Updates on the pathologic diagnosis and classification of mesothelioma

INTRODUCTION

Mesothelioma is a neoplasm that most commonly arises in the pleura, but can also be found in peritoneum, pericardium, or tunica vaginalis of the testes^[1,2]. It is a rare tumor with an annual estimated incidence of 3000 cases per year in the United States; it affects men more than women^[1,3,4]. All mesotheliomas are clinically aggressive and carry a poor prognosis, with studies estimating a median survival of 9-29 months and 5-year survival of 5% or less^[5,6], though patients with localized disease may have better outcomes^[7]. The main diagnostic features of mesothelioma are published in the World Health Organization Classification of Thoracic Tumours^[8]. Mesothelioma remains a challenging entity to diagnose, especially when the pathologist is given a small biopsy. Currently, mesothelioma is divided histologically into three major subtypes (epithelioid, sarcomatoid, and biphasic), which can be further categorized into prognostic groups based on nuclear grade and cytologic and stromal features^[9,10]. The gold standard for the diagnosis of mesothelial lesions as malignant is tissue invasion, which is not always possible on small biopsies^[11]. The diagnosis of mesothelioma can be achieved using ancillary testing, specifically BAP1 and MTAP immunohistochemistry or CDKN2A fluorescence in situ hybridization; furthermore, these markers can also facilitate the identification of mesothelioma in situ prior to the invasive phase of the disease.

EPITHELIOID MESOTHELIOMA

The majority of mesotheliomas (approximately 55%) exhibit an epithelioid growth pattern [Figure 1], and these are further subclassified based on secondary architectural patterns, cytology, and stromal features^[10,12]. While data may be limited with only a few published articles in the literature, studies show that some of these features stratify patients into prognostic groups within epithelioid mesothelioma. Architectural patterns that have been described in epithelioid mesothelioma include tubulopapillary, trabecular, adenomatoid, microcystic, solid, and micropapillary. Cytologic features described in epithelioid mesothelioma include rhabdoid, deciduoid, small cell, clear cell, signet ring, lymphohistiocytoid, and pleomorphic. A subset of mesotheliomas may show myxoid stroma. The pathologist is encouraged to report these features to improve diagnostic accuracy and potentially help with risk stratification^[10,12]. Features associated with a more favorable prognosis in epithelioid mesothelioma include tubulopapillary, trabecular, and adenomatoid architectural patterns, as well as lymphohistiocytoid cytologic features and the presence of myxoid stroma (50% or greater of total tumor)^[10-12]. Unfavorable features that portend a worse prognosis in epithelioid mesothelioma include solid and micropapillary architecture, as well as rhabdoid and pleomorphic cytologic features^[8,10-12]. Grading systems that incorporate nuclear features (atypia, mitotic index) into classification have been shown in medium to large cohorts to determine prognosis^[13,14]. While these grading studies have not been shown to uniformly predict disease performance across histologic subtypes, they are a powerful prognostic tool in epithelioid mesothelioma, and recently proposed high- and low-grade grouping of epithelioid mesothelioma is now recommended^[13-15]. The grading of epithelioid mesotheliomas will be discussed in greater detail below.

Figure 1. Epithelioid mesothelioma characterized by a proliferation of tumor cells with round nuclei, ample cytoplasm, and cellular cohesion (H&E 200×).

SARCOMATOID MESOTHELIOMA

Sarcomatoid mesothelioma [Figure 2] is characterized by the haphazard proliferation of malignant spindled mesothelial cells resembling sarcoma^[8]. The less cellular and densely fibrotic desmoplastic variant has long been recognized as a pattern of sarcomatoid mesothelioma^[8]. Sarcomatoid mesotheliomas are rare, comprising only around 10% of all cases, ranging from 7%-22%^[16,17]. More recently, the application of specific cellular features to the diagnostic classification of sarcomatoid mesothelioma has been proposed, similar to epithelioid mesothelioma^[10,12]. These cytologic features include noting whether cells are pleomorphic, lymphohistocytoid, or transitional; lymphohistiocytoid has been associated with a more favorable prognosis (transitional will be discussed below)^[12]. Pleomorphic features have been observed in both epithelioid and sarcomatoid mesotheliomas, and are genomically similar enough to be classified as either, based upon the cytomorphology of the tumor (epithelioid or sarcomatoid)^[10]. Desmoplasia is classified as a stromal feature^[12].

Figure 2. Sarcomatoid mesothelioma characterized by a haphazard proliferation of malignant spindle cells (H&E 200×).

Transitional mesothelioma (transitional features) presents as a proliferation of elongated epithelioid mesothelial cells with sheet-like, yet discohesive growth and, as previously mentioned, is now considered a cytologic feature observed in sarcomatoid mesothelioma^[8]. While the name and cytomorphology suggest that these cells are transitioning or are halfway between epithelioid and sarcomatoid mesothelioma, transitional mesotheliomas are more similar to the sarcomatoid subtype genomically^[18]. Given their genomic relatedness to sarcomatoid mesothelioma, it is not surprising that transitional mesothelioma is also associated with a worse prognosis, in line with what has been reported in sarcomatoid mesothelioma^[18,19].

BIPHASIC MESOTHELIOMA

Mesotheliomas in which both epithelioid and sarcomatoid morphology are present should be classified as biphasic mesothelioma (approximately 20% of all mesotheliomas; Figure 3), and the specific ratio of epithelioid to sarcomatoid components should be noted. While rather arbitrarily assigned, at least 10% of each component (epithelioid and sarcomatoid) is required for diagnosis in the most recent WHO classification^[8]. This 10% cutoff is only applied to resected mesothelioma. According to the most recent classification schemes, a small biopsy with any amount of both epithelioid and sarcomatoid mesothelioma should be classified as biphasic^[10]. It is known that approximately 20% of biopsies showing epithelioid mesothelioma will, in fact, show biphasic morphology in resection specimens^[20,21]. Biphasic and sarcomatoid morphology on a small biopsy are quite predictive of the tumor having a significant proportion of sarcomatoid morphology at the time of resection^[20,21]. The ratio of each constituent, especially with regard to the sarcomatoid component, may be prognostic. Data is limited, but it is suggested that sarcomatoid predominant biphasic mesotheliomas have worse outcomes, with prognostic cutoffs reported from 20%-50% sarcomatoid morphology^[19,22]. Given that there is some data to suggest a prognostic cutoff around 50% for sarcomatoid morphology, it may be helpful for the patient’s treatment team to be aware that 95% of biopsies with pure epithelioid morphology showed at least 50% epithelioid morphology at the time of resection^[20]. While biphasic mesothelioma has an overall survival in between epithelioid and sarcomatoid mesothelioma, the exact prognostic significance of percent epithelioid and sarcomatoid morphology in biphasic mesothelioma is unclear; nonetheless, pathologists are now encouraged to report these findings on resected mesotheliomas^[12].

Figure 3. Biphasic mesothelioma showing both epithelioid and sarcomatoid growth patterns (H&E 200×).

HISTOLOGIC GRADING

Pathologic grading systems have been described and most robustly studied in epithelioid mesothelioma. The original grading system was developed at Memorial Sloan Kettering and consisted of a nuclear grade score, calculated as a sum of nuclear atypia and mitotic count^[14]. Briefly, nuclear atypia is scored as 1 [Figure 4] for mild cytologic atypia with occasional small nucleoli (mesothelial cells are similar to benign resting or benign reactive mesothelial cells); score 2 [Figure 5] shows increased cytologic atypia with moderate pleomorphism and occasional nucleoli; score 3 [Figure 6] shows marked cytologic atypia with enlarged and highly pleomorphic nuclei and prominent nucleoli. Mitotic count score (per 2 mm²) is scored as 1 (0 to 1 mitotic figures), 2 (2 to 4 mitotic figures), or 3 (≥ 5 mitotic figures). The nuclear atypia and mitotic count scores are added together to yield a three-tier nuclear grade; nuclear grade 1 (sum of nuclear atypia and mitotic count = 2-3), nuclear grade 2 (sum = 4-5), and nuclear grade 3 (sum = 6). This grading scheme was shown to be highly predictive of overall survival. Subsequently, studies from the University of Chicago showed that the addition of necrosis further stratified patients into prognostic groups^[13]. This finding was then incorporated into the most recently proposed grading system^[10]. This grading system, not only recommended by expert consensus but included in the WHO and CAP synoptic reporting for mesothelioma, is a combination of the three-tiered system with the addition of necrosis. This yields a two-tiered (low and high grade) classification scheme in which the mesothelioma is designated as a low grade if it is nuclear grade 1 or nuclear grade 2 without necrosis and designated as a high grade if it is nuclear grade 2 with necrosis or nuclear grade 3. While this system has yet to be seen in widespread clinical use, as it is now a part of the routine classification of epithelioid mesothelioma, one would anticipate it becoming more commonly reported. Nuclear grading in epithelioid mesothelioma can be applied to specimens obtained via biopsy or resection^[23,24].

Figure 4. Epithelioid mesothelioma with nuclear atypia score of 1. The nuclei are relatively small and monotonous without prominent nucleoli (H&E 400×).

Figure 5. Epithelioid mesothelioma with nuclear atypia score of 2. The nuclei display moderate pleomorphism with occasionally prominent nucleoli and more vesicular chromatin when compared to the nuclei in Figure 4 (H&E 400×).

Figure 6. Epithelioid mesothelioma with nuclear atypia score of 3. The nuclei display marked pleomorphism with multinucleated cells and giant tumor cells. Atypical mitotic figures are also obvious (H&E 400×).

At present, there is no consensus on the need to grade sarcomatoid or biphasic mesotheliomas as tumors with sarcomatoid morphology behave aggressively. A number of studies have proposed grading non-epithelioid mesotheliomas, but these are not recommended at present to be utilized in clinical practice^[15,19,25].

ANCILLARY STUDIES AND TESTING

Mesotheliomas often need to be distinguished from other pleural-based malignancies. Tumors, both primary and metastatic to the lung, can show significant pleural involvement and mimic mesothelioma^[26,27]. Clinical context and radiographic imaging should be considered when forming a differential diagnosis. The judicious application of immunohistochemical staining can help with the differential diagnosis^[11]. Markers expressed in mesothelial lesions include WT-1, D2-40 (podoplanin), calretinin, and cytokeratin 5/6^[11,28]. A myriad of other makers have been proposed as indicators of mesothelial differentiation (mesothelin, HBME-1, thrombomodulin) and the reader is referred to other reviews on this topic for additional information^[29,30]. Both carcinoma and mesothelioma often express pancytokeratins (AE1/AE3, CAM 5.2, OSCAR); cytokeratin staining, while nonspecific, may delineate growth patterns within the tumor or exclude other types of cancer from the differential diagnosis^[5]. Historically, numerous “pancarcinoma” or “panepithelial” markers have been proposed with staining in carcinomas but not mesotheliomas. Some of these markers include CEA, B72.3, MOC-31, and Ber-EP4, and while these markers are more often positive in carcinomas than mesotheliomas, these markers can be expressed, even strongly, in mesothelial lesions^[28]. More recently, Claudin-4, an antibody raised against tight junctions of moderate to well-differentiated epithelia, has emerged as the best marker to differentiate carcinoma from mesothelioma^[31-33]. Current recommendations to separate mesothelial lesions from carcinoma are to demonstrate reactivity for two mesothelial markers (CK5/6, calretinin, WT-1, and/or D2-40) and show negativity for two carcinoma markers (claudin 4, MOC-31, Ber-EP4, etc.)^[11]. It should be noted that the mesothelial markers are nonspecific and imperfectly sensitive and that the sensitivity of these mesothelial markers typically drops in sarcomatoid mesotheliomas^[28]. Also of note, the carcinoma/epithelial markers utilized will depend on the carcinoma one is trying to exclude. For example, it would be perfectly reasonable to utilize TTF-1 and p40 immunohistochemical stains in the thoracic cavity, given how commonly non-small cell carcinoma is encountered. It would make less sense to utilize these same markers in the peritoneal cavity, where other carcinomas also need to be excluded (renal cell carcinoma, gynecologic malignancies, etc.) by utilizing another set of immunohistochemical markers. In conclusion, for best results, sufficient yet judicious immunohistochemical studies should be employed utilizing a panel of stains that are interpreted in light of the clinical and radiographic settings.

Once a lesion is confirmed to be mesothelial in origin, the pathologist may still find it challenging to call the lesion malignant, especially if there is limited tissue or a small biopsy and no invasion into underlying tissue identified. Over the past few years, newer immunohistochemical markers have emerged that allow for the separation of benign and malignant mesothelial proliferations. BRCA-associated peptide-1 (BAP1) has been shown to be discriminatory when it comes to separating benign from malignant mesothelial proliferations^[34-36]. Nuclear loss of BAP1 staining [Figure 7] is currently reported as 100% specific for malignancy in mesothelial proliferations. It has long been known that a frequent genetic alteration in mesothelioma is the homozygous deletion of CDKN2A^[37,38]. CDKN2A homozygous deletion can be reliably detected by fluorescence in situ hybridization (FISH)^[39-41]. While FISH testing for CDKN2A deletion is useful, FISH testing is not always as readily available as immunohistochemistry in small laboratories. Thankfully, methylthioadenosine phosphorylase (MTAP) is positioned near CDKN2A and is often co-deleted^[42,43]. Cytoplasmic loss of MTAP [Figure 8] has been shown to be 100% specific for malignancy in mesothelial proliferations. While CDKN2A encodes the protein p16, p16 IHC has historically shown mixed results when judging its utility at predicting homozygous deletion of CDKN2A in mesothelioma, but a recent study shows some promise when evaluating p16 alongside MTAP IHC^[44]. While BAP1 and MTAP are 100% specific for malignancy, unfortunately, BAP1 and MTAP IHC are not 100% sensitive in detecting malignant mesothelial proliferations, but combining BAP1 and MTAP together does increase sensitivity^[45-47]. As data continues to emerge, it appears that a limited panel approach (BAP1, MTAP, NF2, p53, among possible other markers) may be most useful in separating benign from malignant mesothelial proliferations^[36,48]. A possible decisional algorithm that one may follow based on the most commonly used and accepted ancillary studies is provided [Figure 9]. While experts have not yet concluded what is the best panel to utilize, it appears evident that combining multiple stains may eventually reach adequate sensitivity for detection of mesothelioma; however, as new stains are proposed to help in this differential, the reader should be warned that close attention should be paid to the specificity of these newly proposed markers. The last word of caution on this topic pertains to the fact that one must confirm that the lesion is mesothelial before interpreting these markers. For example, BAP1 nuclear loss can also be identified in other malignancies (most notably, some melanomas and some renal cell carcinomas). Also, a recently published study highlighted the lack of specificity of MTAP loss, which can be observed in a number of intrathoracic malignancies^[49].

Figure 7. Epithelioid mesothelioma showing nuclear loss of BAP1. Note small lymphocytes with retained expression (400×).

Figure 8. Epithelioid mesothelioma showing cytoplasmic loss of MTAP. Note inflammatory cells with retained expression (400×).

Figure 9. Decisional algorithm for the work-up of serosal lesions.

The emergence of these markers of malignancy in mesothelial proliferations has implications for the cytologic diagnosis of mesothelioma. Studies do exist showing that loss of BAP1 and MTAP staining can be observed in cytologic fluids from patients with mesothelioma. How this impacts the cytologic examination of fluids is beyond the scope of this review, but this area of cytology is expected to change in the coming years^[50,51].

PERITONEAL MESOTHELIOMA

Peritoneal mesothelioma is the second most common type of mesothelioma, accounting for 10%-15% of all mesotheliomas^[52]. Interestingly, it has been noted that today, peritoneal mesotheliomas appear more frequently in women and younger people than their pleural mesothelioma counterparts^[53-56]. Peritoneal mesothelioma is classified into the same histologic subtypes as pleural mesothelioma (epithelioid, biphasic, and sarcomatoid). Recent data show that peritoneal mesotheliomas appear to have a better prognosis than pleural mesotheliomas^[57]. In some cases, this may be due to different treatment options available based on native anatomical location (hyperthermic intraperitoneal chemotherapy)^[58]. Biphasic and sarcomatoid mesotheliomas are relatively rare compared to their thoracic counterpart but, when identified, show a similar poor prognosis^[59].

As mentioned above, there can be challenges differentiating mesothelioma from carcinoma. This problem is just as evident in the peritoneal cavity as in the thoracic cavity. Therefore, most ancillary studies performed to determine diagnosis concentrate on separating peritoneal mesothelioma from other diffuse peritoneal malignancies, mainly papillary serous carcinoma, using immunohistochemical stains adapted from those used to diagnose mesothelioma in the pleura^[11]. The adaptation for these panels is made to delineate the various potential origins of peritoneal neoplasms. Pleural malignancies generally have a pulmonary origin, but peritoneal malignancies can originate from the ovaries, fallopian tubes, stomach, pancreas, colon, and kidney, and includes metastases from outside the peritoneal cavity (breast, lung carcinomas, etc.).

Peritoneal mesothelioma has somewhat been an orphan disease, with most publications pertaining to mesothelioma focusing on pleural mesothelioma. Excitingly, groups have recently begun to seriously look at the diagnosis and classification of peritoneal mesothelioma^[57,60]. Our group recently published an article highlighting that many of the same pathologic parameters observed in the thoracic cavity can be applied to peritoneal mesothelioma with similar results^[57]. This is an emerging area of exploration, and new and exciting research on this disease will certainly be published in the coming years.

MESOTHELIOMA IN SITU

Mesothelioma in situ (MIS) is the growth of malignant mesothelial cells, either as a monolayer or small papillary growth, without invasion of the underlying tissue and with corresponding negative radiology findings^[8]. MIS, as a concept, is quite old, but could not be proven definitively as a malignant precursor, leading to the controversy surrounding its definition and diagnosis^[61]. In recent years, especially in patients with nonresolving unexplained pleural effusions, loss of BAP1 and/or MTAP by immunohistochemistry, or identification of homozygous deletion of CDKN2A by FISH in noninvasive mesothelial cells, has proven that one can detect signatures of malignancy in mesothelial cells prior to the development of invasive diffuse mesothelioma^[62-66]. MIS is now accepted as an entity in the WHO^[8]. The ability to diagnose MIS is very recent; thus, this diagnosis will most likely have implications in pleural fluid cytology specmens^[67]. The biggest question now is what the oncologist or surgeon should do if this diagnosis is made^[68]. This is a question that remains to be answered. As cases continue to be reported, mesothelioma pathologists will continue to study this disease.

CONCLUSION

Recent advancements in the field of mesothelioma have led to robust prognostic grading schemes that are now included in the routine reporting of mesothelioma specimens. Advancements in ancillary testing have expanded the pathologist’s ability to diagnose mesothelioma on small biopsies and in cytology fluids. The emergence of mesothelioma in situ as a distinct entity has the potential to change how the medical team views mesothelioma and may provide a means to intervene earlier in the disease.