Salivary gland neoplasms are rare lesions in the head and neck (H&N) pathology realm. There are more than 20 malignant and 15 benign salivary gland neoplasms in the 5th edition of the World Health Organization classification of H&N tumors. These neoplasms consist of heterogeneous groups of uncommon diseases that make diagnosis and treatment challenging for the clinical team. Using an algorithmic immunohistochemical approach–defined tumor origin and type has proven to be effective and advantageous. Immunohistochemistry may be used as sort of a “diagnostic looking glass,” not as a positive or negative type tool, but as an indispensable complement to a hematoxylin-eosin morphologic pattern–based approach. Furthermore, the understanding of the novel discoveries of the salivary gland gene fusions and the molecular aspects of these tumors makes the process easier and improve the diagnosis as well as treatment aspects. This review reflects our experience with more recent diagnostic antibodies, which include MYB RNA, Pan-TRK, PLAG1, LEF1, and NR4A3. Each of these is linked with a specific type of neoplasm; for example, gene fusions involving the PLAG1 and HMGA2 oncogenes are specific for benign pleomorphic adenomas, and MYB is associated with adenoid cystic carcinoma.
To review these more recent antibodies, which highly enhance salivary gland neoplasm diagnosis.
The study sources involved literature PubMed searches, including multiple review articles, case reports, selected book chapters, and Geisinger Medical Center cases.
Salivary gland tumors are a rare, varied group of lesions in H&N pathology. We need to have continuous readings and revisions of the molecular consequences of these fusion oncoproteins and their subsequent targets, which will eventually lead to the identification of novel driver genes in salivary gland neoplasms.
As seen in the 5th edition of the World Health Organization classification of head and neck (H&N) tumors, there are more than 20 primary malignant neoplasms of the salivary glands and more than 15 benign lesions (Table).1 Salivary gland tumors constitute the greatest tumor variation of any organ. Further, these tumors are infrequent, with a wide range of biologic behaviors and histologic heterogeneity.
These tumors can arise from the 3 major salivary glands or any of the thousands of minor glands that line the mouth and extend into the nose, pharynx, and esophagus. The 3 pairs of major salivary glands are the parotid, submandibular, and sublingual glands, which are composed of a mixture of serous and mucus acini, with parotid being composed of 100% serous acini and decreasing in proportion, with the submandibular being approximately 50% serous and 50% mucus acini, and the sublingual gland only containing approximately 10% serous acini. About 80% of these tumors arise from the parotid gland, with approximately one-quarter of these presenting as malignant neoplasms. This is in opposition to submandibular, sublingual, and minor salivary glands, from which malignant tumors arise at a higher rate (approximately 45%, 85%, and 75%, respectively).1–3
Tumor diversity and complexity make it difficult to diagnose salivary gland tumors. Therefore, immunohistochemistry (IHC) is a valuable supplement. As always, marker selection and interpretation must be highly related to the clinical history and histologic features. With rapid progress in understanding tumorigenesis by genetic and epigenetic mechanisms, more and more IHC markers are validated in pathology diagnosis. Here, we summarized how we use IHC in salivary gland diagnosis and highlight some updated IHC markers that are highly sensitive and specific.
In the past it was believed that “Any salivary gland tumor can look like any other salivary gland tumor,” and to some that might be true, but we can at least use morphology and color to bunch them into groups. Another thought reflects the user’s lack of understanding of normal salivary gland ultrastructure: “Immunohistochemistry is useless, let alone molecular testing,” but in reality, salivary gland tumors originate from microanatomic sites and are highlighted by IHC. Some of the challenges of salivary glands stem from not understanding clinical and biologic behavior, and the diagnostic difficulty due to their diversity and complexity. But utilizing an algorithmic approach, using immunophenotyping and molecular markers and surrogates, we have come a long way.
In terms of morphology, well-encapsulated or well-circumscribed unilobular lesions are highly favored to be benign, but in situ carcinomas can often originate from such benign-appearing lesions.
However, multinodular lesions such as pleomorphic adenoma with pseudopodia or multinodular basal cell adenoma (BCA) have a higher risk of recurrence and are usually benign. Some multinodular lesions like myoepithelial carcinoma or low-grade mucoepidermoid carcinoma are classified as low-grade malignancies. An infiltrative pattern is considered to be an unequivocal feature of malignancy. Other strong indicators of malignancy are evidence of perineural or angiolymphatic invasion, tumor necrosis, and sclerosis, and desmoplasia also has a relative value as an indicator for malignancy, but benign tumor can always infarct or show desmoplasia, especially as things change after fine-needle aspiration. Believe it or not, in salivary gland tumors, cytologic features, such as pleomorphism or presence of mitotic figures, do not carry such a strong value as in other sites. And it is because these tumors can be so heterogeneous that it is quite difficult to have a definitive diagnosis on limited sample without the use of ancillary studies.
Using the tumor phenotype is a great way to categorize salivary gland tumors. Tumors derived from distal ductal-acinar structures can be monophasic tumors, which are derived from either acinar or ductal epithelium. Examples are acinic cell or salivary duct carcinoma, or basal- or myoepithelial-derived tumors or biphasic tumors that try to recapitulate normal salivary gland structures with an epithelial element and basal cell/myoepithelial component. Proximal excretory duct tumors will have a skin adnexal phenotype. Biphasic tumors are defined by a bilayered arrangement of luminal (ductal/epithelial) cells and abluminal (basal and/or myoepithelial) cells. Both components can often be seen on hematoxylin-eosin and in some tumor types can be sharp and organized. The use of IHC will definitely help pronounce this biphasic phenotype by clearly highlighting the components.
Prototypical biphasic tumors also include benign tumors, such as pleomorphic adenoma, BCA, and Warthin tumors, and malignant tumors, such as epithelial myoepithelial carcinoma, adenoid cystic carcinoma (ACC), etc. Prototypical biphasic tumors are believed to derive from intercalated ducts and striated ducts (Figure 1, A), with some monophasic tumors, such as acinic cell carcinoma, originating from acinar cells, salivary duct carcinoma from ductal cells, and myoepithelial tumors from the most terminal portion of the salivary unit.
A, Hematoxylin-eosin illustrates serous acini, intercalated ducts, and striated ducts. B, Smooth muscle actin highlights abluminal pattern of myoepithelial cells of acini and intercalated ducts. C, CK 5/6 (low–molecular weight keratin) shows strong positive luminal/abluminal immunoreactivity within ductal cells, with weaker variable staining of acinar cells and weak to negative staining within basal and myoepithelial cells. D, CAM5.2 (high–molecular weight keratin) shows strong positive abluminal immunoreactivity within basal and myoepithelial cells and negativity within ductal and acinar cells. E, p63 highlights basal cells and myoepithelial cells and is negative within ductal and acinar cells. F, DOG1 shows apical membranous staining within acinar cells. G, S100 stains basal and myoepithelial cells with weak immunoreactivity of luminal ductal cells. H, SOX10 highlights acinar cells sharply from ductal component (original magnification ×20).
A, Hematoxylin-eosin illustrates serous acini, intercalated ducts, and striated ducts. B, Smooth muscle actin highlights abluminal pattern of myoepithelial cells of acini and intercalated ducts. C, CK 5/6 (low–molecular weight keratin) shows strong positive luminal/abluminal immunoreactivity within ductal cells, with weaker variable staining of acinar cells and weak to negative staining within basal and myoepithelial cells. D, CAM5.2 (high–molecular weight keratin) shows strong positive abluminal immunoreactivity within basal and myoepithelial cells and negativity within ductal and acinar cells. E, p63 highlights basal cells and myoepithelial cells and is negative within ductal and acinar cells. F, DOG1 shows apical membranous staining within acinar cells. G, S100 stains basal and myoepithelial cells with weak immunoreactivity of luminal ductal cells. H, SOX10 highlights acinar cells sharply from ductal component (original magnification ×20).
IHC can help in distinguishing between the different components of the lesion and should not be solely relied on for a definitive positive or negative diagnosis (Figure 1, B through H). Thus, the p63+/p40− immunophenotype (Figures 2 and 3) becomes the group of obvious interest consisting of canalicular adenoma, secretory carcinoma, polymorphous low-grade adenocarcinoma (PLGA), and cribriform adenocarcinoma of the salivary gland. In tumors of ductal phenotype, p63 expression is mainly noted in “intercalated duct” type tumors, such as PLGA/cribriform adenocarcinoma of the salivary gland, canalicular adenoma, and secretory carcinoma, although p40 is uniformly negative, which can help distinguish these entities from truly biphasic tumor types. Remember that IHC should be used to visualize cells and highlight tumor components, not to make a definitive diagnosis based solely on positive or negative results.
A, p40 staining helps differentiate the biphasic pattern of oncocytic mucoepidermoid carcinoma versus (B) what would be considered incomplete p40 abluminal/peripheral or scattered staining of oncocytic carcinoma on the right (original magnification ×20).
A, p40 staining helps differentiate the biphasic pattern of oncocytic mucoepidermoid carcinoma versus (B) what would be considered incomplete p40 abluminal/peripheral or scattered staining of oncocytic carcinoma on the right (original magnification ×20).
A, CAM5.2 highlights an epithelial component of an epithelial myoepithelial carcinoma, with (B) p63 highlighting the myoepithelial overgrowth on the right (original magnification ×20).
A, CAM5.2 highlights an epithelial component of an epithelial myoepithelial carcinoma, with (B) p63 highlighting the myoepithelial overgrowth on the right (original magnification ×20).
New advances in combining the widely used method of protein detection via IHC technique with mRNA in situ hybridization capabilities help us highlight molecular mechanisms to reveal gene tumor expression and regulatory mechanisms. The following are the new neuro markers useful to help the diagnosis.
Pan-TRK
Pan-TRK is expressed in tumors harboring NTRK1/2/3 gene fusions with high sensitivity and specificity ranging from 95% to 100% and 92% to 100% in secretory carcinoma (Figures 4 and 5), respectively.4–8 This is a great marker to help differentiate other low-grade salivary gland tumors that also express mammaglobin (Figure 4, A).5,6 Notably, fluorescence in situ hybridization (FISH) for ETV6 TRK3 fusion should be performed in a case of pan-TRK–negative but still suspicious for secretory carcinoma.
A, Secretory carcinoma with cytoplasmic mammaglobin stain expression, and (B) Pan-TRK stain showing strong nuclear expression (original magnification ×20).
A, Secretory carcinoma with cytoplasmic mammaglobin stain expression, and (B) Pan-TRK stain showing strong nuclear expression (original magnification ×20).
A, Secretory carcinoma expressing (B) a weak, spotty nuclear Pan-TRK stain. ETV6:NTRK3 fusion was detected in this tumor (original magnification ×20).
A, Secretory carcinoma expressing (B) a weak, spotty nuclear Pan-TRK stain. ETV6:NTRK3 fusion was detected in this tumor (original magnification ×20).
Pan-TRK is also expressed in tumors harboring this gene fusion, such as infantile fibrosarcoma and other pediatric NTRK rearranged mesenchymal tumors, cellular congenital mesoblastic nephroma, lipofibromatous-like neural tumor (NTRK1 gene fusions), uterine sarcoma with features of fibrosarcoma harboring NTRK fusions, and other tumors harboring NTRK fusions (colorectal carcinoma, glioblastomas, lung adenocarcinoma, melanoma).4,7–9
MYB
MYB is an oncogene that acted as a DNA-binding transcription regulator to control proliferation and differentiation.10 MYB is not observed in normal salivary gland parenchyma and even normal tissues adjacent to the tumors.11,12
In salivary tumors, MYB IHC (Figure 6, A through C) was up to 94% positive, but specificity was just 54% for ACC. Other salivary tumors and basaloid and sinonasal carcinomas also had more than 50% positivity. MYB RNA ISH provides better specificity for the diagnosis of ACC (95% among other salivary tumors and 81% in basaloid and sinonasal carcinomas), whereas the sensitivity is 92%.12,13
A, Salivary ductal carcinoma expresses androgen receptor stain (B) and MYB stain (C) (original magnification ×20).
A, Salivary ductal carcinoma expresses androgen receptor stain (B) and MYB stain (C) (original magnification ×20).
Many ACCs have t(6;9) chromosomal translocation, resulting in MYB:NFIB fusion (50%–70%). Other fusions included MYBL1:NFIB fusion (up to 15%). MYB or MYBL1 gene fusions are detected in more than 90% of ACCs.14–16 Notably, upregulation of MYB gene was also seen in fusion-negative cases.14–16
Only 1 study showed that MYB overexpression in ACC was significantly poor in terms of overall survival, whereas other studies showed no significant prognostic differences.12,17,18
Our experience has shown MYB RNA antibody will also show similar immunoreactivity in basaloid squamous cell carcinoma as well as other high-grade carcinomas of the H&N.
It was published that the catenin beta 1 (CTNNB1)-PLAG1 that results from the fusion of these 2 genes has been identified in salivary duct carcinoma and myoepithelial carcinoma.19
NR4A3
NR4A3 (Figure 7, A) was not expressed in normal salivary gland tissues20,21 and had high sensitivity and specificity for acinic cell carcinoma (Figure 7, B) with a sensitivity of 82% to 100% and specificity of 93% to 100% by IHC.22–27 Only a small subset of non–acinic cell carcinoma tumors expressed NR4A3, including secretory carcinoma, mucoepidermoid carcinoma, polymorphous adenocarcinoma, and pleomorphic adenoma.25,27 By RNA-seq and confirmation by FISH, t(4;9) is the most common recurrent rearrangement (57%), followed by t(9;12)(19) and t(8;9) (5%), and all activate NR4A3 gene expression.20–22
A, Nuclear receptor subfamily 4 group A member 3 (NR4A3 stain). B, Hematoxylin-eosin stain of acinic cell carcinoma (original magnification ×20).
A, Nuclear receptor subfamily 4 group A member 3 (NR4A3 stain). B, Hematoxylin-eosin stain of acinic cell carcinoma (original magnification ×20).
LEF1
Lymphoid enhancer binding factor 1 (LEF1) is a transcription factor for both B and T lymphoid cells and is expressed by genes transcriptionally regulated by the Wnt/CTNNB1 pathway.28 This pathway regulates and controls the development of airway submucosal glands, mammary glands, teeth, and adnexal structures.29 LEF1 (Figure 8, A) has been used as a immunohistochemical marker for multiple salivary gland tumors, such as BCA (Figure 8, B) and other benign entities. LEF1 will help pathologists increase their accuracy in differentiating basaloid salivary gland neoplasms.30 When comparing benign tumors (pleomorphic adenoma and BCA) with the most common malignant basaloid salivary gland tumor ACC lesion, positive LEF1 favors a benign neoplasm.30 This is especially helpful because ACC, the most common malignant neoplasm, has been reported to be LEF1− on resections.31 Having a positive LEF1 result favors a benign neoplasm, with positive predictive values of 95% pleomorphic adenoma and 97% BC, respectively; however, additional IHC studies with LEF1 are needed.30
A, Lymphoid enhancer binding factor 1 (LEF1 stain). B, Hematoxylin-eosin stain of basal cell adenoma (original magnification ×20).
A, Lymphoid enhancer binding factor 1 (LEF1 stain). B, Hematoxylin-eosin stain of basal cell adenoma (original magnification ×20).
PLAG1
Pleomorphic adenoma (PA) gene 1 (PLAG1) encodes a zing finger transcription factor, and it usually gets triggered by the chromosomal translocation of the 8q12 in salivary gland subgroup PA.32 This proto-oncogene has previously been found in placenta as well as fetal tissues.33 The PLAG1 oncogene’s appearance can be detected by using the PLAG1 stain. The main advantage of using this particular stain in salivary gland tumors is that this marker is mostly limited to pleomorphic adenoma and to carcinoma ex-pleomorphic adenomas. This special marker has good specificity in diagnosing pleomorphic adenomas, and it can be used during signing out of H&N tumor cases.34 It is shown that this marker is negative in most salivary carcinomas. Epithelial-myoepithelial carcinoma is frequently found in preexisting pleomorphic adenoma. In this scenario, IHC staining for HRAS Q61R is recommended to diagnose epithelial-myoepithelial carcinoma, which is reported to be highly specific. However, HRAS Q61R–positive cases were present only in de novo epithelial-myoepithelial carcinomas (54 of 76 cases; 71%) but not in epithelial-myoepithelial carcinoma’s ex pleomorphic adenoma.35 A known limitation of this marker, however, is that it can be expressed in PA mimickers, like PLGA. Moreover, if this marker (PLAG1) is expressed in a malignant case of H&N tumor, you should consider carcinoma ex-PA because it has been shown to have PLAG1+ results.
Some genetic alterations are frequently seen in some malignant salivary gland tumors, but IHC stainings are usually not available. FISH or molecular testing is useful in this setting. A total of 55% to 65% of mucoepidermoid carcinomas had t(11;19)(q21;p13), especially CRTC1-MAML2 fusion, which can be detected by molecular testing; EWSR1-ATF1 fusion can be detected by FISH in most hyalinizing clear cell carcinoma; MEF2C-SS18 fusion can be detected by molecular testing in microsecretory adenocarcinoma; AKT1 E17K mutations can be detected by molecular testing in mucinous adenocarcinoma.36–38
CONCLUSIONS
The diagnostic and treatment challenges of salivary gland tumors are due to its heterogeneous elements of those uncommon diseases. However, continuous research and new discoveries of the translocations and proto-oncogenes in salivary gland neoplasm will make the diagnosis and target treatments easier and will improve the quality of care.