What's New in the Pleura?

Apart from worrying about the diagnosis of malignant mesothelioma, the pleura is one of those anatomic regions that most pathologists tend to ignore. However, the diagnosis of malignant mesothelioma is, in itself, sometimes a complicated and controversial issue, and there is more to the pleura than just malignant mesothelioma.

In this issue of the Archives, 4 articles address the issue of pleural tumors.1–4 Several general points emerge from these articles. The first point is that it is crucial to know the gross appearance when making a diagnosis. Gross appearance can actually be assessed in several different ways. One is the usual naked-eye examination by the pathologist. However, gross appearance is often accurately reflected in a computed tomography scan or equally well understood from the operator's report. Why is this crucial? The answer is that localized pleural tumors, even when cytologically malignant, can often be cured by surgical excision. Thus, there is a substantial salvage rate for malignant forms of localized solitary fibrous tumors, localized malignant mesotheliomas, and localized primary pleural synovial sarcomas; whereas, diffuse malignant mesotheliomas and diffuse primary pleural sarcomas, including diffuse primary synovial sarcomas, have a dire prognosis, as do carcinomas metastatic to the pleura. Indeed, a careful reading of the literature suggests that the overall cure rate for many types of localized pleural malignancies may be as high as 50%.5,6 

The localization of a pleural malignancy is vital information to convey to surgeons and pulmonologists taking care of patients, and it is information that is easy to miss if one only looks at slides. A good example is localized malignant mesothelioma. This rather uncommon tumor microscopically mimics diffuse malignant mesothelioma in every way, but in a recent series from the United States– Canadian Mesothelioma Reference Panel,6 48% of patients were alive without evidence of recurrence for periods up to 11 years, whereas the vast majority of patients with diffuse malignant mesothelioma are dead in 2 years. Only the knowledge that one is dealing with a localized process allows a correct diagnosis.

It is impossible to write about malignant mesothelioma without raising the issue of immunohistochemistry. Ordóñez4 has provided a concise and elegant overview of both new and traditional markers that may help in diagnosis. Perhaps the most important point made by Ordóñez is that “at present … an absolutely specific marker for mesothelioma has not yet been recognized.” In other words, all of the markers that usually stain mesotheliomas sometimes stain carcinomas, and vice versa. This problem can be ameliorated to a certain extent by using several markers. I recommend selecting 2 “mesothelioma” markers and 2 “carcinoma” markers for routine use; in the majority of cases, such a combination will give a clear answer. Which markers should you pick? Immumochemistry is, for better or worse, as much art as science, and, therefore, you should use markers that work in your laboratory. This sounds simplistic, but the literature, even when not overtly contradictory, only provides a theoretic guideline, and blind dependence on published articles may lead to anomalous or simply incorrect results. For example, as noted by Ordóñez, E-cadherin and N-cadherin antibodies do not really seem to work in most hands, although some authors have obtained good results with them; similarly, thrombomodulin immunoreactivity varies enormously between reports. Run a series of established mesotheliomas and carcinomas and see what results you actually obtain in your laboratory. Alternately, refer your cases to a reference laboratory that has established immunoreactivity profiles for specific antibodies.

Another important point that emerges from the paper by Ordóñez is that one must fit the immunochemical profile to the location of the tumor and the sex of the patient. For example, when dealing with tumors arising in the thorax, Wilms tumor 1 (WT1) is a highly specific and sensitive marker of mesothelioma. However, for a diffuse peritoneal tumor in a woman, WT1 may be very misleading, because serous carcinomas of the ovary or peritoneum are immunoreactive for this marker. Similarly, thyroid transcription factor 1 (TTF-1) is very useful when the differential diagnosis is between mesothelioma and metastatic adenocarcinoma from the lung but is of no use in other situations. Ordóñez also discusses the problems created when metastatic renal cell carcinoma is in the differential diagnosis. As much as we hate to admit it, one size does not fit all. Remember, the most useful “special” stain is the hematoxylin-eosin: if it does not look like a mesothelioma on hematoxylin-eosin, no number of special stains will make it one!

A major problem that has arisen in dealing with mesothelial proliferations in the serosal membranes is the separation of benign from malignant processes. This is, in many ways, not only harder than the separation of mesotheliomas from other malignant tumors in the pleura but also much more crucial. At best, a misdiagnosis of a benign reaction as a malignant mesothelioma produces terrible psychologic effects on the patient and, at worst, subjects the patient to radical and/or toxic treatment that has a high morbidity and mortality. Guidelines for making this separation are summarized in the article by Cagle and Churg,3 and more detail is provided in Churg et al.7,8 Unfortunately, experience has shown that morphology frequently fails, immunoreactivity is of no real use in this setting, and, as Cagle and Churg3 point out, even experts in the field often disagree about a given case. What is needed is a molecular marker, and, thus far, no such marker has been described. Therefore, the most important point made in these articles is that, when in doubt, it is far preferable to make a diagnosis of atypical mesothelial proliferation or atypical mesothelial hyperplasia and allow the clinician to decide whether more tissue or watchful waiting is appropriate.

One final area that needs mention is the issue of asbestos exposure and tumors of the serosal membranes. These days, the letters “meso” in a diagnosis line almost always lead to litigation. Two important points about asbestos should be borne in mind. First, a history of asbestos exposure, or the lack thereof, is of no relevance in making a pathologic diagnosis; however, nonetheless, it is common to receive consultations in which the referring pathologist uses that history as an aid to diagnosis or is reluctant to make a diagnosis of mesothelioma because there is no history of exposure. The linkage of mesothelioma and asbestos exposure is not one-to-one; many mesotheliomas are caused by asbestos exposure, but many are not. Would you hesitate to diagnosis a squamous cell carcinoma of the lung on a bronchial biopsy if you were told the patient never smoked? Obviously not! What a history of asbestos exposure may provide is a cause of the tumor in question, but even that is a complicated issue and certainly not a part of a routine diagnosis.

Second, diffuse malignant mesothelioma is the only tumor for which an epidemiologic link to asbestos exposure has been established. Well-differentiated papillary mesothelioma and localized malignant mesothelioma have no such link, and the settings in which these tumors occur, particularly the high proportion of tumors in women, suggest to me that most, and possibly all, are not related to asbestos exposure. Pathologists use the term “mesothelioma” for these latter types of tumors as much for mental convenience as anything else, but the fact that we currently believe that a tumor is of mesothelial origin is not, in and of itself, any indicator of causation. Consider the entity now known as solitary fibrous tumor but originally referred to as “fibrous mesothelioma.” Was it caused by asbestos when it was a fibrous mesothelioma but acquired some other etiology when it became a solitary fibrous tumor?

The bottom line is that the relationship of asbestos exposure and mesothelial tumors is often not straightforward, but this is only a problem for attribution of etiology; it is never a problem for diagnosis.