Primary salivary gland-type lung tumours are rare, accounting for <1% of all primary lung carcinomas (1,2). These tumours are thought to originate from glands in the submucosa of the lower respiratory tract (3); therefore, they usually present as a central space-occupying lung mass extending into the bronchus (4). These tumours exhibit the same histological variability as salivary gland tumours and are therefore classified according to the World Health Organization (WHO) criteria for salivary gland tumours (2). The most common histological subtype of primary salivary gland lung tumours is mucoepidermoid carcinoma, followed by adenoid cystic carcinoma (5). Primary pulmonary epithelial-myoepithelial carcinoma (P-EMC) is a rare and much less common tumour that has been recognized only within the last 2 decades (6), with only case reports and small case series previously published in the literature (4,6-20). Morphologically and immunohistochemically, EMC is a low-grade malignant tumour with biphasic epithelial and myoepithelial morphology. The WHO classifies the tumour as ‘a malignant tumour composed of variable proportions of two cell types, which typically form duct-like structures. The biphasic morphology is represented by an inner layer of duct-lining epithelial type cells and an outer layer of clear myoepithelial cells (1). Many of these lesions are low-grade malignant epithelial neoplasms but often pose diagnostic problems because of their central location, circumscription, low-grade cytological features and occasional focal similarity to other salivary gland tumours, particularly in small endobronchial biopsy samples (21). These tumours usually have a good prognosis after complete surgical resection with clear margins, but recurrence and metastasis have also been reported (15,22,23).

To the best of our knowledge, ~50 cases of P-EMC have been reported worldwide, as detailed in Table SI, but few publications concerning primary P-EMC exist. Due to its rarity and unproven malignant potential, the optimal treatment for P-EMC has not been established. The present study reported a typical case of primary P-EMC with diagnostic challenges and pitfalls associated with small biopsy specimens of this rare tumor and reviewed the literature to comprehensively analyse its clinical features, diagnosis and appropriate treatment.

A 58-year-old nonsmoking female presented to Taihe Hospital (Shiyan, China) after having experienced intermittent haemoptysis for one month. The patient denied other respiratory symptoms, including cough, expectoration, fever, night sweats, fatigue, chest tightness and chest pain. The patient also denied any history of cancer or relevant family history. Physical examination indicated clear breath sounds on auscultation, with no dry or wet rales in both lungs. Chest CT revealed a nodule in the bronchial lumen of the left superior lobe with a maximum diameter of 0.8 cm (Fig. 1A). The hilar and mediastinal lymph nodes were not enlarged. Head and neck magnetic resonance imaging, bone scintigraphy and whole-body CT scans with contrast enhancement revealed no distant metastases or lymphadenopathy. Serum tumour markers for general lung cancer diagnosis were as follows: Neuron-specific enolase, 10.2 ng/ml (normal range, 0-16.3 ng/ml); carcinoembryonic antigen, 4.1 µg/l (normal range, 0-5 µg/l); Cyfra 21-1, 2.9 µg/l (normal range, 0-3.3 µg/l); squamous cell carcinoma antigen, 1.45 ng/ml (normal range, 0-2.7 ng/ml); and ferritin, 164 ng/ml (normal range, 30-400 ng/ml) - all within normal limits.

Bronchoscopy revealed a smooth neoplasm obstructing the lumen at the entrance of the left upper lobe (Fig. 2A). Transbronchial forceps biopsy of the neoplasm was conducted and rapid onsite evaluation (24) suggested non-small cell lung cancer (Fig. 2B). Histopathology indicated a probable diagnosis of mucoepidermoid carcinoma. Positron emission tomography (PET)-CT scan can determine whether there is a possibility of tumor metastasis. Given the patient's financial constraints and the fact that PET-CT scan is not covered by the local health policy (requiring full out-of-pocket payment), the patient declined this procedure. Based on the absence of abnormalities or metastases in the head and neck MRI, bone scintigraphy and contrast-enhanced whole-body CT, along with consensus by the multidisciplinary oncologic pulmonology board that surgical resection was the most appropriate treatment, video-assisted thoracoscopic (VATS) left upper lobectomy plus sleeve resection of the left upper lobe bronchus were performed. Intraoperative frozen section analysis could not definitively diagnose the tumour, but gross examination of the resected mass strongly suggested malignancy based on the mass's firm/rubbery consistency, infiltrative borders (poorly circumscribed) and a variegated (heterogeneous) cut surface showing mixtures of grey-white tumour tissue and dense, glistening white sclerotic areas. Consequently, lymph node dissection was performed at the carinal, hilar, interlobar and Group 12 lymph nodes based on gross examination suggested malignancy and the Chinese Society of Clinical Oncology Non-small Cell Lung Cancer guidelines from 2024(25). Macroscopically, the size of the resected upper lobe of the left lung was 11.5x8.5x4 cm and a mass measuring 1.2x1x0.8 cm was observed 0.6 cm from the bronchial stump (Fig. 3A). The cut surface was grey-white, solid, firm and nodular, with clear demarcation from the surrounding tissue, and the remaining lung tissue was grey-red and soft.

The surgically resected samples were routinely fixed in 10% neutral-buffered formalin, dehydrated, embedded in paraffin and sectioned into 4-µm slices. Routine haematoxylin and eosin staining was performed and the sections were imaged using a CX31 light microscope (Olympus Corp.). Histologically, the tumour exhibited a biphasic pattern of inner ductal and outer myoepithelial cells (Fig. 3B). At low-power magnification, the tumour appeared well-defined but unencapsulated, displaying a multinodular morphology with irregular borders. Fibrosis bands were observed dissecting the tumour. The inner layer consisted of a single row of cuboidal to columnar epithelial cells and was characterized by dense to finely granular cytoplasm surrounded by clear myoepithelial cells. The myoepithelial cells were polygonal or spindle-shaped, with abundant, clear and glycogen-rich cytoplasm. Myoepithelial cells forming solid areas were focally observed and areas of squamous differentiation were also present. These solid patterns blended with more typical tubular patterns, making distinction between them challenging. The luminal spaces contained eosinophilic material. Minimal nuclear pleomorphism and a low mitotic rate were observed. No metastatic cancer was found in the lymph nodes (0/11) and the bronchial incisive margin was negative.

Immunohistochemistry (IHC) was performed on the aforementioned surgical sections using the streptavidin-peroxidase method. The sections were heated at 65˚C for 120 min and then dewaxed by incubation with xylene for 3 min (six times). The sections were rehydrated using a graded alcohol series and antigen retrieval was performed using EDTA repair solution (EDTA·2Na·2H₂O, pH 9.0) with samples heated in a pressure cooker to the boil and then kept warm for 20 min. The sections were incubated with 3% H₂O₂ for 10 min at room temperature and then rinsed with PBS for 3 min to block endogenous peroxidase activity. Sections were incubated with primary antibodies against smooth muscle actin (SMA; cat. no. Kit-0006; Fuzhou Maixin Biotech Co., Ltd.), SOX-10 (cat. no. RMA-0726; Fuzhou Maixin Biotech Co., Ltd.), S-100 (cat. no. Kit-0007; Fuzhou Maixin Biotech Co., Ltd.), CKp (cat. no. M3515; Dako; Agilent Technologies, Inc.), Calponin (cat. no. MAB-0712; Fuzhou Maixin Biotech Co., Ltd.), P40 (cat. no. RMA-1006; Fuzhou Maixin Biotech Co., Ltd.), P63 (cat. no. MAB-0694; Fuzhou Maixin Biotech Co., Ltd.), epithelial membrane antigen (EMA; cat. no. KIT-0011; Fuzhou Maixin Biotech Co., Ltd.), CD117 (cat. no. KIT-0029; Fuzhou Maixin Biotech Co., Ltd.), thyroid transcription factor (TTF)-1 (cat. no. MAB-0599; Fuzhou Maixin Biotech Co., Ltd.), CK7 (cat. no. MAB-0828; Fuzhou Maixin Biotech Co., Ltd.) and Ki-67 (cat. no. MAB-0542; Fuzhou Maixin Biotech Co., Ltd.) for 1 h at room temperature. The sections were then incubated with the EnVision detection system peroxidase/diaminobenzidine (DAB) and rabbit/mouse secondary antibodies (cat. no. K5007; Dako; Agilent Technologies, Inc.) for 30 min at room temperature. DAB from the aforementioned secondary staining kit was added for detection. The sections were counterstained with haematoxylin for 2 min and imaged using a CX31 light microscope (Olympus Corp.). Microscopic observation revealed the following immunohistochemical profile: CKp (+), calponin (myoepithelium weak +), S-100 (myoepithelium +), SMA (-), SOX-10 (myoepithelium +), P40 (myoepithelium +), P63 (myoepithelium +), EMA (ductal cell +), CK7 (ductal cell +), CD117 (ductal cell weak +), vimentin (myoepithelium +), TTF-1 (weak +) and Ki-67 (15% +). Only critical markers for diagnosis were preserved as images-specifically CK7, S-100, SOX10, VIM, P40 and TTF-1 (Fig. 4) and images of the remaining markers are missing.

Based on the above radiological, pathological and immunostaining findings, the patient was diagnosed with P-EMC. The patient did not receive any adjuvant chemotherapy and the patient was discharged 5 days after the surgery with no complications. The patient was referred to the pulmonary oncology clinic for postoperative follow-up and underwent contrast-enhanced CT every 3 months within the first year after surgery, and every 6 months from the second year. The patient was followed up for 5 years, until May 2025. During this period, the oncologist and thoracic surgeon were responsible for performing the follow-up. A recent contrast-enhanced chest CT scan (December 2024) revealed no evidence of recurrence (Fig. 1B) and the patient remained in good health. A close follow-up of every 1 to 2 years is ongoing.

A comprehensive search of the PubMed (https://pubmed.ncbi.nlm.nih.gov/), Google Scholar (https://scholar.nq69.top/) and Web of Science databases (https://www.webofscience.com/) for primary P-EMC was conducted to identify relevant studies published before January 1st, 2025, using the following terms: ‘primary’ AND (‘tracheobronchial’ OR ‘bronchial’ OR ‘endobronchial’ OR ‘trachea’ OR ‘bronchus’ OR ‘lung’ OR ‘pulmonary’) AND (‘epithelial-myoepithelial tumour’ OR ‘epithelial-myoepithelial carcinoma’). Additionally, references from retrieved articles meeting the inclusion criteria were manually searched. Exclusion criteria comprised studies not written in English, metastatic tumours or reports with insufficient data. The extracted data included patient characteristics (geographic region, age, sex, symptoms, smoking history, comorbidities, diagnostic modalities, chest imaging, bronchoscopy, PET-CT, treatment, tumour size, adjuvant chemotherapy, follow-up and survival). Since 2000, 50 global cases have been reported across 39 publications. After applying exclusions-three cases for insufficient data, two non-English reports and one metastatic tumour -44 unique cases met the criteria for analysis. A total of 44 published cases of P-EMC were reviewed with relatively complete clinical data reported from 2000 to 2025 (4,6,7,9,11-18,20,26-43), with the following geographic distribution: The US (n=17), Japan (n=6), the UK (n=6), China (n=4), South Korea (n=3), Australia (n=1), Germany (n=1), Greece (n=1), India (n=1), Italy (n=1), Portugal (n=1), Spain (n=1) and Turkey (n=1). The patients included 20 males and 24 females aged 7-83 years (mean, 55.8 years). A total of 12 patients had a history of smoking, 22 were non-smokers and 10 had an undocumented smoking status. The tumour size ranged from 0.71 to 6 cm (mean, 2.82 cm), with 27 patients presenting with endobronchial masses. Among the 10 patients who underwent PET-CT, only 4 demonstrated normal ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) uptake. Treatment modalities included surgical resection in 37 patients (lobectomy, n=20; pneumonectomy, n=7; wedge resection, n=6; segmentectomy, n=1; atypical resection, n=1; and unspecified resection, n=2); bronchoscopic interventions in 4 patients (electrocautery snare, n=2; curative electrosurgery, n=1; and laser ablation, n=1); and radiation with chemotherapy in 1 patient with nodal metastasis; while 2 patients received undocumented treatment. Adjuvant chemotherapy was administered to only 1 postoperative patient. Among the 44 patients, 32 had follow-up data ranging from 2.5 to 96 months, while 12 had no documented follow-up. Outcome analysis revealed no recurrence or metastasis in 30 patients, death in 2, intrapulmonary metastasis in 1, loss to follow-up in 1 and unknown outcome in 10. Details of the studies retrieved from the above-mentioned literature review are provided in Table SI.

The present case contributed to the existing literature by reporting on the diagnostic challenges and pitfalls associated with small biopsy specimens of this rare tumor and demonstrating favorable outcomes with no recurrence or metastasis during 5 years of long-term follow-up. The study underscores the diagnostic pitfalls in clinical practice and provides evidence that surgical resection can yield favorable clinical outcomes for this rare tumor.

The most common salivary gland tumours in the head and neck are benign, including pleomorphic adenoma, Warthin tumours and basal cell adenoma, whereas the most common malignant tumours are mucoepidermoid carcinoma (MEC) and adenoid cystic carcinoma (ACC). However, salivary gland tumours located in the lungs are mostly malignant, and the common types include ACC, MEC, EMC and acinar cell carcinoma (1). EMC is a rare salivary gland malignancy first described by Donath et al (44) in 1972, accounting for ~1% of salivary gland tumours; EMC in the pulmonary system is even rarer, with only 50 cases reported globally between 2000 and 2025.

The literature review and current case indicate that primary P-EMC is more prevalent in middle-aged and elderly individuals, with women showing greater morbidity than men. The maximum tumour diameter ranged from 0.71 to 6.0 cm (28,30) and the maximum diameter of the resected mass from the patient of the present study was 1.2 cm. Among the 44 reported cases of primary P-EMC, 12 had a history of smoking, 22 had no smoking history and 10 were undocumented. Therefore, it may be speculated that the occurrence of primary P-EMC is not associated with smoking, consistent with the conclusion of Goodwin et al (45). According to Table SI, radiologically, most P-EMCs are characterized as endobronchial masses, or bronchoscopy reveals visible neoplasms or polyps in patients who usually present with airway obstruction symptoms. However, patients with EMC located in the lung parenchyma may have no obvious symptoms in the early stages, and EMC is often detected during chest radiological examinations. The diagnostic value of ¹⁸F-FDG-PET is unclear; only for 4 cases, elevated glucose metabolism in EMC lesions was reported (26,32,33,43); by contrast, the majority of reports indicated normal ¹⁸F-FDG uptake (12,17,28,29,31,34). The diagnosis of EMC depends on histopathology, and a duct-like structure comprising epithelial and myoepithelial cells is the typical pathological feature. These duct-like structures can be categorised into two types: One with a large area with a typical double-layer tubular structure-epithelial cells forming the inner layer and myoepithelial cells forming the outer layer of the lumen; the other type with only few atypical duct-like structures, consisting of epithelial cells that form a single layer of duct-like structure surrounded by large confluent myoepithelial cells. IHC reveals CK positivity in inner epithelial cells, while the outer myoepithelial cells stain positive for SMA and S-100. These two pathological patterns may coexist in varying proportions within the same tumour (6).

Compared with previously reported cases, the pathological diagnosis presented in the current study poses significant diagnostic challenges. Initial bronchoscopic biopsy findings were suggestive of MEC, whereas definitive surgical specimen pathology confirmed P-EMC. The reasons for the misdiagnosis or diagnostic discrepancy in the present case are as follows: Limitations of the small biopsy sample were evident, as the bronchoscopic biopsy consisted solely of fragmented tissue measuring 0.5x0.3x0.2 cm, precluding observation of the characteristic biphasic tubular structure (inner ductal epithelium plus outer myoepithelial layer) of EMC. The biopsy may have sampled only an area of epidermoid or intermediate-type cells, mimicking MEC morphology and leading to misdiagnosis. For instance, abundant myoepithelial cells could be misinterpreted as myoepithelioma or myoepithelial carcinoma, while predominant epithelial cells broaden the differential to include mucinous cystadenocarcinoma, MEC, mucinous adenoma, acinic cell carcinoma, myoepithelioma, myoepithelial carcinoma, clear cell (‘sugar’) tumour, metastatic EMC, and primary or metastatic clear cell carcinomas (6,7,10,30,46). With respect to tumour heterogeneity, EMC frequently exhibits regional variation in differentiation: Sampling a dedifferentiated myoepithelial zone (lacking myoepithelial markers) or a ductal epithelium-predominant region (expressing CK, P40) in small biopsies readily leads to MEC misdiagnosis, whereas comprehensive sampling of resection specimens reveals the typical biphasic differentiation (6). A critical diagnostic pitfall involves P40, which is expressed in myoepithelial cells (particularly basaloid variants), indicating that a small-biopsy P40(+) result likely represents the myoepithelial layer of EMC rather than the squamous cells of MEC. Nevertheless, this P40(+) finding was interpreted as a ‘squamous differentiation’ marker of MEC. Morphologically, MECs and EMCs share overlapping features, including solid areas, clear cell changes and squamoid cells. The absence of definitive EMC biphasic structures or MEC mucinous cells in limited biopsies creates diagnostic uncertainty. Finally, while MAML2 rearrangement demonstrates high specificity for MEC (particularly low- to intermediate-grade), where positivity supports MEC and negativity favours EMC, this test was omitted due to the patient's financial constraints, necessitating reliance on limited immunohistochemistry for the initial diagnosis.

In general, VATS resection emerged as the most prevalent therapeutic approach, with lobectomy predominating (54.1%, n=20), followed by pneumonectomy (18.9%, n=7), wedge resection (16.2%, n=6) and segmentectomy (2.7%, n=1). Of note, Hagmeyer et al (11) reported good outcomes in a young patient who underwent segmental tracheal resection, with no recurrence or metastasis at the 24-month follow-up. Bronchoscopic interventions were reported in 4 patients (9.1%), including electrocautery snare (n=2), curative electrosurgery (n=1) and laser ablation therapy (n=1). Among these, Chao et al (27) reported successful curative electrosurgery for an endobronchial mass in the distal left main bronchus of a 43-year-old woman, demonstrating disease-free survival at the 6-month follow-up. Regrettably, the remaining three bronchoscopically managed patients were lost to follow-up or lacked outcome documentation. All patients treated bronchoscopically and the patient who underwent segmentectomy had sub-2 cm lesions (≤2 cm), suggesting that bronchoscopic management is a viable option for small, endobronchially confined tumours, whereas segmentectomy has favourable outcomes for similarly sized nonmetastatic lesions. By contrast, lobectomy cases predominantly involve lesions >2 cm (4,6,7,12,14,18,20,26,30,33,39,40,42). Among the 6 patients who underwent pneumonectomy, the maximum diameters of the lesions in 4 patients were 2.7, 4.2, 4.5 and 6 cm, while the sizes in 2 patients were not recorded. Among the 6 patients who underwent wedge resection, the maximum lesion size was 4 cm and the minimum was 0.8 cm. The patient of the present study (tumour dimensions: 1.2x1.0x0.8 cm) underwent VATS left upper lobectomy with bronchial sleeve resection and mediastinal lymph node dissection due to initial diagnostic pitfalls on bronchoscopic biopsy. While excellent outcomes with 5-year recurrence-free survival were achieved, retrospective analysis suggested that definitive preoperative diagnosis of low-grade malignant P-EMC might have permitted parenchymal-sparing approaches, such as segmentectomy or wedge resection, as feasible alternatives.

Of note, the biological characteristics of P-EMC and salivary EMC are have remained to be fully elucidated. However, previous reports have shown that the recurrence or metastasis interval of salivary EMC is relatively long, averaging 5 years (47) and 15 years (48), respectively. The longest reported follow-up period to date is 96 months (4). Therefore, the recurrence and metastasis of P-EMC require further investigation through extended follow-up. Although typical EMC is generally considered a low-grade malignancy with a favourable clinical prognosis, numerous studies have shown that P-EMC with high-grade transformation is aggressive, with a poor prognosis and a high rate of distant metastasis (18,30,49). Tajima et al (18) reported that Ki-67 and cyclin D1 were key molecules associated with the high-grade transformation of EMC, suggesting a poor prognosis and the need for close follow-up. Overall, three studies have reported the use of NGS to detect mutations associated with primary P-EMC. Nakashima et al (31) reported on a 54-year-old female patient with P-EMC in the right middle lobe. NGS detected no mutations in the Kirsten rat sarcoma viral oncogene homolog or epidermal growth factor receptor (EGFR) genes. Sharma et al (17) reported a case of P-EMC in a 38-year-old man presenting with a 3-cm diameter homogeneous mass in the lower lobe of the right lung. Genetic testing for mutations in EGFR, anaplastic lymphoma kinase, ROS proto-oncogene 1 tyrosine kinase, B-Raf proto-oncogene, serine/threonine kinase and mesenchymal-epithelial transition factor was performed, but no mutations were found. Charles et al (30) reported on a 72-year-old female patient with P-EMC with focal high-grade transformation. Mutations in DNMT3A, APC, STAT3 and KDM5C were detected. Despite these detected gene mutations, the patient was still treated with a left pneumonectomy along with mediastinal lymph node dissection. Based on this literature review, no specific gene testing is generally recommended for detecting particular mutations associated with P-EMC, including for guiding treatment strategies. The clinical significance of DNMT3A, APC, STAT3 and KDM5C mutations in the diagnosis, prognosis and treatment of P-EMC requires further investigation to be established.

In conclusion, primary P-EMC is a low-grade malignant tumour that most commonly occurs in the tracheobronchial tree. Histologically, a duct-like structure composed of epithelial and myoepithelial cells provides evidence for a definite diagnosis. Based on the present case and a literature review, for endobronchial EMC lesions confined to the bronchus and measuring <2 cm without metastasis, bronchoscopic intervention or VATS segmentectomy are viable therapeutic alternatives; for nonmetastatic central lesions >2 cm, lobectomy or wedge resection should be considered to avoid the more traumatic pneumonectomy as much as possible.