Odontogenic ghost cell lesions, originally described by Gorlin et al. in 1962 , comprise COC, DGCT and GCOC [1, 2]. DGCT is considered the solid counterpart of COC and is occasionally associated with it. These lesions can be classified as central (intraosseous) or peripheral (gingival or alveolar mucosal) based on their clinical presentation, and an extraoral or ectopic DGCT is not yet an established entity [1, 2].
Clinically, most DGCTs occur in the jaw bone (maxilla:mandible = 1:1) and show benign but locally infiltrating behavior [1, 2]. They are more common in men (M:F = 2:1), especially at a younger age (range 11–79 years, mean: 39.7) . The patients usually complain of progressive or slow-growing nodules with swelling, with or without pain [1, 6]. Radiologically, DGCT shows a cystic or solid mass with calcification . In the present case, although the patient was older than the mean age for DGCT occurrence, the imaging findings were consistent with those reported previously [1, 2, 7].
To our knowledge, this the first report to meticulously describe the cytological findings of DGCT. The cell cluster chiefly consisted of basaloid cell proliferation with peripheral palisading. These findings are consistent with those of basal cell adenoma/carcinoma, and we also suspected salivary gland tumors. However, calcification and admixed orange G positive structures without nuclei, similar to ghost cells, are a differential feature and therefore an important cytological feature of DGCT.
The histological features of DGCT include basaloid cell proliferation with ameloblatoma-like epithelial nests resembling the stellate reticulum. Aberrant keratinization was seen with ghost cells having enlarged, polygonal eosinophilic cytoplasm, with or without nuclei, and immature to mature dentinoid or dentino-osteoid structures [1, 2]. Findings indicating the odontogenic nature of the tumor and its transition from a COC are important. The neoplastic cells have been shown to be strongly positive for cytokeratin AE1/3, 5, 7, 14, and 19, but negative for vimentin, desmin, SMA, and CD34. The Ki-67 index has been reported to be < 5% [1, 2]. These histological findings are consistent with those of the present case.
Recently, CTNNB1 mutations and/or nuclear β-catenin accumulation were detected in the histologically analog groups of basaloid tumors with obligate ghost cell differentiation, such as odontogenic ghost cell lesions including COC, DGCT, and GCOC [8, 9], basal cell adenoma/carcinoma [1, 2], pilomatrixoma, pilomatrical carcinoma [10, 11], and adamantinomatous craniopharyngioma . Phosphorylated CTNNB1 activates WNT signaling and promotes nuclear β-catenin accumulation [8, 13], which elicits differentiation of the tumor cells into hair-like cells called ghost cells [8, 14]. We analyzed the hot spots in 50 genes commonly associated with cancer by target next-generation sequencing, and only CTNNB1 mutation was detected.
Considering the anatomical site and histogenetic features, the most important differential diagnosis in the present case is of basal cell adenoma/carcinoma. However, basal cell adenoma/carcinoma exhibits two-cell morphology consisting of CK7-positive ductal structures and p63, SMA, CK5/6, WT-1, or podoplanin positive myoepithelial/basal cell components, unlike the findings of the present case. Moreover, all above-mentioned basaloid tumors with ghost cell differentiation lack histological findings of dentinoid material and precursor COC-like cystic components.
To date, only two cases of ectopic dentinogenic ghost cell-like lesions have been reported. One was a DGCT-like lesion in the ethmoid sinus of an 8-year-old boy , and the other was a GCOC-like carcinoma on the floor of the mouth of a 54-year-old man . Both exhibited characteristic odontogenic epithelium proliferation with ghost cells but lacked the anatomic association to the oral and alveolar mucosa and bone on radiological, intraoperative, and pathological examinations. Further, CTNNB1 mutation was detected in the latter case. The clinicopathological features of the present case were similar to those reported previously. Moreover, the characteristic precursor lesion, COC, was detected, with no history of an odontogenic tumor, trauma, or surgery that could have caused tumor dissemination or metastasis.
Based on this clinical, histological, and genetic evidence, a final diagnosis of extraosseous DGCT arising from a COC in the floor of mouth was confirmed.
The development of DGCT occurs through two major pathways: de novo or from a preceding COC. However, the true etiology of an extraosseous DGCT remains unclear [4, 5, 15]. Peripheral DGCT can originate from oral epithelium following trauma or exposure to an irritating agent [6, 15]. In the present case, these factors were absent, and the lesion had no connection with the oral mucosa. Therefore, ectopic odontogenic epithelium may have been associated with the tumor’s development.
The recurrence rates of central and peripheral DGCT are 73% and 0%, respectively . While segmental resection is indicated for central DGCT, simple excision is recommended for peripheral DGCT. As an ectopic DGCT is extremely rare, the tumor aggressiveness and optimal treatment are unknown. Liu et al.  reported no recurrence of an ectopic DGCT arising from the ethmoid sinus after endoscopic sinus surgery, during a 2-year follow-up. Similar to our findings, they observed that the Ki-67 labeling index was not high, and there was no invasion of the adjacent tissue, vascular, and perineural structures, suggestive of low malignant potential of ectopic DGCT. In our opinion, simple excision of the tumor is therefore justified, and further studies are needed to clarify the nature of the tumor.
This report described a case of DGCT occurring as an ectopic lesion. Despite characteristic histological features, its diagnosis is difficult. Comprehensive clinicopathological examination is important to accurately identify this rare entity to avoid misdiagnosis and overtreatment.