A 60-year-old female with no history of smoking was admitted to the hospital because of a mass under the right ear for 4 days. Physical examination revealed a 3×2.5 cm solid mass in the parotid gland region, characterized by its firm texture, mobility, well-defined borders, and absence of tenderness. Skin temperature over the mass was within the normal range.

Ultrasound examination disclosed a solid mass in the deep lobe of the right parotid gland, displaying uneven internal echoes. Preliminary considerations leaned towards pleomorphic adenoma. Magnetic resonance imaging (MRI) depicted a mass in the deep lobe of the right parotid gland, exhibiting mixed long/short T2 and iso-/slightly long T1 signals. Diffusion-weighted imaging (DWI) showed mixed signals, and enhanced scanning revealed conspicuous uneven enhancement. Septation enhancement was noted within the mass, and localized incomplete capsule formation was observed ( Figure 1 ). The patient then received excision of lesions of parotid gland and parapharyngeal space + parotidectomy III–IV + facial nerve dissection + myocutaneous flap transfer.

Grossly, the surgical resection specimen comprised irregular grey−red tissue measuring 3.5×2×1 cm, displaying a combination of solid and cystic appearance on the cut surface. Histologically, the tumor exhibited well-defined and complex cystic architecture ( Figure 2A ). The cyst wall was lined with stratified epithelium, and the subepithelial stroma demonstrated densely arrangement of lymphocytes with formation of lymphoid follicles ( Figure 2B ). Epithelial cells, polygonal in shape with eosinophilic cytoplasm, exhibited a crowded pattern devoid of obvious polarity. The nuclei were round or oval with small nucleoli. Scattered mucus cells were interspersed in the epithelium ( Figure 2C ). In some areas, sheet-like epithelium hyperplasia was observed, comprising squamoid cells, mucous cells, and intermediate cells ( Figure 2D ). Mitotic figures were infrequent, with an absence of necrosis, nerve, and vascular invasion. Immunohistochemistry demonstrated positive expression of CK7, CK8, and CK18 in the epithelial components ( Figure 2E ). Squamoid cells and intermediate cells expressed CK5/6, P63 and P40, while β-catenin localized to the cell membrane. The Ki-67 index was approximately 1–2%, and calponin expression was absent. Fluorescence in situ hybridization (FISH) revealed MAML2/CRTC1 gene fusion ( Figure 2F ). No metastasis was found in the resected four lateral neck level II lymph nodes. Subsequent to a 7-month follow-up, no recurrence or metastasis was observed.

A 29-year-old non-smoking male sought admission to the hospital because of “tumor recurrence 6 months after submandibular gland tumor resection”. Six months prior to admission, he had undergone submandibular gland tumor resection at another medical facility, with the postoperative pathological diagnosis indicating WT. On physical examination, a 3 cm-long surgical scar was seen on the left mandible, accompanied by a non-tender, non-ulcerated 3×2 cm mass.

Ultrasound examination revealed a 2.7×1.0 cm mass in the left submandibular gland, displaying an oval shape, clear boundaries, hypoechoic interior, uneven distribution, slightly enhanced posterior echo, and abundant blood flow signals. MRI scan illustrated a mass in the left submandibular region, characterized by iso/high signal intensity on T1WI and slightly high signal intensity on T2WI, with uneven enhancement. The lesion exhibited close proximity to the anterior border of the left submandibular gland, prompting consideration of malignancy. Excision of submandibular gland and tumor and functional neck dissection were performed.

Macroscopic examination unveiled a well-defined grey−yellow nodule with a 2 cm diameter, exhibiting clear demarcation ( Figure 3A ). Microscopically, the lesion shared similarities with the previously described Case 1, presenting an epithelial lining and multiple cysts within a background of dense lymphoid stroma ( Figure 3B ). The epithelium was stratified, nonpolar, and interspersed with mucus-secreting cells ( Figure 3C ). Proliferation of epithelial cells in some regions formed nests, comprising squamoid, intermediate, and mucinous cells ( Figure 3D ). Immunohistochemically, epithelial components expressed CK7, while squamoid cells and intermediate cells expressed P63, P40, and CK5/6 ( Figure 3E ). The Ki-67 index was 1%, and calponin and SMA were not expressed. FISH revealed fusion of MAML2/CRTC1 ( Figure 3F ). A lymph node was present in the specimen, and no metastasis was identified. Retrospective examination of slides from the initial operation displayed nearly identical morphological features to those of the recurrent tumor, substantiating the diagnosis of WL-MEC. A subsequent 4-month follow-up post-reoperation revealed no signs of recurrence or metastasis.

WT is frequently observed in elderly male smokers and almost always occurs in the parotid gland. Metaplasia changes in WT are usually focal and accompanied by hemorrhage and fibrosis associated with previous operations, such as biopsy. The epithelium in WL-MEC, as shown in our review, is often multilayered, with crowded cells and variable degree of atypia, lacking the typical bilayer epithelial structure observed in WT which exhibits impressive tall columnar cell morphology, eosinophilic cytoplasm, and polar arrangement. Ishibashi et al. (6) and Zhang et al. (11) both considered the presence of classic bilayer eosinophilic epithelium as the morphological key to differentiating WT from WL-MEC. More importantly, MAML2 gene rearrangement serves as a useful distinguishing factor. As for treatment, WT has no risk of recurrence and metastasis, and only requires local resection.

Lymphoepithelial carcinoma is an undifferentiated cancer, characterized by epithelial nests and sheets distributed in lymphoid stroma, lacking polycystic structures observed in WL-MEC. The epithelium exhibits evident atypia, with vesicular nuclei, prominent nucleoli and abundant mitotic figures. Additionally, the occurrence of this carcinoma is related to Epstein-Barr (EB) virus infection.

Seifert et al. (21) referred to this as the malignant transformation of WT, proposing the following diagnostic criteria: i. pre-existence of a benign WT; ii. existence of transitional zones from benign oncocytic to malignant epithelium; iii. the infiltrating growth in surrounding lymphoid tissue; and iv. exclusion of metastasis.

Also known as lymphoepithelial sialadenitis, it occurs due to diffuse lymphocyte infiltration caused by atrophy of salivary gland tissue. Proliferation of the ductal epithelium and myoepithelium forms characteristic epithelial myoepithelial islands, lacking cellular atypia. The occurrence of these lesions is associated with Sjögren’s syndrome (22).

In addition, WL-MEC also needs to be differentiated from lymphoepithelial cysts, sebaceous lymphadenoma, squamous cell carcinoma with lymphocytic infiltration, and cystic lymph node metastases such as metastatic Warthin tumor-like thyroid carcinoma and metastatic squamous cell carcinoma.

The relationship between MECs and WTs has been a matter of debate for a long time. Squamous metaplasia and mucoid metaplasia are not uncommon morphological changes in WT, and the proportion of metaplasia reported in the literature ranges from 0.2% to 22% (23). When this metaplasia is extensive, it tends to be very similar to MEC in morphology. The occurrence of squamous and mucoid metaplasia in WT, coupled with reports of MECs arising in WT (MEC ex WT) (24, 25), has fueled speculation regarding their interconnection. Previous studies found that MAML2 gene rearrangement occurs in a small number of WT cases, and some scholars have speculated that WT is a precursor of MEC⁸. Rotellini et al. (26) also detected MAML2 gene rearrangement in 2 of 8 cases of WT with metaplasia. Accordingly, they assumed that the metaplastic changes in WT may be a sign of transformation of WT to MEC. However, recent studies have failed to detect MAML2 gene rearrangement in larger samples of patients with WT (8). In addition, using FISH, Ishibashi et al. did not find MAML2 heterogeneity in different morphological regions of whole tumor tissue sections of metaplastic WTs (6). The nomenclature for WL-MEC clarifies this confusion, and the current view is that metaplastic WT with MAML2 gene rearrangement should be reclassified as WL-MEC (10). Therefore, the presence of MAML2 gene rearrangement becomes the key to distinguishing WL-MEC from WT with metaplasia (8). This perspective challenges the recognition of MEC ex WT. Nevertheless, given the limited number of case reports, it is not clear whether the distinction between MEC ex WT and WL-MEC has any clinical significance. However, when squamous and myxoid metaplasia occurs in WT, vigilance should be exercised, and MAML2 gene testing is advisable to exclude the possibility of malignancy (17).

Currently used histological grading systems for MEC include the Armed Forces Institute of Pathology system, modified Healey grading systems, the Brandwein system, and the Memorial Sloan Kettering system (20). Low- and medium-grade MECs exhibit less aggression and a better prognosis, typically treated with total parotidectomy and, if necessary, neck dissection. High-grade MECs may necessitate adjuvant radiotherapy (1, 8). A study by Behboudi et al. showed that MECs with MECT1–MAML2 gene fusion had a better prognosis (27). Notably, WL-MEC, characterized as a low-grade MEC regardless of which grading system is used, exhibits a favorable prognosis. Among the 30 patients with follow-up records that we reviewed, only 3 (10.0%) developed recurrence after 6, 12 and 48 months during a median follow-up of 28 months (1–166 months). No case died of the disease. Therefore, as a low-grade malignant tumor, total parotidectomy with or without neck dissection was recommended for WL-MEC. Highly invasive treatments should be approached with caution.

Since malignant salivary tumors represent a diagnostic challenge because of their rarity and morphologic overlap, machine learning techniques have been applied to the field of pathology to improve diagnostic performance in recent years. Researches revealed promising results, although most applications are in developmental phase. However, in one research, mucoepidermoid carcinomas fell in one of the three histological types that easy to be misclassified by the tree-based machine learning model (28). For this special malignant tumor, more research is still needed.

The CRTC1/3-MAML2 fusion in MEC has been well recognized for many years (2). The fusion encodes a chimeric protein in which the Notch-binding domain of MAML2 is replaced by the CREB-binding coiled-coil domain of CRTC1, and activates transcription of the Notch target gene HES1 independently of Notch ligand (3). Behboudi et al. demonstrated that the CRTC1-MAML2 fusion is a useful marker in predicting the biological behavior of MECs (27). More recently, the study of Chen and colleagues provided direct evidence for CRTC1-MAML2 as a key driver for MEC development and validated CRTC1-MAML2 as a therapeutic target for patients with MEC (29). These studies provide a new direction for the accurate prognosis and treatment of MEC, including WL-MEC.