Diagnostic Pitfalls In Thoracic Tumors

Transcription

1 1376 Diagnostic Pitfalls In Thoracic Tumors Neda Kalhor, MD Cesar A. Moran, MD, FASCP WEEKEND OF PATHOLOGY AMERICAN SOCIETY FOR CLINICAL PATHOLOGY 33 W Monroe Ste 1600 Chicago, IL 60603

2 Program Content and Disclosure The primary purpose of this activity is educational and the comments, opinions, and/or recommendations expressed by the faculty or authors are their own and not those of the ASCP. There may be, on occasion, changes in faculty and program content. In order to ensure balance, independence, objectivity, and scientific rigor in all its educational activities, and in accordance with ACCME Standards, the ASCP requires all individuals in positions to influence and/or control the content of ASCP CME activities to disclose whether they do or do not have any relevant financial relationships with proprietary entities producing health care goods or services that are discussed in the CME activities, with the exemption of non-profit or government organizations and non-health care related companies. These relationships are reviewed and any identified conflicts of interest are resolved prior to the activity. Faculty are asked to use generic names in any discussion of therapeutic options, to base patient care recommendations on scientific evidence, and to base information regarding commercial products/services on scientific methods generally accepted by the medical community. All ASCP CME activities are evaluated by participants for the presence of any commercial bias and this input is utilized for subsequent CME planning decisions. The individuals below have responded that they have no relevant financial relationships with commercial interests to disclose: Course Faculty: Cesar A. Moran, MD Neda Kalhor, MD Commission on Continuing Professional Development Members: Francois Cady, MD, FASCP Steven H. Kroft, MD, FASCP William E. Schreiber, MD, FASCP Daniel D. Mais, MD, FASCP Robert A. Goulart, MD, FASCP Karen A. Brown, MS, MLS(ASCP) Lynnette G. Chakkaphak, MS, MT(ASCP) Amy J. Wendel, SCT(ASCP) The individuals below have disclosed the following financial relationships with commercial interests: Course Faculty: Commission on Continuing Professional Development Members: Michael D. Feldman, MD, PhD, FASCP Aperio Technology, Inc., CRI Stock Options; Advisory Board; Consultant Bioimagene Sponsored Research & License Patents Copyright 2010 by the American Society For Clinical Pathology All Rights Reserved. Printed in the United States of America. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means - electronic, mechanical, photocopying, recording, or otherwise - without the prior written permission of the publisher.

3 1376 Diagnostic Pitfalls in Thoracic Tumors - NEW Directors: Neda Kalhor, MD, MD Anderson Cancer Center Cesar A. Moran, MD, MD Anderson Cancer Center The thoracic cavity may be the site of a gamut of tumors of different etiologies. In addition, the thorax involves structures such as the pleura and mediastinum that in isolation give rise to specific tumors that may pose problems not only in diagnosis but also in classification. Thus, the present course will address new developments in the handling, classification, and diagnosis of diverse conditions from the three compartments of the thorax the mediastinum, the pleura, and the lung. The material selected will allow for a broader assessment of the different problems that may be encountered in these anatomic structures and will include tumor as well as tumor-like lesions that may mimic malignant conditions. The course will focus on solving problems and provide accurate information for patient care. When important and relevant, information will be provided on the use of more sophisticated studies that although not available in some laboratories, the practitioner must know of their usefulness. The course will be case-based with step by step approach from the conventional morphologic approach to the use of ancillary tools. Six cases will be used as a setting to interact with the audience and to elaborate on the work-up of these cases. Following this course, you will be able to: Develop the skills to diagnosed uncommon thoracic tumors Lower malpractice risk through a better understanding of essential imaging findings. Assess the use of immunohistochemical studies to properly work thoracic tumors. Identify the different nomenclatures used in the classification of thoracic tumors.

4 DIAGNOSTIC PITFALLS IN THORACIC TUMORS Neda Kalhor, MD and Cesar A. Moran, MD American Society for Clinical Pathologists Las Vegas, Weekend of Pathology February

5 Case No. 1 A 53-year-old man presented with chest pain of several weeks duration. Chest X-rays revealed the presence of a large anterior mediastinal mass. Complete surgical resection of the mass was performed. PATHOLOGICAL FEATURES Grossly, a tumor mass of approximately 8 cm was resected. The tumor was covered with thin fibro-membranous tissue. At cut surface, the tumor showed a homogenous appearance with a light tan color and soft consistency. No hemorrhage or necrosis was present. Histological sections showed a tumor composed predominantly of a spindle cell proliferation with sprinkle lymphocytes. The spindle cell proliferation was rather homogenous and in some areas it was forming ill-defined lobules separated by thin fibroconnective tissue. At higher magnification, the cells were elongated with a small nucleus but devoid of cellular atypia and mitotic activity. In focal areas, the cellular proliferation appeared to be breaking the capsule. Immunohistochemical studies were performed showing tumor cells positive for broadspectrum keratin, low molecular weight keratin (CAM 5.2), and keratin 5/6. Other stains including EMA, S-100 protein, and Bcl-2 were negative. DIAGNOSIS Spindle cell Thymoma, Invasive THYMIC EPITHELIAL NEOPLASMS Primary thymic epithelial neoplasms have long been a source of controversy in pathology due to their wide spectrum of histologic appearances, biologic behavior, and clinical manifestations. In fact, such variability has been responsible for difficulties in the classification and prognostication of these tumors. This has led to a proliferation of classification systems in recent years and conflicting views on the best approach to the evaluation of these lesions by pathologists. The pathology of thymomas has thus turned into a complex and controversial issue that has generated much confusion for practicing pathologists. 2

6 Historical Considerations The term thymoma, as currently defined, refers to a neoplastic proliferation of thymic epithelial cells. Throughout the years, numerous attempts at classification of these tumors have been presented in the literature. The most widely accepted classification scheme in the Unites States was the one proposed by Dr. Barnatz at. al. from the Mayo Clinic, which classified thymomas according to their relative proportion of lymphocytes and the shape of the neoplastic epithelial cells into predominantly lymphocytic, predominantly epithelial, mixed, and predominantly spindle cell type (Table I). Numerous clinicopathologic studies of thymoma in large series of patients utilizing this histologic approach, however, failed to find any statistically significant correlation between the morphology and the clinical behavior of these tumors. Despite the apparent unreliability of the various morphological classifications of thymoma for predicting biologic behavior in these tumors, it was soon appreciated by several investigators that clinical staging of the lesions based on their status of capsular integrity afforded a better means for assessing their biologic behavior. For this reason, Levine and Rosai in 1978 introduced the concept of defining thymomas on the basis of their capsular status into benign and malignant, depending on whether the tumor was encapsulated or invasive (Table I). In their classification, Levine and Rosai additionally espoused the concept that invasive tumors displaying overt cytologic features of malignancy should be regarded as equivalent with thymic carcinoma (so-called malignant thymoma type-ii) (Table I). More recently, interest in the morphological classification of thymoma was revived by the studies of Marino and Muller-Hermelink, who presented a novel histologic classification of these tumors based on histogenetic considerations. These investigators proposed that thymomas could be divided on the basis of their cytological features into those derived from the cortical or from the medullary epithelium of the thymus into cortical and medullary thymomas. Cases that contained features of both were regarded as mixed. The authors subsequently modified their approach by adding two additional categories, the predominantly cortical or organoid thymoma, and a fifth category designated as well-differentiated thymic carcinoma (Table I). 3

7 Many studies have been presented in the literature that would appear to validate TABLE I: Review of Major Classifications of Thymoma ( ) Barnatz et al (1961) Levine & Rosai (1978) Muller-Hermelink et al (1989) Predominantly epithelial Benign thymoma Medullary thymoma Predominantly lymphocytic - Encapsulated Cortical thymoma Mixed Malignant thymoma Mixed thymoma Spindle cell thymoma - Type I (invasive) Predominantly cortical - Type II (thymic carcinoma) Well-differentiated thymic carcinoma the clinical use of the Muller-Hermelink classification. The proponents of such classification have claimed that the various morphologic subtypes directly correlate with the probability of invasiveness for these tumors, and that histologic subtyping according to this classification is predictive of clinical behavior independent of stage. Despite the undoubted wide appeal of the Muller-Hermelink classification, major objections have been voiced by experts in the field concerning its applicability for clinical practice. Most of the criticism of the studies supporting the Marino & Muller-Hermelink classification has been centered around issues of inadequate sampling, reproducibility, and reliability for 4

8 accurately predicting the prognosis of these tumors, particularly when dealing with limited biopsy samples (i.e., endoscopic biopsies of large mediastinal masses). Given this controversy, a number of new classification schemes have been proposed by different investigators for these tumors in recent years. In the new AFIP Fascicle on Tumors of the Mediastinum, Drs. Shimosato and Mukai propose a complex classification scheme that takes into consideration the extent, histology, cell type and degree of atypia of the neoplastic cells, and incorporates terminology from various other existing classifications (Table II). The most recent classification scheme presented by Dr. Kuo from Taiwan proposes that thymomas be classified according to their cytokeratin expression profiles. Based on a study of 34 immunostained thymomas, and an additional 113 thymomas without the stains, the author proposed that his approach provided a useful method for the clinical evaluation of thymomas (Table II). Table II: Recent Additional Classifications of Thymoma Shimosato & Mukai (1997) Kuo (2000) By extent: Spindle cell thymoma -Circumscribed Small polygonal -Invasive Mixed -With implants or metastasis Organoid By histology: Large polygonal cell -Lymphocytic, mixed, epithelial Squamoid thymoma By cell type: -Spindle, polygonal, polygonal-oval By cell atypia: -Absent, slight, moderate, marked Given the controversy and lack of consensus regarding thymoma classification, the WHO organization commissioned Dr. Juan Rosai to establish a panel for the study of this topic. After many years of deliberation, a compromise, non-commital formula was devised by the WHO panel for the classification of thymic epithelial neoplasms. The new WHO 5

9 classification schema did not introduce a new terminology but simply assigned a combination of letters and numbers to the various histologic types in the existing classifications of thymoma (Table III). The letters A and B represent thymomas predominantly composed of either spindle or round cells, respectively. Type AB are those composed of both spindle and round cells; and type C is reserved for those showing overt features of carcinoma (i.e., thymic carcinoma). In their WHO monograph, the authors state that this schema is not intended as a new histologic classification of thymoma nor is it meant to replace any previous terminology, but rather should be employed as a means for facilitating comparison among the various terms from the existing classifications. Table III: Comparison of WHO Schema With Other Classifications WHO Schema Barnatz et al Muller-Hermelink et al Type A Spindle cell Medullary thymoma Type AB --- Mixed thymoma Type B1 Lymphocyte rich Predominantly cortical Type B2 Mixed Cortical thymoma Type B3 Epithelial rich Well-differentiated thymic carcinoma Type C Thymic carcinoma --- Although the recently introduced WHO schema does not settle the issue of thymoma classification, the WHO monograph is of importance because it supported and stressed certain valuable concepts that were agreed upon by all the panel members in that 6

10 publication: 1) no histogenetic basis has yet been conclusively demonstrated between the normal anatomic compartments of the thymus and any of the different histologic types for any of the existing classifications; 2) thymic epithelial neoplasms represent a spectrum of lesions that may range from histologically benign to malignant, hence the inclusion of thymic carcinoma as thymoma type C ; and 3) the degree of invasiveness relates more closely to recurrence and outcome than the cytoarchitectural features, to the point of markedly reducing the independent prognostic value of the latter. It follows from review of the above that markedly different opinions continue to exist regarding the issue of thymoma classification and that a widely agreed upon classification remains to be devised. Moreover, it has become increasingly obvious to workers in this field that the complexity of thymoma classification continues to increase with every new proposal and with the introduction of new and increasingly complex (albeit colorful) terms. New Concepts Current opinions appear to continue to be divided between those who contend that histologic classification provides a reliable means for assessing prognosis of thymic epithelial neoplasms, and those who claim that staging represents the best if not the only valid parameter for the evaluation of thymoma. We believe the truth probably lies somewhere in-between, and that proper evaluation of these lesions would benefit from an approach that incorporates, as in other tumor systems throughout the body, a combination of histologic grading and clinical staging of the lesions. In a recent review (Am J Clin Pathol, Vol.111: ,1999) we presented a novel conceptual approach to primary thymic epithelial neoplasms that is based on a combination of grading and staging of these tumors. In traditional pathology, most tumor systems follow a stepwise progression in their histologic evolution leading to progressive loss of differentiation of the tumor cells. Thus, for most epithelial tumor systems, the first step in this progression is represented by carcinoma in-situ, followed by welldifferentiated (usually invasive) carcinoma, moderately-differentiated carcinoma, and finally poorly-differentiated carcinoma. The equivalent of such a spectrum has not been yet recognized in the thymus. One of the reasons for this is the tremendous histological variability that these tumors can display. The other reason is the traditional belief that malignancy in thymoma cannot be predicted based on features of differentiation and atypia because of the overwhelmingly bland appearance of the majority of such tumors. 7

11 Thymic epithelial neoplasms characterized by overt cytologic evidence of malignancy (i.e., thymic carcinoma) have thus been traditionally separated from conventional thymomas and felt to represent a totally different and unrelated entity. Careful study of our cases and review of the literature have led us to believe otherwise, and have demonstrated that the lesions which we call thymic carcinoma and thymoma are closely related entities that most likely represent opposite ends of a single spectrum of differentiation. In a recent study we were able to document the existence of tumors demonstrating direct transitions between areas showing the classical features of thymoma and areas showing unequivocal features of thymic carcinoma. Moreover, in several of these cases, we were able to identify different areas within the same tumor showing a spectrum of differentiation that ranged from classical thymic epithelial cells with round to oval, vesicular nuclei with small nucleoli and abundant rim of eosinophilic cytoplasm, to larger cells with well-demarcated cell borders and features of atypical keratinizing epithelium. These observations have led us to believe that thymic epithelial neoplasms form part of a spectrum of closely related lesions that may display varying histological appearances depending on their degrees of differentiation. Unlike other epithelial tumor systems in which a progression from welldifferentiated through poorly-differentiated carcinoma can be easily determined, thymomas have not lent themselves easily to such categorization because of their great variability in cytological composition and architectural growth patterns. Another difficulty involved in establishing the degree of differentiation in these tumors is the fact that the "normal" thymus can differ dramatically in appearance depending on the age of the individual. Thus, the normal mature thymus of a child will look quite different from the normal but involuted thymus in the adult. The mature thymus in childhood and adolescence will show the prototypical features of this organ characterized by a sharp demarcation between the cortex and the medulla and the admixture of thymic lymphocytes with large, round epithelial cells containing vesicular nuclei and indistinct cell borders with abundant amphophilic cytoplasm. The normal involuted thymus of older adults, on the other hand, will often be characterized by a paucity of lymphocytes and the epithelial cells will frequently adopt the shape of small, oval to spindle cells with scant cytoplasm and inconspicuous nucleoli. Although these contrasting appearances are an expression of the functional (i.e., active vs. inactive) state of the organ, they both represent the normal status of this organ at different stages in its evolution. Tumors 8

12 displaying features that closely resemble these two "normal" appearances of the thymus could therefore be regarded as showing a high degree of differentiation. Commonly accepted features of organotypical differentiation in thymomas on routine microscopy include: 1) a well-developed lobular architecture on scanning magnification; 2) dual cell population (neoplastic thymic epithelial cells and thymic lymphocytes) admixed in various proportions; 3) distended perivascular spaces; 4) areas of "medullary" differentiation characterized by discrete rounded foci predominantly composed of epithelial cells surrounded by a population of cells that resembles the normal cortex; and 5) the bland appearance of the epithelial cells, which lack overt cytological features of malignancy. To the above features we would also add the presence of a bland-appearing spindle cell proliferation with scant lymphocytes, cystic and glandular formations, and rosette-like epithelial structures, which are features that are commonly encountered in the regressed thymus of older adults (Table IV). 9

13 Table IV: Organotypical Features of Thymic Differentiation Normal Mature Thymus of Childhood or Adolescence Normal Mature Thymus of the Adult -Low-power lobulation with scant lymphocytes -Bland spindle cell population -Dual (epithelial/lymphoid) cell population -Cystic and glandular formations -Distended perivascular spaces -Rosette-like epithelial structures -Areas of "medullary" differentiation -Lack of cytological features of malignancy Applying the above parameters, tumors that exhibit most or all of the above features could be categorized as well-differentiated, whereas tumors displaying total loss of these organotypical features would be classed as poorly-differentiated neoplasms. Fortunately, the vast majority of thymic epithelial neoplasms will generally fall within the first group, i.e., that of well-differentiated tumors displaying most of the organotypical features of the normal thymus. Such tumors have been designated by convention as thymoma. Thymic epithelial neoplasms in which most or all of the organotypical features of differentiation of the normal thymus have been lost and the tumor cells already display overt cytological features of malignancy would correspond to those traditionally designated in the literature as thymic carcinoma. There exists a third, smaller group of primary thymic epithelial neoplasms that displays features intermediate between thymoma and thymic carcinoma. 10

14 Such tumors are characterized by the presence of cytological atypia of the tumor cells, yet they still retain many if not most of the organotypical features of differentiation of the normal thymus. These tumors can be regarded as representing an intermediate stage in the spectrum of differentiation of thymic epithelial neoplasms (i.e., moderatelydifferentiated tumors). On the basis of the cytological features of atypia displayed by their tumor cells, we have proposed the designation of atypical thymoma for these tumors. Based on the above considerations, we currently believe that thymic epithelial neoplasms can be reliably classified into three simple categories: well-differentiated thymoma, moderately-differentiated or atypical thymoma, and poorly-differentiated tumors (i.e., thymic carcinoma) (Table V). Assigning a given lesion to any of these various categories does not depend on any purported histogenetic considerations, does not require the use of special stains or advanced techniques not readily available to general pathologists in community practice, and simply requires basic familiarity with the organotypical features of differentiation of the normal thymus and attention to the degree of cytological atypia displayed by the neoplastic epithelial cells on routine microscopy. 11

15 Table V: Classification of Thymic Epithelial Neoplasms According to Grades of Differentiation Type Grading Histological Criteria Thymoma Well- -Preservation of organotypical differentiated features of differentiation -No cytological evidence of atypia Atypical Moderately- -Preservation of organotypical thymoma differentiated featured of differentiation -Mild to moderate cytological atypia Thymic Poorly- -Loss of organotypical features of carcinoma differentiated thymic differentiation -Presence of overt cytological evidence of malignancy Prognostic Features The evaluation of prognosis in thymoma remains a controversial issue. Most studies seem to indicate that staging represents the most important parameter for assessing the clinical behavior of these tumors. Many of the proponents of some of the more recent morphologic classifications, however, contend that histologic subclassification of well-differentiated thymoma represents a valuable independent prognostic criterion for determining clinical behavior and guiding therapy in these tumors. Studies using special techniques, such as determination of ploidy by flow cytometry, immunohistochemical determination of proliferative index, assessment of p53 protein expression, etc., have so far yielded conflicting and inconclusive results. The proponents of the Muller-Hermelink classification have maintained that the different histologic types of their histogenetic classification directly correlate with invasiveness, and therefore can be reliably utilized to predict the biologic behavior of the 12

16 lesions. The majority of such studies, however, have failed to demonstrate any correlation with mortality and survival on long-term follow-up. The fact that certain histologic subtypes (such the predominantly epithelial thymoma) were associated with increased invasive properties, and that others (such as spindle cell thymoma) were more often associated with indolent clinical behavior has been well-recognized in several studies predating the histogenetic classification of Muller-Hermelink and colleagues. The same studies, however, also demonstrated that when stratified according to staging, all the different histologic subtypes ultimately behaved in a similar manner. This begs the question: is histologic subclassification of thymoma, particularly the well-differentiated variants, necessary? We believe the answer to this question is that subtyping of well-differentiated thymic epithelial neoplasms offers no distinct advantage for prognostication, and that determination of the status of capsular integrity constitutes the most important step in the evaluation of these tumors. Another factor that renders subclassification of well-differentiated thymoma nearly irrelevant is the fact that these tumors tend to display marked morphologic heterogeneity, with frequent admixtures of different histologic growth patterns and cells types often being observed within the same tumor mass. In a recent study of 630 consecutive thymomas, we compared the final histologic classification of the tumors with the number of sections examined per case. It was found that when the number of sections examined per case increased, more cases were included in the mixed category, and fewer cases could be assigned to either the pure cortical or medullary types. These findings suggest that histologic subclassification of thymoma, although of academic interest, may be of limited practical relevance for the assessment of prognosis, particularly in limited biopsy tissue samples. Staging of thymic epithelial neoplasms also remains a controversial subject. The most popular staging system for these tumors was introduced in 1981 by Masaoka et al, in a study in which statistically significant differences in survival could be appreciated for the different groups of patients depending on the gross and microscopic status of the capsule, spread into adjacent structures, and presence or absence of metastases. Since then, several refinements and modifications to this staging scheme have been introduced by other investigators. A TNM staging system was also introduced by Masaoka in 1991 for thymic carcinoma, but has been felt to be impractical for the well-differentiated variants of thymoma. The most recent staging proposal was presented in the WHO monograph on the Histologic Typing of Tumors of the Thymus (Table VI). We believe until more extensive data becomes available on the discriminatory value for prognosis of 13

17 these various schemas, a more simplified approach should be favored that basically addresses the distinction between encapsulated (non-invasive), locally invasive, widely invasive, and metastatic tumors. In summary, classification of thymoma need not be a cumbersome task for general pathologists and can be readily accomplished by most experienced surgical pathologists in clinical practice. Given the relative rarity of these tumors and the questionable role that complex substratification by histologic type plays in prognostication and treatment of these tumors, we favor a simplified approach that combines histologic grading (based on organotypical features of differentiation of the thymus and cytologic Table VI: Comparison of various staging schemes for thymoma Masaoka et al (1981) WHO (1999) Suster & Moran (1999) Stage I: Encapsulated Encapsulated Encapsulated Stage II: Invasion of capsule Minimally invasive Locally invasive and/or perithymic fat Stage III: Gross invasion Widely invasive Widely invasive of adjacent organs with or without implants Stage IVA: Implants With implants Metastatic Sateg IVB: Metastases Lymph node mets With distant metastases atypia) with staging (status of capsular integrity/presence or absence of metastasis). It is our contention that this simplified approach affords an equal opportunity for proper management of these patients as the more sophisticated systems currently in existence, which generally will require review of the case by an expert in the field for proper typing and categorization. 14

18 Select References Suster S, Moran CA. Primary thymic epithelial neoplasms: current concepts and controversies, in: Fechner RE, Rosen PP (Eds), Anatomic Pathology, 1997, Vol.2, Chicago, ASCP Press, pp Suster S, Moran CA. Thymoma, atypical thymoma, and thymic carcinoma. A novel conceptual approach to the classification of thymic epithelial neoplasms. Am J Clin Pathol 1999; 111: Suster S, Moran CA. Primary thymic epithelial neoplasms: spectrum of differentiation and histological features. Semin Diagn Pathol 1999; 16:2-18. Shimosato Y, Mukai K. Tumors of the mediastinum, Fascicle 21, in: Atlas of Tumor Pathology, 3 rd. series, Washington, D.C., Armed Forces Institute of Pathology, 1997, pp Bernatz PE, Khonsari S, Harrisson EG Jr, Taylor WF. Thymoma: factors influencing prognosis. Surg Clin N Am 1973; 53: Lattes R. Thymoma and other tumors of the thymus: an analysis of 107 cases. Cancer 1962; 15: Salyer WF, Eggleston JC. Thymoma: a clinical and pathological study of 65 cases. Cancer 1976; 37: Bernatz PE, Harrison EG, Clagett OT. Thymoma: a clinicopathologic study. J Thorac Cardiovasc Surg 1961; 42: Masaoka A, Monden Y, Nakahara Z, Tanioka T. Follow-up studies of thymomas with special reference to their clinical stages. Cancer 1981; 48: Levine GD, Rosai J. Thymic hyperplasia and neoplasia. A review of current concepts. Hum Pathol 1978; 9: Marino M, Muller-Hermelink HK. Thymoma and thymic carcinoma: relation of thymoma epithelial cells to the cortical and medullary differentiation of the thymus. Virchows Arch A (Pathol Anat) 1985; 407: Kirschner T, Muller-Hermelink HK. New approaches to the diagnosis of thymic epithelial tumors. Prog Surg Pathol 1989; 10:

19 Ho FCS, Fu KH, Lam SY. Evaluation of a histogenetic classification for thymic epithelial tumors. Histopathology 1994; 25: Quintanilla-Martinez L, Wilkins EW, Choi N et al. Thymoma: histologic subclassification is an independent prognostic factor. Cancer 1994; 74: Kuo T-T, Lo S-K, Thymoma: a study of the pathologic classification of 71 cases with evaluation of the Muller-Hermelink system. Hum Pathol 1993; 24: Kornstein MJ, Curran WJ, Turrisi AT III, Brooks JJ. Cortical vs. medullary thymoma: a useful morphologic distinction? Hum Pathol 1988; 19: Wick MR. Assessing the prognosis of thymomas. Ann Thorac Surg 1990; 50: Pan C-C, Wu H-P, Yang C-F. The clinicopathological correlation of epithelial subtyping in thymoma: a clinicopathological study of 112 cases. Hum Pathol 1994; 25: Shimosato Y. Controversies surrounding the subclassification of thymoma. Cancer 1994; 74: Dawson A, Ibrahim NBN, Gibbs AR. Observer variation in the histopathological classification of thymoma: correlation with prognosis. J Clin Pathol 1994; 47: Moran CA, Suster S. Current status of the histologic classification of thymoma. Int J Surg Pathol 1995; 3: Kuo T-T. Cytokeratin profiles of the thymus and thymomas: histogenetic correlations and proposal for a histological classification of thymomas. Histopathology 2000; 36: Rosai J. Histological typing of tumors of the thymus. In: World Health Organization International Histological Classification of Tumors. Second edition. Springer Verlag, Berlin, Suster S, Moran CA. Primary thymic epithelial neoplasms displaying combined features of thymoma and thymic carcinoma: a clinicopathologic study of 22 cases. Am J Surg Pathol 1996; 20: Moran CA, Suster S. On the histologic heterogeneity of thymoma: impact of sampling in subtyping and classification. Mod Pathol 2000; 13:214A. Moran CA, Suster S. On the histologic heterogeneity of thymoma: impact of sampling in subtyping and classification. Am J Clin Pathol 2000; 114:

20 Case No. 2 Clinical History A 64-year old man presented with shortness of breath and pain in the left chest wall area. Radiographic examination revealed the presence of pleural effusion and pleural thickening in the left pleura. A pleural biopsy was undertaken. Histopathologic Features Fragments of fibroconnective and adipose tissue show an atypical epithelial cell proliferation dissecting the fibroconnective tissue. In some areas, the atypical cellular proliferation forms chords, tubular and papillary features. The cells are rather homogenous without increased nuclear atypia or mitotic activity. No evidence of necrosis or hemorrhage is seen. Special Studies Immunohistochemical studies were performed using keratin cocktail, keratin 5/6, calretinin, CEA, Leu M1, and B72.3. The cellular proliferation showed positive staining for keratin cocktail, keratin 5/6, and calretinin. All other immunohistochemical studies were negative. Diagnosis: MALIGNANT MESOTHELIOMA, EPITHELIAL TYPE. MALIGNANT MESOTHELIOMA Primary malignant tumors of the pleura are in general dominated by malignant mesothelioma. Not only it is the most common malignant tumor in the pleura but also the one that poses more difficulty in the diagnosis. Due to its legal implications and its varied histopathological appearance, mesotheliomas have been the subject of extensive studies. Interestingly, malignant mesotheliomas are unusual tumors and it has been estimated that their incidence in the United States is of approximately 3 to 7 cases per million persons in the population per year. Although mesotheliomas have been associated to the exposure of asbestos fibers, approximately 50% of individuals affected by mesotheliomas do not disclose an asbestos exposure, leading to believe that their 17

21 etiopathology may be a multifactorial one and not one limited to the exposure of asbestos fibers. However, those cases in which there is an association with the exposure to asbestos are the ones that generally gain more attention because are the ones that are usually followed by legal counsel. In general mesotheliomas are more common in adult individuals over 50 years of age. In those cases in which the tumor is associated with asbestos, the patient had been exposed for over 15 years to the asbestos fibers. The stated latency period for asbestos in mesothelioma is of 15 to 60 years, thus, it would be very unusual to find a case in which the tumor is linked to asbestos in a patient with only a few years of exposure. One other debated issue is the type of asbestos fibers that produce mesotheliomas, arguments against some fibers as non-carcinogenic have been made. This latter, issue is one usually brought up in court by the legal system attempting to defend a particular exposure to asbestos fibers. In this review, we will not get into that issue and will reserve it to the legal establishment. Nevertheless, a documented history of asbestos exposure of more than 15 years is a powerful argument in favor of linking asbestos and mesothelioma in a particular case. What is more interesting is the fact that mesotheliomas, as stated before, can occur in the setting of a negative history of asbestos exposure, and examples of it, are the cases that have been described in children and women (house-wives). Other possible etiopathologic causes in the development of mesothelioma includes radiation, chronic inflammation, viral infections, and diethylstilbestrol. However, some of those associations are to some extend easier to establish by clinical means. Clinical Features Clinical and radiological information play a highly important role in the diagnosis of mesothelioma. Ignoring clinical and radiological information may conduce to an erroneous diagnosis even by the most experienced pathologist. History of longstanding exposure to asbestos whether factual or not, should lead to a careful analysis of a particular biopsy material from a particular individual. However, one of the most important aspects in the diagnosis of mesothelioma is the radiological evaluation. Either a plain chest radiograph or more sophisticated studies such as computerized tomography is an important tool in the evaluation of a possible case of mesothelioma. In cases in which a question of mesothelioma arises, one should inquire in the following aspects: 18

22 Diffuse involvement of the pleura (studding by nodules) Intraparenchymal tumor nodules or masses (peripheral) Diffuse thickening of the pleura Encasement of the lung Unilateral or bilateral pleural involvement Pleural base tumor mass A careful analysis of the radiological studies with the appropriate material for histopathological evaluation should lead both clinicians and pathologists to seek the necessary tools in order to arrive at a more specific diagnosis. In this regard, many times the clinical and radiological aspects of the cases are clear but the available material for histopathological examination is not adequate. In such circumstances, one should not make a definitive diagnosis but rather raise the level of suspicion and request additional material if clinically indicated. Part of this rationale is the fact that the surgical treatment for cases of mesothelioma can be extreme, thus, it is imperative for a pathologist to be absolutely sure about the diagnosis. Furthermore, it is well known that there are other pleural conditions of an inflammatory nature that may clinically and radiologically mimic malignant mesothelioma. Therefore, one should use the clinical and radiological information not to make a diagnosis per se but rather to orient one self in the plan to follow with the use of immunohistochemistry and/or electron microscopy. Ultimately the diagnosis of mesothelioma is a pathological one and not a clinical or radiological diagnosis alone. Pathological Features Gross characteristics: mesotheliomas are tumors with a characteristic gross appearance that rarely pose a problem with the diagnosis. The tumor will show diffuse pleural involvement. In some cases, the tumor follows the intrapulmonary septum and in rare instances the tumor may involve the lung parenchyma with small nodules in the surface of the lung parenchyma. However, the presence of a well defined tumor mass in the periphery, even if the tumor also shows diffuse pleural involvement should alert the pathologist about the possibility of an adenocarcinoma with diffuse involvement of the pleura, such tumors have been coded under the designation of Pseudomesotheliomas adenocarcinoma due to its similarity to malignant mesothelioma. 19

23 Histopathologic characteristics: Mesotheliomas may show a variety of histopathological growth patterns. However, traditionally mesotheliomas have been divided into three categories: Epithelial Sarcomatoid Biphasic (combination of epithelial and sarcomatoid) Epithelial Mesothelioma: probably the most common of the three variants, it has been estimated that represents about 70% of all mesotheliomas. Several distinct histopathological growth patterns of epithelial malignant mesothelioma have been described and in some occasions may represent a diagnostic challenge. Among the variants that have been recognized are: Tubulopapillary Epithelioid Deciduoid Clear cell Glandular Myxoid Adenomatoid As it can be seen, it will depend on the histopathologic growth pattern observed, the degree of discomfort with the case in question. It is comforting however that among those previously stated histopathological growth patterns, the epithelioid and tubulopapillary is the most common. Nevertheless, it is very important to be at least theoretically familiar with the other different growth patterns in order to properly orient our differential diagnosis. Regardless of the histopathological growth pattern, whether the tumor shows clear cell change, myxoid areas, glandular differentiation, or an adenomatoid pattern, one can not overlooked the fact that radiologically the tumor is involving diffusely the pleura. This latter fact should alert and lead the pathologist to alt least rule out the possibility of mesothelioma. Thus, the use of histochemical and immunohistochemical studies becomes a crucial factor in the evaluation of such lesions. Histochemical studies: traditionally and before the advent of immunohistochemistry, histochemical studies play an important role in the diagnosis of mesothelioma. Currently in some circumstances they can solve the problem easily in the more banal cases. The 20

24 use of periodic acid-schiff with and without diastase digestion and mucicarmine or on the other hand hyaluronic acid with and without diastase digestion has been used in the past with some success. Although either histochemical technique is very good, one should evaluate the one that one is more familiar with, and only one of those procedures is enough in order to evaluate a particular lesion. Needles to say, if one encounters a difficulty with one of those techniques, the use of an alternative one is indicated. It is important to state that although the positive finding of intracellular mucin is a strong feature of adenocarcinoma, this finding has also been reported in some mesothelioma in up to 5% of the cases. On the other hand, it is also important to note that some mesotheliomas will show abundant presence of extracellular mucin; however, not intracellular mucin. Thus, in the current practice the use of those histochemical techniques many times is by-pass for the use of immunohistochemistry. Immunohistochemical studies: a great deal of information regarding immunohistochemical studies in the evaluation of mesotheliomas is available. Numerous studies attempting to positively identified mesotheliomas have been published, some of them with relative usefulness while other have merely have attempted to identified adenocarcinoma in order to properly rule out the presence of mesothelioma. Thus, the diagnosis of mesothelioma has in the past been considered one of exclusion. However, that is not necessary the case, since if one properly evaluate clinical, radiological, histopathological and immunohistochemical features, the degree of certainty can be fairly high. Although there are essentially dozen of immunohistochemical markers that have been stated to be helpful in the diagnosis of mesothelioma, the fact is that only a few are use in practice. Currently, markers that have been use in the diagnosis of mesothelioma or adenocarcinoma in the pleura include: keratin broad spectrum, keratin 5/6, calretinin, Leu-M1, CEA, B72.3, and Ber-ep4. Some of those markers have been stated to positively stain mesothelial cells (keratin 5/6 and calretinin) while other have been stated to identify cases of adenocarcinoma (CEA, Leu-M1). Broad-spectrum keratin obviously stains both tumors while Ber-ep4 has been found in up to 27% positive in cases of mesothelioma. It is exactly in the setting of an epithelial mesothelioma where all these markers become important in the evaluation of the tumor. However, it is important to mention that even though these immunohistochemical markers are very important in the evaluation of mesotheliomas, their positive staining does not constitute a full proof diagnosis of the tumor. For instance, depending on the antibody use, it may alter the interpretation of it. One of those cases is CEA, which has been demonstrated to show 21

25 some positivity in about 5% of cases of mesothelioma. However, it is possible that this phenomenon may be explained by the use of unabsorbed heteroantisera to CEA. The use of monoclonal CEA appears to be more reliable in this evaluation. In short, we can summarize the immunohistochemical studies in the following manner: if the tumor in question shows positive for either of these antibodies - broad-spectrum keratin, keratin 5/6, and calretinin, then the most likely interpretation is that of mesothelioma. However, if there is positive staining for either CEA, Le-M1, B72.3 or other carcinomatous epitope, then the diagnosis should lean more towards adenocarcinoma. Electron Microscopic studies: the use of ultrastructural studies in cases of mesothelioma is very important due to the specific features that mesotheliomas have. However, the use of ultrastructural studies in many cases is hampered by the lack of material when it is needed the most. In most of the times, the material does not become available until there is a more extensive procedure while in the majority of the cases the initial biopsy which is the one in which one needs to establish the diagnosis is the only material available for diagnosis. Thus it limited use. On the other hand, the use of electron microscopic features is helpful in the better-differentiated tumors while those cases in which the tumor is poorly differentiated; the ultrastructural findings are rarely helpful. Usually in most cases in which immunohistochemistry has failed to provide a clear interpretation of a lesion, electron microscopy will also be questionable. Nevertheless, it can be very helpful and if one should make every effort to obtain a sample for such studies. The finding of long, slender microvilli is a classical feature for mesothelioma. Differential Diagnosis In the setting of an atypical epithelial cellular proliferation the most important conditions to rule out are either an adenocarcinoma that has extended into the pleura, a metastatic epithelial tumor of other origin, or more importantly a mesothelial hyperplasia. If one has concluded that the cellular proliferation in question is malignant, then the use of immunohistochemical studies, namely the use of carcinomatous epitopes, as previously mentioned in the section of immunohistochemistry, will be the next step. Similar steps are adequate in the event of a metastatic epithelial neoplasm from other source to the pleura. However, the interpretation can be more difficult in cases of mesothelial hyperplasia. In this setting, there is no immunohistochemical stain, which can separate a neoplastic cellular proliferation from a hyperplastic one. Thus, even though one has 22

26 followed the necessary steps, it is imperative not only to correlate the histology with the clinical and radiological features but also to properly interpret the results of the immunohistochemical studies. As a matter of fact even electron microscopic studies would fail to separate such cellular proliferations. In essence the diagnosis is a morphologic one and one that requires careful attention to specific features (see tables 1 and 2). TABLE 1 Histopathological features of Mesothelioma vr Mesothelial Hyperplasia Feature Mesothelioma Hyperplasia Penetration into adipose tissue or muscle Stromal invasion Inflammation Cellular atypia + + Mitoses + + Cellular proliferation in surface Granulation tissue Fibrin TABLE 2 Immunohistochemical Features of Mesothelioma and Adenocarcinoma Antibody Adenocarcinoma Mesothelioma CEA Leu-M B Ber-ep /+ Calretinin

27 Keratin 5/ Broad-spectrum keratin EMA Sarcomatoid Mesothelioma: this variant of mesothelioma is less common than the epithelial one and probably represents less than 15% of all these tumors in its pure form. As its name implies, the tumor characteristically has a growth pattern of spindle cells with an elongated nuclei and inconspicuous nucleoli in a manner mimicking a sarcoma of soft tissues. Due to the presence of extensive collagenization there have been recognized some histopathological growth patterns that are very important and includes: Spindle cell type (fibrosarcoma-like or MFH-like) Desmoplastic variant Lymphohistiocytoid Of these variants, the desmoplastic variant offers a special difficulty since it can be incorrectly diagnosed as a benign fibrous pleurisy. However, when the tumor exhibits all the features of malignancy, i.e., necrosis, cellular atypia, mitotic activity with atypical mitoses, stromal and adjacent soft tissue invasion, the most important issue would be to rule out either a sarcomatoid carcinoma with extension into the pleura or a sarcoma of soft tissue. The presence of an atypical spindle cell proliferation with stromal invasion and necrosis is usually a good sign of a sarcomatoid mesothelioma. By far the most difficult diagnosis is the Desmoplastic mesothelioma. In these cases, the pleural biopsy may show more collagenization and only a paucicellular proliferation which may not show marked cytologic atypia; however, in these circumstances, similar morphologic characteristics such as stromal invasion, adjacent tissue invasion, i.e., adipose tissue and muscle are important features. Nevertheless in limited (small) pleural biopsies, the diagnosis may not be apparent and one can only make such suggestion if the clinical and radiological findings are in keeping with such diagnosis. Histochemical studies: the use of histochemical studies such as D-PAS and mucicarmine or hyaluronic acid with and without diastase digestion has no role in the diagnosis of these tumors. 24

28 Immunohistochemical studies: in the setting of a spindle cell mesothelioma whether the tumor is desmoplastic or not, the role of immunohistochemistry is relatively limited since most of the antibodies use in regular epithelial mesotheliomas have no practical use in sarcomatoid mesotheliomas. The use of broad- spectrum keratin is by far the most important of them. All other carcinomatous epitopes are known not to react with sarcomatoid tumors. In addition, the use of calretinin and keratin 5/6 is rather limited since its positivity may vary and negative results do not mean that the tumor in question is not a mesothelioma. Furthermore, because of the lack of more specificity from immunohistochemical markers, the issue of fibrous pleurisy and desmoplastic mesothelioma cannot be solved by immunohistochemistry since both lesions may crossreact with keratin antibodies. In essence, the use of immunohistochemistry is relevant to rule-out other spindle cell tumor of different lineage including leiomyosarcomas, MFH or other mesenchymal tumors. Differential Diagnosis In cases in which there is no doubt about the neoplastic nature of the tumor, the most important differential diagnosis is with another spindle cell neoplasm of mesenchymal origin. In this case, the use of proper immunohistochemical studies and/or electron microscopy will lead to amore appropriate interpretation. In cases of sarcomatoid carcinoma involving the pleura in a diffuse manner, the finding by radiographic studies of an intrapulmonary tumor mass will lead to a more appropriate interpretation. By far the most difficult diagnosis is with fibrous pleurisy of a reactive or inflammatory nature. In these cases, the diagnosis is based on morphologic grounds since immunohistochemistry cannot solve the problem (see table 3). TABLE 3 Histopathological features of Sarcomatoid Mesothelioma and Fibrous Pleurisy Feature Mesothelioma Fibrous Pleurisy Transitions between cellular and acellular Cellular atypia Granulation tissue