizes including too minute will have to be counted.

AgNOR dispersion in the nuclei 

0 = Limited to nucleoli. 

1+ = Occasional dispersion outside nucleoli. 

2+ = Moderate dispersion outside nucleoli. 

3+ = widely dispersed throughout the nucleus.

Data Processing and Analysis

Statistical analysis was performed by applying the SPSS-PC package, version 23 (Statistical Package for Social Science). P values less than 0.05 were considered significant. Analysis of the results of methods (conventional smear and AgNOR staining) was performed using the Chi-square test, Unpaired t-test, ANOVA test, Mann-Whitney U test, and McNEMAR test. The results were published in tables, pie charts, and bar diagrams. 

Table 1: Age distribution of the study cases (n=115).

Fig. 1: Pie chart showing sex distribution of study cases.

Fig. 2: Bar diagram showing the distribution of study cases according to the type of effusion.

Table 2: Mean AgNOR count (mAgNOR) of different groups of study cases diagnosed by conventional smear method (n=115).

ANOVA test was carried out to measure the level of significance.

Table 3: Comparison of study cases diagnosed by conventional smear and AgNOR count method (n=115).

* Figure within the parentheses indicated in percentage

Table 4: Comparison of suspicious cases diagnosed by conventional smear and AgNOR count method (n=9).

Table 5: AgNOR dispersion of different groups of study cases (n=115).

*Chi-Square test was carried out to measure the level of significance 

** Figure within the parentheses indicated in percentage

The diagnosis of malignancy in effusion is often difficult. The challenge is either differentiating malignant cells from macrophage and reactive mesothelial cells or due to subtle cytomorphological features of some malignant neoplasm. Indeed, it is not always possible to distinguish neoplastic cells from reactive mesothelial cells on purely morphologic features. The problem becomes more compounded due to the artifacts that are caused by fixation, preparation, or staining techniques (Gill et al., 2011).  The age range in the current study population was from 11 to 82 years with a mean age of 55.20±12.76 years. The dominant age of the study population was among the age group of 51-60 which constituted 44 (38.3%). These findings are similar to the studies performed in Nepal and Sudan in which the mean age was 52.7 (ranging from 1-88 years) and 52 (ranging from 20-72 years) respectively (Shulbha and Dayananda, 2015). In the current study, the most common fluid was peritoneal fluid 62 (53.9%) followed by pleural 48 (41.7%), pericardial 2 (1.7%), CSF 2 (1.7%) and synovial 1 (0.9%). This data was consistent with the other studies performed in India, Nepal, and Sudan (Shulbha and Dayananda, 2015; Karki et al., 2012 Fagere, 2016). In our study, 115 cases of effusions were evaluated by conventional smear which comprised of benign effusions (48 cases), malignant effusions (58 cases), and a third group consisting of suspicious for malignancy (9 cases). In the latter group, the suspicious cases could not be classified with certainty as to whether they were reactive mesothelial cells or malignant cells. 

This is similar to the study done by Sujathan and her colleagues in India where 37 benign cases, 55 malignant cases, and 8 atypical cases were encountered. In the study of Gill et al. (2011), the number of benign to malignant cases was reversed; i.e. 57 cases were benign and 28 cases were malignant, whereas 15 cases were atypical. The present study validates the diagnostic utility of AgNOR staining of serous effusion. All effusions were subjected to AgNOR staining. The benign group consisted of cells showing 1 to 2 dots which were regular in size and shape. In the malignant group, more than three irregular dots as many as more than twenty dots were observed per cell distributed within the nucleus. The dots had variable size, shape, and irregular contours. In the suspicious group, the reactive mesothelial cells showed 1 to 2 dots and malignant cells showed 3 to 4 irregular dots. In this way, clear separation could be achieved between reactive mesothelial cells and malignant cells. Similar observational views had been put forward by Crocker et al. (1989). After AgNOR counting of three different groups of study cases (diagnosed by conventional smear), the mean ± SD (AgNORs count /100 cells) in malignant effusion cases (5.59 ± 1.05) were found to be higher than benign effusion cases (1.31 ± 0.48) and atypical/ suspicious cases (4.54 ± 1.49). A statistically significant difference is found among the Means of different groups. Sujathan et al. (1996) observed a mean value of 1.92 ± 0.23 for benign, 4.72 ± 0.76 for malignant, and 3.74 ± 1.50 for atypical cases. In a similar study, Gill et al. (2011) found a mean count of 1.53 ± 0.15 for benign cases, 4.03 ± 0.38 for malignant cases, and 3.39 ± 0.59 for atypical cases.  Karki et al. (2012) reported a mean count of 2.12 ± 0.54 in benign effusion, 10.43 ± 0.73 in malignant effusion, and 8.77 ± 2.97 in atypical cases. When compared our findings are in accordance with the findings of other researchers.

A wide variety of abnormalities can produce effusions. Most often effusions are caused by tumors or inflammation of the serous membrane. The role of cytopathology is most significant in the evaluation of fluid for malignant cells. The presence of malignant cells in effusions has an important therapeutic and prognostic implication. Malignant serous effusions are a commonly encountered early clinical manifestation of metastatic disease and often only a clinical clue of an unknown primary. Therefore, assessment effusion has been routinely used in suspected cases of malignancies. The diagnosis of malignancy in effusion is often critical. The challenge is either differentiating malignant cells from macrophage and reactive mesothelial cells or due to fine cytomorphological features of some malignant neoplasm. Indeed, it is not always possible to distinguish neoplastic cells from reactive mesothelial cells on purely morphologic features.

The problem becomes more critical due to the artifacts that are caused by fixation, preparation, or staining techniques. The AgNOR count method has a role in differentiating benign from malignant effusion, they do not supersede the value of routinely used conventional smears as there is very little difference in their accuracy. This method has a supportive value that can be utilized in differentiating malignant effusions from benign ones, especially in suspicious cases. So, our study illustrates that AgNOR staining has a supportive value which can be utilized in differentiating malignant effusions from benign ones, especially in suspicious cases. It can be employed as an inexpensive and rapid additional diagnostic tool for effusions when the cytological diagnosis poses a dilemma.

Before the commencement of this study, the thesis protocol was submitted to the Institutional Review Board (IRB) of BIRDEM General Hospital, Dhaka for approval.

N.R. performed laboratory work, N.R.; and M.R. prepared the manuscript, and S.U.Z. completed data analysis and made the framework.

We acknowledge the stuff of the BIRDEM General Hospital, Pathology laboratory for their cooperation and during data collection.

Authors declare that no competing interest exists to publish the present research work.