Clinicopathological practice in the differential diagnosis of mucoepidermoid carcinoma from neoplasms with mucinous component

Abstract

Background

The differential diagnosis of mucoepidermoid carcinoma (MEC) from neoplasm undergoing mucinous features brings more pitfalls to pathologists. Combining specific MAML2 gene rearrangement and histological characteristics may be the solution.

Methods

Twenty‐five tumors with mucinous components were selected for differential diagnosis of MEC. All the cases were detected for MAML2 gene rearrangement. The cases diagnosed as MEC were classified into four variants: classic, oncocytic, Warthin‐like, and nonclassified, and they were graded using the Brandwein system. The histological characteristics of non‐MECs were summarized for differential diagnosis. Univariate survival analysis was performed on MECs.

Results

There were 16 MECs; 62.5% were MAML2 rearranged. For the low‐, intermediate‐, and high‐grade MECs, the rate of rearrangement was 83.3%, 100%, and 28.6%, respectively. Both the oncocytic and Warthin‐like MECs were MAML2 rearranged. For the classic and nonclassified MECs without MAML2 rearrangement, non‐keratinized squamoid cells and distinctive mucinous cells were essential diagnostic criteria. On survival analysis, all the disease progression occurred in high‐grade MECs (p = 0.038). Nine cases were diagnosed as non‐MECs: pleomorphic adenoma with mucinous metaplasia showed no ex‐capsular involvement; metaplastic Warthin tumor appeared with overt keratinization and residual oncocytic bilayered epithelium; mix squamous cell and glandular papilloma showed an endobronchial papillary growing pattern; adenosquamous carcinoma was accompanied by squamous carcinoma in situ of the overlying mucosa. All the non‐MECs were negative for MAML2 rearrangement.

Conclusion

The application of combining MAML2 rearrangement and histological characteristics is helpful in the differential diagnosis between MEC and other tumors with mucinous components.

INTRODUCTION

Mucoepidermoid carcinoma (MEC) is the most common salivary gland malignancy, comprised of a variable combination of three major cell types. Meanwhile, MEC is unique because its cell morphology is more variable than we ever thought, including but not limited to mucinous cells, squamoid cells, clear cells, oncocytic cells, and spindle cells, which are mixed in different proportions to form cystic and solid nest structures. Recently some new types of MEC have been reported such as ciliated MEC and MEC with acinar differentiation. Thus the diagnosis of MEC is always challenging for pathologists when nonclassical features appear as major components. The accurate diagnosis of MEC has been greatly improved by the discovery of the unique genetic alteration, CRCT1/CRTC3–MAML2 translocation. However, the grading of MEC and the diagnosis of MEC without MAML2 gene rearrangement remain controversial. MEC has been described as low-frequency in non-head-and-neck sites such as the bronchus, breast, thymus, skin, uterine cervix, and rare in the esophagus. Even recently a case of MEC with MAML2 gene rearrangement in the liver has been reported. The differential diagnosis of MEC and more common tumors at these sites will bring more pitfalls to the routine pathological practice. In retrospect of previous studies, lack of overt keratinization has been agreed to to some extent as a clue to the diagnosis of MEC. Besides, according to the definition, MEC should contain recognizable mucous-secreting cells. Here we are going to share our clinical experience of differential diagnosis between MEC and tumors with mucinous components, combining MAML2 gene rearrangement and histological characteristics.

MATERIALS AND METHODS

Case selection and follow-up

A review of pathological archives from our institution was conducted to search for tumors with mucinous components requiring differential diagnosis with MEC from 2017 to 2022. Our study collected 25 cases that meet the above requirement. The progression-free-survival (PFS) was calculated from the day of surgery to the day of disease progression, death, or last contact (April 2022). The clinical information and outcomes of patients were obtained from digital medical record databases and telephone interviews.

Histological classification of MEC and non-MECs

Hematoxylin–eosin-stained slides were reviewed by three experienced pathologists. Immunohistochemical staining of p63, p40, CK7, CK5/6, S100, and Calponin was performed for auxiliary diagnosis (Table ). Periodic acid Schiff (PAS) staining, alcian blue (AB) staining, and mucicarmine staining were performed for determination of mucous-secreting cells. The diagnosis of MECs and non-MECs followed the criteria described in the literature. The histological variants of MEC were classified as follows: Classic variant at least focally contained mucinous cystic components. The oncocytic variant contained >50% oncocytic tumor cells and in which epidermoid cells are indistinctive which could be determined by the focal expression of p63. A warthin-like variant would be considered with a lack of orderly, bilayered, overtly oncocytic epithelium at least focally. Nonclassified MEC referred to those who do not meet the above three types of cases, in which nonkeratinized squamoid cells and clear cells were prominent with less mucinous cells. The grading of MEC was performed using the criteria that Brandwein et al. described.

Table 1 Antibodies for immunohistochemical staining

Antibodies	Clone	Source
Calponin	EP63	ZhongShan-GoldenBridge, Beijing, China
p63	4A4	Roche Diagnostics, Shanghai, China
CK7	OV-TL12/30	Maxim, Fuzhou, China
CK5/6	MX040	Maxim, Fuzhou, China
S100	4C4.9	Maxim, Fuzhou, China
p40	MX008	Maxim, Fuzhou, China

Fluorescence in situ hybridization analysis of MAML2 gene rearrangement

Fluorescence in situ hybridization (FISH) was performed on a formalin-fixed paraffin-embedded section using commercial MAML2 dual color break-apart probe (Anbiping Medical Technology). The interpretation criteria were based on the literature; one hundred tumor cells were counted. The cells without MAML2 gene rearrangement showed two yellow signals. FISH results were considered positive if >15% of all nuclei showed a splitting signal consisting of an orange and a green signal, or if a deletion of the orange or green signal was observed.

The sections were de-paraffinized in xylene twice for 15 min, dehydrated in 100% ethanol for 5 min, rehydrated in a series of ethanol solutions (100%, 85%, and 70% for 2 min each), and then placed into deionized water for 5 min. The sections were treated using deionized water at 90°C for 20 min and then digested in protease solution (1 mg/ml) in 0.01 mol/L HCl at 37°C for 15–30 min according to the sample conditions. The sections were then washed in deionized water for 5 min, dehydrated in a series of ethanol solutions (70%, 85%, and 100% for 2 min each), and air-dried. After applying the probe, the sections were covered and sealed with rubber cement, then incubated in the ThermoBrite instrument (Abbott) with co-denaturation at 83°C for 5 min and hybridization at 42°C/16 h. After carefully removing of the coverslips, the sections were washed in 0.1% NP40/2× SSC solution at 46°C for 5 min then in 70% ethanol at room temperature for 3 min. The sections were air-dried in the dark, and counterstained with DAPI (Abbott). Analysis was performed using CytoVision Imaging Workstation (Leica Biosystems).

Statistical analysis

All statistical analyses were performed using SPSS software (version 19.0, IBM SPSS Statistics). Differences in categorical data between MAML2 gene rearrangement and histologic diagnosis were determined using the χ² test and Fisher exact test. The survival curves were generated from the Kaplan–Meier method and log-rank test. Two sides p < 0.05 were considered statistically significant.

RESULTS

Clinical features of MECs

Of the 25 selected cases, 16 were diagnosed as MEC (9 females and 7 males). The median age was 51.0 years and the mean age was 43.7 years (range, 10–65 years). Most of the MECs originated from the major and minor salivary glands (12, 75.0%), and one of them was considered to be originated from an ectopic minor salivary gland in the left superior lymph node because no evidence of primary foci was found. In addition, two cases were derived from the bronchus, and one case each derived from the esophagus and thymus. All the cases were from primary radical excision except one from the biopsy. None of the patients received preoperative chemotherapy except one patient, who had lymph node metastasis according to the pathological consultation of local excision from other hospitals. The characteristics of MECs are shown in Table . The microscopic histological features of MECs are presented in Figures  and . Other than the classic features as the major histological criteria, non-keratinized squamoid cells and identified scattered mucinous cells would be considered as the minor histological appearances of MEC.

Table 2 Clinicopathological features of MECs

Characteristics	n (%)
Gender
Male	7 (43.8)
Female	9 (56.3)
Age groups (years)
≤51	7 (43.8)
>51	9 (56.3)
Specimen type
Excision	15 (93.8)
Biopsy	1 (6.2)
Primary sites
Major salivary	6 (37.5)
Minor salivary	6 (37.5)
Bronchus	2 (12.5)
Esophagus	1 (6.3)
Thymus	1 (6.3)
Variant
Oncocytic	5 (31.3)
Warthin like	2 (12.5)
Classic	5 (31.3)
Nonclassified	4 (25.0)
Brandwein grade
Low grade	6 (37.5)
Intermediate grade	3 (18.8)
High grade	7 (43.8)
LNM (n = 11)
Negative	8 (72.7)
Positive	3 (27.3)
MAML2 gene rearrangement
Yes	10 (62.5)
No	6 (37.5)
Margin status (n = 15)
≥1 mm	13 (86.7)
<1 mm	2 (13.3)
Treatment (n = 15)
Surgery only	10 (66.7)
Surgery + RT	2 (13.3)
Surgery + CRT	3 (20.0)

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

Histological characteristics, MAML2 gene rearrangement, and grade of MECs

Ten of the 16 MECs (62.5%) were positive for MAML2 gene rearrangement and 100% of the oncocytic and Warthin-like variants were positive. From a grading point of view, five of the six low-grade MECs (83.3%) and all three intermediate-grade MECs (100%) were MAML2 gene rearranged. Nevertheless, MAML2 gene rearrangement occurred in only two of the seven high-grade MECs (28.6%). The data are shown in Table .

Table 3 Comparison of variant and tumor grade in MAML2 gene rearranged and not-rearranged mucoepidermoid carcinoma

Characteristics	MAML2 gene rearrangement	p value
Yes, n (%)	No, n (%)
Variant			0.053
Oncocytic	5 (100.0)	0 (0.0)
Warthin-like	2 (100.0)	0 (0.0)
Classic	2 (40.0)	3 (60.0)
Nonclassified	1 (25.0)	3 (75.0)
Brandwein grade			0.042
Low grade	5 (83.3)	1 (16.7)
Intermediate grade	3 (100.0)	0 (0.0)
High grade	2 (28.6)	5 (71.4)

For the 10 MAML2 genes rearranged MECs, both the oncocytic and Warthin-like variants were classified as a low-intermediate grade. Oncocytic MECs contained fewer mucinous cells and squamoid cells (shown by p63 staining) than classic MECs (Figure ). Warthin-like MECs contained extensive cystic lumen surrounded by dense lymphoid stroma mimicked Warthin tumor (Figure ), but lacked typical bilayer oncocytic epithelium (Figure ). Among the five classic MECs, two were MAML2 gene rearranged, one of which was of intermediate grade and the other was of high grade. The high-grade classic MEC contained both low-grade mucinous cystic components (Figure ) and high-grade non-keratinized solid nests (Figure ); splitting signals on FISH assay could be detected in both areas of this case. All four nonclassified MECs were of high grade, and the only one with MAML2 gene rearrangement was from the thymus. This case contained a partial epithelial–myoepithelial morphology (Figure ) and was dominated by clear cells (Figure ), with focal mucinous lumen formation and a single green signal on FISH assay (Figure ).

Three classical MECs and three nonclassified MECs were MAML2 genes not-rearranged. Among the three classical MECs, one originated from the cervical lymph node ectopic salivary gland and was of low grade (Figure ). The remaining two were of high grade, showing varying degrees of mixed squamoid and clear cells, all of which were focal primary foci and lymph node metastasis with definite mucinous features (Figure ). Of the three nonclassified MECs, one was composed of clear cells nests with bizarre nuclei and scattered mucinous cells with AB positive staining (Figure ). Another one was an endobronchial tumor composed of cords dominated by nests of squamoid cells with stromal sclerosis and scattered mucinous lumen formation (Figure ). The last one was composed of nests of high-grade non-keratinized squamoid cells and scattered mucinous cells (Figure ), and the tumor cells showed strong expression of both CK5/6 and CK7 (Figure ). The above three nonclassified MECs all had the histological features of non-keratinized squamoid cells and identifiable mucinous cells.

Clinical outcome of MECs

The median follow-up time of MECs was 22.0 months (4.0–59.5 months). Four of the seven high-grade MECs had recurrence or metastasis after surgery and one of them died of the disease. No disease progression occurred in low-intermediate grade MECs. There was a statistically significant difference in PFS between high-grade and low-intermediate grade MECs (p = 0.038). There was no statistical difference in PFS between MAML2 gene rearranged and not-rearranged MECs (p = 0.93). The survival curves are shown in Figure .

[IMAGE OMITTED. SEE PDF]

Clinicopathological features and MAML2 gene rearrangement of non-MECs

Of the nine non-MECs, two were pleomorphic adenoma (PA) with mucinous metaplasia, characterized by confining to the tumor capsule (Figure ) and low-grade morphology of the mucinous components (Figure ). Three were metaplastic Warthin tumors, of which oncocytic bilayered epithelium still could be found in the tumor (Figure ) and overt keratinization of squamous epithelium existed (Figure ). One was a bronchial mix of squamous cell and glandular papilloma, which was limited to the bronchial lumen without invasion (Figure ), and characterized by the prominent papillary structure (Figure ). Three were adenosquamous carcinoma (Figure ). One patient with adenosquamous carcinoma of the infra-temporal fossa, who had a history of radiotherapy for nasopharyngeal carcinoma 19 years ago, recurred 19 months after this palliative surgery. Another patient with adenosquamous carcinoma of the epiglottis received postoperative radiotherapy and did not relapse during a 49 month follow-up. The last patient with esophageal adenosquamous carcinoma and the remaining six patients with benign tumors were treated with surgery only, and no disease progression was observed during follow-up. The microscopic histological characteristics of non-MECs are shown in Figure  and clinicopathological features are listed in Table . MAML2 gene rearrangement did not occur in any of the non-MECs.

[IMAGE OMITTED. SEE PDF]

Table 4 Clinicopathological features and diagnosis of non-MECs

No.	Sex	Age (years)	Site	Diagnosis	Histological features inconsistent with MEC	MAML2 rearranged	Treatment	Follow-up (months)	Disease progression
1	F	66	Parotid	PA with mucinous metaplasia	Lack of ex-capsular involvement, low-grade morphology	No	Surgery only	56.0	No
2	F	26	Parotid	PA with mucinous metaplasia	Lack of ex-capsular involvement, low-grade morphology	No	Surgery only	46.5	No
3	M	63	Parotid	Metaplastic WT	Consist of oncocytic bilayered epithelium, overt keratinization	No	Surgery only	13.5	No
4	M	56	Parotid	Metaplastic WT	Consist of oncocytic bilayered epithelium	No	Surgery only	8.0	No
5	M	58	Parotid	Metaplastic WT	Consist of oncocytic bilayered epithelium	No	Surgery only	6.0	No
6	F	37	Bronchus	Mix squamous cell and glandular papilloma	Endobronchial papillary growing without invasion, keratinizing squamous epithelium	No	Surgery only	56.0	No
7	F	62	Epiglottis	ASC	Overlying mucosa with squamous carcinoma in situ, identified adenocarcinomatous components	No	Surgery+RT	49.0	No
8	M	35	Infra-temporal fossa	ASC	Overt keratinization, identified adenocarcinomatous components	No	Surgery only	19.0	Yes, relapse
9	F	73	Esophagus	ASC	Overlying mucosa with squamous carcinoma in situ, identified adenocarcinomatous components	No	Surgery only	11.5	No

DISCUSSION

Although the pathological description of MEC has a long history, the diagnosis remains challenging. In this study, 25 cases of tumors with mucinous components were collected for differential diagnosis of MEC. In the study of Cipriani et al., they compared four grading systems for MEC, and only the Brandwein high grade was a statistically significant univariate risk factor for recurrence. We also agree that it is important to meet the clinical need to treat tumors at risk of progression. Therefore, we chose the Brandwein grading system, and the disease progression only occurred in high-grade cases (Figure , p = 0.038), which is consistent with Cipriani et al. In contrast to the high rate of MAML2 gene rearrangement in high-grade MECs of their series, less than 30% of our high-grade MECs could be detected with MAML2 gene rearrangement. Besides, the status of MAML2 rearrangement was not statistically significant in predicting PFS (Figure , p = 0.93). Considering the sensitivity limit of FISH detection and sample quality, we believe that the diagnosis of MEC requires a combination of MAML2 gene rearrangement and histological characteristics.

Our result exhibited that MAML2 gene rearrangement had a subtype tendency that was close to statistical significance (Table , p = 0.053). All the oncocytic and Warthin-like MECs were positive for MAML2 gene rearrangement in our study. Oncocytic MEC has a limited amount of both mucous-secreting cells and squamoid cells and was previously difficult to be distinguished from oncocytic carcinoma, acinar cell carcinoma, and adenocarcinoma-NOS (not otherwise specified). Since the application of the FISH assay, the MAML2 gene rearranged oncocytic MECs have reached 100% both in the literature and in our study, which greatly assists in the accurate diagnosis of this variant. Warthin-like MEC is another variant that caused years of diagnostic confusion for pathologists. Ishibashi et al. speculated that Warthin-like MEC is a de novo tumor rather than is derived from an existent metaplastic Warthin tumor because CRCT1-MAML2 fusion could be detected from every type of epithelium of the Warthin-like MEC from their study. Due to the dense lymphoid stroma and the ambiguous invasion, MAML2 gene rearrangement is a powerful tool for the diagnosis of the Warthin-like variant. In addition, the lack of a typical oncocytic bilayered epithelium and the clinical information including younger age as well as the non-parotid site could also remind us to rule out Warthin-like MEC. These have been proved useful in our study as well.

For the five classical MECs, MAML2 gene rearrangement was not detected in three of them. One low-grade classical MEC derived from the cervical lymph node could be distinguished from lymphadenoma. Because the former contained three major cell types of MEC with cellular dysplasia, the latter contains benign basaloid and glandular cells but not mucous-secreting cells. The other two high-grade classical MECs could be distinguished from hyalinizing clear cell carcinoma because both of them definitely contained mucous cells which accumulated or formed cystic lumen. These characteristics could not be seen in hyalinizing clear cell carcinoma according to the literatures. For the four nonclassified MECs, MAML2 gene rearrangement was detected in only one of them. The MAML2 gene rearranged nonclassified MEC had a mixed morphology of MEC and epithelial–myoepithelial carcinoma, and molecular aberration of MAML2 gene rearrangement confirmed its essential diagnosis of MEC rather than a collision tumor. The remaining three nonclassified MECs without MAML2 gene rearrangement contained non-keratinized squamoid cells, scattered mucinous cells, or mucinous lumen. What is more, all three cases lacked squamous dysplasia or carcinoma in situ of overlying epithelium, which is an important diagnostic criterion for adenosquamous carcinoma. Two of them originated from the parotid excluding metastatic disease, and one was endobronchial without dysplasia of the respiratory epithelium. In addition, in one case from the parotid, the strong expression of both CK5/6 and CK7 was more consistent with the diagnosis of MEC than squamous cell carcinoma, according to the previous literature that CK7 is only expressed focally in squamous cell carcinoma of the head and neck, although there were so few mucinous cells in this tumor that they could only be identified by mucicarmine stain. In summary, most of the MECs that were not rearranged in our study came from high-grade and nonclassified cases, which are also diagnostic difficulties in pathological practice. Therefore, we believe that morphological characteristics are also important, and the identification of classic components, non-hyperkeratinized squamoid cells, and definite mucinous cells, combined with immunohistochemical staining and mucin staining, are the key to giving a reliable diagnosis of MEC.

This study also included nine non-MECs with difficult differential diagnoses in clinical practice (Table ). Cases 1 and 2 were PA with mucinous metaplasia, which should be differentiated from carcinoma ex PA. Microscopically, the low-grade mucinous components in the two cases lacked ex-capsular involvement. According to previous records, intra-capsular invasive components of carcinoma ex PA are always the high-grade components such as ductal carcinoma; and myoepithelial cells of accompanied intraductal carcinoma were considered benign. Associating with the evidence of characteristic RET fusion of myoepithelial cells to support myoepithelial cells to be neoplastic in intercalated duct-type intraductal carcinoma, we assumed that if the mucinous cells in PA were MEC-derived, MAML2 gene rearrangement would be strong evidence. Actually, both the two cases were not MAML2 gene rearranged, therefore the molecular alteration to support MEC is lacking. Besides, MEC ex PA reported in other literature was characterized by ex-capsular involvement and accompanied by high-grade morphology, which is also contrary to the morphology of our cases, supporting our diagnosis of PA with mucinous metaplasia rather than MEC for the two cases. Negative MAML2 gene rearrangement, overt keratinization, and residual oncocytic bilayered epithelium have made it easier to identify metaplastic Warthin tumors in Cases 3–5. Although the endobronchial mass in Case 6 was large (maximum diameter, 7.5 cm), adequate sampling, combined with the microscopic images of papillary growth without invasion, and clear keratinization without cellular dysplasia, enabled a correct diagnosis of mixed squamous cell and glandular papilloma other than MEC. In Cases 7–9, the diagnosis of adenosquamous carcinoma was consistent with previous criteria, including overlying mucosa with squamous carcinoma in situ, identified adenocarcinomatous components, and possibly accompanied overt keratinization. Last but not least, MAML2 gene rearrangement, as a specific genetic aberration of MEC, did not occur in any of these cases, which strongly supported the diagnosis of non-MEC.

Despite the relatively short follow-up time of the patients and mere FISH assay for MAML2 gene alteration, we still tried our best to explore the practical diagnostic implications of combining MAML2 gene rearrangement and histological characteristics for MEC.

In conclusion, although MAML2 is not a good prognostic marker for PFS compared with tumor grade, our study confirmed that MAML2 gene rearrangement is very helpful for some special subtypes of MEC which deviate from conventional appearances such as oncocytic MEC and Warthin-like MEC. Meanwhile, diagnosis of high-grade MEC is always difficult, and holding on to classic histological features is crucial. Furthermore, understanding the characteristics of other lesions resembling MEC is also important, particularly adenosquamous carcinoma and other tumors with mucinous components.

AUTHOR CONTRIBUTIONS

Haizhen Lu, Jianming Ying, and Yuelu Zhu designed this study. Jiali Mu, Haifeng Zhang, and Xin Li enrolled the patients and collected the clinical data. Yuelu Zhu, Haizhen Lu, and Jianming Ying reviewed the pathological sections. Yan Li, Lei Guo, and Wenbin Li analyzed the FISH assay. Yuelu Zhu, Jiali Mu, Haifeng Zhang, and Xin Li conducted the statistical analyses. The manuscript was drafted by Yuelu Zhu, Yan Li, Lei Guo, Wenbin Li, Jianming Ying, and Haizhen Lu. All authors participated in the interpretation of the results.

ACKNOWLEDGMENTS

We thank Yun Zhang for technical support of the FISH assay, Ying Wang, Yiyuan Dou, and Wenchao Liu for immunochemical staining, and Lu Yu for data presentation advice. This study was supported by the Beijing Hope Run Special Fund of the Cancer Foundation of China, No. LC2018A19.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

The data analyzed during the current study are available from the corresponding author on reasonable request.

ETHICS STATEMENT

This study was approved by Ethics Committee of National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (No. 21/041-2712). Written informed consent for participation was not required for this retrospective study. All the procedures were accordance with the Helsinki Declaration of 1975, as revised in 2000.

References

Cheuk W, Chan JKC. Tumors of the salivary glands. In: Fletcher CDM, ed. Diagnostic Histopathology of Tumors. 5th edn. Elsevier; 2021:301‐410.

Bishop JA, Cowan ML, Shum CH, Westra WH. MAML2 rearrangements in variant forms of mucoepidermoid carcinoma: ancillary diagnostic testing for the ciliated and warthin‐like variants. Am J Surg Pathol. 2018;42(1):130‐136. [DOI: https://dx.doi.org/10.1097/PAS.0000000000000932]