A Prospective Analysis Comparing Serous Fluid Samples Processed Using Liquid-Based Cytology Versus Conventional Cell Block and Smear Classified by the International System for Reporting Serous Fluid Cytopathology in a Philippine Public Tertiary Hospital

Introduction 

Background of the Study 

Serous effusion is the accumulation of excess fluid in the pleural, pericardial or peritoneal space. Common causes of effusion include malignancy, infection, inflammation, and mechanical injury1. A diagnosis of malignant effusion indicates end-stage disease, which, in turn, precludes surgical therapy in favor of chemotherapy or radiotherapy. In many cases, the presence of an effusion affects its AJCC tumor staging. Apart from detecting a malignancy, cytology can contribute to the prognostication and treatment of a disease by identifying its primary site or cell lineage, alongside ancillary techniques 4. An accurate cytologic diagnosis is tantamount to appropriate prognostication and treatment planning. 2 Its diagnostic accuracy, however, depends on the type of fluid submitted, the volume of the sample, the type of specimen processing, and the experience of the cytopathologist reviewing the case. This can lead to poor reliability of cytologic diagnoses across different laboratories, or even among different pathologists. 

In the Philippines, cytology laboratories typically make use of direct smear with cell block or the liquid-based cytology methods for effusion cytology diagnosis. While majority of samples for both gynecological and non-gynecological cases are processed through the former, several papers reported that despite the higher cost, using liquid-based cytology (LBC) yielded a cleaner background, better cell distribution, and better-preserved cytomorphology. 

These laboratories also reported reduced screening time, better-preserved cells in solution for longer storage, and significantly reduced air-drying artifacts than a direct smear.5, 6, 7, 8 Despite these benefits, implementing LBC requires trained cytotechnologists who are adept in instrument operation as well as experienced pathologists. One pitfall brought about by inadequate experience of pathologists and visual fatigue is that long-term microscopic examination can result in increased false positives and false negatives.9 These problems have been dealt with successfully in other areas of cytopathology and are not unique to serous effusion cytology. 

Following the footsteps of prior established reporting systems in cytology, the International System for Reporting Serous Fluid Cytopathology (ISRSFC)11 was established with the aim to develop a clear, evidence-based reporting system that would better align with recommended management. The reporting system, similar to the previously established systems, consists of five diagnostic categories4. These are: nondiagnostic (ND), negative for malignancy (NFM), atypia of undetermined significance (AUS), suspicious for malignancy (SFM), and malignant (MAL) - primary and secondary. The diagnostic categories alone should hopefully be adequate to guide treatment, but a descriptive interpretation should be reported alongside the category, whenever possible12. Each diagnostic category has clear, strict, well- defined criteria to reduce interobserver variability. Clear separation between the diagnostic categories allows easy statistical analysis of diagnostic accuracy between individuals, as well as identifying general diagnostic trends. It is hoped that the development and wide adoption of a standardized reporting system with well-defined diagnostic categories will result in reporting consistency, improved interobserver agreement, and ultimately translate to a better standard of care for patients10. 

The study by Farahani and Baloch10 serves as a landmark in compiling the most 

comprehensive data on the risk of malignancy and accuracy of each diagnostic category. This was done through a meta-analysis of the available literature and re-assignment of the outcomes utilizing the ISFRSFC categories. It is noteworthy to understand that the traditional methods of calculating ROM, such as by tissue follow-up, has its inherent flaws. Calculating ROM based on tissue follow-up tends towards overestimation due to selection bias since not all effusion cytology samples have a follow-up tissue biopsy. In addition, ROM tends to vary between individual institutions or (freelancing) cytopathologists, because of the difference in their patient base. Since this is the case, it is ideal for each laboratory to conduct their own ROM assessment based on annual statistics and follow-up clinical information, in conjunction with cytohistopathologic correlation15. In the Philippines, the severe shortage of ROM statistics for effusion cytology calls for an effort to establish institutional and thereafter, local baseline statistics. It is hoped that more refined ROM statistics may be achieved in the future with studies focusing on a specific tumor type and body site.

General Objective

The study aimed to examine serous fluid samples comparing liquid-based cytology versus routine cell block and smear and to determine the rates of malignancy when classified according to the International System for Reporting Serous Fluid Cytopathology.

Specific Objectives

Specifically, this study aimed:

1.      To classify serous fluid samples processed using both methods according to the International System for Reporting Serous Fluid Cytopathology;

2.      To describe demographics and clinicopathologic profile of patient samples;

3.      To compare the frequency of serous fluid samples in LBC and cell block and smear under each diagnostic category of the International System for Reporting Serous Fluid Cytopathology;

4.      To determine the rate of malignancy of serous fluid samples in LBC and cell block and smear under each diagnostic category of the International System for Reporting Serous Fluid Cytopathology.

5.      To compare the rate of malignancy of samples processed using routine cell block and smear with that of liquid based method in each diagnostic category of the International System for Reporting Serous Fluid Cytopathology.

Conceptual Framework

Significance of the Study

Knowing the serous fluid rates and risks of malignancy for each of the diagnostic categories within the International System for Reporting Serous Fluid Cytopathology (ISRSFC) provides clinically relevant and useful inherent information to guide appropriate clinical management and follow-up. This study impacts primarily patients with serous effusion that may have malignant etiology and hopes to guide treatment decisions by those directly involved in cancer care.

Being familiar with the International System for Reporting Serous Fluid Cytopathology diagnostic categories, risks of malignancy, diagnostic parameters and prognosis benefits the practicing pathologists and pathologists-in-training by increasing their diagnostic acumen, improving approach to the diagnosis, broadening their clinical perspective, and becoming more mindful of what their diagnoses entail.

Policy makers and stakeholders in the health sector, including various health maintenance organizations will find the study useful in drafting institutional clinical pathways involving cytology.

This study helps to promote wider adoption and acceptance of the International System among Filipino medical community particularly pathologists, oncologists, and surgeons.

This study contributes valuable information representing the Philippine demographic profile to the current available literature and serves as groundwork and resource for future researchers to build upon, who might investigate other avenues concerning serous fluid cytopathology.

Scope and Limitation

This study determined the rate of malignancy for serous fluid samples received at the Department of Laboratories, Philippine General Hospital. Samples submitted for pleural, pericardial, and peritoneal fluid cytology have been included. A liquid-based cytology method (ThinPrep, Hologic, Ca.) was compared in parallel to a conventional direct smear with cell block method as the primary means of slide preparation. For equivocal cases, the researchers utilized immunohistochemistry to further evaluate the sample for a more accurate diagnosis and classification. This study was limited to serous fluid samples submitted for cytopathologic examination to the Department of Laboratories of the Philippine General Hospital.

Methods

Study Design

This was a prospective cohort study which determined the rate of malignancy of serous effusion cytology specimens prepared using two different methods, conventional cell block with smear, and liquid-based cytology. The samples were classified under the different categories of the International System for Reporting Serous Fluid Cytopathology. The rate of malignancy was determined for the individual diagnostic categories using the International System for Reporting Serous Fluid Cytopathology.

Study Setting

The study was conducted at the Cytology Unit of the Department of Laboratories, University of the Philippines-Manila - Philippine General Hospital. A portion of serous fluid specimens from both inpatient and outpatient sent for examination at the Cytology Unit was utilized for specimen collection. These samples were split between LBC and routine cell block and smear which were prepared and processed by the staff cytotechnologists using equipment and facilities of the Cytology Unit.

Study Period

Specimen collection, processing, and microscopic examination of the samples was conducted from February 2022 to April 2023. Statistical analysis of the results and final manuscript completion concluded in August 2023.

Study Population

Inclusion Criteria

The study includes pleural, pericardial, and peritoneal fluid cytology samples submitted for examination at the Cytology Unit of the Department of Laboratories, Philippine General Hospital from February 2022 to April 2023. Samples included in the study were subject to the acceptance criteria set by the Department, namely: 1) Specimens should be submitted fresh (without any alcohol or formalin); 2) Specimens should be transported to the laboratory at the soonest possible time (should not be beyond 8 hours); 3) Specimens should be labelled correctly and placed in a tightly sealed, clean container; 4) Submit as much of the fluid as reasonable (volume should be at least 10 mL to qualify for processing using both methods); and 5) Laboratory request forms should be filled appropriately (eg. Complete name of the patient, date of birth, ward, source of specimen, brief clinical history, specific test/s to be performed, requesting physician’s name, and contact number). Only samples from patients above 18 years old were included in the study.

Exclusion Criteria

Specimens that do not comply with standard acceptance criteria set by the Cytology Unit of the Department of Laboratories were excluded from the study population. This includes improperly labelled serous fluid samples and samples with added fixative such as ethanol. Specimens stored at temperatures outside of 8 degrees Celsius for more than 3 days were excluded from the study.

Sampling Procedure

The study utilized systematic random sampling wherein all even-numbered serous fluid cytology specimens received and accessioned at the Cytology Unit of the

Department of Laboratories from February 2022 to April 2023 were included in the study sample. All serous fluid samples deemed acceptable for processing and have satisfied the inclusion criteria were included in the study until the goal sample size of 147 was reached.

Sample Size and Power Calculations

To arrive at the adequate number of specimens included in the study, the sample size formula16 for prospective cohort studies was used:

Following statistical consult regarding sample size, the researchers determined a total of 147 serous fluid samples must be collected to derive a 95% confidence interval

with a power of 80%. Hence, a group of 147 serous fluid samples was processed using liquid-based and conventional direct smear with cell block methods in parallel. Allowance for processing errors and problematic specimens was set at 10% of the sample size.

Operational Definition of Terms

Serous Fluid

Serous fluid (pleural, pericardial, and peritoneal fluid) serves as a lubricant to facilitate normal physiologic sliding motion between mesothelial surfaces17. The small volume of lubricating fluid is maintained by a delicate balance of hydrostatic and oncotic pressure and peripheral sulcal lymphatic drainage. Disturbances in any of these mechanisms would lead to fluid accumulation within the cavity, hence an effusion.

In this study, serous fluids includes pleural, pericardial, or peritoneal fluid specimens from which diagnosis of a possible malignant etiology of a serous effusion may be rendered.

Liquid-Based Cytology

Liquid-based cytology (LBC)18 refers to a method of preparing samples for cytopathologic examination wherein the sample is deposited into a small bottle of preservative liquid which is treated to remove other elements such as mucus before a layer of cells is placed on a slide.

In this study, the researchers utilized ThinPrep 5000 Processor as the primary LBC method of processing pleural fluid samples and subsequent slide preparation for cytopathologic examination.

Routine Cell Block and Smear

Conventional cytologic smear with cell block preparation19 is the routine procedure for processing serous fluids which begins by shaking the container to disperse the cells evenly, then spinning the aliquot in a centrifuge. After the supernatant is discarded, the sediment is used to prepare smears, and cytocentrifuge preparations. Afterwards, the slides are usually alcohol-fixed and Papanicolaou-stained.

In this study, conventional direct smears and cell block refers to the routine processing method of serous fluid samples for cytopathologic examination.

The International System for Reporting Serous Fluid Cytopathology

The International System11 is a reporting system for serous fluid cytopathology, later endorsed by the International Academy of Cytology (IAC) and the American Society of Cytopathology (ASC) with an evidence-based diagnostic system along with management recommendation. There are five diagnostic categories within TIS outlined in the Appendix Table 1.1, namely: non-diagnostic (ND), negative for malignancy (NFM), atypia of undetermined significance (AUS), suspicious for malignancy (SFM), and malignant (MAL).

In this study, TIS was utilized as the primary format of rendering the definitive diagnostic interpretation after cytopathologic examination. The resulting diagnoses were classified according to a diagnostic category within the TIS, and a descriptive diagnosis were below the diagnostic category in compliance to the prescribed format.

Randomization Procedure

Systematic random sampling was utilized such that every even-numbered sample from the pool of serous fluid samples received at the Cytology Unit was included as part of the study population. Odd numbered serous fluid samples received were included in the study.

Blinding

A triple blind procedure was utilized to reduce bias from the cytotechnologists processing the sample, cytopathologists who evaluated the slides through microscopic examination, and the statistician involved in data processing and analysis. Anonymity of the samples was ensured by numerical coding in the data collection forms.

Data Collection Method and Tools

Maneuvers

Fig 2. Flow chart illustrating the sequence in which the research methods ensued.

Specimen Collection and Preservation

Serous fluids were obtained primarily by thoracentesis, pericardiocentesis, or paracentesis whether blind or ultrasound guided. Sometimes, fluid may also be procured by suction during cardiothoracic or abdominal surgery. Submitted unfixed fluid samples were allocated into two wide mouthed bottles for processing using liquid-based and routine cell block and smear. These aliquots were refrigerated at 4℃ until the time of slide preparation.

Specimen Processing

ThinPrep

Procedure for processing body fluids using ThinPrep 5000 begins with sample collection of fresh pleural fluid into a CytoLyt Solution. Fluids collected in CytoLyt Solution requires a CytoLyt solution wash prior to instrument processing. Next, the solution is concentrated by centrifugation (600g for 10 minutes or 1200g for 5 minutes). The supernatant is then discarded while the remaining cell pellet is resuspended by vortexing. The cell pellet appearance is then evaluated. If there are remaining visible traces of blood, 30 mL of CytoLyt solution is added, centrifuged, and the supernatant is discarded until free of blood. Depending on the size of the cell pellet, an appropriate amount of specimen is added to the PreservCyt Solution vial which is then left to stand for 15 minutes. The mixture is then run on ThinPrep 5000 processor using Sequence Non-Gyn. The prepared slides are then fixed, stained, and are ready for evaluation. A more detailed explanation of the procedure is included in the Appendix for reference.

Conventional Direct Smear and Cell Block

The procedure begins with transferring the sample in a conical tube followed by addition of 95% ethanol (1:1 ratio). The mixture is then centrifuged for five minutes at 2500

rpm. The supernatant is afterwards discarded. The bottom of the tube is inspected for solid formation to proceed with cell block formation. 10% Neutral Buffered Formalin is added to the tube followed by centrifugation for five minutes. The cell block is then wrapped in filter paper and processed as a tissue sample, paraffin embedded, sectioned using a microtome, and stained with hematoxylin and eosin. Smears are then made of the remaining precipitates and subsequently stained using Papanicoloau stain, mounted and coverslipped, ready for microscopic examination.

Cytopathologic Examination

Microscopic examination of the slides prepared using both conventional and liquid- based methods were performed by three pathologists who classified each sample according the TIS diagnostic category and added subsequent descriptive diagnosis as appropriate. For cases with discordant cytologic interpretation, a second-look review was done and the category in which two out of three pathologists agree upon was considered the final disposition for cytology. Cases processed within one month was collated and sent to the assigned pathologist for the month with the accompanying data collection form. The pathologists were given a turnaround time of one month to review each case before the batch of slides, along with the interpretation, were returned for collation. The slides were then given to another pathologist for review, until all pathologists have reviewed all cases.

In compliance with the department’s usual workflow, routinely processed cell block and smears were signed out by the cytology resident rotator with the pathologist on-duty for the day. Except for cases requiring another pathologist’s signature for concurrence, intradepartmental consultation, or referral for expert opinion, results for routine cases were released after sign-out to adhere to the department’s recommended turnaround time. Hence, the study did not hinder releasing of official results for cytology cases.

Biosafety Procedures

Airborne and universal precautions were observed by cytotechnologists responsible for receiving and processing freshly submitted pleural fluid samples. This included wearing of a properly fitted N95 respirator, gown, gloves, and face shield during handling of samples for cytologic processing. Sample containers were stored for a maximum of fourteen days at 4℃ prior to disposal along with other laboratory wastes following hospital safety disposal procedures.

Data Processing and Analysis

To determine the rate of malignancy (ROM), a malignant diagnosis needed to be confirmed by a previous and/or subsequent effusion fluid from the same location or a tissue biopsy from the same general location (pleura, pericardium, or peritoneum) in conjunction appropriate immunohistochemistry panels, as available. Frequency and percentage were used to characterize patient demographics and clinicopathologic profile. To calculate rate of malignancy, percentage and ratio were utilized. Statistical analysis was performed using the SPSS Descriptive Statistics Package.

Data Protection Plan

Sample anonymity was ensured by assigning unique numerical codes during specimen accessioning prior to processing. Enhanced protection was afforded by the triple blind procedure which aimed to minimize bias in the performance of the processing steps, microscopic examination, and data analysis. Complete access to the data was exclusive to the investigators, while the statistician only accessed data which underwent coding and blinding procedures. Data storage was set for a maximum of 10 years after study completion before

disposal of the physical records through incineration and permanent deletion of electronic data. Collected data was used only for research purposes. Only the principal investigator and the co-investigators were granted full access to the anonymized data. Collected data will not be utilized for marketing purposes. Furthermore, data gathering, processing, and analysis were performed by the investigators of the study.

Storage of Archival Specimens

Slides for routinely processed cell block and smear for which an official report was issued were returned to the Cytology Unit, Division of Surgical Pathology of the Department of Laboratories for storage and archiving. Length of time the slides and blocks are stored was covered under the department’s retention policy and custodianship. Future studies which require access to such samples has to undergo appropriate review and approval from the department through regulatory and administrative channels.

Ethical Considerations

The study does not pose known risks to individual patients since patient identifiers, demographic data, and other clinical information in the medical record were not gathered. All data gathered were confidential and access was limited only to investigators and statistician directly involved in data analysis.

The study was submitted to the UP-Manila Research Ethics Board for review and approval prior to implementation.

Informed Consent and Confidentiality

In accordance with the National Ethical Guidelines for Health and Health-Related Research in 2017, a waiver of consent was obtained from the research ethics committee for

this study in the context of the study presenting with no more than minimal risk, the waiver did not adversely affect the rights and welfare of the participants. The study cannot be practicably carried out without the waiver. Clinical and demographic data collected from the surgical pathology consultation form and surgical pathology reports were de-identified, coded, and anonymized. No personal data such as name, address, and contact information were collected from the chart, consultation forms, or issued reports.

Risks of the Study

The study poses only minimal risks to individual patients. This included decreased availability of serous fluid sample for routine processing and further ancillary work-up due to sample allocation for liquid-based cytology processing. However, this risk was partially offset by the added benefit of pursuing further ancillary studies which help confirm a definitive diagnosis on the difficult cases in the samples included in the study. Since the study involved handling of glass slides prepared from actual patient samples, there was also the risk for loss or damage to the glass slides. To mitigate this risk, the slides were labelled with coded, de- identified, and anonymized study labels. These were placed in designated slide boxes, which minimized impact and were labelled as fragile. These slide boxes, along with data collection forms were personally handed out to the study co-investigators who performed the microscopic examination to ensure chain of custody. After they are done reviewing the cases, the slide box, along with the filled-out data collection forms, were returned and subsequently given to another pathologist co-investigator for evaluation.

Benefits of the Study

This study was able to benefit patients whose samples were included in the study in instances wherein morphologic evaluation alone precludes a definitive diagnosis. For such

cases, a repeat processing or further immunohistochemistry studies were done to arrive at a more definitive diagnosis at no additional cost to the patient.

Declaration of Conflict of Interest

This study was investigator funded and was not sponsored by any third-party manufacturer. The investigators have no financial arrangements with Hologic which distributes the ThinPrep system, whereby the value of the compensation to the investigators for conducting the study could be influenced by the outcome of the study. The investigators do not have significant financial interests such as receiving payments for services and/or equity nor do they have any proprietary interest in the research for which application for patent, trademark, copyright, or licensing will be made. Finally, the investigators have no significant personal nor family interest with the sponsor or the results of the study.

Dissemination/Data Sharing Plans

Collected data was used only for research purposes and was utilized for marketing purposes. Only the principal investigator and the co-investigators was granted full access to the anonymized data.

Results of the study may be disseminated through oral and/or poster presentation through the following channels: 1) Annual Research Forum of the Department of Laboratories;

2) Possible publication either in local or international pathology or cytology journals; 3) Presentation in oral and/or poster formats in professional gatherings including but not limited to conventions hosted by the Philippine Society of Pathologists and conventions hosted by other medical societies.

RESULTS

Demographic and Clinicopathologic Profile

Table 1 shows the demographic and clinicopathologic profile of pleural fluid samples processed using the liquid-based method. When grouped according to age, most of the pleural fluid samples are from young adults aged 19-40 (n=67, 66.34%). When grouped according to sex, majority are from males (n=54, 53.47%) while the remaining are from females (n=47, 46.53%). With regard to the history of immunocompromised state, most samples are from immunocompetent patients (n=70, 69.31%), while the remaining are from patients with a known history of immunocompromised condition (n=31, 30.69%). Regarding the history of malignancy, most of the pleural fluid samples are from patients with no history of malignancy (n=91, 90.10%), while the remaining are from patients with a known history of malignancy (n=10, 9.90%).

Table1. Demographic and Clinicopathologic Profile of Pleural Fluid Samples Processed Using Liquid-Based Cytology

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL=malignant, adolescents=12-19 years, young adults=19-40 years, middle adults=40-65 years, older adults=>65.

Table 2 shows the demographic and clinicopathologic profile of pleural fluid samples processed using routine cell block and smear. When grouped according to age, majority of the samples are from young adult patients aged 20-39 years old (n=67, 66.33%). When grouped according to sex, most of the samples are from males (n=54, 53.46%) while the remaining are from females (n=47, 46.53%). With regard to the history of immune compromised state, majority of the samples are from immunocompetent individuals (n=70, 69.30%) while the remaining patients have a history of immune compromised state (n=31, 30.69%). Concerning history of malignancy, majority of the samples have no known history of malignancy (n=91, 90.09%).

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL=malignant, adolescents=12-19 years, young adults=19-40 years, middle adults=40-65 years, older adults=>65.

Table 2. Demographic and Clinicopathologic Profile of Pleural Fluid Samples Processed Using Routine Cell Block and Smear

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL=malignant, adolescents=12-19 years, young adults=19-39 years, middle adults=40-65 years, older adults=>65.

Table 3 shows the demographic and clinicopathologic profile of peritoneal fluid samples processed using liquid-based cytology. In terms of age, majority of the peritoneal fluid samples are obtained from young adults aged 20-39 years old (n=43, 63.23%). In terms of sex, majority of peritoneal fluid samples submitted are from females (n=37, 54.41%). As regards history of immunocompromised condition, majority of samples are from patients who do not have a history of an immunocompromised state (n=37, 54.41%) and the remaining have a known history or are currently being managed for a concomitant immunocompromised condition (n=31, 45.58%). In terms of history of malignancy, most of the peritoneal fluid samples are obtained from patients without a known history of malignancy (n=45, 66.17%) whereas the remaining samples are from patients with a documented history of malignancy (n=33.82%).

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL=malignant, adolescents=12-19 years, young adults=19-39 years, middle adults=40-65 years, older adults=>65.

Table 4 shows the demographic and clinicopathologic profile of peritoneal fluid samples processed using routine cell block and smear. In terms of age, majority of the samples are from young adults aged 20-39 years old (n=43, 63.24%). In terms of sex, most of the samples are from females (n=37, 54.41%) while the remaining are from males (n=31, 45.58%). Concerning history of immunocompromised state, majority of the samples have no history of immunocompromised state (n=37,54.41%) while the remaining have a known history of immunocompromised state (31, 45.58%). In terms of history of malignancy, majority of the samples have no known history of malignancy (n=45, 66.17%) whereas the remaining samples are from patients with a documented history of malignancy (n=23, 33.82%)

Table 4. Demographic and Clinicopathologic Profile of Peritoneal Fluid Samples Processed Using Routine Cell Block and Smear

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL=malignant, adolescents=12-19 years, young adults=20-39 years, middle adults=40-65 years, older adults=>65

Diagnostic Categories

Non-diagnostic (ND)

Table 5. Pleural and Peritoneal Fluid Samples Processed Using Liquid Based and Routine Cell Block and Smear Methods Classified Under the Nondiagnostic Category

Notes: ND=nondiagnostic, LBC=liquid-based cytology

For the pleural fluid samples classified in the nondiagnostic category, a higher number of samples fall under the nondiagnostic category in those processed using routine cell block and smear (n=18, 17.8%) than the liquid-based cytology (n=1, 1%). For peritoneal fluid samples, there are more samples processed using routine cell block and smear (n=9, 68%) falling under the nondiagnostic category than those processed using the liquid-based method (n=1, 1.47%).

Negative for Malignancy (NFM)

Table 6. Pleural and Peritoneal Fluid Samples Processed Using Liquid Based and Routine Cell Block and Smear Methods Classified Under the Negative for Malignancy (NFM) Category

Notes: NFM=negative for malignancy, LBC=liquid-based cytology

For pleural fluid samples classified in the negative for malignancy category, a higher number of samples fall under the negative for malignancy category in those processed using liquid-based method (n=74, 73.3%) than those processed using routine cell block and smear (n=62, 61.4%). For peritoneal fluid samples, there are more samples processed using the liquid- based method (n=46, 67.65%) that fall under the negative for malignancy category than those processed using routine cell block and smear (n=42, 61.8%).

Atypia of Undetermined Significance

Table 7. Pleural and Peritoneal Fluid Samples Processed Using Liquid Based and Routine Cell Block and Smear Methods Classified Under the Atypia of Undetermined Significance (AUS) Category

Notes: AUS=atypia of undetermined significance, LBC=liquid-based cytology

For pleural fluid samples classified under the atypia of undetermined significance category, a higher number of samples fall under the atypia of undetermined significance category in those processed using the routine cell block and smear (n=13, 12.87%) than LBC (11, 10.90%). For peritoneal fluid samples, there are more samples processed using the routine cell block and smear (n=14, 20.59%) that fall under the atypia of undetermined significance category than those processed using the liquid-based method (n=9, 13.24%).

Suspicious for Malignancy

Table 8. Pleural and Peritoneal Fluid Samples Processed Using Liquid Based and Routine Cell Block and Smear Methods Classified Under the Suspicious for Malignancy Category

Notes: SFM=suspicious for malignancy, LBC=liquid-based cytology

For pleural fluid samples classified under the suspicious for malignancy category, a higher number of samples fall under the suspicious for malignancy category in those processed using the liquid-based method (n=7, 6.93%) than those processed using routine cell block and smear (n=6, 5.94%). For peritoneal fluid samples, there are more samples processed using the liquid-based method (n=6, 8.82%) that fall under the suspicious for malignancy category than those processed using routine cell block and smear (n=3, 4.41%).

Malignant

Table 9. Pleural and Peritoneal Fluid Samples Processed Using Liquid Based and Routine Cell Block and Smear Methods Classified Under the Malignant Category

Notes: MAL=malignant, LBC=liquid-based cytology

For pleural fluid samples classified under the malignant category, samples processed using the liquid-based method (n=8, 7.92%) showed higher number of samples that fall under the malignant category than those samples processed using routine cell block and smear (n=2, 1.98%). For peritoneal fluid, samples processed using the liquid-based method (n=6, 8.82%) have a higher number of samples classified under the malignant category than those processed using routine cell block and smear (n=0, 0%).

Rate of Malignancy

Notes: ND =nondiagnostic, NFM=negative for malignancy, AUS=atypia of undetermined significance, SFM=suspicious for malignancy, MAL/M=malignant, B=benign, E=equivocal

For pleural fluid samples processed using routine cell block and smear, the rates of malignancy for each diagnostic category were calculated as follows: 27.78% in the nondiagnostic category (n=5, 27.78%), 14.2% in the negative for malignancy category (n=9, 14.2%), 82.62% in the atypia of undetermined significance category (n=11, 82.62%), 100% in the suspicious for malignancy category (n=6, 100%), and 100% in the malignant category (n=2, 100%).

For pleural fluid samples processed using liquid-based method, the rates of malignancy for each diagnostic category were calculated as follows: there were no samples classified under the nondiagnostic category which turned out to be malignant (n=0, 0%), 9.46% in the negative

for malignancy category (n=7, 9.46%), 100% in the atypia of undetermined significance category (n=11, 100%), 100% in the suspicious for malignancy category (n=7,100%), and 100% in the malignant category (n=8, 100%).

For peritoneal fluid samples processed using routine cell block and smear, the rates of malignancy for each diagnostic category were calculated as follows: 22.22% in the nondiagnostic category (n=2, 22.22%), 9.52% in the negative for malignancy category (n=4, 9.52%), 71.43% in the atypia of undetermined significance category (n=10, 71.43%), and 66.67% in the suspicious for malignancy category (n=2, 66.67%). There were no peritoneal fluid samples processed using routine cell block and smear classified in the malignant category.

For peritoneal fluid samples processed using liquid-based method, the rates of malignancy for each diagnostic category were calculated as follows: there were no samples classified under the nondiagnostic category (n=0, 0%), 6.52% in the negative for malignancy category (n=3, 6.52%), 33.3% in the atypia of undetermined significance category (n=2, 33.3%), 66.67% in the suspicious for malignancy category (n=2, 66.67%), and 100% in the malignant category (n=6, 100%).

Comparing the two methods of processing pleural fluid samples for cytology per diagnostic category showed the following: liquid-based cytology (n=0, 0%) has no samples classified in the nondiagnostic category compared to routine cell block and smear (n=5, 27.78%). In the negative for malignancy category, liquid-based cytology (n=7, 9.46%) has a lower rate of malignancy compared to routine cell block and smear (n=9, 14.52%). In the AUS, SFM and MAL category, both liquid-based, and routine cell block and smear have similar rates of malignancy (AUS: n=11, 100%, SFM: n=6, 100%, MAL: n=8, 100%).

Comparing the two methods of processing peritoneal fluid samples for cytology per diagnostic category showed the following: in the nondiagnostic category, liquid-based cytology (n=0, 0%) has a lower rate of malignancy than routine cell block and smear (n=2, 22.22%). In the negative for malignancy category, liquid-based cytology (n=3, 6.52%) has a lower rate of malignancy than routine cell block and smear (n=4, 9.52%). In the atypia of undetermined significance category, liquid-based cytology (n=3, 33.33%) has a lower rate of malignancy than routine cell block and smear (n=10, 71.43%). In the suspicious for malignancy category, liquid-

based cytology (n=6, 100%) has a higher rate of malignancy than routine cell block and smear (n=2, 66.67%). In the malignant category, liquid-based cytology (n=6, 100%) has a higher rate of malignancy (n=6, 100%) with all samples turned out to be malignant on tissue follow-up and ancillary studies than routine cell block and smear (n=0, 0%) in which none of the processed peritoneal fluid samples were classified in the malignant category.

DISCUSSION

Serous effusion cytology is a vital tool in the initial evaluation of effusions. It is a cost- effective, minimally invasive, simple, and safe procedure that is very helpful in determining the origin of malignancy in patients with fluid accumulation in a body cavity as the initial presentation. The study by Farahani et al. found that malignancy is the underlying cause of serous effusion in 10%-61% of cases.²¹ Given this high rate of malignancy, it is comforting to know that cytology plays a crucial role in diagnosis of serous effusion and is the gold standard in determining metastatic cancer at the body cavity.²⁷ This is because biopsy of focal lesions on the serosal surface may miss, which may account for false negatives. Meanwhile, malignant cells coming from the entire surface of the cavity exfoliate and accumulate in an effusion.

Hence, surgical biopsy becomes less sensitive than effusion cytology for detecting serosal malignancy.²⁷ Majority of malignant effusions are secondary to metastatic adenocarcinoma from sites such as lung, breast, gastrointestinal, and female genital tract. ^{20,21,27, 28}

The International System for Reporting Serous Fluid Cytopathology was established with the goal of standardizing the reporting terminology used for serous effusion cytology for a more consistent reporting across many countries. Placing each case in categories allows for better statistical analysis for the laboratory. This practice is useful in determining diagnostic rates as well as to follow laboratory diagnostic trends. For the clinicians, diagnostic categories enable easier determination of the next steps in management and follow-up frequency. Thus, cytologic examination of serous effusions provides clinically relevant information that guides therapeutic management in malignant cases and saves the patient unwarranted invasive procedures in benign effusions.

To do so, best practices in fluid processing techniques and how ancillary studies are best utilized for equivocal cases should also be investigated. Recently, the use of molecular studies, immunohistochemistry, and immunocytochemistry are great adjuncts to the cytopathologist’s arsenal and have greatly contributed to the improvement of malignancy diagnosis as well as in differentiating tumor subtypes.

This study sought to classify serous fluid samples processed using cell block and smear and liquid-based cytology (ThinPrep) according to the International System for Reporting Serous Fluid Cytopathology. A split sample of fresh specimen was processed using both methods since the liquid-based method is not yet widely adopted in the institution possibly due to its relatively higher cost. The study initially included pericardial fluid/effusion samples for analysis, however, the lack of fresh pericardial fluid/effusion samples received was probably a reflection of the clinical management of these situations. As such, asymptomatic effusions and effusions without hemodynamic compromise were usually not drained.¹

Several studies found the mean prevalence of malignancy was 48.2% ±3.7% across all diagnostic categories. The study by Farahani et al.¹⁰ examined 11,799 cases of pleural fluids and classified them into categories which accounted for the following: ND (0.3%, mean ROM of 25.7%), NFM (69.2%, mean ROM of 22.3%), AUS (0.6%, mean ROM of 71.8%), SFM (3.5%,

mean ROM of 75.9%), and MAL (26.4%, mean ROM of 99.2%). Likewise, the distribution of cytology diagnoses in peritoneal fluid specimens are as follows: ND (0.3%, mean ROM of 0%), NFM (59.7%, mean ROM of 27.1%), AUS (0.4%, mean ROM of 40.5%), SFM (2.6%, mean ROM of 84.6%), and MAL (37%, mean ROM of 98.4%).

Fluid samples classified under the nondiagnostic (ND) category have insufficient cellular elements for a cytologic interpretation.²⁰ Findings in this study showed more nondiagnostic samples in the routine cell block and smear compared to liquid-based preparations in both pleural (cell block and smear: n=18, 17.8%; LBC: n=1, 1%) and peritoneal fluid samples (cell block and smear: n=9, 68%; LBC: n=1, 1;47%). These incidence rates were higher compared to those reported in the literature which should vary between 0.2 and 1%.²⁰ In this category, the rate of malignancy (ROM) for pleural fluid samples is 27.8% while there were no nondiagnostic samples which turned out to be malignant in peritoneal fluid samples. The

results were within the range of reported ROMs of ND which vary widely between 0 and 100% but were slightly higher than the reported mean ROM of 17.4% (±8.9%). ^20, 21

Samples classified under the negative for malignancy (NFM) category implies absence of malignant cells in a specimen deemed appropriate for diagnosis.²⁰ In this study, there were more samples in the LBC group classified under NFM as opposed to the cell block and smear (pleural fluid: LBC, 73%; cell block and smear, 61.4%; peritoneal fluid: LBC, 67.65%; cell block and smear, 61.8%). Diagnosis of NFM was expected to comprise the bulk of diagnoses rendered by a laboratory. The results of LBC in this study approximated the expected incidence of 70-80%²⁰ while that of cell block and smear were below the expected range. In this category, the ROM for pleural fluid samples in the LBC group was 9.46% while an ROM of 14.2% was seen in the cell block and smear group. For the peritoneal fluid samples, the ROM in the LBC group was 6.52% while an ROM of 9.52% was seen in the cell block and smear group. The results were within the range of reported ROMs for NFM which also vary widely between 0 and 80% but were slightly lower than the reported mean ROM of 21% (±0.3%).^20, 21

Atypia of Undetermined Significance (AUS) represents a true gray zone in serous fluid cytology. This category encompasses specimens in which the expectation of finding malignancy is low but that contains cells of uncertain nature, due to either qualitative or quantitative factors.²⁰ In this study, 10.9% of pleural fluid samples in the LBC group fall under AUS compared to cell block and smear with 8.9%. On the other hand, 20.58% of peritoneal fluid samples in the cell block and smear group were classified under AUS compared to 19.12% in the LBC group. These results were higher than the expected incidence of 0.6-1.6% in some studies which reported more frequent occurrence in pericardial and pleural specimens compared to pleural.²⁰ In this category, the ROM for pleural fluid samples in the LBC group was 100% while an ROM of 82.62% was noted in the cell block and smear group. For peritoneal fluid samples, the ROM in the LBC group was 33.3% while 71.43% in the cell block and smear group. The results were still within the range of reported ROMs for AUS which vary between 13% and 100% but were considerably higher than the reported mean ROM of 66% (±10.6%).^20, 21

Suspicious for Malignancy (SFM) specimens highlight cytologic features usually found in malignant lesions but are insufficient in either quality or quantity to make a definitive

diagnosis of malignancy.²¹ In this study, 6.93% of pleural fluid samples in the LBC group fall under SFM compared to cell block and smear with 5.94%. In contrast, 8.82% of peritoneal fluid samples in the LBC group are classified under SFM compared to 4.4% in the cell block and smear group. These results were higher than the expected incidence of around 2%.²⁰ However, some studies reported higher incidence ranging from 3.6-6.3% attributed to the inherent sampling bias associated with patient population in cancer centers and oncological hospitals.²⁴ In this category, ROM for pleural fluid samples in both the LBC and cell block and smear groups was 100%. Meanwhile, the ROM for peritoneal fluid samples in both the LBC and cell block and smear groups was 66.67%. The results were within the range of reported ROMs for SFM which vary between 0 and 100% but were lower than the reported mean ROM of 82% (±4.8%).^20, 21

The malignant (MAL) category includes specimens showing cytomorphologic features that either alone or in combination with ancillary studies, are diagnostic of a primary or secondary malignancy.²⁰ In this study, 7.9% of pleural fluid samples in the LBC group fall under MAL compared to cell block and smear with 2.2%. On the other hand, 8.82% of peritoneal fluid samples in the LBC group were classified under MAL. There were no samples from the cell block and smear group of peritoneal fluid classified under MAL by cytologic evaluation.

These results were lower than the reported expected incidence of around 30% in peritoneal and pleural effusions and more than 50% in pericardial effusions.²⁰ In this category, the ROM for pleural fluid samples in both the LBC and cell block and smear group was 100%. In addition, the ROM for peritoneal fluid samples in the LBC group was 100%. The results were within the range of reported ROMs for MAL which vary between 87% and 100%, and higher than the mean ROM of 99% (±0.1%).^20, 21 Compared with several studies in the literature which reported the percentage of about 10.4-20.5%, the relatively lower rate of cytology specimens classified under MAL may be attributed to sample bias since large number of samples submitted for cytology are ethanol-fixed which were excluded from the study population since liquid-based processing requires the sample to be received fresh. Institution-wide practice improvement in specimen collection and prompt submission to the laboratory for proper storage and/or processing may help increase the number of eligible samples and promote wider adoption of LBC as the method of choice for cytology preparations.

Comparing the two methods of preparing cytology smears, this study showed that in both pleural and peritoneal fluid samples, LBC has no samples classified under the ND category while cell block and smear has an ROM of 27.78% ROM. NFM category has lower ROM in LBC (pleural: 9.46%, peritoneal: 6.52%) than cell block and smear (pleural: 14.52%, peritoneal: 9.52%). Although similar ROMs were observed in the AUS, SFM, and MAL categories in pleural fluids (ROM: 100%, respectively), LBC had a lower ROM in the AUS category (33.33%), and a higher ROM in the SFM and MAL categories (100% respectively) than cell block and smear counterparts (AUS: 71.43%, SFM: 66.67%, MAL: 0%) of peritoneal fluid samples.

LBC offers numerous advantages and has a diagnostic accuracy that is equal to or surpasses that of conventional cell block and smear.²³ There is no smearing involved on the part of the clinician as samples are directly collected into appropriate preservative bottles; Transport is more seamless as samples are suspended in a stable collecting media; Specimen processing is automated along with standardized procedure within the laboratory and across laboratories; The specimen undergoes homogenization, filtration or sedimentation resulting in decrease of obscuring elements; Screening is easier on the part of the cytotechnologist and pathologist as the area to be screened is standardized to 20-mm diameter circle for all cellular contents.^23, 24 On the slide, LBC preparations permit a more even distribution of cells over the slide area, have reduced obscuring background elements, and have better preserved nuclear detail and cytoplasm spread evenly in a thin layer.^20, 21 Hence, LBC may account for the dramatic decrease in unsatisfactory or nondiagnostic rates by means of reducing obscuring elements. Due to the reduction of obscuring elements, the uniform distribution of the cells within a thin layer and the optimal preservation of the cells, LBC preparations are relatively easier to screen, and the atypical cells are more readily identified, allowing a more definitive diagnosis of malignancy in many cases.^26, 33, 35 A study by Fischler and Toddy has shown that processing cytology using LBC was able to decrease the global unsatisfactory rate from 17% to 1%.³³ In addition, the systematic review by Farahani et al. showed that liquid-based preparations of serous effusions have significantly higher sensitivity, diagnostic odds ratio, positive likelihood ratio, and lower negative ratio when compared to conventional preparations.²⁰

In terms of slide quality, several studies showed LBC was superior in overall cellularity and quality of the slides, as well as the quantity of diagnostic cells. The implementation of LBC shifted the slide preparation from the clinicians to the cytology laboratory led to improvement in slide quality.³⁴ Hence, the superior performance of LBC preparations is due to the elimination of blood and mucus as well as adequate preservation, contributing to good nuclear detail. In addition, liquid-based preparation techniques are hands-off, standardized, timesaving, and permit application of ancillary studies.²⁰ These features could justify the preference for LBC by most high-volume cytology laboratories over the conventional cell block and smear method since the former affords to bridge the future application of molecular and other ancillary techniques to cytology.³⁰

The accuracy of diagnosis is dependent on the fluid type, specimen volume, type of specimen processing, and the expertise of the cytopathologist reviewing the case. Serous fluid cytopathology may pose diagnostic challenges brought about by the notorious atypia of mesothelial cells, cytomorphologic overlap of malignant cells and mesothelial cells or simply the presence of benign glandular cells. Therefore, reporting cytopathologic findings of these specimens are crucial in patient’s clinical management. Differentiating reactive from malignant effusions by cytologic examination is often, but not always possible. Reactive or degenerative changes in mesothelial cells and/or macrophages may be difficult to distinguish from malignant epithelial cells based on morphology alone. Hence, ancillary testing such as immunohistochemistry is usually performed when trying to arrive at a definitive diagnosis, to better delineate cell lineage, and to distinguish benign reactive and/or degenerative changes from malignancy. The use of these ancillary studies in conjunction with cytomorphology has shown to significantly improve sensitivity up to 93% to 100%.⁴ Technical issues may plague interpretation of IHC stains: these include paucity of cells for staining, heterogenous expression of markers, or an altered immunoprofile of tumor cells. In these cases, it is best to correlate findings with clinical features and radiologic findings to further support the diagnosis.

CONCLUSION

This study classified serous fluid samples processed using both liquid-based cytology and routine cell block and smear methods into the different categories of the International System for Reporting Serous Fluid Cytopathology. In both pleural and peritoneal fluid samples, majority are from young adults aged 20-39 years old. The pleural fluid sample population is comprised of slightly more males than females whereas the peritoneal fluid sample population is comprised of slightly more females than males. Most of the samples from both sample types are received from immunocompetent patients with no history of malignancy.

In pleural fluid samples, LBC had a significantly lower rate of samples in the ND category than cell block and smear. Conversely, LBC showed higher rates of AUS, SFM, and MAL compared to samples processed using cell block and smear. In peritoneal fluid samples, LBC also showed lower rate of ND samples, and higher rates of samples classified under SFM and MAL than cell block and smear which had relatively higher rate of AUS.

Comparing the ROM of pleural fluids, LBC has a lower rate of malignancy in the ND and NFM categories than routine cell block and smear. Both methods have similar ROM at 100% in the SFM and MAL categories. However, in the AUS category, LBC has a higher ROM than routine cell block and smear. In the peritoneal fluid group, LBC also has a lower ROM in the ND, NFM, and AUS categories. However, LBC consistently showed higher ROM than routine cell block and smear in the SFM and MAL categories. Compared to routine cell block and smear in both sample groups, LBC was able to reduce the number of nondiagnostic samples and was able to improve detection of atypical and malignant cells.

RECOMMENDATIONS

Serous fluid cytology comprises a range of samples from different body cavities. To compare the two methods of cytology preparation, fresh specimens were required during sample collection to be eligible for processing using the liquid-based method. Given the advantages of a liquid-based method, instituting a hospital-wide policy to submit fresh fluid samples directly to the laboratory would help in increasing the number of samples eligible for liquid-based processing.

To aid in adoption of the liquid-based method for cytology in general, information dissemination such as mini-lectures and issuing informative memorandum by the laboratory which highlights the advantages of LBC may be done to encourage test utilization.

Pathologists are also encouraged to adopt the International System when issuing reports for cytology cases. This aids in fostering familiarity of the reporting system by both clinicians and pathologists which may lead to improvement in report standardization, data collection, and management outcomes.

When use of the reporting system is ubiquitous, further studies on ROM may be done retrospectively and with larger sample size for a more representative picture of the cytology laboratory’s caseload. In addition, more clinico-demographic data and performance characteristics such as sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratio may be better assessed with a larger sample size.

Other ancillary studies which may help clinch the diagnosis such as flow cytometry and immunocytochemistry may be performed when cell block material is not sufficient. Currently, these techniques are limited to a few laboratories and have yet to see a wider adoption.

REFERENCES

1.    Rodriguez, E. F., Jones, R., Gabrielson, M., Santos, D., Pastorello, R. G., & Maleki, Z. (2020). Application of the International System for Reporting Serous Fluid Cytopathology (ISRSFC) on Reporting Pericardial Effusion Cytology. Acta cytologica, 64(5), 477–485. https://doi.org/10.1159/000507311

2.    Kim GE, Kim JH, Kim YH, Choi C, Lee JS. Utility of promoter hypermethylation for differentiating malignant and benign effusions in liquid-based cytology specimens. Korean J Pathol. 2010;44:315–21.

3.    Shidham VB (2007). Introduction. In: Shidham VB, Falzon M, editors. Cytopathologic diagnosis of serous fluids. Philadelphia: Saunders; pp 1-18.

4.    Hou, T., Landon, G., Stewart, J., & Roy-Chowdhuri, S. (2021). The value of a tiered cytology diagnostic reporting system in assessing the risk of malignancy in indeterminate serous effusions. Cancer cytopathology, 129(1), 75–82. https://doi.org/10.1002/cncy.22345

5.    Veneti S, Daskalopoulou D, Zervoudis S, Papasotiriou E, Ioannidou-Mouzaka L (2003). Liquid-based cytology in breast fine needle aspiration. Comparison with the conventional smear. Acta Cytol, 47, 188-92.

6.    Gabriel C, Achten R, Drijkoningen M (2004). Use of liquid-based cytology in serous fluids: a comparison with conventional cytopreparatory techniques. Acta Cytol, 48, 825-35.

7.    Nandini NM, Nandish SM, Pallavi P, et al (2012). Manual liquid-based cytology in primary screening for cervical cancer-a cost effective preposition for scarce resource settings. Asian Pac J Cancer Prev, 13, 3645-51.

8.    Sigurdsson K (2013). Is a liquid-based cytology more sensitive than a conventional Pap smear?. Cytopathol, 24, 254-63.

9.    Konofaos P, Tomos P, Malagari K, Karakatsani A, Pavlopoulos D, Lachanas E, Flessas M, Kostakis A, Karakitsos P: The role of ThinPrep cytology in the investigation of lung tumors. Surg Oncol 2006;15:173–178.

10. Farahani, S. J., & Baloch, Z. (2019). Are we ready to develop a tiered scheme for the effusion cytology? A comprehensive review and analysis of the literature. Diagnostic cytopathology, 47(11), 1145–1159. https://doi.org/10.1002/dc.24278

11. Chandra A, Crothers B, Kurtycz D, Schmitt F. Announcement: The International System for Reporting Serous Fluid Cytopathology. Acta Cytol. 2019;63:349-351.

12. Crothers BA, Chandra A. Proceedings of the American Society of Cytopathology companion session at the 2019 United States and Canadian Academy of Pathology Meeting part 1: toward an International System for Reporting Serous Fluid Cytopathology. J Am Soc Cytopathol. 2019;8:362-368.

13. Saab J, Hoda RS, Narula N, et al. Diagnostic yield of cytopathology in evaluating pericardial effusions: clinicopathologic analysis of 419 specimens. Cancer Cytopathol. 2017;125(2):128-137.

14. Ronco G, Cuzick J, Pierotti P, et al. Accuracy of liquid based versus conventional cytology: overall results of new technologies for cervical cancer screening: randomised controlled trial. BMJ (Clin Res Ed). 2007; 335(7609):28.

15. Gabriel C, Achten R, Drijkoningen M. Use of liquid-based cytology in serous fluids: a comparison with conventional cytopreparatory techniques. Acta Cytol. 2004;48(6):825-835.

16. Rosner B. Fundamentals of Biostatistics. 7th ed. Boston, MA: Brooks/Cole; 2011.

17. Skok K, Hladnik G, Grm A, Crnjac A. Malignant pleural effusion and its current management: a review. Medicina (Kaunas). 2019;55(8):490.

18. Ylagan LR, Zhai J. The value of ThinPrep and cytospin preparation in pleural effusion cytological diagnosis of mesothelioma and adenocarcinoma. Diagn Cytopathol. 2005;32(3):137-144.

19. Collins GR, Thomas J, Joshi N, Zhang S: The diagnostic value of cell block as an adjunct to liquid-based cytology of bronchial washing specimens in the diagnosis and subclassifica- tion of pulmonary neoplasms. Cancer Cyto pathol 2012;120:134–141.

20. Pinto, D., Chandra, A., Crothers, B. A., Kurtycz, D. F. I., & Schmitt, F. (2020). The international system for reporting serous fluid cytopathology—diagnostic categories and clinical management. In Journal of the American Society of Cytopathology (Vol. 9, Issue 6, pp. 469–477). Elsevier Inc. https://doi.org/10.1016/j.jasc.2020.05.015

21. Lobo CC, J; Petronilho, S; Monteiro, P; Leça, L; Schmitt, F. Cytohistological correlation in serous effusions using the newly proposed International System for Reporting Serous Fluid Cytopathology: experience of an oncological center. Diagn Cytopathol. Apr 2020;

22. Rossi ED, Bizzarro T, Schmitt F, Longatto-Filho A. The role of liquid-based cytology and ancillary techniques in pleural and pericardic effusions: an institutional experience. Cancer Cytopathol. 2015;123(4):258-266.

23. Cox JT. Liquid-based cytology: evaluation of effectiveness, cost-effectiveness, and application to present practice. J Natl Compr Canc Netw. 2004;2(6):597-611.

24. Wu GP, Wang EH, Li JH, Fu ZM, Han S: Clinical application of the liquid-based cytological test in cytological screening of sputum for the diagnosis of lung cancer. Respirology 2009; 14: 124–128.

25. Gerbes AL, Jungst D, Xie YN, Permanetter W, Paumgartner G. Ascitic fluid analysis for the differentiation of malignancy-related and non-malignant ascites. Proposal of a diagnostic sequence. Cancer. 1991;68 (8):1808-1814.

26. Heffner JE, Klein JS. Recent advances in the diagnosis and management of malignant pleural effusions. Mayo Clin Proc. 2008;83(2):235-250.

27. Awadallah SF, Bowling MR, Sharma N, Mohan A. Malignant pleural effusion and cancer of unknown primary site: a review of literature. Ann Transl Med. 2019;7(15):353.

28. Allen VA, Takashima Y, Nayak S, Manahan KJ, Geisler JP. Assessment of false- negative ascites cytology in epithelial ovarian carcinoma: a study of 313 patients. Am J Clin Oncol. 2017;40(2):175-177.

29. Valerio E, Nunes W, Cardoso J, et al. A 2-year retrospective study on pleural effusions: a cancer centre experience. Cytopathology. 2019;30:607-613.

30. Michael, C. W., & Bedrossian, C. C. W. M. (2014). The implementation of liquid- based cytology for lung and pleural-based diseases. In Acta Cytologica (Vol. 58, Issue 6, pp. 563–573). S. Karger AG. https://doi.org/10.1159/000369198

31. Fashoyin-Aje LA, Brahmer JR (2014). Malignant effusions. In: Nieder JE, Armitage JO, Doroshow JH, et al, editors. Abelloff’s clinical oncology. 5th ed. Philadelphia: Saunders, pp 794-805.

32. Nance KV, Shermer RW, Askin FB (1991). Diagnostic efficacy of pleural biopsy as compared with that of pleural fluid examination. Mod Pathol, 4, 320-4.

33. Fischler DF, Toddy SM: Nongynecologic cytology utilizing the ThinPrep processor. Acta Cytol 1996; 40: 669–675.

34. Wang HH, Sovie S, Trawinski G, Garcia LW, Abu-Jawdeh GM, Upton M, Werneke S: ThinPrep processing of endoscopic brushing specimens. Am J Clin Pathol 1996; 105: 163–167.

35. Hees K, Lebeau PB: Comparison of conventional and ThinPrep preparations of mucoid cytology samples. Diagn Cytopathol 1995; 12:181–185.

APPENDIX

A.  BUDGET