Abstract
Primary lung tumours are rare in paediatric patients. Mucoepidermoid carcinoma (MEC), typically low-grade and diagnostically straightforward, is the second most common tumour of the bronchus after carcinoid tumours. However, rare MEC may show divergent differentiation, be misdiagnosed as low-grade adenocarcinoma, not otherwise classified, and pose clinical challenges, especially when mastermind-like protein 2 (MAML2) gene arrangement is negative by fluorescence in situ hybridisation (FISH). Here, we report an MAML2 FISH-negative low-grade bronchial tumour in a juvenile patient that demonstrates both mucoepidermoid and acinar differentiation based on morphology and immunophenotype. Next-generation sequencing identified a CREB regulated transcription coactivator 3 (CRTC3::MAML2) fusion gene, located upstream of traditional translocation points and potentially undetectable by currently available FISH probes. This tumour appears to be a novel presentation of a bronchial tumour with dual mucoepidermoid and acinar differentiation, first described as mucoacinar carcinoma—a newly proposed subtype of MEC, originally described in the major salivary gland.
Keywords: Genes, Neoplasm; Lung Neoplasms; Pediatrics
Introduction
Primary lung tumours are uncommon in children. A retrospective study revealed carcinoid tumours and mucoepidermoid carcinoma (MEC) as the predominant malignant entities, comprising 55% and 33% of cases, respectively.1 MEC, often found in salivary glands, can also arise in the submucosal minor salivary glands of the bronchus2 and is usually low-grade when seen in younger patients.3 Molecular genetics and next-generation sequencing (NGS) have discovered CRTC1::MAML2 and CRTC3::MAML2 fusions in the majority of MEC, especially in low grade MEC.4
A recent study identified 11 subtype MEC cases with serous acinar differentiation, labelled as ‘mucoacinar carcinomas (MACs)’, all found in adult salivary glands and positive for mastermind-like protein 2 (MAML2) translocation via fluorescence in situ hybridisation (FISH).5 Among the MACs, two cases tested by NGS harboured a CRTC1::MAML2 fusion. Here, we present a unique case of a right bronchial tumour with both mucoepidermoid and acinar elements and unconventional CRTC3::MAML2 fusion identified by RNA-targeted NGS panel, with negative MAML2 FISH results.
Case report
A juvenile patient initially presented to the emergency department with a history of recurrent pneumonias, poor growth, anaemia, chronic cough, reduced oxygen saturations to SpO2 readings of 91–96%, intermittent fevers every 2 weeks with temperatures ranging from 100 to 101°F and shortness of breath over the past year and a half. A pulmonary examination showed diminished air flow, prompting a chest CT scan that detected a concerning mass in the right middle and lower bronchi, leading to pneumonia with associated bronchiectasis in the right middle and lower lobes (figure 1A). A bronchoscopy provided a clearer view of a friable fungating mass (figure 1B) which completely blocked the bronchus intermedius. It was removed through electrocautery snare. The remaining abnormal endobronchial tissue was ablated by cryotherapy and thermotherapy to open the bronchus intermedius. Significant purulent secretions were noted in the middle lobe and the basilar and superior segments of the lower lobe, which were suctioned after recanalisation.
Figure 1. CT, bronchoscopic and histopathological findings. The chest CT (A) before the surgery showed that a mass (thick arrow) in the right bronchus intermedius, measuring 2.7 cm (black double arrow) × 2.0 cm (white double arrow). Additional mass was seen in the parabronchial tissue (thin arrow) representing potential lymph node metastasis. The ‘drowned lung’ finding (arrowhead) was consistent with consolidation to recurrent pneumonia. Bronchoscopic view from the right mainstem bronchus showed that the intermediate bronchus was completely blocked by the tumour (B, white arrow). The right upper lobe bronchus (arrowhead) was patent. After the endoscopic excision, the intermediate bronchus was no longer obstructed by the tumour (C, white arrow). The CT after the endoscopic (D) tumour excision demonstrated the recanalisation of the right bronchus intermedius, with the remainder of the tumour still visible (thick arrow). The parabronchial tumour (thin arrow) was still evident and the return of ventilation to the right lower lobe was clearly visible (arrowhead). The white arrowhead in B and C points to the right upper lobar bronchus. Medium power view (E) of the excised endobronchial tumour on H&E stain demonstrated cystic (arrows) and solid (arrowheads) areas underneath the bronchial mucosa (stars). Higher power view (F) showed bubbly cytoplasm with eosinophil and clear areas. Mucicarmine stain highlighted extracellular and intracellular mucin in goblet cells (G). Strong positive CK7 (H) staining confirmed the epithelial nature of the tumour. The acinar cell element of this tumour was highlighted by SOX10 (I), NR4A3 (J), LMO2 (K), focal) and DOG1 (L), focal). There were only focal areas positive for CK5/6 (M) and p63 (N). Histological examination of the parabronchial mass (O) submitted as ‘lymph node’ revealed that the tumour invaded the bronchial wall with the cartilage and extended into the parabronchial connective tissue without evidence of lymphoid tissue. Lower power view (P) showed glandular areas of the tumour. (E, 100×; F, 400×; G, 400×; H, 100×; I, J, K, 200×; L, 400×; M, 100×; N, 200×; O, 20×; P, 40×).
The patient’s condition improved significantly 4 days post-bronchoscopy, with enhanced oxygen saturation to an SpO2 of 100% and cessation of fevers. Post-discharge, the patient continued to recover on amoxicillin, with dyspnoea also resolving. The initial electrocautery resection only removed the endobronchial part of the tumour (figure 1C and D) and so the remaining tumour was removed by the resection of the right middle and lower lobes that were non-functional due to recurrent pneumonia.
Sections of the endobronchial tumour removed by bronchoscopy revealed an epithelial neoplasm with glands, solid and cystic areas (figure 1E). The nuclei were round and uniform with evenly distributed chromatin. The cytoplasm was granular and eosinophilic with focal clearing and foamy appearance (figure 1F). No necrosis or significant increase in mitosis was identified. Both intracellular and extracellular mucins were identified with mucicarmine stain (figure 1G). No definitive zymogen granules were identified with Periodic Acid-Schiff (PAS) and PAS with Diastase (PAS-D) stains (not shown). The tumour cells stained positively for cytokeratin AE1/AE3 (not shown), cytokeratin 7 (figure 1H), SOX10 (figure 1I) and NR4A3 (figure 1J), which has been recognised as a specific and sensitive biomarker for acinic cell carcinoma.5,7 Additionally, focal staining was observed for CD56 (not shown), LMO2 (figure 1K), DOG1 (figure 1L), CK5/6 (figure 1M) and p63 (figure 1N). No expression of mammaglobin, S100, TTF-1, HMB45, MART1, Napsin A, chromogranin, synaptophysin, GATA3 and smooth muscle myosin was identified in the tumour cells. The Ki-67 labelling index was approximately 5%.
As immunostains were not typical for any specific salivary gland tumour, FISH was performed with multiple probes. No MAML2 gene rearrangement, typically seen in MEC, was identified with MAML2 break-apart probe. It was also negative for the EWSR1 gene rearrangement characteristic of hyalinising clear cell carcinoma, the ETV6 gene rearrangement typical of secretory carcinoma and the NR4A3 gene rearrangement seen in about 70% of acinic cell carcinoma.6 Additionally, no genetic mutations were detected by a DNA-based NGS panel in a commercial laboratory. Based on overall morphological features and both acinar (positivity for SOX10, NR4A3, LMO2 and DOG1) and mucoepidermoid (positivity for mucin, CK5/6 and p63) differentiation, we considered that this low-grade tumour was a newly identified subtype of MEC, which shows dual mucoepidermoid and acinar differentiation.
The peribronchial tissue was submitted as suspicious ‘level 7 subcarinal lymph node’. Microscopic examination revealed similar tumour morphology to the endobronchial tumour. It involved peribronchial fibroadipose with adjacent cartilage with an expansile nodular growth pattern (figure 1O and P). There were some lymphoid aggregates, but no morphological features of a lymph node were present. These findings suggest that the tumour was localised to the bronchial wall and nearby connective tissue, with no indications of metastasis to the lymph nodes.
No residual tumour was identified in the lobectomy specimen after extensive sampling. There was bronchiectasis with acute and chronic inflammation. The lung parenchyma showed diffuse fibrosis with chronic inflammation, lymphoid aggregates and foamy macrophages, consistent with recurrent obstructive pneumonia.
MAML2 break-apart probe FISH testing at a reference laboratory showed no gene rearrangement. We identified a fusion of CREB regulated transcription coactivator 3 (CRTC3) exon 2 with MAML2 exon 2 on the resected tumour from the peribronchial soft tissue by RNA-based NGS. RNA sequencing was performed using the custom Solid Tumour HotSpot and Fusion OncoPanel (ArcherDx) on the Illumina MiSeq platform. Library preparation used anchored multiplex PCR with molecular barcoding, and data were analysed using the Archer Analysis Pipeline. The tumour predominantly expressed CRTC3 exon 2 spliced with MAML2 exon 2 (1070 reads). The sample had very high-quality RNA, and a few lower frequencies of expressed splice variants were also detected, all under 100 reads. The traditional CTRC3 exon 1 and MAML2 exon 2 transcripts described in the catalogue of somatic mutation in cancer (COSMIC) were detected at a low frequency (74 reads). Interestingly, a CRTC3 exon 1 and MAML2 exon 2 out-of-frame transcript occurring within the protein coding portions of both exons (ie, not at splice sites) was detected with 58 reads. Assimilating the data from these transcripts, the dominant expressed transcript is CRTC3 exon 2 spliced with MAML2 exon 2. Based on detected alternatively spliced transcripts (two with the entire coding region of CRTC3 exon 2, and one with part of the coding region of MAML3 exon 1), the genomic fusion may have occurred between CRTC3 exon 2 (within intron 2) and the interior of MAML2 exon 1 or upstream of exon 1. The precise genomic breakpoint locations are not known for this tumour or described in the literature, but the detected fusion transcripts in this case point to a more upstream MAML2 genomic fusion, which may explain the negative MAML2 FISH findings. Additionally, a pathogenic PIK3CA p.Glu545Ala mutation (13% VAF, COSM12458, chr3:178 936 092 A>C) was also detected, which may be targetable in the setting of a clinical trial.
The patient was followed for 6 months without evidence of recurrence. Subsequent follow-up was unsuccessful despite repeated contact attempts, limiting assessment of long-term outcome.
Discussion
This case has many novel attributes, including the first described case of MAC of the bronchus, the first observed case in a child and the first observation of the CRTC3 fusion partner in the newly described MAC. Of note is the unusual dominant expressed fusion transcript of CRTC3 exon 2, while the fusion of exon 1 is predominantly reported in the literature. It is important to note the negative MAML2 FISH study and positive RNA-based NGS CRTC3::MAML2 along with the sequencing data pointing to a possible fusion location within or 5’ of MAML2 exon 1. This is likely a more 5’ fusion than what is typically found in MEC. While it is unknown if this 5’ MAML2 fusion location is standard within MAC, the observed variety of fusion transcripts in this case is of note. The CRTC3::MAML2 t(15q26;11q21) fusion gene product is likely oncogenic because it brings the CREB-binding domain of CRTC3 next to the MAML2 transcriptional activation domain. In one study, a similar fusion involving the CREB-binding domain of the homologous CRTC1 fused to MAML2 resulted in activation of CREB-dependent transcription and transactivation of AP-1 and MYC.8 CRTC3::MAML2 fusions described in COSMIC involve exon 1 of CRTC3 fused to exon 2 of MAML2 (13 cases, COSF1104), and CRTC1::MAML2 fusions also involve exon 1 of CRTC1 and exon 2 of MAML2 (253 cases, COSF1102).9 In addition to the well-described MAML2 fusions typically seen in MEC, additional variants have been identified in TP53, CDKN2A, CDKN2B, TERT, BAP1, PTEN, HRAS, MTAP and PIK3CA, among others.10 11 PIK3CA alterations, as seen in this case, represent a potential therapeutic target, and have been described in 5.9–16.9% of MECs of the salivary gland.10 11 In one study, the presence of an RAS or PIK3CA alteration correlated with higher MEC clinical stage.11 This is the first documentation of a PIK3CA pathogenic alteration in an MAC of the bronchus.
Typical MEC is comprised of intermediate, mucous and epidermoid cells and may display sclerotic stromal changes and clear or oncocytic cellular morphology. It is usually strongly positive for cytokeratin 7, p63 and cytokeratin 5/6. With characteristic morphology and confirmatory immunoprofile, there is usually no diagnostic challenge or need for molecular testing. However, this case was only focally positive for p63 and cytokeratin 5/6. Therefore, additional immunostains were performed and revealed that the tumour cells expressed acinar cell markers: NR4A3, LMO2, SOX10 and DOG1.12 SOX10 is usually positive in acinic cell carcinoma but often negative in MEC. However, a small group of MEC with polygonal epithelial cells, pale-to-eosinophilic cytoplasm and colloid-like dense eosinophilic material is positive for SOX10.13 It is not clear whether this group of MEC shows acinic differentiation. FISH was negative for any specific gene rearrangement in this case. Approximately 20–25% of MEC lacks the MAML2 gene arrangement by FISH. NR4A3 gene arrangement is detected in about 70% of acinic cell carcinoma. NR4A3 immunostain is positive in over 90% of acinic cell carcinoma and has been shown to be a more specific and sensitive marker than NR4A3 FISH for acinic cell carcinoma.14 This diagnosis was further corroborated by the CRTC3::MAML2 fusion identified with our NGS data and positivity for NR4A3 by immunostaining. If NR4A3 rather than MAML2 gene arrangement is detected in this tumour, squamoglandular subtype of acinar cell carcinoma should be considered. This novel subtype, as reported by Dr Shah and Seethala,15 demonstrated predominant basaloid squamoid proliferation with bland morphology, scattered foci of serous acinar differentiation, rare mucous cells and tubules. The tumour was diffusely positive for NR4A3, cytokeratin and DOG1, while the squamoid components were positive for p40. NR4A3 rather than MAML2 and MSANTD3 gene rearrangements were detected by fluorescence in situ hybridisation.
In summary, this case lacked MAML2 and NR4A3 gene arrangements by FISH, which are typical of MEC and acinic cell carcinoma, respectively. However, the morphological features and immunostaining patterns did demonstrate acinic and mucoepidermoid differentiation. More importantly, the CRTC3::MAML2 fusion was identified with our NGS data. Therefore, this low-grade tumour is best classified as a salivary gland type carcinoma with acinic and mucoepidermoid differentiation or MAC, a newly proposed subtype of MEC which demonstrates dual mucoepidermoid and acinic differentiation.
Footnotes
Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Handling editor: Munita Bal.
Patient consent for publication: Not applicable.
Provenance and peer review: Not commissioned; externally peer reviewed.
References
- 1.Roby BB, Drehner D, Sidman JD. Pediatric tracheal and endobronchial tumors: an institutional experience. Arch Otolaryngol Head Neck Surg. 2011;137:925–9. doi: 10.1001/archoto.2011.153. [DOI] [PubMed] [Google Scholar]
- 2.Markovitz M, Monforte H, Shieh HF, et al. Endobronchial mucoepidermoid carcinoma in a pediatric patient: A case report. Radiol Case Rep. 2022;17:2034–7. doi: 10.1016/j.radcr.2022.03.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Shen C, Che G. Clinicopathological analysis of pulmonary mucoepidermoid carcinoma. World J Surg Oncol. 2014;12:33. doi: 10.1186/1477-7819-12-33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fehr A, Werenicz S, Trocchi P, et al. Mucoepidermoid carcinoma of the salivary glands revisited with special reference to histologic grading and CRTC1/3-MAML2 genotyping. Virchows Arch. 2021;479:975–85. doi: 10.1007/s00428-021-03146-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bundele M, Weinreb I, Xu B, et al. Mucoacinar Carcinoma: A Rare Variant of Mucoepidermoid Carcinoma. Am J Surg Pathol. 2021;45:1028–37. doi: 10.1097/PAS.0000000000001752. [DOI] [PubMed] [Google Scholar]
- 6.Shi Q, Zhang B, Bsirini C, et al. NR4A3 fluorescence in situ hybridization analysis in cytologic and surgical specimens of acinic cell carcinoma. Hum Pathol. 2022;127:86–91. doi: 10.1016/j.humpath.2022.06.007. [DOI] [PubMed] [Google Scholar]
- 7.Wong KS, Mariño-Enriquez A, Hornick JL, et al. NR4A3 Immunohistochemistry Reliably Discriminates Acinic Cell Carcinoma from Mimics. Head Neck Pathol. 2021;15:425–32. doi: 10.1007/s12105-020-01213-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chen Z, Ni W, Li J-L, et al. The CRTC1-MAML2 fusion is the major oncogenic driver in mucoepidermoid carcinoma. JCI Insight. 2021;6:e139497. doi: 10.1172/jci.insight.139497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.COSMIC gene fusion database. Accessed 2024. https://cancer.sanger.ac.uk/cosmic/fusion Available.
- 10.Bou Zerdan M, Kumar PA, Zaccarini D, et al. Molecular Targets in Salivary Gland Cancers: A Comprehensive Genomic Analysis of 118 Mucoepidermoid Carcinoma Tumors. Biomedicines. 2023;11:519. doi: 10.3390/biomedicines11020519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Morita M, Murase T, Okumura Y, et al. Clinicopathological significance of EGFR pathway gene mutations and CRTC1/3-MAML2 fusions in salivary gland mucoepidermoid carcinoma. Histopathology. 2020;76:1013–22. doi: 10.1111/his.14100. [DOI] [PubMed] [Google Scholar]
- 12.Haller F, Skálová A, Ihrler S, et al. Nuclear NR4A3 Immunostaining Is a Specific and Sensitive Novel Marker for Acinic Cell Carcinoma of the Salivary Glands. Am J Surg Pathol. 2019;43:1264–72. doi: 10.1097/PAS.0000000000001279. [DOI] [PubMed] [Google Scholar]
- 13.Hsieh MS, Lee YH, Chang YL. SOX10-positive salivary gland tumors: a growing list, including mammary analogue secretory carcinoma of the salivary gland, sialoblastoma, low-grade salivary duct carcinoma, basal cell adenoma/adenocarcinoma, and a subgroup of mucoepidermoid carcinoma. Hum Pathol. 2016;56:134–42. doi: 10.1016/j.humpath.2016.05.021. [DOI] [PubMed] [Google Scholar]
- 14.Griffith CC, Pai RK, Schneider F, et al. Salivary gland tumor fine-needle aspiration cytology: a proposal for a risk stratification classification. Am J Clin Pathol. 2015;143:839–53. doi: 10.1309/AJCP-2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Shah AA, Seethala RR. Squamoglandular Variant of Acinic Cell Carcinoma: A Case Report of a Novel Variant. Head Neck Pathol. 2022;16:870–5. doi: 10.1007/s12105-021-01399-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

