Approach to Diagnostic Cytopathology of Serous Effusions

GENERAL APPROACH

Aspiration of an effusion is a relatively simple procedure. Cancer is one of the frequent causes of effusions and it may be the first manifestation of advanced cancer. Consequently, any general pathology laboratory may receive effusion fluid for cytologic evaluation. As a consequence, all general pathologists and cytopathologists should be conversant with the diagnostic challenges and pitfalls of effusion fluid cytology. These facts emphasize the significance of understanding the approach to effusion fluid cytology by all practicing pathologists.

Finding neoplastic cells in effusion specimens not only reveals that a patient has cancer but also denotes the advanced nature of the disease, which, at this stage, is almost always incurable. Other than the detection of cancer cells in cerebrospinal fluid, no other exfoliative cytology specimen carries such an ominous prognostic significance for the detection of cancer cells. Apart from the finding of cancer cells, cytopathologic examination of serous effusions may reveal inflammatory conditions, parasitic infestations, bacterial, fungal, or viral infections, and certain other non-neoplastic conditions.

• Other than high-grade neoplasms with pleomorphic cells, the standard cytologic criteria of malignancy based on evaluation of single cell morphology are not applicable for most of the effusion cytology specimens. Since cells in a fluid medium ‘round up’ because of surface tension, the native shapes of cancer cells cannot be a dependable guiding feature for deciding the primary origin of malignant cells. As effusion fluids are nutrient rich, the cancer cells have the potential to proliferate and may continue to divide and form proliferation spheres. It is crucial to consider these factors while interpreting effusion cytology.

Although adenocarcinomas are the most common neoplasm to spread to the peritoneum, pleura, and pericardium, almost any type of cancer may do so. Any neoplasm including carcinomas, melanomas, hematopoietic neoplasms, sarcomas, and mesotheliomas may involve serous cavities. Carcinomas are almost always adenocarcinomas. In general, sarcomas metastatic to serous cavities are rare. • The morphologic features of most of the cancer cells in effusion smears are different from those seen in exfoliative, brushing, and fine-needle aspiration cytology.

Reactive mesothelial cells are a consistent component of effusion fluids, and they may have significant morphologic overlap with malignant cells. Consequently, the interpreter must be aware of the wide range of cytomorphologic appearances of reactive mesothelial cells in effusion fluids.

To avoid various diagnostic pitfalls and increase the diagnostic accuracy, it is essential to understand multifactorial nuances in the setting of effusion fluid cytology. If one is familiar with the cytomorphologic features and the pitfalls associated with the effusion cytology, diagnostic interpretation may be facilitated, even without having to resort to the ancillary studies [Figures 1,2].

Close

Figure 1:

Algorithm for evaluation of a ‘second foreign population’.

Export to PPT

Close

Figure 2:

Cytologic evaluation of effusion fluids for primary site.

Export to PPT

Proper cytopathologic evaluation of effusion fluids thus needs a special approach at various stages from specimen processing to final interpretation level.

PROCESSING APPROACH

• The backbone of effusion cytopathology is to organize the specimen processing in such a manner that it accomplishes the following objectives:

  1. Detection of any ‘second population of foreign cells’ other than the mesothelial and inflammatory cells [Figure 8].

  2. Studying the nuclear details of the ‘second population’.

  3. Semiquantitative evaluation of individually scattered mesothelial cells and large groups of three-dimensional mesothelial cells, as in mesotheliomas [Figure 1[5]].

  4. Objective confirmation of the ‘second foreign population’ and the differential diagnosis of their primary neoplasm.

Any approach that does not address these objectives may lead to misinterpretation and suboptimal results.

In our experience:

• The 1^st objective is best evaluated in Diff-Quik (DQ)-stained (Romanowsky-stained) air-dried Cytospin smears (or an air-dried direct smear prepared from the cell pellet).

• The 2^nd objective is best evaluated in Papanicolaou (PAP)-stained smears (such as wet fixed direct smears or rehydrated post-fixed Air-dried direct smears from the cell pellet, liquid-based cytology preps: SurePath, ThinPrep, Cytospin smears, or filters).

• The 3^rd objective is best evaluated in a DQ-stained direct smear prepared from the effusion fluid prior to concentration.

• The 4^th objective is evaluated better if a cell-block is prepared for elective ancillary tests, including immunostaining, as indicated.

Close

Figure 3:

DQ-stained direct smear of effusion fluid prior to concentration.

Export to PPT

Close

Figure 4:

DQ-stained Mega-funnel Cytospin smear of concentrated effusion fluid (metastatic ovarian carcinoma, peritoneal fluid).

Export to PPT

Close

Figure 5:

PAP-stained SurePath (AutoCyte) smear (metastatic pulmonary carcinoma, pleural fluid).

Export to PPT

Close

Figure 6:

PAP-stained ThinPrep (Reactive mesothelial cells (red arrows), pleural fluid).

Export to PPT

Close

Figure 7:

Hematoxylin and eosinstained sections of cell-block of an effusion (metastatic colonic adenocarcinoma, pleural fluid).

Export to PPT

Close

Figure 8:

Metastatic adenocarcinoma (pleural fluid). An example with morphologically identifiable unequivocal ‘second foreign population’ (red arrow NC) other than mesothelial cells (blue arrow RM) and inflammatory cells (brown arrow IC) in DQ- and PAP-stained smears. This ‘second population’ of cells (red arrow NC) is easier to identify with the DQ stain (a) than with the PAP stain (b). IC, inflammatory cells; RM, reactive mesothelial cells; NC, neoplastic cells. [a, Diff-Quik-stained Cytospin smear; b, PAP-stained ThinPrep smear (a,b, 100X zoomed).]

Export to PPT

Based on this approach, effusion fluids may be processed by preparing a DQ-stained direct smear of unconcentrated effusion specimen [Figure 3], DQ-stained Megafunnel Cytospin smear from the concentrated cell pellet [Figure 4], PAP-stained SurePath preparation [Figure 5], PAP-stained ThinPrep preparation [Figure 6], and PAP stained direct smears (Wet-fixed or Air-dried rehydrated postfixed).^[1] In addition, the remaining effusion fluid may be processed for cell-block preparation for hematoxylin and eosin (HE)-stained sections [Figure 7], and for elective immunocytochemistry.^[2]

It is recommended to process fluids contaminated with significant fresh blood after removing the erythrocytes. Such an approach allows optimum display of cells crucial for cytologic interpretation of corresponding effusion fluids by preventing interference from the contaminating erythrocytes, especially in DQ-stained Cytospin preparations. However, in our experience with Sure Path preparations, erythrocytes are lysed during the processing and are not a significant problem, unless the specimen is predominantly blood In ThinPrep, however, blood in the specimen may interfere with the final cellularity in the cytology preparation by occluding the pores of the ThinPrep filter. Another simple approach is removal of blood with ammonium chloride based lyzing reagent such as BloodLyz^[3,4] used for flow cytometry. The erythrocyte depleted sediments can be used for preparing cytology preparations and for cell-block.

Both liquid based cytology (LBC) methodologies: SurePath [Figure 5] and ThinPrep [Figure 6] affect the thickness of individuals cells due to partial fixation in the LCB preservative/ fixative. This may compromise the cytomorphological evaluation of cellular/nuclear details. This introduces challenges related to optimal imaging of cellular details. However, both methods usually achieve concentration of cellular components in paucicellular specimens in a defined area on the slide.

• Proper clinical details are a critical component for optimum interpretation of effusion fluid cytology. Availability of clinical information is important, even during the stage of specimen processing, so that the appropriate protocol is implemented. With a clinical history suggestive of cancer, preparation of cell-block for elective immunocytochemical evaluation is recommended as a routine protocol.

The initial evaluation of both DQ- and PAP-stained smears, even at low magnification, provides a general impression regarding the morphologic features of the cells. This further narrows the differential diagnosis. Examination at a higher magnification may assist in the final interpretation with or without tests such as histochemical stains, immunostains, flow cytometry, or other ancillary studies. With increased experience in cytomorphology, ancillary studies may not be required in straightforward cases without sacrificing accuracy of the final interpretation [Figure 1].

INTERPRETATION APPROACH [FIGURES 1, 2]^[5-26]

• In general, the standard diagnostic criteria of malignancy may not be applicable to single cells in serous fluids except in specimens with high-grade neoplasms [Figures 12, 13]. Final interpretation as to the cells in effusion fluids being a metastatic neoplasm has to be based on collective information deduced from various features mentioned below:^[27]

1. The cytomorphology of reactive mesothelial cells overlaps significantly with neoplastic cells^[28-30] [Tables 1]. Different clinical conditions [Table 2] may induce remarkably florid reactive changes in mesothelial cells, which may lead to the pitfall of misinterpreting them as malignant cells.^[30]

2. A single population of cells with a wide spectrum of morphologic continuum between all cells may be considered as evidence of reactive mesothelial cells or mesothelioma cells. However, in rare cases, metastatic cancer cells may outnumber the mesothelial cells, resulting in a single population of neoplastic cells [Figure 9].

3. • Demonstration of a ‘second foreign population’ other than mesothelial and inflammatory cells in effusion fluids is generally consistent with metastatic neoplasm [Figure 8].

4. For neoplasms such as some sarcomas, this ‘second foreign population’ may be obvious and is usually easy to interpret. But sarcomas are rare in serous effusions; a previous history of sarcoma is almost always known.^[28,29]

5. • The Romanowsky group of stains (e.g. DQ) highlight the ‘second population’ of cells with significant ease as compared to the PAP stain [Figure 8].

6. • Any ancillary test such as immunocytochemistry, with a properly organized immunopanel to distinguish and identify the ‘second population’ of cells, facilitates objective interpretation.

7. • Identical orientation of serial sections of cell-blocks on all slides facilitates precise evaluation of a ‘subtractive coordinate immunoreactivity pattern’ (SCIP)^[35].

8. Most mesothelioma cells do not show marked atypia except for their increased size, which is easy to overlook. The important morphologic clues are quantitative (the presence of numerous isolated scattered cells of mesothelial lineage in a hypercellular smear) and qualitative (many large groups of three-dimensional cells of mesothelial lineage) [Figure 1[5]]. Both features are evaluated better in DQ-stained smears.

9. Apoptosis in unequivocally non-inflammatory cells, especially if associated with mitotic figures, correlates with malignancy. However, unequivocal interpretation of such apoptotic cells as non-inflammatory cells with apoptosis may not be possible in some specimens, especially if the cells are solitary, small in size, and scant in numbers.

Close

Figure 9:

Pulmonary adenocarcinoma (pleural fluid). The non-cohesive metastatic cancer cell population is the predominant cell population without being seen as a ‘second population’ (a–d). Some apoptotic tumor cells (arrows in c and d) are present. If indicated, immunocytochemistry would facilitate confirmation of these cells as non-mesothelial. The predominant cell population shows immunoreactivity for thyroid transcription factor-1 (TTF-1), consistent with lung primary (e,f). Most of the cells without nuclear immunoreactivity for calretinin are carcinoma cells. A rare mesothelial cell (arrows in g,h) with nuclear (and cytoplasmic) immunoreactivity is seen as an internal positive control. [a–d: PAP-stained Cytospin smear, e–h: immunostained on cell-block sections (a, 10X; b, 40X; c,d, 100X; e,g, 40X; f,h, 100X).]

Export to PPT

Close

Figure 10:

Proliferation spheres (metastatic mammary carcinoma, pleural fluid). Cellular and nuclear details are better seen in a PAP-stained SurePath preparation (blue arrow in c), especially at the periphery under higher magnification (blue arrow in c), as compared to DQ-stained Cytospin smear (d, e, f). [a–c, PAP-stained Cytospin smear, d–f, DQ-stained cytospin smear (a, 20X; b, 100X; c, 100X zoomed; d, 20X; e, 100X; f, 100X zoomed).]

Export to PPT

Close

Figure 11:

Proliferation spheres (metastatic mammary carcinoma, pleural fluid). They may be round (a,d,e,f) or oblong (b,c). Conglomeration of more than one sphere (a) and oblong shape (b,c) of others may lead to irregular configurations, resembling papillary-like structures, especially in cytology preparations. They may be solid (a–f) or hollow (g–i). Some proliferation balls associated with metastatic adenocarcinoma may show formations of gland-like spaces (arrows) (j–l). [a–l, HE-stained cell block sections (a -c 40X; d -l, 100X).]

Export to PPT

Close

Figure 12:

Metastatic poorly differentiated non-small-cell carcinoma of lung (pleural fluid). The specimen predominantly shows non-cohesive solitary cells (a–d). The neoplastic cells (NC inset d) are pleomorphic, with unequivocal features of malignancy. Mitotic figures (MF in c) are present in concert with apoptotic cells, which show apoptotic bodies (AP in d). Solitary cells (NC in d) may resemble high-grade lymphoma cells. DQ-stained preparation and immunocytochemistry on cell-block sections may help in cases with unknown primary. This patient had poorly differentiated adenocarcinoma of lung. AP, apoptotic cells; MF, mitotic figure; NC, neoplastic cells. [a–d, PAP-stained SurePath preparation (a, 10X; b, 40X; c,d, 100X; d inset, 100X zoomed).]

Export to PPT

Close

Figure 13:

Metastatic poorly differentiated adenocarcinoma (peritoneal fluid). Specimen predominantly shows neoplastic cells (NC in insets c,d) with unequivocal features of malignancy. Mitotic figures (MF in c) are present in concert with apoptotic cells with apoptotic bodies (AP in d). Some cells show unequivocal cohesive pattern (arrowheads), consistent with carcinoma. Such specimens do not need ancillary help of DQ stain or cell-block sections for immunocytochemistry, unless a search for unknown primary is indicated. AP, apoptotic cells; MF, mitotic figure; NC, neoplastic cells. [a–d, PAP-stained SurePath preparation (a, 10X; b, 40X; c,d, 100X; insets of c,d, 100X zoomed).]

Export to PPT

CYTOLOGIC PARAMETERS^[5-23]

Various cytomorphologic features applicable during interpretation of effusion cytology are summarized below:

IMMUNOCYTOCHEMICAL APPROACH

If a ‘second foreign population’ is suggested in a DQ-stained preparation but cannot be confirmed with PAP stain, it may be analyzed further by immunostaining of cell-block sections [Figure 2[2^a]]. This may be needed in effusions secondary to the well-differentiated adenocarcinomas, especially of the ovary and breast. In addition to the confirmation of the second population by evaluation with SCIP approach in cell-block sections^[43], immunocytochemistry also helps in determining the primary site. Examples include prostate-specific antigen in prostatic adenocarcinoma, CDX2 in colonic adenocarcinoma, calcitonin in medullary carcinoma of the thyroid, GATA 3 in breast carcinoma, and thyroglobulin in thyroid carcinoma etc.

REPORTING SYSTEM

A five-tier cytopathology terminology system practiced in other organ system is also recommended for reporting serous fluid cytopathology findings as mentioned below:^[43,44]

• Non-diagnostic (ND)

• Negative for malignancy (NM)

• Atypia of undetermined significance (AUS)

• Suspicious for malignancy (SFM)

• Positive for malignancy (PM) – primary and metastatic

As discussed before, some cases may require ancillary testing. This is especially applicable to a significant number of specimens in atypia and suspicious ‘holding’ preliminary categories for objective confirmation or ruling out ‘second’ foreign population of metastatic cancer and narrow the grey-zone with minimum reporting in the atypical-suspicious categories. In the definitive malignant category, ancillary testing may be required for differential diagnosis of primary site including unknown primary and for evaluating prognostic markers such as Her2/Neu, Estrogen receptor, etc.

Most of the causes which lead to grey-zone reporting may be related to quantitatively and qualitatively sub-optimal specimens. Lower specimen volume and scant cellularity may lead to sub-optimal cell-blocks without the ability to perform effective ancillary testing such immunohistochemistry. In addition, in some cases the quality of the specimen may be sub-optimal due to in vivo degenerative changes associated with long-standing effusion, improper collection, transportation, or storage of the specimen leading to poor preservation related compromization.

Diagnostic criteria for atypical and suspicious categories are not methodically studied and show significant interinstitutional and interobserver variation. Because positive for metastatic disease implies late disease with decision not to plan for curative management but initiate palliative management, it is recommended to be conservative with efforts to avoid false positives. In addition, collection of most of the effusion fluids is relatively less complex and so repeating a properly collected specimen in optimum quantity may be performed for definitive interpretation as clinically indicated.

Most of the limitations related to this problem may be minimized by including an appropriate comment with such equivocal reports. The comment should explain the probable reason for the gray-zone interpretation along with a recommendation how to resolve it.

The freshly collected, unfixed effusion specimen should be transported to the laboratory without delay. If delay is expected, it should be submitted on ice without freezing. The fluid specimen may be refrigerated at 4°C until processing up to 24 hours period. The specimen is usually stable up to 7 days at 4°C for some testing including cell-block if required.^[45,46]

Entire collected effusion fluid volume, up to 1000 ml, should be sent for cytopathological evaluation. Although, the minimum quantity is variable and depends on the cellularity of the specimen, it should be at least 50 ml so that cell-blocks can be prepared. However, 20–30 mL of cellular serous fluids may be processed but with relatively sub-optimal cytology preparation processing with potential to compromise the opportunity to make cell-blocks.