The morphological similarities and diagnostic challenges between Warthin tumor (WT) with metaplasia and mucoepidermoid carcinoma (MEC) with lymphoid stroma have recently been discussed.^(1–3) Warthin tumor is the second most common benign salivary gland tumor,⁽⁴⁾ whereas MEC is the most prevalent malignant neoplasm.⁽⁵⁾ Typically, WT is characterized by abundant lymphoid stroma, cysts lined by a double-layered oncocytic epithelium, and occasionally a papillary architectural pattern.⁽⁶⁾ Conversely, MEC often shows a prominent cystic component, epithelium with variable proportions of squamous, intermediate, and mucous cells, desmoplastic stroma, and a non-exuberant lymphoid infiltrate.^(7,8) However, WT can occasionally exhibit squamous and/or mucous metaplasia, making differentiation from MEC challenging.^(9,10) Additionally, MEC may present with a substantial lymphocytic infiltrate in the stroma, termed Warthin-like MEC.^{(1,3,11–13)}

Currently, there is no clear histological definition for distinguishing these two types of tumors.^(1–3) Establishing histopathological criteria is crucial due to the clinical implications of misdiagnosing benign (i.e., WT) and malignant (i.e., Warthin-like MEC) tumors.

To analyze the clinicopathological and immunophenotypic characteristics of a series of morphologically unconventional Warthin tumors and mucoepidermoid carcinomas.

This study was approved by the Research Ethics Committee of the Universidade Estadual de Campinas (CAAE: 11521019.0.0000.5404; # 3.357.193). All procedures involving human participants were conducted in strict accordance with the ethical standards of the institutional research committee and the principles outlined in the Declaration of Helsinki (1964) and its subsequent amendments. Informed consent was obtained from all the participants. Consent was obtained in written form, ensuring that patient anonymity was rigorously maintained throughout the manuscript and publication process.

All cases were retrieved from the archives of the Department of Pathological Anatomy of the Escola de Ciências Médicas of the Universidade Estadual de Campinas and the Clinical Hospital of the same university to ensure consistency in sample processing and diagnostic evaluation.

In this study, we conducted a comparative analysis of 11 low-grade epithelial salivary gland tumors with Warthin-like features. Group A comprised cases showing a double-layered oncocytic epithelium involving less than 50% of the total epithelial cells, whereas Group B included tumors lacking this morphological feature but exhibiting glandular and/or cystic structures within abundant lymphoid stroma. For reference, we compared these tumors with cases of conventional WT and MEC.

The primary criterion for inclusion in Group A was the presence of a double-layer oncocytic epithelium. However, upon closer examination, we identified other distinguishing features that were more prevalent in Group A than in Group B. These included ciliated and apocrine cells as well as the lymphoepithelium. Tumor-associated fibrosis was more commonly observed in Group B, indicating a heightened stromal response to dysregulated epithelial growth. The presence of intermediate cells (with and without squamous differentiation) and mucous cells varied in both groups, but was more pronounced in cases lacking the double-layer oncocytic epithelium.

Our study revealed significant similarities in the immunohistochemical positivity patterns, particularly for CK7 and mitochondrial antigens. In Group A and WT controls, CK7 was predominantly expressed in luminal columnar/oncocytic cells, a trend that was also observed with anti-mitochondrial antibodies. The mitochondrial antigen demonstrated a robust response in the luminal cells, which was distinct from the patterns observed in the basal and intermediate cells. In contrast, Group B and MEC controls exhibited diffuse CK7 expression across all cell types. Mitochondrial antigen staining indicated predominantly uniform moderate-intensity positivity throughout the epithelial layers, with one case in Group B showing focal yet distinct luminal cell positivity.

The basal cell markers CK14 and CK5/6 demonstrated consistent immunoprofiles across all four groups, with expression in the peripheral cells within the tumor nodules and glandular structures. CK14 was particularly valuable for identifying focal linear basal cells, even in Group A cases lacking the characteristic double-layered oncocytic epithelium. CK5/6 exhibited positivity in tumors with higher proportions of intermediate and squamous cells, a pattern that was particularly pronounced in MEC controls. p63 immunoreactivity was similar across Groups A and B and MEC controls, and was primarily observed in the basal layer, with increased expression in the intermediate cells. According to García et al.,⁽¹⁾ anti-p63 positivity favors classification of the lesion as MEC. These observations can be attributed to a distinct immunoprofile, revealing a notable abundance of intermediate cells beneath the columnar and mucous cells, which aligns with our findings. Interestingly, the WT controls exhibited contrasting basal cell expression patterns, with most tumors displaying fewer intermediate cells.

Notable morphological and immunohistochemical parallels were observed between Group A and WT controls, including the presence of a double-layer oncocytic epithelium and an immunophenotype characterized by CK7, CK14, and mitochondrial antigens. One case from Group B showed distinct luminal positivity with anti-mitochondrial antibody staining, suggesting a closer resemblance to Group A and WT controls. However, most cases in Group B and MEC controls exhibited similar immunohistochemical profiles for CK7 and mitochondrial antigens.

The morphological and immunohistochemical findings of this study support the classification of Group A as metaplastic WT. Conversely, the absence of the typical double-layered epithelium and the frequent presence of desmoplastic stroma in Group B suggest that these should be classified as Warthin-like variants of MEC. Although Warthin-like MEC may present with a double-layered epithelium, this feature is often atypical. The subjectivity of visual assessment complicates the diagnosis when based solely on morphological criteria, especially in limited samples.^(15,16)

In this context, molecular analysis, particularly the detection of MAML2 gene rearrangements, becomes essential. The chromosomal translocation t(11;19)(q21;p13), responsible for the CRTC1-MAML2 fusion, is present in 38–82% of MEC cases and serves as a more precise diagnostic marker. Fluorescence in situ hybridization (FISH) is the recommended method for identifying this fusion and confirming MEC diagnosis.^(15,17) Lei et al.⁽¹³⁾ reported two cases initially diagnosed as metaplastic WT that were reclassified as Warthin-like MEC after detection of MAML2 gene rearrangements by FISH. These findings emphasize the need for a combined diagnostic approach to avoid errors in cases with challenging morphological features.

Immunohistochemistry is an important complementary tool, particularly when molecular analyses are unavailable. Previous studies have highlighted the importance of markers such as CK5/6, CK14, and p63 in differentiating salivary tumors such as WT and MEC. However, our findings suggest that the inclusion of CK7 and mitochondrial antigens may expand the diagnostic options, particularly in morphologically atypical cases. Yang et al.⁽¹⁸⁾ observed that the eosinophilic double-layered epithelium was absent in 73.5% of Warthin type MEC cases, making morphological diagnosis difficult. In the remaining cases, the epithelium was flattened and less oncocytic. While immunohistochemistry with markers, such as p40, p63, CK5/6, CK7, and CA125, is useful for differentiation, the analysis of genetic rearrangements, particularly the CRTC1-MAML2 fusion identified by FISH, is crucial for the definitive and accurate diagnosis of MEC.^(13,15,17)

This study has several limitations. Molecular analyses targeting MAML2 rearrangements were not performed in all cases. To ensure the accuracy of the diagnostic classification of WT and MEC, it is crucial to perform molecular analyses to confirm the presence or absence of MAML2 rearrangements in both tumor groups. This approach would provide a more robust foundation for distinguishing these lesions and complement the histological and immunohistochemical findings. Moreover, while the results with CK7 markers and mitochondrial antigens are promising, the limited number of cases analyzed may affect their reliability. Validation of these markers in a larger sample size is necessary to confirm their specificity and sensitivityand to strengthen the conclusions. Expanding the sample and incorporating molecular testing would enable a more comprehensive and reliable diagnostic approach for these lesions.

Despite these limitations, our findings have important implications for clinical practice. Differential diagnosis between WT and MEC can be challenging, especially in atypical cases. Distinct immunohistochemical patterns, such as CK7 expression and mitochondrial antigens in WT, provide valuable diagnostic clues that may assist pathologists even in the absence of molecular tests. Accurate differentiation is essential as it directly affects treatment decisions, which varies depending on whether the lesion is benign or malignant. These results provide a basis for future investigations exploring the specificity and sensitivity of these markers to improve diagnostic accuracy.

Our results indicate that CK7 and mitochondrial markers are useful for differentiating Warthin tumor from mucoepidermoid carcinoma, particularly based on their distinct expression patterns in luminal cells. However, validation using larger sample sizes is needed to confirm their specificity and sensitivity. The analysis of MAML2 rearrangements is crucial for accurate molecular diagnosis and to define the appropriate treatment. The combination of established basal markers, such as CK5/6, CK14, and p63, with newer indicators, such as CK7 and mitochondrial markers, represents a step forward in improved diagnosis. Further studies with larger samples and more detailed molecular analyses are needed to consolidate these findings.