Reticular Myxoid Odontogenic Neoplasm with Novel STRN::ALK Fusion: Report of 2 Cases in 3-Year-Old Males

Abstract

Odontogenic tumors represent a collection of entities ranging from hamartomas to destructive benign and malignant neoplasms. Occasionally, pathologists encounter gnathic lesions which clearly exhibit an odontogenic origin but do not fit within the confines of established diagnoses. Here, we describe two such odontogenic tumors, both affecting 3-year-old males. Each case presented as a destructive, radiolucent mandibular lesion composed of mesenchymal cells, some with unique multi-lobed nuclei, frequently arranged in a reticular pattern and supported by a myxoid stroma with focal laminations. Production of odontogenic hard tissues was also seen. Because of their unique microscopic features, both cases were investigated by next-generation sequencing and found to harbor the same STRN::ALK oncogene fusion. To our knowledge, these cases represent the first report of an odontogenic tumor with a STRN::ALK gene rearrangement. We propose the possibility that this neoplasm could be separate from other known odontogenic tumors. Both patients were treated with surgical resection and reconstruction. The prognosis of patients with this entity is currently uncertain but shall become more apparent over time as more cases are identified and followed.

Introduction

Odontogenic tumors are a diverse group of lesions exhibiting a broad spectrum of appearances related to the same complex interactions that occur between epithelial and mesenchymal tissues during tooth development [1]. Several distinct entities have been defined, and reclassifications occur as new data emerges. The 5th Edition (2022) World Health Organization Classification of Head and Neck tumors outlines Essential and Desirable Diagnostic Features for specific entities [2], however some lesions may exhibit overlapping or inconsistent features that create a diagnostic challenge. Molecular diagnostics, including fluorescent in situ hybridization (FISH) and next-generation sequencing (NGS), have recently emerged as tools for refining the diagnosis of odontogenic tumors in addition to identifying targeted therapies as possible treatment options. For example, genetic alterations involving the mitogen-activated protein kinases/extracellular signal-regulated kinases (MAPK/ERK) signaling pathway have been detected in both benign and malignant odontogenic tumors [3]. Among the MAPK/ERK mutations found in odontogenic tumors, BRAF V600E is the most prevalent [3]. Here, we report two cases of an aggressive odontogenic tumor, both of which involved the mandible in 3-year-old male patients. The cases exhibited strikingly similar histopathologic features which are not compatible with those of other known odontogenic tumors.

Case Reports

Case 1

A 3-year-old white male presented to the oral and maxillofacial surgery clinic after evaluation by pediatric dentistry for a swelling in the anterior mandible and a radiographic alteration. His past medical history was significant for a congenital hernia repair at 6 months of age with no complications, and he took no medications. A panoramic radiograph and CBCT showed a multilocular radiolucency extending from teeth K through R, and displacing developing teeth (Fig. 1a). Buccal and lingual expansion of the mandible was appreciated (Fig. 1b).

Fig. 1.

a Panoramic radiograph showing a multilocular radiolucent lesion involving the anterior and left mandibular body. b Axial CT view showing buccal cortical expansion and erosion

During surgery, an incisional biopsy showed white gelatinous tissue in the bony cavity. Microscopic examination revealed a mesenchymal neoplasm comprised of small, primitive-appearing, ovoid, round, and stellate-shaped cells with hyperchromatic nuclei and focal pseudonuclear inclusions (Fig. 2a). Many of the cells contained multiple nuclei or multi-lobed nuclei with 2–5 lobes joined by slender connections (Fig. 2b). Mitotic figures were not appreciated, although irregularity in cell size and shape was noted. The cells were set within a basophilic myxoid and reticular stroma interspersed with thin bands and wispy strands of collagen (Fig. 2c and d). The myxoid stroma vaguely imparted a rippled or laminated effect, alternating between waves of deeply basophilic stroma and lightly basophilic stroma (Fig. 2e and f). Foci of disorderly tooth formation, comprised of enamel and dentinoid material, were appreciated. The dentinoid material had a lobular periphery in some areas with entrapped enamel, ameloblasts, and stellate reticulum (Fig. 2g), and was directly abutted by a moderately-cellular zone of ovoid/round cells (Fig. 2h). The tumor was signed out as “myxomatous mesenchymal tumor with complex odontoma.”

Fig. 2.

a (200 × original magnification) Tumor exhibiting spindled/stellate-shaped, discohesive cells and scattered vascular channels. b (400 × original magnification) Rounded tumor cells with abundant eosinophilic cytoplasm and hyperchromatic, multi-lobed nuclei or multinucleation. Some nuclei demonstrate central vacuoles. c (20 × original magnification) Areas of relatively cellular stroma alternate with zones of paucicellular myxoid stroma. d (200 × original magnification) Scattered ropey collagen within the tumor stroma. e (20 × original magnification) The majority of the tumor is comprised of a proliferation of loose, basophilic, myxoid mesenchymal stroma exhibiting hyperchromatic ripples. Focal formation of dentinoid is noted at the periphery. f (200 × original magnification) Rippled basophilic stroma with variable staining intensity containing small clusters of rounded cells, some of which are bi- or multinucleated. g (20 × original magnification) Although not the majority of the tumor, foci of aberrant tooth formation are noted. h (200 × original magnification) Hard-soft tissue interface shows (from left to right) proliferative basophilic myxoid stroma with primitive stellate-shaped cells, eosinophilic dentinoid material, ameloblasts, and stellate-reticulum

Given the clinical aggressiveness, a segmental resection was undertaken. The defect was reconstructed with an allogenic bone graft and resorbable plates. The final resection exhibited similar histomorphologic features, and negative margins were achieved.

Case 2

A 3-year-old black male presented with asymptomatic enlargement of the left mandible. His medical history was significant only for autism; he took no daily medications and had no prior surgeries. Antibiotics were prescribed and there was no improvement after two weeks. The patient was then referred to an oral and maxillofacial surgeon. Imaging revealed a large multilocular radiolucent lesion in the left mandible associated with four displaced developing permanent teeth (Fig. 3a). The lesion extended from the midline to just anterior to the developing permanent first molar and exhibited buccal and lingual cortical expansion and erosion (Fig. 3a and b).

Fig. 3.

a Panoramic image showing a radiolucent left mandibular lesion containing multiple developing permanent teeth within. b Coronal CT image demonstrates buccal and lingual cortical expansion and erosion

Incisional biopsy of the lesion revealed a neoplastic proliferation of cells exhibiting a variable appearance with alternating hypercellular and stroma-rich zones (Fig. 4a and b). The tumor cells were set within a myxoid stroma and exhibited either a sheet-like arrangement or were organized in a distinct, reticular pattern (Fig. 4c). Individual cells ranged in appearance from epithelioid/plasmacytoid to mesenchymal with indistinct boundaries (Fig. 4c and d). Scattered multinucleated and apparent syncytial cells were noted and frequently had eccentric nuclei (Fig. 4e and f). Nuclear pleomorphism was observed however no mitotic figures were appreciated; Ki67 staining revealed a proliferation rate of approximately 1%. The mesenchymal cells were negative for immunohistochemical staining with pancytokeratin, S100, desmin, and CD138. Deposits of amorphous and acellular eosinophilic material were identified (Fig. 4g) and stained negatively with Congo Red. Products of odontogenesis were identified including focal dental papilla-like tissue (Fig. 4h). Within the more stroma-rich areas of the tumor were rosette-like deposits of eosinophilic hard tissue compatible with dentinoid (Fig. 4i). The dentinoid had a globular periphery and contained entrapped strands and adjacent islands of odontogenic epithelium, some of which were characterized by columnar, ameloblastic peripheral cells (Fig. 4j). Production of basophilic hard tissue was also identified and most closely resembled dystrophic calcifications (Fig. 4k and l). The case was diagnosed descriptively as “myxoid odontogenic neoplasm with atypia and complex odontoma.” A comment was made that the neoplasm was interpreted as most likely benign but with aggressive biologic potential.

Fig. 4.

a (20 × original magnification) The variable appearance of the tumor is evident from low power. A central stroma-rich area is lined by zones of prominent hypercellularity exhibiting reticular groupings of cells. b (40 × original magnification) A hypercellular, sheet-like zone of tumor cells (left) transitions to a more paucicellular, stroma-rich zone with production of eosinophilic amorphous material (right). c (100 × original magnification) A myxoid stroma supports tumor cells with a variable appearance and arrangement. Spindled, discohesive cells (left) abut plasmacytoid cells grouped in clumps and reticular strands (right). d (200 × original magnification) An area of the tumor exhibiting a sheet-like arrangement of tumor cells with indistinct boundaries and hyperchromatic round to ovoid nuclei which are relatively uniform in size. e (100 × original magnification) Abundant wispy amphiphilic stroma containing scattered round tumor cells, some of which are multinucleated. f (400 × original magnification) High power view of tumor cells exhibiting a rounded appearance, eccentric and vesicular nuclei, and binucleated cells. g (200 × original magnification) Deposits of focally wavy eosinophilic material within the tumor were negative with Congo Red staining and interpreted as immature collagen. h (20 × original magnification) Production of dental papilla-like tissue (bottom) adjacent to a hypercellular zone of tumor (top). i (20 × original magnification) Production of dentinoid in a rosette-like pattern within loose, myxoid tumor stroma. j (100 × original magnification) Strands of odontogenic epithelium entrapped within eosinophilic dentinoid exhibiting a globular peripheral morphology. k (40 × original magnification) Production of basophilic hard material resembling dystrophic calcifications (right) in close proximity to areas of dentinoid production (left). l (200 × original magnification) Basophilic calcifications ranging from irregular fragments to particulate granules exhibiting a vague “chicken wire” pattern

Given the destructive clinical behavior of the lesion and reports of successful management of large ameloblastomas with BRAF inhibitors, BRAF mutation analysis was requested by the treating clinician after shared decision-making with the patient’s family. The specimen was forwarded to an outside laboratory for BRAF V600 mutation analysis by bi-directional Sanger sequencing with a reported lower limit of detection of 1%; the mutation was not detected. The patient received a segmental mandibular resection via a transcervical approach. A mandibular reconstruction plate was utilized to bridge the defect until final histopathology results were available. The resection specimen exhibited microscopic features similar to the incisional biopsy and the anterior surgical margin was positive. A second surgery was performed; an additional 1 cm margin was resected, a new mandibular reconstruction plate was applied, and the defect was reconstructed with bone morphogenic proteins and allogenic bone graft. Given the age of the patient, future treatment will include explantation of the mandibular reconstruction plate to prevent growth restriction.

Next-Generation Sequencing

Given that the microscopic features of cases 1 and 2 were similar to one another but inconsistent with those of the other known odontogenic tumors, both were submitted for next-generation sequencing.

Targeted RNA sequencing was performed in the University of Texas Southwestern NGS Clinical Lab at the Once Upon a Time Human Genomics Center. Briefly, whole-slide tissue sections were cut at 10 µm, and Qiagen AllPrep kits (Qiagen) were used for DNA and RNA isolation. DNA and RNA sequencing libraries were generated using IDT xGen (Integrated DNA Technologies) baits targeting 1505 genes. Sequencing was performed on the NextSeq. 550 (Illumina) with a minimum of 6 million mapped reads for RNA and minimum 100 × coverage for DNA. DNA and RNA analysis was done as described previously [4]. Germline results were filtered using gnomAD and all variants were manually reviewed via the Integrated Genomics Viewer (Broad Institute).

Both, cases 1 and 2, harbored the same STRN::ALK oncogenic fusion. DNA analysis only found variants of uncertain clinical and biological significance. We did not perform paired germline DNA testing, and therefore, it is likely that most or all observed single-nucleotide variants are rare germline polymorphisms. STRN::ALK fusion transcript was observed in both DNA and RNA libraries. The fusion transcript linked STRN exon 3 and ALK exon 20.

To our knowledge, this is the first report of a STRN::ALK fusion being identified in an odontogenic cyst or tumor. Subsequent immunohistochemical staining with ALK was performed. Case 1 showed focal, weak cytoplasmic reactivity with clone ALK01 (Ventana) and was interpreted as equivocal. Case 2 was negative utilizing clone ALK1 (Dako). Both cases were negative upon immunohistochemical staining for ALK with the D5F3 clone.

Discussion

Recent advances in molecular genetics are providing an increased understanding of the etiology and pathogenesis of certain odontogenic tumors. Benign neoplasms are now known to have oncogenic mutations once believed to be unique to cancers. Characterization of the molecular processes involved in tumor pathogenesis is important for drug discovery and accordingly, an increasing number of targeted therapies are being developed.

Gene mutations within the MAPK/ERK-signaling pathway have been discovered in both epithelial and mixed odontogenic tumors [3]. The genetic landscape of ameloblastoma has been extensively studied, with BRAF V600E being the most commonly identified mutation. A recent systematic review and meta-analysis showed BRAF V600E mutation is present in 70% of ameloblastomas. [5]. Various additional somatic mutations such as SMO (10.6%), RAS, FGFR2, and others have also been identified in ameloblastoma but at much lower frequency than BRAF [6]. In addition to ameloblastoma, BRAF mutation has been found in ameloblastic fibroma and ameloblastic fibrosarcoma [1]. Laser capture microdissection has found the BRAF mutation predominantly in the mesenchymal portions of mixed odontogenic tumors [1]. Knowing the mutation status of aggressive tumors is useful as BRAF-targeted therapy could serve as a neoadjuvant treatment to reduce tumor size and surgical morbidity [2].

The RAS oncogene family is composed of KRAS, NRAS, and HRAS. RAS mutations, which activate the MAPK/ERK pathway, are found in 30% of human malignancies with KRAS being the most often mutated oncogene [3]. KRAS codon 12 mutations have recently been discovered as a marker of adenomatoid odontogenic tumors (AOTs) [3]. A recent systematic review by Marin et al. found that 75% of AOTs have KRAS mutation, with all studies reporting missense mutations affecting codon 12 [6]. KRAS codon 12 mutations have also been reported in 25–30% of non-small cell lung cancers and nearly all pancreatic ductal adenocarcinomas [7, 8]. Development of effective therapies targeting KRAS has been elusive however this is a high priority for cancer research. It is notable that AOTs, which are clinically indolent, harbor the same KRAS mutation as certain malignancies having extremely poor prognoses.

Chromosome rearrangements involving anaplastic lymphoma kinase (ALK) were first reported in anaplastic large cell lymphoma and have since been found in multiple tumor types [9]. The ALK gene has many fusion partners, all of which serve as oncogenic drivers. However, almost all ALK fusions share a breakpoint at exon 20, just upstream of the tyrosine kinase domain [10]. STRN::ALK gene fusions have been identified in a small percentage of thyroid, lung, renal, colorectal, prostate and breast cancers, as well as in mesotheliomas and pancreatic intraductal tubulopapillary neoplasms [11–23]. Two recent reports also show STRN::ALK fusion is a driver in some cases of salivary intraductal carcinoma; one arose within an intraparotid lymph node and the other was a frankly invasive carcinoma ex-intraductal carcinoma [24, 25]. STRN is the most common ALK fusion partner in thyroid cancers overall, particularly in papillary thyroid carcinomas (PTC) and poorly differentiated thyroid carcinomas [14, 26]. Among the pediatric population, STRN::ALK fusion has been identified in PTC and the follicular variant of PTC [14, 21, 22, 27]. To our knowledge, the two mandibular odontogenic lesions described here represent the only other STRN::ALK fusion positive neoplasms reported in the pediatric population.

ALK fusions activate the MAPK pathway and have led to the development of targeted therapy in the form of ALK tyrosine kinase inhibitors (TKIs) such as crizotinib, alectinib, and others [12, 28]. Proper detection of various ALK fusions and rearrangements is therefore crucial to treatment decisions. Diagnostic methods include NGS, FISH, real-time PCR, and immunohistochemistry (IHC). Updated guidelines for the selection of lung cancer patients for treatment with targeted TKIs suggest IHC as an equivalent alternative to FISH for ALK testing [29]. However, there is variation in staining outcomes among the available ALK monoclonal antibodies. In the U.S., only clone D5F3 is FDA-approved as an IHC assay for selection of lung cancer patients to receive ALK TKIs. Clone ALK1 (used initially for both cases reported here) is less sensitive and not recommended for this purpose [29, 30]. The equivocal (case 1) and negative (case 2) results obtained by IHC staining with the ALK1 clone were attributed to the antibodies lacking optimal sensitivity. However, on subsequent staining with the D5F3 clone, both cases were negative, suggesting that ALK IHC is an unreliable diagnostic tool for this lesion.

Although our cases have overlapping features with other odontogenic tumors, such as primordial odontogenic tumor, ameloblastic fibro-odontoma, and odontogenic myxoma, they have a distinct cellular morphology and reticular arrangement not observed in the aforementioned entities. With limited information on the genetic characterization of these microscopically similar odontogenic tumors, we suggest the possibility that this neoplasm is a new entity but acknowledge that it could represent an unusual variant of a known odontogenic tumor.

Primordial odontogenic tumor (POT), a rare neoplasm originally described in 2014, shows some similarity to our cases. POT is composed of mesenchymal, dental papilla-like tissue surrounded at the periphery by cuboidal to columnar epithelium. There are some reports of focal calcifications in POT however formation of dental hard tissues has not been described [31]. Our case 2 had only focal dental papilla-like tissue (Fig. 4h) and none was observed in case 1. Additionally, the odontogenic epithelium in our cases did not line the peripheral tumor stroma as in POTs, but rather was arranged into focal islands or strands near the dental hard tissues only. POTs have been analyzed by NGS for numerous cancer- and odontogenesis-associated genes (including ALK), and no mutations have been detected [31, 32].

Ameloblastic fibro-odontoma (AFO) has overlapping features with our two cases, both clinically and microscopically, however there are important differences. To our knowledge, a reticular arrangement of mesenchymal cells has not been reported previously in either ameloblastic fibroma (AF) or AFO. A recent 20-year review and assessment of published AF and AFO cases for unusual microscopic features did not reveal any reports of this finding (histopathologic features included ghost cells, pigmentation, and association with calcifying odontogenic cyst) [33]. Additionally, the odontogenic epithelium in both of our cases was only present closely surrounding or entrapped within dentinoid or enamel; there were no proliferative odontogenic epithelial islands within the stroma, which is a characteristic finding in both AF and AFO. Lastly, a comprehensive study of AFOs by Buchner et al. revealed 80% of AFOs occurred in the posterior jaws and 90% were unilocular, whereas both of our cases were multilocular lesions involving the anterior mandible [34]. In 2017, the WHO removed AFO from its classification of odontogenic tumors, arguing that once dental hard tissues are produced, the lesion likely represents a developing odontoma [35]. Significant debate over this decision exists among head and neck specialists, as there are well-documented examples of large and disfiguring AFOs, whereas odontoma is considered to be hamartomatous and clinically indolent [34]. Furthermore, some cases of AFO have BRAF V600E mutation and are therefore genetically distinct from odontoma, which does not have BRAF V600E mutation [1]. While it is generally accepted that immature odontomas and AFO have the same histopathologic features, there is currently no unequivocal way to distinguish the two [34]. In a recent comparison of AFOs and odontomas, Soluk-Tekkesin and Vered suggested a combination of age and lesion size to differentiate the two entities [36]. Their statistical results showed that a patient age of less than 13.5 years and lesion size greater than 2 cm were likely to represent true neoplasia (AFO) rather than odontoma [36]. It remains to be seen whether future molecular studies will be able to solve this conundrum.

A third lesion that demonstrates intersecting features with our two cases, based on abundant myxoid stroma, is the odontogenic myxoma. The main difference is that odontogenic myxomas do not produce dental hard material, although small biopsy samples of our currently described tumor likely would not always contain odontogenic hard tissue. Odontogenic myxomas typically contain a sparse population of stellate-shaped to spindled cells with rare to absent islands of odontogenic epithelium. The average age of odontogenic myxoma occurrence is 28.6 whereas both of our cases presented in young children [37]. Santos et al. performed NGS on nine cases of odontogenic myxoma and identified no BRAF or any other pathogenic mutations [38].

The best of our knowledge, no other previously described odontogenic tumor contains ovoid-to-round cells with bi- or multi-lobed nuclei arranged in a reticular pattern amongst a basophilic myxoid stroma. As we recognize that we cannot exclude the possibility that this neoplasm represents a variant of a previously described odontogenic tumor, we emphasize the need for further molecular characterization of currently classified tumors and the foresight to preserve gross samples of lesions in formalin (without decalcification) so that genetic testing may be performed. Based on our limited sample size, ALK IHC appears to be an insufficient substitute for ALK molecular testing.

Conclusion

We have presented two cases of an unusual odontogenic neoplasm with microscopic features that appear unique when compared to other existing tumors. It is unknown at this time whether they represent a new entity or a variant of a known odontogenic tumor. Given the discovery of a novel STRN::ALK fusion, which has not been reported in other odontogenic tumors, we propose that this could be a distinct neoplasm. The prognosis for these two patients is also unclear given that only a year has elapsed since both cases were treated. Historically, STRN::ALK fusions have been predominantly identified in malignant neoplasms; however the cytologic features of both our cases suggest a benign entity with uncertain clinical behavior. Close and long-term follow-up is planned for both patients. It is our hope that the publication of these two cases will facilitate the discovery and identification of additional examples of this lesion by other colleagues, and ultimately lead to a better understanding of this entity.

Author Contributions

M.R. wrote the main manuscript text. M.R., M.S., K.C, J.B. were involved in providing the microscopic descriptions, diagnosis, and photomicrographs for the two cases. J.B. and J.G. were involved in the NGS procedure which identified the same fusion in both specimens. N.C. is the treating clinician for Case 1 and provided the associated specimen. W.Z. is the treating clinician for Case 2 and provided the associated specimen. M.R, M.S., K.C. prepared figures. All authors reviewed the manuscript

Funding

Jane B. and Edwin P. Jenevein M.D. Distinguished Chair in Pathology Endowment.

Data Availability

All data underlying the results are available as part of the article and no additional source data are required.

Declarations

Competing interests

The authors declare no competing interests.