SEARCH

SEARCH BY CITATION

Keywords:

  • mesothelioma;
  • mesothelial hyperplasia;
  • adenocarcinoma;
  • pleural effusion;
  • cytology

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

BACKGROUND

The differentiation of malignant epithelial mesothelioma, nonmucinous pulmonary adenocarcinoma, and mesothelial hyperplasia by direct light microscopic examination is challenging. The detection of long microvilli, an ultrastructural feature unique for malignant epithelial mesothelioma, requires immunolabeling of epithelial membrane antigen to be visible by light microscopy. The purpose of this study is to report that this feature is now made conspicuous enough to be detected by light microscopy by a different approach in cytologic processing.

METHODS

The study includes 9 cases of mesothelioma, 50 cases of mesothelial hyperplasia, and 50 cases of nonmucinous pulmonary adenocarcinoma. Direct smears were made from the sediments of fresh or refrigerated effusions, air dried, rehydrated with normal saline for 30 seconds to create retraction halos, and processed with the Ultrafast Papanicolaou stain (UFP).

RESULTS

Long microvilli in thick bundles as well as short, fine projections radiating from the free surface of the epithelial mesothelioma cells were observed within the retraction halos in pleural effusions processed by UFP. These features were not observed in the UFP-processed hyperplastic mesothelial cells or in the pulmonary nonmucinous adenocarcinoma cells.

CONCLUSION

Direct visualization of long microvilli via UFP may aid in the distinction of malignant mesothelioma from mesothelial hyperplasia and adenocarcinoma before immunohistochemical and electron microscopic studies are performed. Cancer (Cancer Cytopathol) 2003;99:17–22. © 2003 American Cancer Society.

Effusion cytology offers an opportunity for the early detection of mesothelioma, because the first sign is commonly an unexplained bloody pleural effusion. Distinguishing mesothelioma from adenocarcinoma and mesothelial hyperplasia is of paramount importance, but is difficult using cytology or small pleural biopsy. This difficulty is reflected by the countless studies in pathology journals to search for new antibodies to distinguish mesothelioma from mesothelial hyperplasia and adenocarcinoma as reviewed by Ordonez.1 The positive immunomarkers for adenocarcinomas include E-cadherin,2 carcinoembryonic antigen (CEA), LeuM1, and TTF-1,3 which is particularly useful to identify those of lung origin. The positive immunomarkers for mesothelial cells include cytokeratin 5/64 and calretinin.5 Calretinin, with its characteristic nuclear and peripheral cytoplasmic staining pattern, is particularly useful in labeling mesothelial cells. Malignant mesothelial cells may be distinguished from hyperplastic mesothelial cells by the “thick” membranous staining by epithelial membrane antigen (EMA),which was first advocated by Leong et al.6 EMA was immunolocalized ultrastructurally along the microvilli, but not along the nonvillus flat cell membrane of mesothelioma cells as reported by van der Kwast et al.7 The thickness of EMA membranous staining reflects the length of microvilli associated with malignant mesothelial cells. If the immunohistochemical results are equivocal, direct visualization of long, often branching, microvilli by transmission electron microscopy is the gold standard used to diagnose mesothelioma.8, 9 In an ultrastructural study, Sakuma et al.10 reported that the length/diameter ratios of microvilli measured 19.1 ± 7.0 for 12 cases of mesothelioma, 9.2 ± 2.4 for 11 cases of mesothelial hyperplasia, and 9.1 ± 2.2 for 12 cases of adenocarcinoma. Scanning electron microscopy11 and atomic force microscopy12 are also used to identify the long microvilli of mesothelioma cells and to differentiate those cells from secretory blebs of adenocarcinoma cells.

This study reports a method that enables direct visualization, by light microscopy, of the long microvilli of mesothelioma cells in pleural effusions. A review of literature (Medline search from 1966 to May 2002, reviews, books, and atlases) found no published reports of the cytologic feature of mesothelioma cells described in this study.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Pleural effusions from 9 patients with mesothelioma, 50 patients with reactive mesothelial hyperplasia, and 50 patients with nonmucinous pulmonary adenocarcinomas were processed by the Ultrafast Papanicolaou (Pap) stain (UFP) protocol at the Cytopathology Laboratory at New York University Medical Center from 1996 to 2001. They form the basis of this study. Direct oval smears were made from the sediments from fresh or refrigerated pleural effusions. They were air dried, rehydrated with normal saline for 30 seconds, and fixed in alcoholic formalin for 10 seconds. They were stained with Hematoxylin II and Cytostain (Richard-Allan Scientific, Kalamazoo, MI) as previously described for UFP,13, 14 except that the time for Hematoxylin II and Cytostain was increased to 20 and 40 seconds, respectively. The smears were first scanned using a 4× objective lens to locate retraction halos in areas of monolayer spread, which were usually along the periphery of the oval smear. Cells that directly adhered to the glass slide were examined using a 40× lens to identify potential spikes and hairy projections and confirmed by examination using a 60× or a 100× objective lens.

The UFP-stained smears were examined independently for the presence or absence of long microvilli by two rotating residents who had no previous knowledge of the diagnosis. Immunohistochemical study was performed on cell blocks in all nine cases of mesothelioma, whereas electron microscopic study was used only in three cases.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The clinical, cytologic, immunohistochemical, electron microscopic, and histologic features of the nine patients with mesothelioma are shown in Table 1. The results of the visualization of long microvilli in UFP-processed pleural effusions are shown in Table 2. Long microvilli were observed in eight of nine cases of mesothelioma, but not in 50 cases of mesothelial hyperplasia or 50 cases of nonmucinous pulmonary adenocarcinoma (Fig. 1). Based on nine cases of mesotheliomas, the sensitivity for long microvilli as a marker for mesothelioma is 88.9% and the specificity is 100%. Two types of long microvilli were observed: fine hairy projections (Fig. 2A–D) radiating from the free surface of the mesothelioma cells and bundles of extra long microvilli (Fig. 2E–H) projecting from the free cell surface. In Case 1, long microvilli were absent on both UFP and electron microscopy and the EMA immunostain was negative. The final diagnosis of mesothelioma was made by the second pleural biopsy over a 4-month period and by the same loss of heterozygosity in the spindly tumor cells microdissected from the biphasic mesothelioma cells found in the pleural biopsy and in the epithelial tumor cells microdissected from the cell block.

Table 1. Clinical, Cytologic, Immunohistochemical, Histology, and Electron Microscopic Findings of Nine Patients with Mesotheliomasa
PatientAgeGenderHistoryNo. of thoroco-centesesClassic cytologic featuresLong microvilli on UFPThick membrane on EMAaElectron microscopyOther ImmunostainsaNo. of pleural biopsiesDefinitive diagnosis
  • ND: not done; UFP: Ultrafast Pap stain; EMA: epithelial membrane antigen; SOB: shortness of breath; LOH: loss of heterozygosity; CEA: carcinoembryonic antigen.

  • a

    Immunohistochemistry was performed on cell blocks. LOH study performed on epithelial tumor cells microdissected from cell blocks made from pleural effusion. Spindly tumor cells of the biphasic mesothelioma microdissected from pleural biopsy showed similar alterations.

167MaleS.O.B. Bloody effusions× 5YesNoNoShort microvilliAE1/AE3+CEA-LeuM1-×2Biopsy 2: +invasion LOH
254FemalePPD(+),R/ o TB Yellow effusions× 3YesYesYesNDCalretinin+ CEA-LeuM1-×4Biopsy 4: +invasion
352MalePPD(+),R/ o TB Yellow effusions× 3YesYesYesNDCalretinin+ MIB1: +40% LeuM1-×1Cytology and biopsy
465FemaleS.O.B. Bloody effusions× 3YesYesYesNDCalretinin+ LeuM1-×1Cytology and biopsy: +invasion
574FemaleS.O.B. Bloody effusions× 1YesYesYesLong microvilliCalretinin+ LeuM1-×1Cytology and biopsy
679FemaleColon carcinoma Yellow effusions× 3Yes, but small bland nuclei, chronic inflammationYesYesLong microvilliCalretinin+ CEA-LeuM1-×1Biopsy: +invasion
757Male+Asbestos +Tobacco Bloody effusion× 1YesYesYesNDCalretinin+ MIB1: +40%NDCytology
872Male+Tobacco Bloody effusions× 2YesYesYesNDCalretinin+ CEA-LeuM1-NDCytology
938FemaleMesothelioma Bloody effusion× 1YesYesYesNDCalretinin+ CEA-LeuM1-×1Previous pleural biopsy: +invasion
Table 2. Visualization of Long Microvilli in UFP-Processed Pleural Effusions
 Mesothelial hyperplasia (%)Mesothelioma (%)Adenocarcinoma (%)
  1. UFP: Ultrafast Pap stain.

Long microvilli (−)50 (100.0)1 (11.1)50 (100.0)
Long microvilli (+)0 (0.0)8 (88.9)0 (0.0)
Total50 (100.0)9 (100.0)50 (100.0)
thumbnail image

Figure 1. Comparison of look-alike tissue fragments with scalloped borders. Mesothelial hyperplasia (A), mesothelioma identified by the microvilli projected from free cell surface (B), and pulmonary adenocarcinoma (C). (Ultrafast Papanicolaou stain; original magnification × 1000 (A–C).

Download figure to PowerPoint

thumbnail image

Figure 2. Spectrum of microvilli of mesothelioma cells demonstrated by Ultrafast Papanicolaou stain protocol. (A–D) Mesothelioma cells showing long microvilli on free surface. One of them is branched (A, top microvillus). (E–H) Thick bundles of extra long microvilli conspicuously project from the free surface of mesothelioma cells. Some show unipolar toupee-like projection in addition to shorter, hairy projections along the circumference of the mesothelioma cells (H). Original magnification × 1000 (A–H).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The cytologic features of mesothelioma processed using the Romanovsky stain were described by Lopes-Cardoso.15 The cytologic features of mesothelioma processed as described by Klempman,16 Naylor,17 Tao18 and Ehya,19 are still widely used today. These features include high cellularity, clusters with scalloped borders, abundant dense cytoplasm, surface blebs, cell engulfment, intercellular windows, round nuclei with macronucleoli, and giant mesothelial cells. Stevens et al.20 reported that when they used true papillary aggregates, multinucleation with atypia, and cell-to-cell apposition as the variables, a logistic model correctly predicted 95.4% of 44 cases of malignant mesothelioma versus 100% of 46 cases of adenocarcinoma. Using variables of nuclear pleomorphism, macronucleoli, and cell-in-cell engulfment, the logistic model correctly predicted 100% of 44 mesothelioma cases versus 90% of 30 mesothelial proliferation cases. Tao21 illustrated blebs on the surface of some mesothelial cells, but similar structures were also present in some hyperplastic mesothelial cells and some adenocarcinoma cells.

Increased resolution by UFP has been previously reported.22 Air drying allows the cells to settle on the glass slide and become flattened. Rehydration with normal saline, a technique developed by Chan and Kung in 1988,23 allows water to enter the flattened air-dried cells through osmosis, restoring some of their original shape. The cell clusters retract from the original air-dried border, creating halos (Fig 3A). The retraction may have a “straightening” effect on the microvilli, which then appear clearly as spikes and hairy projections (Fig. 3B) within the retraction halos. The blebs found on the surface of mesothelioma cells may have been long microvilli processed by the Pap method,21 but the resolution of traditional processing was not high enough to differentiate those blebs from the secretory blebs of adenocarcinoma cells or hyperplastic mesothelial cells.

thumbnail image

Figure 3. Rationale for conspicuous microvilli in Ultrafast Papanicolaou stain protocol is attributed to the retraction halos (A) caused by rehydration of air-dried cells by normal saline. Notice that the width of the halo is proportional to the diameter of the tissue fragments and single cells. (B) The centrifugal force of retraction of mesothelioma cells and tissue fragments from the surrounding cells pulled and straightened the microvilli so that they are visible within the clear background of the retraction halos. Original magnification × 1000 (A); × 400 (B).

Download figure to PowerPoint

A cautionary note is necessary, as not all cases of mesothelioma possess long microvilli and not all mesothelial cells in the same case possess long microvilli. In this small study of nine mesotheliomas, already one case lacks the characteristic long microvilli in UFP-stained smears, in EMA-immunostained cell blocks, and in an ultrastructural study. For obvious reasons, the number of cases involved in an ultrastructural study of mesotheliomas are fewer than those included in immunohistochemical studies. Because thick membranous EMA labeling indicates the presence of long microvilli, the percentage of mesothelioma cases possessing long microvilli can be estimated by employing data generated from large series of EMA immunohistochemical studies. In a large study of 141 cases of mesothelioma, Wolanski et al.24 reported that 73% of cases exhibited thick EMA immunolabeling of cell membranes. Of their 73 cases of hyperplastic mesothelial cases, 3 had equivocal EMA membranous staining, but none were positive. In a comprehensive review, Whitaker25 stated there are no morphologic criteria that are entirely specific for mesothelioma. Using the morphology of spikes and hairy projections found in UFP-processed pleural effusions in combination with traditional cytologic features may aid in the identification of malignant mesothelial cells.

The pleural effusions of the first seven patients were examined, worked up, and signed out by the author. As an electron microscopist, I always search for ultrastructural features that can secure a difficult cytologic diagnosis. As a result, the long microvilli were deliberately searched and found.

The number of mesothelioma cases in this study are too small to allow a meaningful statistical analysis. The purpose of this study is primarily to call attention to this observation so that cytopathologists in regions with a high incidence of mesothelioma, such as western Australia (one new case each week in average),25 can try this processing method. A test of the sensitivity and specificity of this observation in a large study comprising hundreds of cases may show this cytologic feature to be far less sensitive and specific than that is reported in this study.

The UFP stain is cost-effective. Hematoxylin II ($81.14 per 4 pints) lasts up to a month before losing its potency and Cytostain ($87.71 per gallon) lasts up to 6 months if refilled whenever the solution evaporates. Little maintenance is required because filtering in not needed for either of the two stains, in contrast to traditional Pap stains. Laboratories equipped with a centrifuge with a rotor for 50-mL conical tubes can try this protocol. The staining steps are as simple as those of frozen sections and can be performed by pathologists. The smears are easily screened because the background is clean, the focal plane is closer to the surface of the glass slides than the Pap method, and the tissue fragments are highlighted by the retraction halos.

The need for ancillary tests depends on the experience and confidence level of the cytopathologist. Experienced cytopathologists who specialize in effusion cytology may be able to make definitive and accurate diagnoses of mesothelioma without ancillary tests. However, most cytopathologists may require cell blocks with immunostains for confirmation. Visualization of long microvilli on UFP-processed smears at the initial evaluation of a pleural effusion may give the confidence to the cytopathologist to limit the number of immunostains to two. i.e., calretinin to confirm the mesothelial nature and MIB1 to confirm the high proliferation index. If the immunostains are equivocal, additional immunostains and electron microscopy may be used as needed.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The author thanks Dr. Jerome Holaysan and Dr. Joon Yim for participating as observers.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES