Advances in Anatomic Pathology:
Translocation-associated Salivary Gland Tumors: A Review and Update
Weinreb, Ilan MD, FRC(P)C*,†
*Department of Pathology, University Health Network
†Department of Pathobiology and Laboratory medicine, University of Toronto, ON, Canada
The author has no funding or conflicts of interest to disclose.
Reprints: Ilan Weinreb, MD, FRC(P)C, Department of Pathology, University Health Network, 200 Elizabeth Street, Room 11E403, Toronto, ON, Canada M5G 2C4 (e-mail: email@example.com). All figures can be viewed online in color at http://www.anatomicpathology.com.
In recent years the discovery of translocations and the fusion oncogenes that they result in has changed the way diagnoses are made in the salivary gland. These genetic aberrations are recurrent and reproducible and at the very least serve as powerful diagnostic tools in salivary gland diagnosis and salivary gland classification. They also show promise as prognostic markers and hopefully as targets of therapy. Many of these fusions have been found in other tumor types that show little to no overlap with their salivary gland counterparts, but effectively they are specific within the salivary gland. In this review the 5 tumors currently known to harbor translocations will be discussed, namely pleomorphic adenoma, mucoepidermoid carcinoma, adenoid cystic carcinoma, mammary analog secretory carcinoma, and hyalinizing clear cell carcinoma. The discovery and implications of each fusion will be highlighted and how they have helped reshape the current classification of salivary gland tumors.
Translocations are believed to account for about 20% of all cancers. Until recent years, translocations and their resultant fusion oncogenes have been thought to be rare in epithelial tumors. Typically, these somatic genetic aberrations were found in hematolymphoid and soft-tissue neoplasms and thought to be relatively rare in other tumor lineages. However, much of the lag time in finding epithelial tumors with translocations probably relates to the difficulty in growing carcinomas in culture and therefore in getting karyotypes on these tumors. Although PLAG1 rearrangements in pleomorphic adenomas (PAs) have been known for some time, it has only been recently that multiple salivary carcinomas have been found to have translocations as well. These include mucoepidermoid carcinoma (MEC), adenoid cystic carcinoma (ACC), the new entity mammary analog secretory carcinoma (MASC), and hyalinizing clear cell carcinoma (HCCC). This is likely an incomplete list that will grow over the next few years, especially with the emergence of next-generation sequencing technologies that have demonstrated the ability to predict gene fusions by transcriptome or whole-genome sequencing.1 Salivary carcinomas are good candidates for finding translocations, as most of them are low grade and show a relative homogeneity from case to case. In other words, every tumor of a specific subtype essentially looks alike. They also appear unlike the normal salivary gland elements. This is in contrast to tumors that contain complex genomic abnormalities, many of which arise in a background of preneoplastic dysplasia and essentially look like the malignant counterpart of the normal tissues that is, squamous cell carcinoma looks like squamous epithelium and arises in a stepwise manner from dysplasia. Using this general schema, one might expect tumors such as epithelial-myoepithelial carcinoma (EMC) and basal cell adenocarcinoma to also have translocations. This review will focus on the salivary gland tumors that are known to harbor recurrent translocations at this point and will focus on their pathologic features and how the translocations may impact us as pathologists diagnostically or clinically. It will attempt to address when they can or should be tested for rather than focus on the specifics of their molecular biology or how they lead to cancer formation, much of which is not known at this time. The review is not intended to provide an exhaustive description of each entity’s morphologic features, most of which are well known to pathologists.
Pleomorphic Adenoma (Benign Mixed Tumor)
PA, also known as benign mixed tumor, is a benign salivary gland neoplasm composed of combinations of bilayered ducts, myoepithelial cells and matrix (Fig. 1A). The latter is composed of myxoid matrix, chondroid, and other elements and tends to be directly associated with the myoepithelial cells or modified myoepithelial cells. The ducts tend to be bland and reminiscent of normal intercalated ducts and may show squamous, sebaceous, or other metaplasias. The myoepithelial cells are highly variable with epithelioid, spindled, plasmacytoid, and clear cell forms. PAs are known to harbor translocations involving PLAG1 or HMGA2 in a sizable percentage of cases.2 This tumor is the most common salivary gland neoplasm at all sites and can occur anywhere in the upper and lower respiratory tract, as well as in the skin and soft tissues, where it is known as chondroid syringoma and soft-tissue myoepithelial tumor (SMET), respectively. In the salivary glands they are well known for significant morbidity2 and are therefore enigmatic among benign tumors. This includes very frequent recurrence, particularly if not resected with a wide margin (enucleation surgery). When recurrent they may grow in small nodules, seed nerves, and/or invade the skin. Sometimes they regrow in such a morbid manner that they can no longer be resected or may even require radiation. In addition, a very small subset has been known to metastasize while still appearing benign. This is the so-called “metastasizing pleomorphic adenoma,” which while it appears benign can cause death due to disease.2 Finally, about 6% of PAs can transform to a frank carcinoma of any salivary type (Fig. 1B).2 These are often aggressive malignancies that have a high mortality rate when widely invasive. Transformation is not part of the natural history of the skin and soft tissue counterparts for the most part. They (PA) perhaps would be better thought of as a tumor of low malignant potential or as a precursor to malignancies. Most clinicians in this author’s experience consider this to be a tumor that should always be resected if possible.
Numerous cytogenetic studies over the years have shown abnormalities in the 8q12 region in PAs.3 Some have shown a balanced reciprocal translocation t(3;8)(p21;q12) involving PLAG1 and others have shown a wide array of abnormalities with numerous known and possibly unknown fusion partners (Fig. 1C).3,4 The most common fusion partner is CTNNB1, which is the gene encoding β-catenin and promoter swapping between these genes leads to upregulation and expression of PLAG1 protein and downregulation of its partner. This PLAG1 upregulation can be measured by immunohistochemistry but does not appear to be specific to those cases with a translocation involving PLAG1 and ultimately has limited diagnostic utility.3,5 The PLAG1 protein expression has been shown by Northern blot analysis in tumors harboring typical fusions, those with rearrangements of 12q13-15, and even those with an apparently normal karyotype.6 This highlights the importance of this gene in the development of PAs in general.6 Occasional cases with rearrangements in this general region do not involve PLAG1 at all but rather downstream elements to this gene.7 A second important fusion partner, namely LIFR has also been identified,3 but seems to be unrelated to CTNNB1 despite causing similar upregulation. Other less common partners, such as SII have also been cloned.6 The existence of ring chromosomes in some PAs have been described which fuse FGFR1 to PLAG1 and yet other cases with gain in whole PLAG1 copy number have been highlighted which lead to similar appearing tumors.8 In addition to PLAG1, the HMGA2 gene has been shown to be involved in a significant subset of PAs with identical morphology.7,9 The gene HMGA2 is often amplified with numerous copies being found in double minute chromosomes and homogeneously staining regions (Fig. 1D). This is often accompanied by MDM2 amplification with coexistent rearrangement of HMGA2 with fusion of WIFI or other gene partners.9,10 This has been found in both PAs and carcinoma ex PA.9,11 Interestingly both PLAG1 and HMGA2 have been described in other benign mesenchymal tumors, although none of these have the same recurrence or transformation potential as PA. This suggests that the combination of the function of these aberrant genes and the site or cell of origin, are critical in the impact of these genetic changes. More importantly, PAs with these various aberrations and those with no identifiable aberrations, essentially have the same appearance and morphologic spectrum. This suggests that performing fluorescence in situ hybridization (FISH) for these genes has limited diagnostic utility. There is no prognostic significance for these mutations in terms of PA behavior. FISH for these genes has proved useful in answering research questions however.
A study by Martins et al12 looked at PLAG1 FISH in PAs with known 8q12 abnormalities and found it to be rearranged in the majority of cases. Moreover, they were able to show that the gene was rearranged in both the ductal and myoepithelial components, which proved that both cell types were neoplastic and that they arose in a single progenitor cell capable of differentiating down both cell lines.12 Carcinomas ex PA have been shown to retain the same genetic abnormality as their benign PA precursor5 and presumably a second or multiple additional aberrations are what lead to the carcinoma. PLAG1 FISH and other techniques have also helped prove that mixed tumors of skin and soft tissue have a similar origin to their salivary counterparts, with frequent rearrangement of the gene.13,14 One soft tissue case showed a LIFR-PLAG1 fusion, similar to some salivary PAs.13 However, most cases did not and none showed the more common CTNNB1 fusion partner. There have been no soft tissue cases with HMGA2 abnormalities described thus far.13 In addition to this link, studies of carcinoma ex PA have been performed showing similar genetic abnormalities. This can serve to link the carcinoma to the PA element as both populations harbor the PLAG1 or HMGA2 rearrangement.5 However, this tells us nothing about the cause of the transformation in most cases. Antonescu et al13 described a subset of myoepithelial carcinoma ex PA with almost all cases having either PLAG1 rearrangement and amplification, or HMGA2 rearrangement and amplification. Those with PLAG1 abnormalities were consistently associated with FGFR1 gene amplification.13 More interestingly, the de novo myoepithelial carcinomas did not show any of these changes, suggesting no occult PA and more importantly that the mechanism of these tumors may be different from cases arising in a PA.13 Bahrami et al5 examined a cohort of 22 carcinomas ex PA and showed that a number of different carcinoma types can arise in PLAG1-positive and HMGA2-positive PAs, including myoepithelial carcinomas and adenocarcinomas. We recently also showed that salivary duct carcinoma (SDC) has frequent rearrangement or amplification of the PLAG1 and/or HMGA2 genes.15 Importantly, this was also found in de novo SDC, suggesting that unlike myoepithelial carcinomas, some de novo SDC actually arose in a PA that was obliterated by tumor growth.15 More research will undoubtedly find other “de novo” carcinoma types, likely high grade, which originally arose in an unapparent PA. It is possible that those carcinoma cases that arose in PAs have the same basic genetic changes as de novo cases albeit with the additional PA genetic aberrations. For instance, in our study of SDC, we found carcinomas harboring both PLAG1 rearrangement and Her-2 gene amplification, the latter presumably an oncogenic driver of the carcinoma component,15 which has been described in about 20% of de novo SDC as well.2
MEC is the most common salivary gland carcinoma and occurs approximately equally in major and minor salivary gland sites.2 Like PAs they can arise in other locations including the tracheobronchial tree, cervix, skin, and upper respiratory tract. MEC is believed to be a tumor of large duct (striated or excretory) origin. They are composed of 3 cell types: the mucinous cells, often large and goblet like, which often line cystic spaces, the epidermoid cells which are nonkeratinizing and may even look frankly squamous, and finally the intermediate cells which are more basal or cuboidal (Fig. 2A). The latter are believed to show features of the other 2 cell types and can transform into either, although this author believes they are simply a smaller variant of the epidermoid cell. The idea that all 3 cell types must be found to make this diagnosis is false and the identification of intermediate cells is largely subjective. Atypia is unusual and the epidermoid/squamoid cells tend to be very bland akin to normal mucosal epithelium. Even the lowest grade squamous cell carcinoma has more atypia than a conventional MEC and atypia in a MEC should alert the pathologist to at least think of alternative diagnoses.
The differential diagnosis of this tumor includes anything that would mimic the mucinous component, the epidermoid component or would have cystic changes. In the low-grade end of the spectrum this includes duct ectasia and cystadenomas. Intermediate-grade MEC may mimic MASC and HCCC (see later). In the high-grade end of the spectrum it may mimic squamous cell carcinoma, adenosquamous carcinoma, and SDC. The grading may be based on a number of different systems, which will not be discussed in detail. They include the armed forces institute of pathology Brandwein, and modified Healey grading systems. All essentially take into account cystic content, anaplasia, mitotic activity, necrosis, perineural invasion, and variably may include such criteria as invasive tumor fronts, lymphovascular invasion, and other findings (Fig. 2B). Until the discovery of the CRTC1-MAML2 fusion, there were no useful prognostic or diagnostic markers for MEC and in particular there was difficulty in triaging which of the intermediate-grade tumors would behave negatively. Low-grade and high-grade tumors were generally easy to stratify on the other hand and predictably did well and poorly, respectively.
Tonon et al16 first described the fusion gene after many years of cytogenetic studies showing a recurrent t(11;19)(q14-21;p12-13) in these tumors (Fig. 2C). Using BAC clones, they used a FISH mapping approach in cell lines of MEC to narrow down the rearranged sequences on chromosome 11 and found the novel MAML2 gene. 5′RACE was used to discover the novel CRTC1 gene and an in-frame fusion between exon 1 of CRTC1 and exon 2 of MAML2 was found.16 This fusion was predicted to disrupt the NOTCH signaling pathway. They also found the fusion in a pulmonary example of MEC.16
Excitement with this translocation began when Behboudi et al17 showed a prognostic role for this fusion. A total of 55% of a retrospective cohort contained the abnormality and the fusion-positive cases tended to have a better outcome.17 Fusion-positive cases were smaller, were found in younger patients, and had less recurrence, metastases, and tumor-related mortality than fusion-negative cases. They also tended to be low grade and sometimes intermediate grade, but not high grade.17 This suggested a role for prognostication especially for the intermediate group. Similar findings by Martins et al18 were seen, although the rate of positivity in this cohorts was 70%. A second fusion, CRTC3-MAML2 was discovered by Fehr et al19 and later found in about 6% of a large cohort by Nakayama et al20 (Fig. 2D). The latter study found that this rarer fusion was similarly prognostic and tended to occur in a younger age group than the traditional CRTC1-MAML2 fusion.20 The MAML2 rearrangement has been found in all grades in more recent studies,21,22 has been found in central jawbone, cervix, and thyroid sites,22–24 and in at least 1 variant, namely oncocytic MEC.21,25 We have also seen positive cases with prominent lymphoid stroma and others with partial clear cell differentiation (unpublished observations). The fusion has not been found in adenosquamous carcinomas, squamous cell carcinomas, SDC, or any other salivary gland carcinoma.21 It is therefore specific to MEC among malignant tumors of the head and neck.
The presence of aggressive behavior or higher grade in some translocation-positive MECs in recent studies has raised questions as to the true utility of MAML2 FISH as a prognostic test.21 The use of variable grading systems may account for the discrepancy with older studies which found no high-grade cases. In addition, it is likely that at least some high-grade MECs included in most cohorts were actually something other than MEC.21 Seethala et al21 point out in their recent retrospective study that 46% of high-grade MECs were translocation positive and that those tumors that were positive had minimal anaplasia and were high-grade due to other factors, most of which were architectural criteria. In contrast, their high-grade translocation-negative MEC group tended to have significant anaplasia and may actually have represented other tumor types according to the authors (eg, adenosquamous carcinoma or SDC, which have frequent mucinous differentiation). This is also highlighted in a study by Chenevert et al,26 which examined a large group of putative MECs at the University of Pittsburgh Medical Center and found that about half were not really MEC but rather adenosquamous carcinomas, squamous cell carcinomas, and SDC. These reclassified cases accounted for the majority of the lymph node metastases and mortality in this cohort.26 If the literature on MAML2 rearrangement is muddied by non-MEC cases it would greatly change the prognostic value of this biomarker. It would also suggest a higher sensitivity for MAML2 FISH and perhaps a diagnostic role rather than a prognostic role. In Seethala et al’s21 cohort, the fusion was associated with better outcome but less convincingly than other previous studies. For now, MAML2 FISH is considered potentially prognostic, although its clinical utility has yet to be proven sufficiently. Incidentally, it is not a biomarker that clinician’s request at this author’s institution.
There has also been great controversy as to whether Warthin tumors (WT) have this identical fusion.27 Initially it was found in a small subset of cases, however, subsequent studies could not reproduce this finding outside of an occasional “metaplastic Warthin tumor.”28 These were considered suspicious for MEC by the authors.28 Other authors have described MEC arising in WT with both elements being positive, suggesting that the t(11;19) can transform the WT into a MEC.27 However, another interpretation could be that these represent the “Warthin-like” MECs with prominent lymphoid stroma that have rarely been described25 and that there was never a benign component in these cases. Martins et al,18 Seethala et al,21 and Skalova et al,29 have all found no WTs with the MAML2 fusion in a large number of cases, including metaplastic examples.18,21,29 The current evidence (and the opinion of this author) is that WT does not carry the t(11;19) or have MAML2 rearrangement. The final word on this controversy, however, still awaits and until resolved, the fusion cannot be considered specific to MEC among salivary tumors. MAML2 FISH can, however, aid diagnostically in some challenging cases.
Adenoid Cystic Carcinoma
ACC is the second most common salivary gland carcinoma after MEC.2 It is believed to be of intercalated duct origin, hence the participation by both ductal and myoepithelial cells.2 In fact, it is predominantly a myoepithelial tumor with variable duct formation. The classic ACC shows a combination of tubular and cribriform elements (Fig. 3A). The tubules are bilayered with a visible outer myoepithelial layer that can show clear cell differentiation, overlapping with EMC. The cribriform areas are predominantly myoepithelial with a basaloid appearance and have cells with angulated hyperchromatic nuclei. The cribriform pattern is formed by mucoid and/or hyalinized matrix formed by the myoepithelial cells. Both components are generally monotonous, with bland nuclei and minimal mitotic activity. Solid areas that are largely myoepithelial form less matrix material, but generally are still relatively monotonous with minimal mitotic activity (although they can variably show increasing mitoses and atypia). The reduction in matrix could be because the cells lose the ability to make the mucoid or hyalinized substance or alternatively that the growth rate is too rapid to form matrix and impart the cribriform architecture. This untested hypothesis would also explain why solid architecture increases the grade to II or III, depending on the percent involvement in the tumor.
ACC is important because it is the only low-grade salivary carcinoma that causes significant mortality. Although the tumor grows slowly and only metastasizes late (often >10 y after presentation), ACC is considered relentless and causes significant local morbidity with perineural extension and intracranial involvement and causes death when metastasizing to lungs, liver, and other sites.2 Recurrence is the rule rather than the exception and the first resection is the only real chance of cure if margins are negative. Achieving a negative margin is rare as the tumor is difficult to delineate, because of frequent wide invasion and lack of a desmoplastic response at the advancing tumor front (Fig. 3B). Despite this general behavior, ACC only very rarely involves cervical lymph nodes and therefore there is little role for elective neck dissection. This makes a correct preoperative diagnosis very useful in planning the extent of surgery.
In recent years a t(6;9)(q22-23;p23-24) has been seen in a number of karyotypes of ACC. In 2009 Persson et al30 cloned the fusion showing involvement of the MYB oncogene and NFIB transcription factor gene and this was seen in 100% of the cases tested, which involved salivary glands, other head and neck sites, and breast. The fusion was shown to upregulate MYB protein expression as well, which was believed to be the oncogenic driver of this tumor (Figs. 3C, D). MYB expression was shown in MYB-NFIB containing tumors and multiple variant fusion breakpoints were shown.30 Mitani et al31 confirmed this fusion in ACC, including finding multiple variant fusions, and that it was specific to ACC among salivary gland tumors. A single case of polymorphous low-grade adenocarcinoma has shown the fusion, however, they are largely negative and this could alternatively represent a misdiagnosed ACC.32 In addition, the fusion has subsequently been found in cylindromas of skin, a tumor with significant morphologic overlap to ACC.33 The significance of this finding is unknown at this time.
Mitani et al31 only found the fusion in about one third of cases in their series, unlike the 100% finding in the original study by Persson et al.30 This larger study by Mitani et al31 did, however, show that fusion-negative cases also overexpressed MYB, presumably through different mechanisms. Subsequent studies of promotor hypermethylation have been negative suggesting alternative mechanisms of MYB activation are required in fusion-negative cases.34 MYB overexpression was also shown by RT-PCR with greater expression in fusion-positive cases and was confirmed with immunohistochemistry for MYB protein.31 The fusion was only associated with age over 50 years and was not prognostic or associated with any other clinical parameter31; however, in a subsequent study by the same group the expression of MYB RNA was associated with worse prognosis.35 Interestingly not all fusion-positive cases show fusion transcript formation.35 West et al36 found the MYB-NFIB fusion in about half of ACC cases and MYB rearrangement without NFIB in 16%, for a total of 65% of cases showing abnormal MYB patterns. This suggests alternative fusion partners for MYB. A single NFIB-AIG1 fusion has been found as well, also suggesting that MYB itself is not always required for ACC development.35 A trend towards worse outcome in fusion-positive cases was also found by West et al.36 Nuclear MYB protein expression was found in two thirds of cases by immunohistochemistry in this study and interestingly was accentuated in the basal component of tumor nests.36 The MYB protein expression was negative or only focal in non-ACC salivary cases, making strong MYB expression specific to ACC in the salivary gland, although not very sensitive overall. Brill et al37 showed a very high rate of fusion positivity in frozen tissue samples (86%) but a much lower rate with paraffin-embedded samples (44%) implicating testing methods with fusion positivity rates in ACC. They also found that 89% of ACCs, with or without the fusion, showed MYB RNA overexpression, similar to the finding by Mitani et al,31 confirming the critical role for MYB in all ACCs, irrespective of fusion status.37 Persson et al38 have shown 86% of frozen ACC samples positive for the fusion and 97% positive for MYB RNA expression. All anatomic sites have shown fusion positivity in ACC.30,37
At this point, it is not clear whether the translocation in ACC will be useful as a diagnostic marker, a prognostic marker (poor prognosis) or as a target of therapy. As most ACCs are easily diagnosed it is unlikely to be useful diagnostically and it appears that the MYB protein is not sensitive enough for this purpose. The potential role as a target of therapy is not only interesting but also may finally offer some hope for this aggressive tumor. There is no drug that targets this fusion and there may be some challenges to targeted therapy associated with the slow growth of the tumor even when a drug becomes available.
Mammary Analog Secretory Carcinoma
MASC was first described by Skalova et al39 in 2010 as a rare salivary gland carcinoma mimicking acinic cell carcinoma (AcCC). This tumor shared the granular cytoplasm, microvacuoles, and microcystic growth of AcCC and occurred mainly in the parotid gland (Fig. 4A). The hallmark features include solid and cystic growth, hyalinization, and papillary structures.39 The cyst wall may contain a single cell lining in some areas,40 and the cells in these foci often show hobnailing (Fig. 4B). The lumina of the cysts may show muciphages, hemosiderin, and granular eosinophilic debris. Because of cyst rupture, portions of the tumor often have fibrosis, an inflammatory reaction and cholesterol clefts. Many of these features may suggest a benign diagnosis. In other cases there can be virtually pure solid growth, obvious infiltrative architecture, and perineural invasion, and in these cases malignancy will not be in doubt.
MASC was initially recognized as different from AcCC on the basis of 3 factors. First, it showed no basophilia in the cytoplasm, which is the hallmark of AcCC due to zymogen granules. Second, MASC had a completely different immunohistochemical profile with frequent expression of S100, mammaglobin, vimentin, and MUC4, all of which were rare in AcCC (Fig. 4C). Finally, unlike AcCC, it was found to harbor an ETV6-NTRK3 fusion gene due to a t(12;15)(p13;q25) in most cases tested (Fig. 4D). The near 100% rate of fusion positivity in MASC has been confirmed by other studies40,41 and is unlike the variable rates of the translocations discussed in the preceding 3 tumors. More importantly the morphology, immunohistochemical, and molecular profiles were all identical to secretory carcinoma of the breast. Secretory carcinoma of the breast had long been recognized as an indolent tumor arising mainly in juvenile and adolescent girls.42 It showed essentially 100% survival. Tognon et al43 first identified the ETV6-NTRK3 fusion in virtually all cases of breast secretory carcinoma and in only rare cases of usual ductal breast carcinoma. This translocation fuses the ETV6 transcriptional regulator to the NTRK3 membrane receptor kinase.43 The fusion is identical to that found in cellular mesoblastic nephroma and congenital fibrosarcoma, and more importantly is also identical to that found in MASC,39 hence, “mammary analog secretory carcinoma.”
This fusion gene clearly must have different effects in different organs. Secretory carcinoma shares essentially nothing in common with congenital fibrosarcoma, but there are also significant differences between the breast and salivary tumors. For instance, MASC is seen in all age groups with a predilection for young adults and is more common in males than in females, both of which are different than the breast counterpart.39–41 The behavior of MASC is of a low-grade carcinoma with about 15% to 20% recurrence, similar rates of metastases to lymph nodes and occasional mortality.39,41 This is also unlike the indolent behavior of breast secretory carcinoma. In terms of salivary gland mimics there is less of an impact to this diagnosis. AcCC and MASC essentially have similar clinical outcomes, with only a slight tendency towards more aggressive behavior in MASC.41 MASC can therefore be considered more of an academic interest relative to its most common histologic mimic. However, ETV6 FISH is still useful to make the diagnosis in difficult cases and to exclude other benign or more indolent carcinomas. For instance, some MASC are unicystic with a focal papillary lining and can mimic a benign cystadenoma. Some malignant salivary gland tumors, such as low-grade cribriform cystadenocarcinomas, may mimic MASC as well, but essentially have no metastatic potential when they are noninvasive or minimally invasive.44 Some cystadenocarcinomas are considerably more aggressive than MASC and so the diagnosis is still important to make correctly.
In addition to this, the ETV6-NTRK3 fusion and ETV6 FISH, in particular, have helped clarify some important histologic features that impacts on the classification of salivary gland tumors. This includes the realization that MASC can show focal or even diffuse intracellular mucin and can be positive for high–molecular-weight keratins, including 34βE12.40,41 This raises the differential diagnosis of MEC in many cases. The key distinction is that MEC should not be S100 positive and is almost always p63 positive, whereas MASC shows the reverse findings.40 The molecular diagnostics have also shown that MASC is a lot more common in minor salivary gland sites than previously recognized by Skalova et al39 in their original description.40 Bishop et al45 recently published their experience with “intraoral acinic cell carcinomas” in their archives and showed that the vast majority were actually MASC with ETV6 FISH positivity. This highlights the fact that the current WHO blue book description of almost 20% of AcCCs arising in the oral cavity is probably not accurate, and that these mostly represent MASC cases.2,40,45 In addition, some variants of AcCC, namely “papillary-cystic” and “thyroid follicular-like,” probably largely represent MASC.40ETV6 FISH has also shown that some tumors may have intraductal growth and therefore have origin in striated ductal cells, unlike the presumed acinic and intercalated duct origin of AcCC.40 This focal intraductal growth may also raise the differential diagnosis of low-grade cribriform cystadenocarcinoma with an invasive component (as mentioned earlier), and the immunohistochemical profile is identical in these 2 tumors.40 In addition, an occasional S100-negative case has been described and this warrants ETV6 FISH confirmation before making a diagnosis.40 In most cases of MASC, the combined histologic and immunohistochemical features are very typical, however, and in this author’s opinion, molecular dianostics are not necessary in these “classical” cases.
Currently, there is no available functional work on what the fusion gene does in these tumors and there is no prognostic or therapeutic role. However, the availability of commercial FISH probes for the ETV6 gene makes it an invaluable diagnostic tool and this is helping to reshape the classification of low-grade salivary gland tumors with “pink” cytoplasm.
Hyalinizing Clear Cell Carcinoma
Similar to the role of ETV6 FISH in “pink” tumors of salivary glands, EWSR1 FISH is helping reshape the classification of “clear cell” tumors. HCCC was first described by Milchgrub et al in 1994.46 The classification of clear cell salivary gland tumors until that point consisted of clear cell variants of usual salivary gland tumors such as myoepithelial carcinoma, AcCC, and MEC, as well as biphasic or monophasic (clear cell predominant) EMC. Milchgrub et al46 recognized that pure clear cell carcinomas existed and that they were unrelated to myoepithelial tumors such as myoepithelial carcinoma and EMC. She recognized this tumor’s lack of myoepithelial differentiation based on invariably negative S100 and actin stains in HCCC46 and was convinced of the distinctiveness of this tumor type even on pure morphologic grounds (Sara Milchgrub, 2013, oral personal communication). This is in contrast to the WHO’s recognition of this entity as a clear cell carcinoma, “not otherwise specified.” The original description and subsequent reports have shown strong evidence of squamous differentiation with diffuse p63 and 34βE12 positivity in almost all cases, desmosomes, glycogen, and tonofilaments on ultrastructural examination, and occasional overt squamous pearl formation.46–49 Mucinous differentiation may be focally seen or occasionally even diffusely and this may prompt a diagnosis of clear cell MEC. The distinction can be made by the negativity for MAML2 rearrangement by FISH or by EWSR1 FISH positivity (see below). HCCC may or may not show predominance of clear cells or have significant hyalinization; however, the majority have a lot of both of these features, hence the preferred name “hyalinization clear cell carcinoma” (Figs. 5A, B).
Although often indolent, HCCC shows occasional recurrence and/or metastases to lymph nodes in the neck.50,51 They have a tendency to invade bone and show prominent perineural invasion in the order of 50% (Fig. 5C).50 This can occasionally lead to pain or muscle atrophy.49 Very rare distant metastases have been reported and occasional death because of the disease has been described as well.49,52,53 These rates of morbidity and mortality are probably comparable with most low-grade salivary gland carcinomas. This may make a correct diagnosis mostly of academic interest similar to MASC; however, there are also benign entities on the differential diagnosis including clear cell oncocytomas and myoepitheliomas, and this distinction is obviously more important. In most instances the diagnosis is readily apparent, however, cases with only a small biopsy or which have only focal clear cell differentiation often require additional ancillary testing. This has made EWSR1 FISH of great value, particularly because it is widely available in many laboratories for sarcoma testing.
Our group first described the EWSR1-ATF1 fusion oncogene in HCCC in 2011 (Fig. 5D).50 This was recognized because of the morphologic overlap with some SMET, known to harbor EWSR1 rearrangements.54 The typical SMET partner genes, namely PBX1, ZNF444, and POU5F1 were all negative, however, and 3′RACE was used to identify the ATF1 gene as the partner in HCCC.50 This molecular signature was found in 82% of HCCC cases and in none of its salivary gland clear cell mimics, including EMC, MEC, and myoepithelial carcinoma. We have performed EWSR1 FISH on a number of cases of HCCC since the description of the fusion, and found it to be present in nearly 100% of all cases and in none of its mimics. In fact, 2 of the negative cases in our original description have subsequently been reclassified as “clear cell variant of calcifying epithelial odontogenic tumor” and “clear cell squamous cell carcinoma,” making the initial 82% positivity an underestimate. The one exception to the specificity of the fusion is with the so-called “clear cell odontogenic carcinoma (CCOC),” which was initially considered a mimic of HCCC but has been found to have the identical EWSR1 rearrangement and EWSR1-ATF1 fusion as HCCC.55 This makes CCOC the odontogenic equivalent of HCCC as they also share morphologic and immunohistochemical overlap. Otherwise the EWSR1 FISH has been found to be negative in all head and neck epithelial mimics of HCCC.49,50,53 Of course EWSR1 is not specific to HCCC. It has been found to be rearranged in tumors of hematolymphoid, mesenchymal, melanocytic, and epithelial differentiation, including cutaneous hidradenomas and SMET.54,56 Similarly, the EWSR1-ATF1 fusion gene is not specific to HCCC, having first been recognized in clear cell sarcoma, angiomatoid fibrous histiocytoma and the so-called “malignant gastrointestinal neuroectodermal tumor,” which was originally recognized as a gastrointestinal variant of clear cell sarcoma.57 An angiosarcoma has even recently been described in the parotid with a EWSR1-ATF1 fusion.58 This of course makes a good morphologic diagnosis paramount before interpreting molecular results in establishing a diagnosis.
The current recommendation is that EWSR1 FISH is not necessary in classic mucin-negative cases of HCCC. However, when abundant mucin is present or when there is minimal clear cell differentiation or hyalinization, confirmation can be useful, as MECs showing similar features will generally have a higher grade than HCCC and potentially different treatment.49,50 This is because most HCCCs lack cystic content, have highly infiltrative tumor fronts, and often show perineural invasion. With mucinous differentiation, these would be considered at least intermediate grade, or even high grade when using most traditional MEC grading schemes.49,50
In the current salivary classification system, there are 5 tumors that harbor recurrent translocations and these include the most common benign tumor (PA) and the 2 most common malignant tumors (MEC and ACC). This is not surprising given that salivary tumors are not typically high grade or morphologically heterogeneous and they do not seem to arise in a stepwise manner from a precursor lesion such as ductal carcinoma in situ. In fact, salivary tumors generally look “weird” and appear unlike anything normal in the gland (the exception being AcCC). With that in mind, it is likely that the list of translocations will grow. Many of these will serve as diagnostic markers such as those in HCCC and MASC. Just as in the latter diagnosis they may serve to identify as yet undiscovered tumor types or to reclassify some tumors as different from one another when currently lumped together. They may do the opposite as is the case with CCOC and HCCC. Some of these markers may be prognostic; such as MAML2, a possibly good prognostic marker in MEC, or MYB, a possibly bad prognostic marker in ACC. However, this requires refinement in the diagnosis before being fully accepted. Translocations and their fusions may serve as targets of therapy in the future, although most salivary carcinomas are slow growing and low grade and are unlikely to benefit from such therapies. Also given the rarity of most of these tumors, it is unlikely that therapies could ever be clinically proven efficacious in a trial setting. For now, the presence of these fusions is of academic interest and can serve as a powerful diagnostic tool when used appropriately.
1. Robinson DR, Wu YM, Kalyana-Sundaram S, et al .Identification of recurrent NAB2-STAT6 gene fusions in solitary fibrous tumor by integrative sequencing.Nat Genet. 2013; 45:180–185.
2. Barnes L, Eveson JW, Reichart P, et al .World Health Organization Classification of Tumours. Pathology and Genetics of Head and Neck Tumours. 2005; .Lyon:IARC Press.
3. Voz ML, Aström AK, Kas K, et al .The recurrent translocation t(5;8)(p13;q12) in pleomorphic adenomas results in upregulation of PLAG1 gene expression under control of the LIFR promoter.Oncogene. 1998; 16:1409–1416.
4. Kas K, Voz ML, Röijer E, et al .Promoter swapping between the genes for a novel zinc finger protein and beta-catenin in pleiomorphic adenomas with t(3;8)(p21;q12) translocations.Nat Genet. 1997; 15:170–174.
5. Bahrami A, Dalton JD, Shivakumar B, et al .PLAG1 alteration in carcinoma ex pleomorphic adenoma: immunohistochemical and fluorescence in situ hybridization studies of 22 cases.Head Neck Pathol. 2012; 6:328–335.
6. Aström AK, Voz ML, Kas K, et al .Conserved mechanism of PLAG1 activation in salivary gland tumors with and without chromosome 8q12 abnormalities: identification of SII as a new fusion partner gene.Cancer Res. 1999; 59:918–923.
7. Röijer E, Kas K, Behrendt M, et al .Fluorescence in situ hybridization mapping of breakpoints in pleomorphic adenomas with 8q12-13 abnormalities identifies a subgroup of tumors without PLAG1 involvement.Genes Chromosomes Cancer. 1999; 24:78–82.
8. Persson F, Winnes M, Andrén Y, et al .High-resolution array CGH analysis of salivary gland tumors reveals fusion and amplification of the FGFR1 and PLAG1 genes in ring chromosomes.Oncogene. 2008; 27:3072–3080.
9. Persson F, Andrén Y, Winnes M, et al .High-resolution genomic profiling of adenomas and carcinomas of the salivary glands reveals amplification, rearrangement, and fusion of HMGA2.Genes Chromosomes Cancer. 2009; 48:69–82.
10. Geurts JM, Schoenmakers EF, Röijer E, et al .Identification of NFIB as recurrent translocation partner gene of HMGIC in pleomorphic adenomas.Oncogene. 1998; 16:865–872.
11. Stenman G .Fusion oncogenes and tumor type specificity—insights from salivary gland tumors.Semin Cancer Biol. 2005; 15:224–235.
12. Martins C, Fonseca I, Roque L, et al .PLAG1 gene alterations in salivary gland pleomorphic adenoma and carcinoma ex-pleomorphic adenoma: a combined study using chromosome banding, in situ hybridization and immunocytochemistry.Mod Pathol. 2005; 18:1048–1055.
13. Antonescu CR, Zhang L, Shao SY, et al .Frequent PLAG1 gene rearrangements in skin and soft tissue myoepithelioma with ductal differentiation.Genes Chromosomes Cancer. 2013; 52:675–682.
14. Bahrami A, Dalton JD, Krane JF, et al .A subset of cutaneous and soft tissue mixed tumors are genetically linked to their salivary gland counterpart.Genes Chromosomes Cancer. 2012; 51:140–148.
15. Bahrami A, Perez-Ordonez B, Dalton JD, et al .An analysis of PLAG1 and HMGA2 rearrangements in salivary duct carcinoma and examination of the role of precursor lesions.Histopathology. 2013; 63:250–262.
16. Tonon G, Modi S, Wu L, et al .t(11;19)(q21;p13) translocation in mucoepidermoid carcinoma creates a novel fusion product that disrupts a Notch signaling pathway.Nat Genet. 2003; 33:208–213.
17. Behboudi A, Enlund F, Winnes M, et al .Molecular classification of mucoepidermoid carcinomas-prognostic significance of the MECT1-MAML2 fusion oncogene.Genes Chromosomes Cancer. 2006; 45:470–481.
18. Martins C, Cavaco B, Tonon G, et al .A study of MECT1-MAML2 in mucoepidermoid carcinoma and Warthin’s tumor of salivary glands.J Mol Diagn. 2004; 6:205–210.
19. Fehr A, Röser K, Heidorn K, et al .A new type of MAML2 fusion in mucoepidermoid carcinoma.Genes Chromosomes Cancer. 2008; 47:203–206.
20. Nakayama T, Miyabe S, Okabe M, et al .Clinicopathological significance of the CRTC3-MAML2 fusion transcript in mucoepidermoid carcinoma.Mod Pathol. 2009; 22:1575–1581.
21. Seethala RR, Dacic S, Cieply K, et al .A reappraisal of the MECT1/MAML2 translocation in salivary mucoepidermoid carcinomas.Am J Surg Pathol. 2010; 34:1106–1121.
22. Tirado Y, Williams MD, Hanna EY, et al .CRTC1/MAML2 fusion transcript in high grade mucoepidermoid carcinomas of salivary and thyroid glands and Warthin’s tumors: implications for histogenesis and biologic behavior.Genes Chromosomes Cancer. 2007; 46:708–715.
23. Bell D, Holsinger CF, El-Naggar AK .CRTC1/MAML2 fusion transcript in central mucoepidermoid carcinoma of mandible—diagnostic and histogenetic implications.Ann Diagn Pathol. 2010; 14:396–401.
24. Lennerz JK, Perry A, Mills JC, et al .Mucoepidermoid carcinoma of the cervix: another tumor with the t(11;19)-associated CRTC1-MAML2 gene fusion.Am J Surg Pathol. 2009; 33:835–843.
25. García JJ, Hunt JL, Weinreb I, et al .Fluorescence in situ hybridization for detection of MAML2 rearrangements in oncocytic mucoepidermoid carcinomas: utility as a diagnostic test.Hum Pathol. 2011; 42:2001–2009.
26. Chenevert J, Barnes LE, Chiosea SI .Mucoepidermoid carcinoma: a five-decade journey.Virchows Arch. 2011; 458:133–140.
27. Bell D, Luna MA, Weber RS, et al .CRTC1/MAML2 fusion transcript in Warthin’s tumor and mucoepidermoid carcinoma: evidence for a common genetic association.Genes Chromosomes Cancer. 2008; 47:309–314.
28. Fehr A, Röser K, Belge G, et al .A closer look at Warthin tumors and the t(11;19).Cancer Genet Cytogenet. 2008; 180:135–139.
29. Skalova A, Vanecek T, Hauer L, et al .CRTC1/3-MAML2 fusions are not seen in metaplastic Warthin tumors and metaplastic pleomorphic adenomas of salivary glands.Mod Pathol. 2013; 26:S2 301–315.
30. Persson M, Andrén Y, Mark J, et al .Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck.Proc Natl Acad Sci USA. 2009; 106:18740–18744.
31. Mitani Y, Li J, Rao PH, et al .Comprehensive analysis of the MYB-NFIB gene fusion in salivary adenoid cystic carcinoma: Incidence, variability, and clinicopathologic significance.Clin Cancer Res. 2010; 16:4722–4731.
32. Persson F, Fehr A, Sundelin K, et al .Studies of genomic imbalances and the MYB-NFIB gene fusion in polymorphous low-grade adenocarcinoma of the head and neck.Int J Oncol. 2012; 40:80–84.
33. Fehr A, Kovács A, Löning T, et al .The MYB-NFIB gene fusion-a novel genetic link between adenoid cystic carcinoma and dermal cylindroma.J Pathol. 2011; 224:322–327.
34. Shao C, Bai W, Junn JC, et al .Evaluation of MYB promoter methylation in salivary adenoid cystic carcinoma.Oral Oncol. 2011; 47:251–255.
35. Mitani Y, Rao PH, Futreal PA, et al .Novel chromosomal rearrangements and break points at the t(6;9) in salivary adenoid cystic carcinoma: association with MYB-NFIB chimeric fusion, MYB expression, and clinical outcome.Clin Cancer Res. 2011; 17:7003–7014.
36. West RB, Kong C, Clarke N, et al .MYB expression and translocation in adenoid cystic carcinomas and other salivary gland tumors with clinicopathologic correlation.Am J Surg Pathol. 2011; 35:92–99.
37. Brill LB II, Kanner WA, Fehr A, et al .Analysis of MYB expression and MYB-NFIB gene fusions in adenoid cystic carcinoma and other salivary neoplasms.Mod Pathol. 2011; 24:1169–1176.
38. Persson M, Andrén Y, Moskaluk CA, et al .Clinically significant copy number alterations and complex rearrangements of MYB and NFIB in head and neck adenoid cystic carcinoma.Genes Chromosomes Cancer. 2012; 51:805–817.
39. Skálová A, Vanecek T, Sima R, et al .Mammary analogue secretory carcinoma of salivary glands, containing the ETV6-NTRK3 fusion gene: a hitherto undescribed salivary gland tumor entity.Am J Surg Pathol. 2010; 34:599–608.
40. Connor A, Perez-Ordoñez B, Shago M, et al .Mammary analog secretory carcinoma of salivary gland origin with the ETV6 gene rearrangement by FISH: expanded morphologic and immunohistochemical spectrum of a recently described entity.Am J Surg Pathol. 2012; 36:27–34.
41. Chiosea SI, Griffith C, Assaad A, et al .Clinicopathological characterization of mammary analogue secretory carcinoma of salivary glands.Histopathology. 2012; 61:387–394.
42. Karl SR, Ballantine TV, Zaino R .Juvenile secretory carcinoma of the breast.J Pediatr Surg. 1985; 20:368–371.
43. Tognon C, Knezevich SR, Huntsman D, et al .Expression of the ETV6-NTRK3 gene fusion as a primary event in human secretory breast carcinoma.Cancer Cell. 2002; 2:367–376.
44. Weinreb I, Tabanda-Lichauco R, Van der Kwast T, et al .Low-grade intraductal carcinoma of salivary gland: report of 3 cases with marked apocrine differentiation.Am J Surg Pathol. 2006; 30:1014–1021.
45. Bishop JA, Yonescu R, Batista D, et al .Most nonparotid “acinic cell carcinomas” represent mammary analog secretory carcinomas.Am J Surg Pathol. 2013; 37:1053–1057.
46. Milchgrub S, Gnepp DR, Vuitch F, et al .Hyalinizing clear cell carcinoma of salivary gland.Am J Surg Pathol. 1994; 18:74–82.
47. Bilodeau EA, Hoschar AP, Barnes EL, et al .Clear cell carcinoma and clear cell odontogenic carcinoma: a comparative clinicopathologic and immunohistochemical study.Head Neck Pathol. 2011; 5:101–107.
48. Dardick I, Leong I .Clear cell carcinoma: review of its histomorphogenesis and classification as a squamous cell lesion.Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2009; 108:399–405.
49. Tanguay J, Weinreb I .What the EWSR1-ATF1 fusion has taught us about hyalinizing clear cell carcinoma.Head Neck Pathol. 2013; 7:28–34.
50. Antonescu CR, Katabi N, Zhang L, et al .EWSR1-ATF1 fusion is a novel and consistent finding in hyalinizing clear-cell carcinoma of salivary gland.Genes Chromosomes Cancer. 2011; 50:559–570.
51. Solar AA, Schmidt BL, Jordan RC .Hyalinizing clear cell carcinoma: case series and comprehensive review of the literature.Cancer. 2009; 115:75–83.
52. O’Regan E, Shandilya M, Gnepp DR, et al .Hyalinizing clear cell carcinoma of salivary gland: an aggressive variant.Oral Oncol. 2004; 40:348–352.
53. Shah AA, LeGallo RD, van Zante A, et al .EWSR1 genetic rearrangements in salivary gland tumors: a specific and very common feature of hyalinizing clear cell carcinoma.Am J Surg Pathol. 2013; 37:571–578.
54. Antonescu CR, Zhang L, Chang NE, et al .EWSR1-POU5F1 fusion in soft tissue myoepithelial tumors. A molecular analysis of sixty-six cases, including soft tissue, bone, and visceral lesions, showing common involvement of the EWSR1 gene.Genes Chromosomes Cancer. 2010; 49:1114–1124.
55. Bilodeau EA, Weinreb I, Antonescu CR, et al .Clear cell odontogenic carcinomas show EWSR1 rearrangements: a novel finding and a biological link to salivary clear cell carcinomas.Am J Surg Pathol. 2013; 37:1001–1005.
56. Möller E, Stenman G, Mandahl N, et al .POU5F1, encoding a key regulator of stem cell pluripotency, is fused to EWSR1 in hidradenoma of the skin and mucoepidermoid carcinoma of the salivary glands.J Pathol. 2008; 215:78–86.
57. Stockman DL, Miettinen M, Suster S, et al .Malignant gastrointestinal neuroectodermal tumor: clinicopathologic, immunohistochemical, ultrastructural, and molecular analysis of 16 cases with a reappraisal of clear cell sarcoma-like tumors of the gastrointestinal tract.Am J Surg Pathol. 2012; 36:857–868.
58. Gru AA, Becker N, Pfeifer JD .Angiosarcoma of the parotid gland with a t(12;22) translocation creating a EWSR1-ATF1 fusion: a diagnostic dilemma.J Clin Pathol. 2013; 66:452–454.
translocation; mucoepidermoid carcinoma; pleomorphic adenoma; hyalinizing clear cell carcinoma; adenoid cystic carcinoma; mammary analog secretory carcinoma
Copyright © 2013 by Lippincott Williams & Wilkins
Highlight selected keywords in the article text.