Immunocytochemistry of effusion fluids: Introduction to SCIP approach

UNIQUENESS OF EFFUSION IMMUNOCYTOCHEMISTRY

Intricacies in finding and locating the cells of interest in cell-block sections of effusion fluids is a very important limiting factor, with the potential to adversely affect the final results. If it is not approached with special consideration, it may lead to improper immunocytochemical interpretation and, eventually, a suboptimal final result. Although it is not unique for effusion fluid cytology, it is quite common in this scenario to face the challenge associated with evaluation of coordinate immunoreactivity because of the higher frequency of small cell groups and solitary cells in effusion specimens. While processing and interpreting effusion fluid immunocytochemistry with objectivity, the following attributes should be considered:

1. Application of SCIP (Subtractive co-ordinate Immunoreactivity Pattern) approach: The pattern of cancer cells in effusions varies from solitary scattered cells to small, cohesive groups. It may be more difficult to find the same cells or groups of cells in adjacent serial sections on different slides. Most of the individually scattered, abnormal cells, however, will be present in at least a few 4 μm thick serial sections [see Figures 1, 2, 3,4, 6,9 ]. It is important to know the sequence of these serial sections and to orient them identically on the slides to identify the same cells (or small group of cells) more precisely on the adjacent serial sections for proper evaluation of coordinate immunoreactivity pattern [Figure 1]. To achieve this, we routinely orient all serial sections identically on slides and label them sequentially [Figures 2, 3, 4]. This simple approach can help in some unexpected situations where the immunostaining pattern is not straightforward. If this SCIP approach is practiced routinely, it expedites and simplifies the immunocytochemical evaluation of effusion fluids in all cases.

2. The most effective approach for diagnosis of metastatic effusions is the confirmation of a ‘second-foreign’ noninflammatory population of cells other than mesothelial cells. But this does not help to distinguish between reactive and neoplastic mesothelial cells. The interpretation of epithelioid mesothelioma should be based on other features after demonstrating the abnormal cells to be mesothelial in nature. • For reproducible results, it is important to select an immunopanel which will fundamentally highlight all of the mesothelial and inflammatory cells to create the basic map for confirmation of a ‘second-foreign’ population by the subtractive coordinate immunoreactivity pattern (SCIP) approach [see Figure 1].

3. The challenge of distinguishing cells of epithelioid mesothelioma from reactive mesothelial cells has to be approached differently and must be based on the quantity (numerous vs a few) and the quality (numerous large groups vs a few small groups) of abnormal cells with proper clinical and radiologic correlation. Once it is determined that the cytologic features favor mesothelioma rather than reactive mesothelial cells, it is relatively simple, with the aid of immunocytochemistry to exclude adenocarcinoma and more recently with molecular pathology testing [Figures 5, 19].^[5]

4. The final interpretation of any immunoprofile is the result of comparative evaluation of the database accumulated from the information in reported studies that were performed predominantly on formalin-fixed paraffin-embedded tissue sections. The immunoreactivity pattern for a variety of immunomarkers may not remain the same when other fixation and processing protocols are used.^[6] • Consequently, it is prudent to perform immunocytochemistry on formalin-fixed cell-block sections only and to avoid other protocols such as the evaluation of inappropriately processed cell-block with exposure to alcohol and alcoholic fixatives (e.g. CytoLyt) prior to formalin-fixation^[7,8] or various cytology smears (direct smears: wet fixed in alcohol or acetone; air-dried fixed with alcohol; air-dried smears rehydrated and postfixed in formol alcohol; liquid-based cytology smears: SurePath or ThinPrep, Cytospin smears, etc.).^[9,10]

   Thus, for reproducible results, a standardized protocol with steps comparable to the processing of formalin-fixed paraffin-embedded tissue sections is essential.^[11-14] Deviating from such a practice may lead to suboptimal results with loss of reproducibility, which is frequently observed in clinical practice and related publications.

5. For the objective confirmation of an immunoprofile highlighting different types of cells in the effusion, it is important to evaluate the coordinate immunoreactivity pattern of in the same cells [see Figure 1]. • This is not possible with cytology smears since the same cells cannot be followed on different smears. In contrast, serial sections of cell-block allow evaluation of immunoreactivity by different immunomarkers in same cells in adjacent serial sections (coordinate immunoexpression).[15,16]

6. The proteinaceous effusion fluid around suspended cells may contribute to unexpected non-specific immunoreactivity. The non-specific staining of adjacent inflammatory cells, especially when it is substantial, may hinder the evaluation of membranous immunostaining patterns.

7. Discrepant results between formalin-fixed paraffin-embedded tissue sections of surgical pathology material and effusion fluid cell-block sections are not uncommon. The variables responsible for such discrepancies include sample size (tiny cell groups or single cells), selection of fixatives, antigen retrieval methods (i.e. heat-induced epitope retrieval, enzyme digestion, etc.), antibody clones used, antibody titer, and other variations in immunostaining protocols. Additional causes for variable results reported by different studies include variation in laboratory sensitivities and study size.^[17,18] Furthermore, both the qualitative (pattern of immunostaining: membranous, cytoplasmic, nuclear, etc.) [Table 1] and quantitative criteria for interpretation of immunoreactivity may vary between pathologists and different institutions.^[19]

Close

Figure 1:

Basic immunopanel for evaluation by subtractive coordinate immunoreactivity pattern (SCIP) approach.

Export to PPT

Close

Figure 2:

Metastatic colonic adenocarcinoma, peritoneal fluid. a. Serial sections of cell-blocks oriented identically on all slides in a sequential order. The respective numbers identify their relative position with each other: 1, HE; 2, vimentin; 3, cytokeratin; 4, LCA (CD45); 5, calretinin; 6, WT-1; 7, CD-X2. b. Enlarged view showing slide numbers 4 and 5, emphasizing identical orientation (arrows) of cell-block sections on all glass slides. Identical orientations of all serial sections facilitate immunocytochemical evaluation with the SCIP approach. If not properly organized in this fashion, it may be impossible in effusion specimens with a predominance of small cell groups and solitary cells to evaluate the coordinate immunoreactivity pattern in relation to various immunomarkers.

Export to PPT

Close

Figure 3:

Cell-blocking (with NGCBTM kit) and AV marker: subtractive coordinate immunoreactivity pattern (SCIP) approach. Identical orientation of all sections on glass slides with proper labeling of their exact sequence in relation to each other.

From Open Access publication: CytoJournal 2021;18:2. https://dx.doi.org/10.25259/Cytojournal_83_2020

Export to PPT

Close

Figure 4:

Serial numbering and orientation of cell-block sections.

Export to PPT

Close

Figure 5:

Algorithm for immunocytochemical evaluation of effusions (in conjunction with Figure 1).

Export to PPT

Close

Figure 6:

Metastatic colonic adenocarcinoma, peritoneal fluid. The neoplastic cells (red arrow NC) are immunoreactive for pan-cytokeratin (B) and non-immunoreactive for vimentin (A), CD45 (C), calretinin (D), and WT-1 (E). They have nuclear immunoreactivity for CDX2 (F), which is consistent with a colonic primary. The reactive mesothelial cells [with immunoreactivity for calretinin (nuclear-cytoplasmic) (D), WT-1 (nuclear-cytoplasmic) (E), vimentin (A)] and inflammatory cells [with immunoreactivity for vimentin (A) and LCA (C) ] in the background can be subtracted from neoplastic cells to deduce a diagnostic coordinate immunoreactivity pattern. LCA, leukocyte common antigen; NC, neoplastic cell; RM, reactive mesothelial cell; WT-1, Wilms’ tumor1. [A–F, Immunostained cell-block sections.]

Export to PPT

Close

Figure 7:

Metastatic ovarian carcinoma, peritoneal fluid. The neoplastic cells (red arrow NC) are immunoreactive for pan-cytokeratin (B), BerEP4 (D), and non-immunoreactive for vimentin (A) and calretinin (C). The reactive mesothelial cells [with immunoreactivity for calretinin (nuclearcytoplasmic) (C), vimentin (A), and cytokeratin 7 (E)] and inflammatory cells [with immunoreactivity for vimentin (A)] in the background can be subtracted from neoplastic cells to deduce a diagnostic coordinate immunoreactivity pattern. With a history of ovarian carcinoma, coordinate immunoreactivity for cytokeratin 7 (E) and non-immunoreactivity for cytokeratin 20 (F), it is consistent with metastatic ovarian carcinoma. In smears, the morphology of the carcinoma cells overlapped that of reactive mesothelial cells. NC, neoplastic cell; RM, reactive mesothelial cell. [A–F, Immunostained cell-block sections.]

Export to PPT

Close

Figure 8:

Metastatic lung carcinoma, pleural fluid. The specimen shows scattered solitary neoplastic cells (red arrows NC) with rare mesothelial cells (blue arrow RM). The neoplastic cells are immunoreactive for pan-cytokeratin (B) and non-immunoreactive for vimentin (A), LCA (C), and calretinin (D). Rare reactive mesothelial cells (blue arrow RM) [with immunoreactivity for calretinin (nuclear-cytoplasmic) (D)] and inflammatory cells (arrowhead) [with immunoreactivity for LCA (C)] in the background can be subtracted from the neoplastic cells to deduce a diagnostic coordinate immunoreactivity pattern. This is consistent with poorly cohesive singly scattered carcinoma cells, which were also non-immunoreactive for the melanoma marker. Nuclear immunoreactivity for TTF-1 (E) with a clinical history of lung mass is consistent with a lung primary. NC, neoplastic cell; RM, reactive mesothelial cell; TTF-1, thyroid transcription factor-1. [A–E, Immunostained cell-block sections.]

Export to PPT

Close

Figure 9:

Metastatic mammary adenocarcinoma, pleural effusion. Specimen with proliferation spheres. The neoplastic cells (red arrows NC) are immunoreactive for cytokeratin 7 (D) and BerEP4 (E). They are non-immunoreactive for vimentin (A), CD68 (B), and calretinin (C). Rare mesothelial cells (blue arrow RM) [with immunoreactivity for vimentin (A), calretinin (nuclear-cytoplasmic) (C), and cytokeratin 7 (D)] and inflammatory cells (arrowheads) [with immunoreactivity for vimentin (A) and CD68 (C)] in the background can be subtracted from the neoplastic cells to deduce a diagnostic coordinate immunoreactivity pattern. Cytokeratin 7 immunoreactivity with a clinical history of mammary carcinoma is consistent with a breast primary. NC, neoplastic cell; RM, reactive mesothelial cell. [A–E, Immunostained cell-block sections.]

Export to PPT

Close

Figure 10:

Metastatic mammary adenocarcinoma, pleural effusion. Specimen with solitary neoplastic cells as the predominant population. The neoplastic cells (red arrows NC) are immunoreactive for BerEP4 (D) and estrogen receptors (E). They are non-immunoreactive for vimentin (A), CD68 (B), and calretinin (C). Rare mesothelial cells (blue arrow RM) [with immunoreactivity for vimentin (A) and calretinin (nuclear-cytoplasmic) (C)] and inflammatory cells (arrowheads) [with immunoreactivity for vimentin (A) and CD68 (B)] in the background can be subtracted from the neoplastic cells to deduce a diagnostic coordinate immunoreactivity pattern. Immunoreactivity for ER with a clinical history of mammary carcinoma is consistent with a breast primary. ER, estrogen receptors; NC, neoplastic cell; RM, reactive mesothelial cell. [A–F, Immunostained cell-block sections.]

Export to PPT

Close

Figure 11:

Metastatic small cell carcinoma, pleural fluid. Specimen with proliferation spheres. The cytology preparations demonstrated morphologic features of small cell carcinoma (see Figure 5). The neoplastic cells (red arrows NC) are immunoreactive for cytokeratin 7 (A), TTF-1 (nuclear) (C), and various neuroendocrine markers (D,E,F). Rare mesothelial cells and inflammatory cells were present in the background. They could be subtracted from these cells by constructing a coordinate immunoreactivity pattern based on findings in different serial sections. The immunoreactivity for cytokeratin 7 (A) and TTF-1 (C) with nonimmunoreactivity for cytokeratin 20 (B) is consistent with a lung primary. NC, neoplastic cell; TTF1, thyroid transcription factor-1. [A–F, Immunostained cell-block sections.]

Export to PPT

Close

Figure 12:

Large B-cell lymphoma, peritoneal fluid. PAP-stained Cytospin smears (a–c) showed non-cohesive solitary cells (red arrow NC in c) with many apoptotic bodies (arrowheads in c). These cells in PAP-stained smears resemble non-cohesive poorly differentiated non-small cell carcinoma with a solitary cell pattern (compare with Figure 8). They are present in the HE-stained cell-block section as small groups (d). By the SCIP approach, neoplastic cells could be separated out from cytokeratin 7 (A) and calretinin (B) immunoreactive mesothelial cells (blue arrow RM). After following them in serial sections, the neoplastic cells (red arrow NC) were immunoreactive for CD20 (C) and nonimmunoreactive for CD3 (E). They expressed Bcl2 (D). Flow cytometry demonstrated the monoclonal nature of the CD20 immunoreactive cells. The patient had lymphoma of the small bowel. This was a consult case and a DQ-stained smear, which is usually helpful to evaluate lymphoma cells, was not available. NC, neoplastic cell; RM, reactive mesothelial cell. [a–c, PAP-stained Cytospin smear; d, HE-stained cell-block section, A–E, immunostained cell-block sections.]

Export to PPT

Close

Figure 13:

Adenocarcinoma, peritoneal fluid. Neoplastic cells are immunoreactive for EMA (a) and HBME-1 (b) with a cytoplasmic immunostaining pattern (arrows) with focal blotchy immunostaining along the cell membrane. Compare Figure 16 with membranous immunostaining highlighting the microvilli in mesothelial cells. [a,b, Immunostained cell-block sections (a,b, 100X zoomed).]

Export to PPT

Close

Figure 14:

SCIP with dual color immunostaining (metastatic mammary carcinoma, effusion fluid). Reactive mesothelial cells (blue arrowheads in a,b) are immunoreactive for both vimentin (brown) and cytokeratin 7 (red) (inset of a). The metastatic carcinoma cells (arrows), both isolated (c) and small cohesive groups (d), do not show immunostaining for vimentin (no brown color) but only immunostaining for cytokeratin 7 (red) (inset of c). The corresponding cells in adjacent serial section with dual color immunostaining were also negative for calretinin (brown) but immunoreactive for BerEP4 (red) (not shown here). As expected, the background inflammatory cells are highlighted with vimentin (brown) immunostaining only; they are negative for cytokeratin 7 (red) (a,b,c,d). [a–d, Dual color immunostained cell-block sections—vimentin (brown) followed by cytokeratin 7 (red), (a–d 100X).]

Export to PPT

Close

Figure 15:

Metastatic mammary adenocarcinoma, pleural fluid. Metastatic carcinoma cells (arrows in a–c) are immunoreactive only for cytokeratin 7 (red) without vimentin (brown) immunostaining (b). For additional confirmation, these cells are weakly immunoreactive for BerEP4 (red) and negative for calretinin (brown) (c). As expected, the background inflammatory cells (blue arrowhead in b) are highlighted with vimentin (brown) only; they are negative for cytokeratin 7 (red) (b). As noted in c, the reactive mesothelial cells were rare in this specimen. [a, HE-stained section; b, vimentin (brown) followed by cytokeratin 7 (red) (dual color immunostained cell-block section); c, calretinin (brown) followed by BerEP4 (red) (dual color immunostained cell-block section); (a–c, 40X).]

Export to PPT

Close

Figure 16:

Two color BerEP4 (Brown) with vimentin (Red): the neoplastic cells (red arrows) are immunoreactive for BerEP4 (Brown) and non-immunoreactive for vimentin (Red). The background inflammatory cells and rare reactive mesothelial cells are immunoreactive for vimentin (blue arrowheads). Two color immunostaining for BerEP4 with vimentin highlights the basic map of the effusion fluid cell-block section and facilitates detection of ‘second foreign population’ consistent with metastatic tumor cells. These would be followed in serial sections (Figure 1) to evaluate coordinate immunoreactivity for other immunomarkers used depending on the clinical scenario for the differential diagnosis of primary site. (Metastatic adenocarcinoma, pleural fluid, cell-block section).

Export to PPT

Close

Figure 17:

Metastatic adenocarcinoma: Mullerian primary (pleural fluid). A. HE-stained cell-block section: Section shows of proliferation of the neoplastic cells (red arrow). B. Two color BerEP4 (Brown) with vimentin (Red): the neoplastic cells (red arrow) are immunoreactive for BerEP4 (Brown) and non-immunoreactive for vimentin (Red). The background inflammatory cells and rare reactive mesothelial cells are immunoreactive for vimentin (Red). C. Cytokeratin 7: the neoplastic cells (red arrow) show cytoplasmic immunoreactivity. D. PAX8: the neoplastic cells (red arrow) show nuclear immunoreactivity in most of the neoplastic cells. E. Estrogen Receptor (ER): the neoplastic cells (red arrow) show nuclear immunoreactivity in some neoplastic cells. Two color immunostaining for BerEP4 with vimentin highlights the basic map of the effusion fluid cell-block section and facilitates detection of ‘second foreign population’ consistent with metastatic tumor cells. These cells are followed in serial sections to evaluate coordinate immunoreactivity for other immunomarkers used depending on the clinical scenario for the differential diagnosis of primary site. In this case, the tumor cells showed immunoreactivity for Cytokeratin 7 with nuclear immunoreactivity for PAX8. The tumor cells also showed nuclear immunoreactivity for ER in some tumor cells. This immunoprofile is consistent with Mullerian primary (patient had history of ovarian carcinoma).

Export to PPT

Close

Figure 18:

Metastatic gastric adenocarcinoma, peritoneal fluid. The metastatic carcinoma cells are only immunoreactive for cytokeratin 7 (red) (brown arrows in b). Vimentin (brown) immunoreactive inflammatory cells (blue arrows in b) and the reactive mesothelial cells with calretinin nuclear immunoreactivity (brown) (blue arrow) are seen in the background (c). These inflammatory cells (blue arrows in b) and the reactive mesothelial cells (blue arrow) are negative for cytokeratin 7 (red) and BerEP4 (red) in b and c, respectively. Discussion in brief. Numerous isolated neoplastic cells as predominant population (a) suggested a differential diagnosis of high-grade large cell lymphoma and mesothelial cells with florid reactive changes. Immunoreactivity for cytokeratin 7 (brown arrow) in b excluded lymphoma and concurrent lack of vimentin (brown) immunostaining excluded mesothelial cells. The patient had a history of gastric carcinoma. The pattern was consistent with metastatic gastric carcinoma. A few vimentin immunoreactive (brown) inflammatory cells were present in the background (b). Rare reactive mesothelial cells with brown nuclear and cytoplasmic immunostaining (blue arrows) were also present in the background with inflammatory cells (c). The metastatic cells were also weakly immunoreactive for BerEP4 by one color immunostaining, but this weak immunoreactivity was not detectable in the dual color immunostained sections even after using the same titer of antibody dilution used for routine one color immunostaining (c) ^[30a] [a, HE-stained section; b, vimentin (brown) followed by cytokeratin 7 (red) (dual color immunostained cell-block section); c, calretinin (brown) followed by BerEP4 (red) (dual color immunostained cell-block section); (a–c, 40X).]

Export to PPT

Close

Figure 19:

Algorithmic approach for diagnostic work-up for evaluation of mesothelial proliferations.^[61-63]

Export to PPT

Table 1. Characteristic immunostaining pattern with immunomarkers in effusion immunocytochemistry (modified from 26,27)

Immunomarker	Diagnostic immunomorphological (immunostaining) patterns				Remark
	Nuclear	Cytoplasmic	Membranous	Microvillous
vimentin		X
Ber-EP4 Claudin-4		X	X	X1
Calretinin	X2
*I. Other nuclear immunomarkers	X*				See notes below
**II. Other cytoplasmic immunomarkers		X**			See notes below

1Rare cases of mesothelioma.

22Diagnostic pattern is nuclear, but synchronous cytoplasmic immunostaining is common. *I. Nuclear immunostains: BAP1, beta-catenin (may be with cytoplasmic), CDX2, Estrogen receptor (ER), GATA3, MDM2, PAX2, PAX8, NKX3.1, progesterone receptor (PR), S-100 protein (may be with cytoplasmic), SATB2, SPC10, TFE3, TTF-1, WT1, etc. **II. Cytoplasmic immunostains: Arginase, CD31,CD34,CD117, Cytokeratin (CK) 7, CK 5/6, CK 20, Claudin 4, DOG1, DPC4(SMAD4), GCDFP-15, Glypican 3, HepPar-1, Her2/Neu, Inhibin, LCA (CD45), Melan A (MART1), Mammaglobin, Mesothelin (Cytoplasmic/membranous), PAP, PLAP, PSA, RCC, Thyroglobulin, Uroplakin III, Villin (brush border- Apical), etc.

Other immunomarkers with immunostaining patterns: Bile canalicular immunostaining: polyclonal CEA (pCEA, not mCEA: monoclonal CEA) and CD10 in hepatocytes

PERSPECTIVE

Cell-blocks are the preferred choice for immunocytochemical evaluation of effusions. However, if only very scanty material is present, it is not advisable to perform immunocytochemistry on such material. In such a setting, an adequate additional specimen is recommended to be submitted. • Because malignant effusions usually reaccumulate quickly, acquiring a new sample is generally not a problem. It is not uncommon to submit only a small fraction of a large volume of effusion fluid collected.

Therefore, it should be specifically communicated in the final cytopathology report as a recommendation with a comment: ‘Recommend submission of most of the drained effusion fluid (up to 1000 mL). Larger specimen volume facilitates retrieval of adequate cellular material in cell-block sections for evaluation with ancillary studies including immunocytochemistry.’

Cell-blocks offer dual benefits including precise immunomorphologic interpretation and better standardization, with results comparable to those with surgical pathology specimens. Multiple sections from a single cell-block allow for multiple immunomarkers to be evaluated. Additionally, the cell-blocks may be archived and made available for other types of testing including molecular pathology and proteomics as indicated at later stage.^[8,20]

The frequency of using cytology smears for immunocytochemical testing for different applications is increasing.^[21-23]

With the advent of heat-induced epitope retrieval techniques, the quality of immunostained cytology smears has improved remarkably. Immunocytochemical evaluation of effusion smears has also been reported and discussed.^[6,10,24] However, most experts do not recommend immunostaining of cytology smears of effusion fluids as a routine practice.^[6,24]

On rare occasions, however, performing immunocytochemistry on effusion smears may be justified: e.g. patients with a known primary neoplasm showing a distinct immunoreactivity pattern. Immunostaining may be performed on cytology smears in such cases, if only a scant effusion specimen is available and cell-blocks are not possible. However, applicable limitations, such as the inability to evaluate the coordinate immunoreactivity pattern in cytology smears and possible lack of cross-verification should be weighted prior to final interpretation.

Straightforward positive–negative interpretation of immunostained sections is an artificial simplification, but it may serve its purpose in many situations. However, this approach needs to be refined in complex situations such as evaluation of cell-block sections of effusion fluids, which are affected by many variables not applicable to other specimens. Some of these variables are:

1. The cells to be evaluated in the cell-block sections may not be obvious. They usually have significant morphologic overlap with the reactive mesothelial cells which form the predominant background population in the cell-block sections of effusion fluids.

2. In addition to reactive mesothelial cells, a variety of inflammatory cells are also intricately intermingled with the cells under scrutiny.

3. It is challenging to follow the same cells in different cell-block sections on different slides for the evaluation of immunoreactivity for various immunomarkers to study the coordinate immunoexpression [see Figure 1]. This challenge is more apparent if the cells in the effusion are predominantly singly scattered cells.

4. The protein-rich effusion fluid may introduce some nonspecific immunoreactivity, not otherwise observed when the corresponding immunomarker is applied to routine surgical pathology tissue specimens.

5. The protocols for cell-block preparation may vary and introduce remarkable variations at different steps, such as:

   • the fluid may not have been submitted fresh but in a fixative

   • these fixatives themselves may be of different types, such as ethanol or other proprietary fixatives, without any data on the exact composition

   • transportation conditions: time and temperature variation

   • delayed processing of unfixed specimen after collection

   • variation in methods of cell-block preparation: agar gel method, picric acid method, thrombin clot, and so on. However, recently there is opportunity to standardize the methods. For example currently available ready to use NextGen CelBloking (NGCB kits)^[25] has introduced processing comparable to surgical pathology specimens by allowing the fresh-unfixed specimen to be fixed directly in 10% formalin to generate formalin-fixed paraffin-embedded cell-block.^[25-27]

Some of these variables are of less significance than others. For example, submission of effusions in fixative, delayed transportation at room temperature, and the picric acid method for cell-block preparation are deleterious for most of the immunomarkers.

For optimal results, the immunostained cell-block sections of effusion fluids should be evaluated immunohistomorphologically in a manner similar to histomorphological evaluation of hematoxylin and eosin (HE)-stained sections, where we consider multiple qualitative and quantitative morphologic features to reach a final interpretation. • All aspects of individual and complementary immunomarkers should be considered collectively [see Figures 1 and 5], rather than applying a reflexive positive–negative approach.

Other immunomarkers

1. Some immunomarkers are neither positive nor negative mesothelial markers, but they have typical immunostaining patterns [see Table 1]. A thick, membranous ruffled pattern highlighting microvilli favors mesothelial cells [see Figure 20]. Adenocarcinoma cells usually show diffuse, coarse cytoplasmic or flimsy, membranous immunostaining patterns [Figure 13]. The immunomarkers in this group are EMA and HBME-1 (and very rarely other non-mesothelial immunomarkers such as BerEP4). Without considering this immunomorphologic correlation of immunostaining patterns, these immunomarkers may lead to misinterpretation.

2. Immunomarkers highlighting various components of effusion fluid are:

   • Non-epithelial (and mesothelial) component by vimentin [including two color BerEP4 (brown) for adenocarcinoma with vimentin (Red) for mesothelial cells and inflammatory cells]^[45]

   • Epithelial component by pan-cytokeratin (e.g. mixture of AE1/AE3 and CAM5.2) or other cytokeratin, including coordinate application of cytokeratin 7 and cytokeratin 20

   • Inflammatory cell component by LCA (leukocyte common antigen, CD45) or PGM1 (CD68) or mixture of LCA and PGM1. LCA (CD45) is relatively strong and sensitive. It does not require an antigen retrieval step. As compared to this, PGM1 (CD68) needs an antigen retrieval step. If LCA and PGM1 are combined, then the titer of LCA antibody should be adjusted to accommodate the antigen retrieval used for PGM1. For selection of CD68 antibody, in our experience PGM1 usually lacks non-specific immunoreactivity, which is observed frequently with KP1, and so PGM1 may be preferred over KP1.

Close

Figure 20:

Diffuse malignant mesothelioma of epithelial type (DMME), pleural fluid. Neoplastic cells are immunoreactive for EMA (a,b) and HBME-1 (c,d,e) with a membranous immunostaining pattern (arrows) highlighting long, slender microvilli (arrowheads). [a–e, Immunostained cell-block sections (a, 100X; b, 100X zoomed; c, 40X; c, 100X; e, 100X zoomed).]

Export to PPT

This group of immunomarkers creates a basic map for evaluation, localization, and identification of neoplastic cells [see Figures 1 and 5] by the SCIP approach (see below).

THE ANTIBODY PANEL

• In general, monoclonal antibodies offer better specificity with less non-specific background staining. Recent availability of Rabbit monoclonal antibodies^[50] for immunohistochemistry has further increased the sensitivity for many immunomarkers. As immunoexpression patterns may be heterogeneous, it is prudent to select at least two immunomarkers. Immunoprofiles of various neoplasms and lesions may overlap significantly. Consequently, it is advisable to use the optimum number of immunomarkers for appropriate objective interpretation. • Using a suboptimal restricted immunopanel may lead to spurious interpretation, with the possibility of undesirable consequences for patient care.

Depending on the institutional protocol and the algorithm used, immunocytochemical evaluation with different immunomarkers may be sequential (to economize) or concurrent (for rapid turnaround time) [see Figure 5].

If the presence of neoplastic cells (a ‘second-foreign’ population) is confirmed with the SCIP approach [see Figures 1 and6 through 12] in correlation with the cytomorphology, the next task is to characterize these neoplastic cells for their type and primary site as indicated by clinical scenario [see Figure 5].

1. Vimentin and two color BerEP4 (brown) with vimentin (red), Claudin-4, pan-cytokeratin (mixture AE1/AE3 and CAM5.2), and LCA (CD45) or PGM1 (CD68) or preferably a mixture of LCA and PGM1 help to create the basic map for SCIP approach for identifying a second-foreign population [see Figures 1, 6A-C, 7A,B, 8A-C, and 9A-C].

2. The addition of calretinin and WT-1 (or cytokeratin (CK) 5/6, CK 7), as positive mesothelial markers, identifies both reactive and neoplastic mesothelial cells in the basic map [Figures 6D,E, 7C, 8D, and 9C].

3. If, after microscopic subtractive assessment of serial sections, non-mesothelial/non-inflammatory cells are present as a second foreign population, they are usually consistent with a metastatic neoplasm.

   • If this population is immunoreactive for cytokeratin (rarely also for vimentin), it is consistent with carcinoma [see Figures 1B, 5B, 6B, 7B, 8B, and 9D]. This population is usually immunoreactive for one or all ‘negative’ mesothelial markers such as, BerEP4, B72.3, Claudin-4, CEA, and/or MOC-31 [see Figures 5, 7D, 9E].

   • If the second population is non-immunoreactive for cytokeratin but immunoreactive for vimentin, it may represent sarcoma, melanoma (immunoreactive for melanoma markers), or lymphoma (immunoreactive for LCA and other lymphoma markers, with flow cytometry or immunocytochemistry as indicated) [see Figures 5, 12]. Except for lymphoma, these are rare causes of malignant effusion.

In summary, the initial immunopanel should include two color BerEP4 (brown) with vimentin (red), cytokeratin 7, calretinine, and claudin-4 for evaluation of SCIP to detect or confirm an initial suspicion of a second population of neoplastic cells in cytology smears.

Modify the initial immunopanel depending on the clinical scenario:

• Possible primary neoplasms known—add a short panel, such as TTF-1 for lung, PSA and NKX3.1 for prostate, GATA-3+ER+ mammaglobin+CRxA-01 for breast, CK 20+CDX2+SATB2+ for colon, PAX8 for Mullerian primary, CA IX for renal primary, appropriate lymphoma panel for lymphoma, S-100 protein and other melanoma markers for melanoma, and so on [see Table 2].[23,30,31,46,47,51-53]

• If the primary neoplasm is unknown: add a suitable immunopanel (in concert with CK 7 with CK 20 for evaluation of CK7/CK20 coordinate immunoreactivity patterns) for many broad category of immunomarkers [see Table 2].^{[16,26,27,29,32,54]}

In cases with numerous, solitary, scattered, non-mesothelial, malignant cells exhibiting negligible numbers of mesothelial cells, the second-foreign population of cells may be the only population with non-mesothelial immunoprofile (nonimmunoreactivity for nuclear calretinin and vimentin with immunoreactivity for BerEP4) [see Figure 8]. Depending on cytomorphology and clinical history, these cells may demonstrate an immunoprofile consistent with a carcinoma (cytokeratin+. usually vimentin-) [see Figures 8, 10], lymphoma (CD45+), sarcoma (vimentin+, cytokeratin−), or melanoma (S-100+ with one of the melanoma markers).^{[16,29,30,47,55,56]}

• The availability of a past history of cancer remarkably improves the interpretation by proper modification of the basic immunopanel. The challenge is relatively complex in cases with malignant effusions secondary to neoplasms without a known primary site. In such cases, a wider range of site-specific immunomarkers may have to be included [see Table 2].^{[16,26,27,29,32,54]}

Most metastases to serous cavities are secondary to adenocarcinoma and majority of which are vimentin nonimmunoreactive except a few listed in Table 3. Because of this, some cases may show immunoreactivity for both vimentin and BerEP4 with red and brown color immunostaining in the same metastatic tumor cells.

On similar lines, although a vast majority of adenocarcinomas are Ber-EP4 immunoreactive, a few listed in Table 4 may be Ber-EP4 non-immunoreactive. Other immunomarkers including Claudin-4 would cover many of the tumor types in this group.

It is important to highlight that, even if immunocytochemistry does not demonstrate a second foreign population of nonmesothelial and non-inflammatory cells, mesothelioma is not excluded. The diagnosis of mesothelioma is based on qualitative and quantitative cytologic features in concert with a clinical and radiologic correlation (discussed in another review article in the series). Reports have suggested that p53 and OV632 are definitive markers for distinguishing between mesothelioma and reactive mesothelial cells, but such a magic immunomarker does not exist at this date.^[57,58] However, after evaluating clones E29 and Mc5 of EMA antibodies, E29 has been reported to distinguish between reactive mesothelial cells and neoplastic mesothelial cells. In this study, 75% of malignant mesotheliomas showed immunoreactivity for EMA with clone E29, but reactive mesothelial cells were nonimmunoreactive in all cases.^[59] Malignant mesothelioma demonstrates high Ki-67 index.^[60] Mesothelioma cells demonstrate strong membranous GLUT-1 immunoreactivity with strong nuclear immunoreactivity for p53.^[60] Lack of desmin immunoreactivity with EMA immunoreactivity favors malignant mesothelioma cells over reactive mesothelial cells, which usually show immunoreactivity for desmin with lack of EMA immunoreactivity.[60]

A few additional immunomarkers applied in a algorithmic manner along with molecular testing is suggested for distinguishing mesothelioma versus reactive mesothelial cells [Figure 19].^[61-63] Loss of BAP1 immunoreactivity in mesothelial cells immunoreactive for mesothelial immunomarkers is consistent with mesothelioma. Some cases with BAP1 immunoreactive mesothelial cells with loss of MTAP cytoplasmic immunoreactivity and deletion on CDKNA FISH are consistent with mesothelioma. Other patterns with undetermined patterns should be tested further with molecular test/Next Gen Sequencing (NGS) [Figure 19].^[61-63]

ADENOCARCINOMA VERSUS MESOTHELIOMA/ REACTIVE MESOTHELIAL CELLS

An increasing number of antibodies that could potentially be applied for the evaluation of effusions have been developed. Many of these are already commercially available and are applied directly to effusion cytology.^{[26,27,32,71]} Those not listed may be added during the immunocytochemical evaluation of effusions by individual pathologists, depending on the clinical scenario, personal experience, and institutional resources.^[6,28,31]

Historically, the interpretation of mesothelial cells (both reactive and neoplastic) has been based on their nonimmunoreactivity for immunomarkers such as BerEP4, Claudin-4, B72.3, monoclonal carcinoembryonic antigen (mCEA), and LeuM1. These immunomarkers have been labeled as ‘negative’ mesothelial markers.^{[4,18,19,72-75]} However, unexpected immunoreactivity of mesothelial cells for these immunomarkers has been reported in a small percentage of cases.^[65,76,77] The final interpretation in such cases can be reached if an optimum immunopanel is applied and the immunoreactivity patterns (such as cytoplasmic vs membranous due to bushy microvilli) are also taken into consideration [Figure 20]. Meanwhile, the search for additional, definitive immunomarkers for mesothelial cells continues. Some of the commonly used antibodies in the differential diagnosis of adenocarcinoma from mesothelial cells are listed in Table 5.

Until recently, satisfactory immunomarkers for mesotheliomas were not available for their ‘positive’ identification. The mesothelioma cells were usually interpreted by a ‘negative’ immunoprofile, but this approach has been less reliable for the definitive identification of mesothelioma cells. Recently, some ‘positive’ mesothelioma markers such as calretinin and WT-1 have been described. Depending on the clinical differential diagnosis, these immunomarkers show promising results if properly interpreted with the help of the SCIP approach [see Figure 1]. The immunopanel may be more effective in cases with a known primary than in cases with an unknown primary neoplasm. Other immunomarkers, including CK 5/6, HBME-1, N-cadherin, OV632, thrombomodulin, mesothelin, CD44S, AMAD-2, and the recently reported D2-40, podoplanin,^{[31,34,36,37]} have also been claimed as ‘definitive’ for mesothelial cells with variable reports.[17-19,41-43,57,64,66,69,70,73-76,78-97]

Sections immunostained with HBME-1 and EMA show thick, bushy, membranous immunostaining of mesothelial cells. This immunostaining pattern along the cell borders is due to the presence of slender microvilli, which are seen easily under an oil immersion lens [see Figure 20b,e, Table 1]. If adenocarcinoma cells are immunoreactive, they usually show a diffuse, coarse, cytoplasmic or flimsy, membranous immunostaining pattern [see Figure 13].^[98] Proper understanding of the immunoreactivity pattern with these immunomarkers usually obviates the need for electron microscopy (EM) to demonstrate microvilli in malignant cells, which is difficult or impossible with effusion specimens. The most significant challenge is that the malignant cell(s) cannot be identified unequivocally in EM sections of effusion fluids to be selected for evaluation of microvilli. This may lead to misinterpretation of background reactive mesothelial cells with microvilli as mesothelioma cells.

The approach for immunocytochemical evaluation of malignant effusion for an unknown primary neoplasm is comparable to metastases to other sites and is summarized in Table 2.^[8,29,54]

Although the basic approach with SCIP evaluation would be the same, the suggested combination of immunomarkers may be modified for a given laboratory, depending on the pathologist’s experience with a particular immunomarker, the quality of immunostaining, and the accessibility of resources with additional availability of improved immunomarkers in future.^[6,31,99]

ABBREVIATIONS (IN ALPHABETIC ORDER)

DMEM - Diffuse malignant epithelioid mesothelioma

EM - Electron microscopy

HE - Hematoxylin and eosin

Neg – Negative

NGCB kits - NextGen CelBloking kits

Pos – Positive

SCIP - Subtractive coordinate immunoreactivity pattern