Current Concepts and Occurrence of Epithelial Odontogenic Tumors: II. Calcifying Epithelial Odontogenic Tumor Versus Ghost Cell Odontogenic Tumors Derived from Calcifying Odontogenic Cyst

Abstract

Calcifying epithelial odontogenic tumors (CEOTs) and ghost cell odontogenic tumors (GCOTs) are characteristic odontogenic origin epithelial tumors which produce calcifying materials from transformed epithelial tumor cells. CEOT is a benign odontogenic tumor composed of polygonal epithelial tumor cells that show retrogressive calcific changes, amyloid-like deposition, and clear cytoplasm. Differentially, GCOTs are a group of transient tumors characterized by ghost cell presence, which comprise calcifying cystic odontogenic tumor (CCOT), dentinogenic ghost cell tumor (DGCT), and ghost cell odontogenic carcinoma (GCOC), all derived from calcifying odontogenic cysts (COCs). There is considerable confusion about COCs and GCOTs terminology, but these lesions can be classified as COCs or GCOTs, based on their cystic or tumorous natures, respectively. GCOTs include ameloblastomatous tumors derived from dominant odontogenic cysts classified as CCOTs, ghost cell-rich tumors producing dentinoid materials as DGCTs, and the GCOT malignant counterpart, GCOCs. Many authors have reported CEOTs and GCOTs variably express keratins, β-catenin, BCL-2, BSP, RANKL, OPG, Notch1, Jagged1, TGF-β, SMADs, and other proteins. However, these heterogeneous lesions should be differentially diagnosed to allow for accurate tumor progression and prognosis prediction.

Headnote

Calcifying epithelial odontogenic tumors (CEOTs) and ghost cell odontogenic tumors (GCOTs) are characteristic odontogenic origin epithelial tumors which produce calcifying materials from transformed epithelial tumor cells. CEOT is a benign odontogenic tumor composed of polygonal epithelial tumor cells that show retrogressive calcific changes, amyloid-like deposition, and clear cytoplasm. Differentially, GCOTs are a group of transient tumors characterized by ghost cell presence, which comprise calcifying cystic odontogenic tumor (CCOT), dentinogenic ghost cell tumor (DGCT), and ghost cell odontogenic carcinoma (GCOC), all derived from calcifying odontogenic cysts (COCs). There is considerable confusion about COCs and GCOTs terminology, but these lesions can be classified as COCs or GCOTs, based on their cystic or tumorous natures, respectively. GCOTs include ameloblastomatous tumors derived from dominant odontogenic cysts classified as CCOTs, ghost cell-rich tumors producing dentinoid materials as DGCTs, and the GCOT malignant counterpart, GCOCs. Many authors have reported CEOTs and GCOTs variably express keratins, β-catenin, BCL-2, BSP, RANKL, OPG, Notch1, Jagged1, TGF-β, SMADs, and other proteins. However, these heterogeneous lesions should be differentially diagnosed to allow for accurate tumor progression and prognosis prediction.

Key Words: Calcifying; Dentinoid; Ghost cell odontogenic tumors

Odontogenic tissue is programmed to produce dentin and enamel due to active interactions between odontogenic mesen- chyme and epithelium. Tooth formation is achieved via odonto- genic mesenchyme and epithelium stage- and spatial-specific differentiation from early tooth development to late matura- tion.1,2 Therefore, when odontogenic tissue becomes undifferen- tiated and undergoes tumorous change, it has the potential to produce abnormal calcifications with enameloid, dentinoid, and cementum-like material histologic features. For this reason, these odontogenic calcifications are important odontogenic tu- mor characteristics and occasionally are accompanied by odon- togenic epithelium ghost cell change and amorphous odonto- genic mesenchyme hyalinization.3

Ghost cells are anucleate epithelial cells with swollen homo- geneous pale eosinophilic cytoplasm and pale to clear central areas occupied by the previous nucleus. In addition, they are highly keratinized with degenerated nuclei shadow images. Ghost cells are frequently found in calcifying odontogenic cyst (COC) derivatives, including craniopharyngioma, pilomatrico- ma, etc., in association with calcified foci.4

Calcifying epithelial odontogenic tumors (CEOTs) and COCs produce calcifying materials by transforming odontogenic epi- thelial cells. In particular, COC tumorous growth can lead to ghost cell odontogenic tumor (GCOT) development, compris- ing the calcifying cystic odontogenic tumor (CCOT), dentino- genic ghost cell tumor (DGCT), and ghost cell odontogenic carcinoma (GCOC), in which many ghost cells emerge. How- ever, for prognostic reasons, epithelial odontogenic tumors pro- ducing calcifications and/or ghost cells should be differentially diagnosed based on tumor cellular differentiation.

CALCIFYING EPITHELIAL ODONTOGENIC TUMOR

General features

CEOT is an uncommon, benign odontogenic lesion that ac- counts for less than 1% of all odontogenic tumors. Dr. Pindborg (1958)5 first described this tumor, and thereafter, CEOT was termed Pindborg tumor by Shafer et al.6,7 CEOT mainly presents discrete polyhedral eosinophilic epithelial cell islands, strands, or sheets in a fibrous stroma (Fig. 1A). Its tumor cells produce eo- sinophilic deposits resembling amyloid, which react positively to Congo red and calcify gradually. The epithelial cellular out- lines are distinct, and intercellular bridges may be seen. Nuclei show considerable variation, and giant nuclei may be seen. Some tumors show considerable nuclear pleomorphism, but this fea- ture is not considered to indicate malignancy (Fig. 1C).8,9

CEOT is usually encountered in the posterior mandible, and is most common in patients between 30 and 50 years of age, with no gender predominance. It usually presents as a slow- growing, painless swelling, and is asymptomatic, but locally invasive.10 CEOT manifests as an intraosseous lesion (central type) in the majority of cases (95%). Extraosseous or peripheral lesions account for fewer than 5% of cases.8,11,12 It can be associ- ated with an impacted tooth and radiographically simulate a dentigerous cyst. Most CEOTs are solid masses, and rarely show cyst spaces. They appear either as monocystic or polycystic ra- diolucent lesions with radiopaque spots or masses, and also ex- hibit peripheral destruction upon imaging. Calcifications are usually scattered within tumors,13 and are a distinctive feature. They develop within amyloid-like materials and form concen- tric rings (Liesegang ring calcifications) (Fig. 1B, C), which tend to fuse to form large, complex masses. These amyloid-like materials stain positively with Congo red or thioflavine T. Fur- thermore, after Congo red staining, amyloid exhibits an apple- green birefringence in polarized light.8,11

Types and variants

Several CEOT variants may appear with clear cell focal areas, cementum-like components, abundant Langerhans cells, com- bined epithelial odontogenic tumor (adenomatoid odontogenic tumor), and abundant myoepithelial cells.12,14

The extraosseous CEOT clear cell variant is extremely rare, and shows polyhedral cells, hyaline materials, and calcified ar- eas. The hyaline materials are usually positive for Congo red, crystal violet, and Lugol's iodine, but negative for Coomassie blue. The clear polyhedral cells often exhibit stained granules after periodic acid-Schiff staining. Therefore, these hyaline ma- terials are considered amyloid-like. Furthermore, clear cell pres- ence in extraosseous CEOT imply aggressive tumor growth in vascular connective tissue.15

The intraosseous CEOT noncalcifying variant is also very rare, and commonly appears in association with Langerhans cells. The noncalcifying, Langerhans cell-rich variant has only been reported in four cases to date,16 and most were located in the anterior and premolar maxillary regions, in contrast to clas- sical CEOTs, which are usually located in the mandible molar and ascending ramus areas. Furthermore, the intraosseous CEOT noncalcifying variant is believed to be aggressive.17

A few peripheral (extraosseous) CEOT cases have been re- ported, and manifest as nonspecific, sessile gingival masses, usually on the anterior gingiva. In some cases, this lesion is as- sociated with underlying bone cupped-out erosion.18

The adenomatoid odontogenic tumor and CEOT combina- tion is rare; 22 cases have been published.19 Because many ade- nomatoid odontogenic tumors contain variable sized CEOT- like areas, CEOT-like change may be considered a normal fea- ture in the adenomatoid odontogenic tumor histomorphologi- cal spectrum. The combination variant is benign, and repre- sents odontogenic epithelium dedifferentiation.19

The multifocal CEOT variant shows multiple CEOT lesions in intraosseous and extraosseous jaw regions, including mandi- ble, maxilla, and gingiva.20 Each multifocal tumor mass must be carefully evaluated for clinical and histologic evidence of neopla- sia, because they may exhibit different tumorigenic propensities.

Pathogenesis

CEOT epithelial tumor cells bear close morphologic resem- blance to enamel organ stratum intermedium cells (Fig. 1C). However, based on its anatomic distribution in jaw, some in- vestigators have recently suggested that the tumor arises from dental lamina remnants.8

The CEOT polyhedral epithelial cells commonly express lam- inins 1 and 5, fibronectin, cytokeratins, and vimentin, whereas its amyloid-like materials do not.21 Tumor epithelial cells are slightly positive or negative for PKK1 (specific for the 44, 46, 52, and 53 kD keratins) detectable keratins, but slightly to strongly positive for KL1 (specific for the 55-57 kD keratins) and TK (41-65 kD keratins). Tumor epithelium is slightly posi- tive for vimentin but negative for desmin.22 Notably, CEOT ep- ithelial tumor cells exhibit alkaline phosphatase and ATPase lo- calization in their membranes, indicating active membrane transport.23 These high alkaline phosphatase and ATPase activi- ties may be relevant to abnormal eosinophilic materials (amy- loid-like) and irregular calcification production.24

At the electron microscope level, CEOT epithelial tumor cells consist of polyhedral epithelial cells and myoepithelial-like cells, containing large numbers of electron-dense tonofilament bun- dles, electron-dense bodies, and fine lamina densa filaments.25 Amyloid-like materials compose fine filament sheets measuring 10-12 nm in diameter and lamina densa fragment aggregates, which are probably secreted by polyhedral epithelial and myo- epithelial-like cells. The fine filamentous materials are a form of amyloid, and their formation results from lamina densa material degradation.26 Recently, the odontogenic ameloblast-associated protein (ODAM) fibril-forming region was found in CEOT am- yloid-like materials, thus suggesting ODAM might have amy- loidogenic potential.27

A number of dendritic cells, which are frequently found among CEOT epithelial sheets, are strongly positive for S-100 and CD-1a antisera. Ultrastructurally, these dendritic cells show indented nuclei and Birbeck's granules similar to Langerhans cells. Thus, they are likely to be Langerhans cells, and play a role in antigen presentation from epithelial tumor cell abortive products.21,28

The fact that an ameloblastin mutation was found exclusively in CEOT suggests that alterations to this gene may be relevant to CEOT pathogenesis.29 However, both ameloblastin trans- genic and ameloblastin-null mice showed amelogenesis imper- fecta features. In particular, ameloblastin-null mice developed unidentified type odontogenic tumors.30 More mutational anal- yses are required to elucidate the mutation nature in CEOT odontogenic genes.

Prognosis

CEOTs are generally considered benign, but can be locally aggressive, and exhibit 10% to 15% recurrence rates.31 Maxil- lary CEOTs tend to be more aggressive and spread more rapidly to possibly involve surrounding vital structures than mandibu- lar CEOTs.32 Therefore, maxillary CEOT should be treated more aggressively, and five years is considered the absolute min- imum follow-up period.12

Although it was originally believed that CEOT is similar to ameloblastoma biologic behavior, growing evidence indicates it tends to be less aggressive.8,12 Thus, due to CEOT's relatively indolent biological behavior, mutilating procedures, such as wide resection or mandible hemisection, seem unwarranted. Accordingly, enucleation within macroscopically normal tissue is recommended for CEOT involving the mandible.8 However, repeated local recurrence may imply malignant transformation and tumor metastasis with vascular invasion and spread to a cervical lymph node.33 Contrary to primary CEOT benign his- tological features, malignant CEOT shows nuclear pleomor- phism with frequent mitotic figures and increased proliferative activity, as assessed by immunostaining for Ki-67.34 Therefore, recurrent CEOT carries the threat of malignant transformation, and may justify radical surgery, and adjuvant radiotherapy.35

CALCIFYING ODONTOGENIC CYST

COC was first described by Gorlin et al. in 1962,36 and since then several hundred cases have been reported.37 COC exhibits considerable histopathologic diversity and variable clinical be- havior. Cyst epithelium is usually thin and tends to be detached easily, but is focally thickened by keratinized epithelial cells and ghost cells. Hyperplastic basal cells often grow into fibrous cyst walls, resulting in daughter cysts. Unlike other odontogenic cysts, COCs may contain highly-differentiated lining epitheli- um composed of columnar cells or stellate reticulum-like cells associated with ghost cells, and undifferentiated lining epitheli- um resembling reduced enamel epithelium (Fig. 2A).38,39

Ghost cell masses frequently fuse to form large amorphous, acellular sheets. Furthermore, ghost cells are commonly associ- ated with dystrophic calcification (Fig. 2B, C), and rarely with aberrant ossification.40 Confocal laser scanning microscopy ob- servations have disclosed that COC ghost cells autofluoresce. However, autofluorescence intensities are variable, possibly due to hard keratin presence.4

Immunohistochemically, amelogenin protein is expressed chiefly in ghost cells,41 whereas cytokeratin 19 and B-cell leuke- mia/lymphoma-2 (BCL-2) are expressed chiefly in odontogenic cyst lining epithelial cells. Lining epithelial cells are sporadically positive for Ki-67 antigen, especially in proliferative and amelo- blastomatous epithelium. These findings suggest that ghost cells or lining epithelial cells show ameloblastic cytodifferentiation or odontogenic epithelial characteristics and consistently proliferate and increasingly survive under cystic conditions.42

COC is predominantly an intraosseous lesion, although in 13% to 30% of reported cases, COC has manifested as a pe- ripheral lesion. About 65% of cases are found in the incisor or canine areas. Mean age is 33 years, and most cases are diagnosed in the second and third decades of life. However, COCs that are associated with odontomas tend to occur in younger patients (mean age, 17 years).43,44

Central COC is usually a unilocular, well-defined radiolucen- cy, although the lesion may occasionally appear multilocular. Radiopaque structures within lesions, either irregular calcifica- tions or tooth-like densities, are present in about one third to one half of cases. The radiolucent lesion is associated with an unerupted tooth, usually a canine, in approximately one third of cases.45,46

Extraosseous COCs are usually localized sessile or peduncu- lated gingival masses with no distinctive clinical features, and can resemble common gingival fibromas, gingival cysts, or pe- ripheral giant cell granulomas. They appear as variably sized odontogenic epithelial islands in a fibrous stroma composed of peripheral palisading columnar cells and central stellate reticu- lum, reminiscent of ameloblastoma. However, ghost cell nests are present within epithelium, and juxtaepithelial dentinoid is commonly present. These features differentiate this lesion from peripheral ameloblastoma.46

COC prognosis is relatively good, and few recurrences have been reported after simple enucleation. Peripheral neoplastic COC appears to have the same prognosis as a peripheral amelo- blastoma with minimal recurrence chance after simple surgical excision. Accordingly, when a COC is associated with some oth- er recognized odontogenic tumor, such as, ameloblastoma, treat- ment and prognosis are likely to be the same as that of the asso- ciated tumor.47

The term COC is consistent with a unicystic lesion with or without associated odontoma, while other related lesions iden- tified as benign or malignant tumors must be classified sepa- rately.44 Although the COC term represents a cyst, some inves- tigators prefer to classify it as a neoplasm.8 The cyst lining uni- focal or multifocal epithelial proliferation into the lumen may resemble ameloblastoma, but the COC is intermixed with varying ghost cell numbers. About 20% of COCs are associated with odontoma, but the neoplastic solid COCs are relatively uncommon, and account for 2% to 16% of all COCs in report- ed series.48 The 2005 World Health Organization (WHO) Classification of Odontogenic Tumors divided COC neoplastic lesions into GCOTs comprised of CCOT, DGCT, and GCOC.49

COC may have more heterogeneous components than envis- aged, and thus, careful pathological observation and long-time follow-up are required. Given the uneventful outcomes for 35 cases filed in our Gangneung-Wonju National University Den- tal Hospital over 18 years, we consider the diagnostic term, COC, is still useful. However, some authors have described COC as CCOT, or calcifying ghost cell odontogenic cyst. Therefore, these heterogeneous lesions can be classified as COCs when benign cystic lesions, and as GCOTs when tumorous.

GHOST CELL ODONTOGENIC TUMORS

According to the WHO International Histological Classifi- cation of Odontogenic Tumors guideline (2005) GCOTs com- prise CCOT, DGCT, and GCOC.50 In this guideline, CCOT is defined as a benign cystic neoplasm of odontogenic origin, characterized by ameloblastoma-like epithelium with ghost cells that may be calcified,51 and DGCT as a locally invasive neoplasm characterized by epithelial cell ameloblastoma-like is- lands in mature connective tissue stroma. Aberrant ghost cell keratinization may be found in association with varying dys- plastic dentin (so called dentinoid materials) amounts.52 Fur- thermore, previously reported odontogenic ghost cell carcinoma and malignant epithelial odontogenic ghost cell tumor cases were re-named by WHO as GCOC (Table 1).53,54

GCOTs usually occur within maxillary and mandibular bones (central GCOTs), and in gingival soft tissues (peripheral GCOTs). Central GCOTs can be associated with odontoma arising from an impacted tooth, but are gradually replaced by solid epithelial odontogenic tumor tissue. On the other hand, peripheral GCOTs may present on gingiva as a painless nodule that clinically appears as a reactive hyperplastic mass.55 Periph- eral GCOTs are usually treated by simple excision and show good prognosis with few recurrences, but central GCOTs should be treated carefully depending on neoplastic epithelium infiltrative growth.56

A clear cell GCOT variant was reported without further at- tempt at classification.57 It showed glycogen sheets and islands containing clear epithelial cells separated by a thin fibrous con- nective tissue stroma. Furthermore, both ameloblastic and clear cells were immunopositive for cytokeratin 19 and AE1/3.58 Al- though it is unclear whether this tumor represented a pre-exist- ing GCOT clear cell change or a separate and distinct neoplasm derived de novo from odontogenic epithelium, the clear cell components were the most prominent clear cell GCOT distin- guishing feature.44

Previous reports indicate GCOTs have wide neoplastic po- tential.59 CCOT is a cystic, painless, slowly growing tumor that commonly presents as a well-defined radiolucent or combined lesion that rarely recurs, whereas DGCT is aggressive and recur- rences may be expected, and GCOC is in actuality a malignant neoplasm.50 Furthermore, the term GCOT is useful for describ- ing a solid neoplastic COC variant characterized by amelobla- stomatous epithelial components accompanied by abundant ghost cell clusters and dentinoid materials (Table 1).

CALCIFYING CYSTIC ODONTOGENIC TUMOR

General features

CCOT is a benign odontogenic origin cystic neoplasm charac- terized by an ameloblastoma-like epithelium and ghost cells. Its cystic epithelial lining shows a well-defined columnar cell basal layer and an overlying layer often composed of many epithelial cells, which may resemble the enamel organ stellate reticulum. CCOT usually contains ghost cell masses that may be located within the epithelial lining or in the fibrous capsule. The CCOT neoplastic epithelium is closely associated with COC, and exhib- its ghost cells that may undergo calcification (Fig. 2F).60,61

CCOT usually presents as a painless, slow-growing mass in- volving both maxilla and mandible, primarily in the anterior area (incisors and canines). It generally affects young adults in the third to fourth decades and has no gender predominance.62 CCOT may occur in an intraosseous or extraosseous area. Pe- ripheral CCOT accounts for about 26% of all reported cases.63

CCOT typical microscopic features include ameloblastoma- tous epithelium containing ghost cells clusters and the simple COC lesion (Fig. 2G-I). The cystic lesion may sometimes dom- inate and be associated with a hard dental tissue area resem- bling odontoma. However, a limited dysplastic dentin amount can usually be found (Fig. 2D-F).61

CCOT calcification appears as a thin radiopaque line, and discrete radiopaque foci, which contrast with those in adenoma- toid odontogenic tumors, which exhibit numerous dispersed or clustered radiopaque foci.64 These characteristic radiopacities may facilitate differential diagnosis by revealing internal calcifi- cations in cone-beam computed tomography images.65 CCOT is also associated with benign odontogenic tumors other than odontoma, such as, ameloblastic fibroma, ameloblastic fibro- odontoma, adenomatoid odontogenic tumor, or odontoamelo- blastoma.66

Types and variants

Although CCOT is designated histologically as a compound lesion composed of COC and tumorous odontogenic epitheli- um, it can be classified into four sub-types using the following predominant findings, 1) simple cystic (Fig. 2A-C), 2) odonto- ma associated (Fig. 2D-F), 3) ameloblastomatous proliferating (Fig. 2G-I), or 4) CCOT associated with a benign odontogenic tumor other than odontoma.54,66

The simple CCOT cystic type resembles ordinary COC, but with polygonal odontogenic epithelium and ghost cell tumor- ous growth. Furthermore, this cystic lesion may have ghost cells without odontoma features.54

Odontoma-associated CCOT contains discrete, abnormal tooth structures originating from odontogenic mesenchyme. Dentin, enamel, cementum, and pulpal tissues are present as a complex or compound odontoma. Furthermore, CCOT enamel epithelium is relatively well differentiated, which implies this CCOT type has a good prognosis.54

Ameloblastomatous proliferating CCOT shows predominant ameloblastoma-like epithelial proliferation features, producing anastomosing epithelial strands with palisading basal cells (Fig. 2G-I). Its epithelium is less infiltrative than ameloblastoma, and is usually localized at subepithelial connective tissue with sparse odontogenic myxoid mesenchyme. Ameloblastomatous prolifer- ating CCOT is characterized by intraluminal or capsular plexi- form growths resembling those seen in the cystic ameloblastoma plexiform variant. Sometimes follicles simulating solid amelo- blastoma are observed in the connective tissue capsule.54

The existence of benign odontogenic tumor-associated CCOT implies CCOT's heterogeneous nature. Hybrid odontogenic tu- mors can occur. To date, five CCOT adenomatoid odontogenic tumor cases,67 and one CCOT plexiform ameloblastoma case have been reported.68

Pathogenesis

CCOT tumor epithelium shows positive reactions for kera- tin-14 and keratin-10/13 in its basal and upper cell layers, respec- tively, which indicate its epithelium differentiates towards the squamous type.69 Of its epithelial basement membrane compo- nents (laminins 1 and 5, collagen type IV, and fibronectin) lam- inin 5 is found faintly in suprabasal cells, but expressed strongly in ghost cells. Therefore, it is presumed that laminin 5 is involved in ghost cell formation and in calcification initiation.70

CCOT calcified masses demonstrate positive immunoreactivity to human bone sialoprotein (BSP) antibodies, and hybridization signals for BSP are located in cells near calcified particles. In COC, BSP signals are also seen in cells surrounding the characteristic ghost cell nests, which often subsequently calcify. BSP may play an important role in ghost tumor formation and differentiation involving pathological calcification.71

Ghost cells demonstrate Notch1 and Jagged1 overexpression, suggesting Notch1-Jagged1 signaling subserves the main trans- duction mechanism responsible for ghost cell fate decision in CCOT. Protein localizations are largely membranous and/or cy- toplasmic. In particular, mineralized ghost cells are strongly pos- itive, which suggests ghost cell calcification processes might be associated with Notch1 and Jagged1 upregulation.72 However, both polygonal epithelial and stromal cells show positive immu- noreactivity for nuclear factor kappa-B (NF-κ B; RANK), RANK ligand (RANKL), and osteoprotegerin receptor activa- tion.73

The transforming growth factor-β/SMAD signaling pathway is less activated in CCOT than in ameloblastoma, indicating CCOT may show less cellular proliferation and less advanced cellular differentiation than ameloblastoma.74 Furthermore, CCOTs express fewer matrix metalloproteases (MMPs), tissue inhibitors of metalloproteinases (TIMP), growth factors, epider- mal growth factor receptor, and extracellular signal-regulated kinase than odontogenic keratocysts (keratocystic odontogenic tumors). Therefore, it is believed that CCOT is less aggressive than odontogenic keratocysts.75 In addition, CCOT has been re- ported to show lower MMPs, TIMPs, and growth factor expres- sion than ameloblastoma.76

NF-κB, Ki-67, and MMP-9 are weakly stained in CCOT, but nevertheless, may consistently affect its progression and lo- cal invasiveness.77 Podoplanin, a representative immunohisto- chemical marker for lymphatic endothelial cells, is localized in both CCOT basal and polyhedral cells, which are coincidentally positive for proliferating cell nuclear antigen and integrin β1. Thus, podoplanin-positive cells may be closely associated with extracellular matrix signaling and cell proliferation in CCOT.78

Prognosis

CCOT is a noninvasive odontogenic neoplasm with amelo- blastomatous epithelium with fewer aggressive features than ameloblastoma. CCOT is usually uneventful after simple enu- cleation and curettage, whereas extensive radical surgery is rec- ommended for ameloblastoma. CCOTs associated with odonto- ma, a supernumerary tooth, or a dentigerous cyst of embedded canine may show good prognosis after simple enucleation.65

DENTINOGENIC GHOST CELL TUMOR

General features

DGCT is a rare benign odontogenic solid tissue tumor de- rived from a neoplastic COC variant, but neoplastic epithelium is more predominant than the cystic component, which is con- trary to that observed in CCOT.54 The peripheral DGCT vari- ant appears as a well-circumscribed mass mimicking a nonspe- cific gingival enlargement, whereas intraosseous DGCTs are more aggressive than extraosseous DGCTs and have a high pro- pensity for local recurrence.79,80

DGCT is histologically characterized by odontogenic epithe- lium islands showing ameloblastoma-like features in fibrous tissue, where dysplastic dentin and ghost cells are frequently observed. The neoplastic DGCT epithelia are considered to be locally invasive. A microscopic dentinoid formation with ghost cells may provide a definitive DGCT diagnosis. Van Giesson histochemical stain further confirms the dentinoid-like materi- als (Fig. 3).81 DGCT is sometimes accompanied by COC, and the DGCT dentinoid materials are eosinophilic, amorphous, and sometimes undergo transformation into osteodentin.82 Fur- thermore, dendritic cells representing Langerhans cells and me- lanocytes often infiltrate into tumoral epithelium.83

Types and variants

DGCT can be separated into central aggressive and peripher- al non-aggressive types.54 As peripheral (extraosseous) lesions commonly show gingival swelling, they can be easily confused with other gingival lesions, such as, reactive or inflammatory lesions, or with other peripheral odontogenic tumors. Central lesions have the potential to invade locally, thus may have vari- able recurrence rates.79

DGCT is defined as a solid neoplastic growth formed by epi- thelial cell groups and islands visualized as an ameloblastoma- like basal cell layer admixed with variable quantities of dentin- like materials and ghost cell groups.54 These DGCT amelobla- stomatous features are sometimes difficult to distinguish from ameloblastoma. In one case, ameloblastoma was associated with DGCT with a predominant proliferating neoplastic epithelium feature.84

Pathogenesis

The DGCT odontogenic tumor epithelium shows variable reactions for keratins, and demonstrates PKK1 staining in pe- ripheral tumor cells, staining for KL1, and involucrin in cen- trally located cells. Lectin binding in amorphous dentinoid ma- terials was reported to be comparatively strong for PNA and SBA, moderate for WGA, RCA-1, and UEA-1, and slight for DBA and ConA. In the same study, amorphous dentinoid ma- terial elemental analysis by electron probe X-ray microanalysis showed a pattern similar to that found in normal dentin.82

β-Catenin and lymphoid enhancer-binding factor 1 are fre- quently positive in odontogenic epithelial cell cytoplasm and nuclei adjacent to ghost cells, which suggests involvement of the Wnt signaling pathway during ghost cell formation.85 Nu- cleated tumor cells adjacent to ghost cells are occasionally ter- minal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay positive. Furthermore, β-catenin mutations have been described in DGCT and in COC, suggesting β-catenin plays an important role in DGCT tumorigenesis.86

Prognosis

DGCT is a locally invasive neoplasm characterized by amelo- blastoma-like epithelial islands, ghost cells, and dentinoid.87 Peripheral DGCT is absolutely benign with no features of inva- sive growth and is invariably uneventful after surgical removal. On the other hand, intraosseous DGCT is potentially invasive and tends to recur, and hence requires radical surgery.54,79,80

In one report on seven intraosseous DGCTs, five treated by conservative surgery recurred, but the two treated by aggressive local resection did not.88 Thus, intraosseous DGCT should be treated by resection with an adequate safety margin of at least 0.5 cm (as is recommended for ameloblastoma), and adequately followed for signs of recurrence.80,88,89

GHOST CELL ODONTOGENIC CARCINOMA

General features

GCOC is an uncommon malignant neoplasm exhibiting prominent mitotic activity, nuclear atypia, cellular pleomor- phism, ghost epithelial cell groups, necrosis, and sometimes scarce mineralized or dentin-like materials. GCOC has an infil- trative growth pattern with regional metastasis. Furthermore, as tumor cells undergo malignant processes, ghost cell may be difficult to find.90

This neoplasm has a locally aggressive, destructive behavior, and at times metastatic deposits can be found. It may be diag- nosed as CCOT, DGCT, or GCOTs before its malignant trans- formation, although it often co-occurs with benign variant odontogenic tumors derived from COC.91

In one 30 case GCOC study, mean age at diagnosis was 40.3 years, 22 (73%) involved the maxilla, and 12 (40%) were sec- ondary benign CCOT malignant manifestations, or DGCTs. Five patients died of recurrence or distant metastasis.92 It has also been reported that GCOC is more prevalent in Asians (12/18) than in other racial groups.93 A radiological study found a mixed radiolucent and radiopaque lesion pattern was most fre- quent (14/19) compared to a radiolucent lesion pattern (5/19). In addition, 89% (17/19) showed poorly defined borders and 11% (2/19) showed well defined borders. Root resorption was reported in 31% (6/19) of patients, and tooth displacements in 21% (4/19).93-95

Types and variants

Three GCOC variants are identified according to their patho- genic mechanisms, as follows: 1) GCOC arising de novo is a ma- lignant neoplasm not associated with previous DGCT or CCOT, although sometimes DGCT-like areas may be present;96 2) GCOC ex-CCOT is considered when a GCOC is found to be derived from previously diagnosed CCOT, or when a GCOC is associated with benign CCOT;54,92 and 3) GCOC ex-DGCT is a recurrent malignant neoplasm previously diagnosed as benign DGCT, in which occasional dentin-like material areas are dem- onstrated.90

Pathogenesis

Although GCOC has different histological presentations and unpredictable indolent or locally aggressive growth behaviors, the tumor usually arises due to preexisting benign COC (CCOT) malignant transformation, and sometimes from other odonto- genic tumors.93 However, immunohistochemical reactivities for Ki-67 and MMP-9 are significantly stronger in GCOC than in CCOT, or in DGCT.77,97 In particular, nucleated cells adjacent to ghost cells in GCOC express cytokeratins, involucrin, and BCL- 2 associated X (BAX), and they are frequently positive for Ki-67 and p53, and occasionally TUNEL positive.98 These findings support the observation that GCOC exhibits significantly greater proliferative activity and invasive growth than CCOT or DGCT.

Despite their retrogressive and apoptotic status, GCOC ghost cells consistently express BAX and vitronectin receptor pro- teins, and are strongly positive for the tartrate-resistant acid phosphatase reaction.99 These facts indicate anucleated ghost cells still play important roles in the extensive bone resorption observed in GCOC.91,98

Prognosis

GCOCs may grow slowly and be locally invasive or grow rap- idly and be infiltrative. Therefore, wide local excision with his- tologically clean margins is recommended for GCOCs. Further- more, close long-term surveillance for recurrence is required for GCOCs arising from long-standing benign COCs or GCOTs.100

To date 31 GCOC cases have been described in English pub- lications, and three cases involved Koreans.91,98 As was suggest- ed by another author,50 we recommend more GCOC cases be reported (using the WHO tumor classification) to allow their variable malignant potential to be better defined.

SUMMARY

This review of the currently available CEOT and GCOTs lit- erature shows that CEOT and GCOTs are uncommon tumors exhibiting variable calcification features and ghost cells pro- duced during odontogenic epithelium tumorigenic differentia- tion. We advise CEOT and GCOTs differential diagnosis should be undertaken carefully, despite the confusion and controversy associated with histological diagnoses of these heterogeneous and transient odontogenic tumors. Furthermore, in order to more precisely define their biological and pathological behav- iors, further advanced studies are required, and publication should be encouraged to provide comprehensive data.

Conflicts of Interest

No potential conflict of interest relevant to this article was reported.

References

REFERENCES

1. Main JH, Waheed MA. Epitheliomesenchymal interactions in the proliferative response evoked by polyoma virus in odontogenic ep- ithelium in vitro. J Natl Cancer Inst 1971; 47: 711-26.

2. Slavkin HC, Cummings E, Bringas P, Honig LS. Epithelial-derived basal lamina regulation of mesenchymal cell differentiation. Prog Clin Biol Res 1982; 85(Pt B): 249-59.

3. Badger KV, Gardner DG. The relationship of adamantinomatous craniopharyngioma to ghost cell ameloblastoma of the jaws: a his- topathologic and immunohistochemical study. J Oral Pathol Med 1997; 26: 349-55.

4. Lucchese A, Scivetti M, Pilolli GP, Favia G. Analysis of ghost cells in calcifying cystic odontogenic tumors by confocal laser scanning microscopy. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2007; 104: 391-4.

5. Pindborg JJ. A calcifying epithelial odontogenic tumor. Cancer 1958; 11: 838-43.

6. Shafer WG. Cysts, neoplasms, and allied conditions of odontogenic origin. Semin Roentgenol 1971; 6: 403-13.

7. Singh N, Sahai S, Singh S, Singh S. Calcifying epithelial odontogen- ic tumor (Pindborg tumor). Natl J Maxillofac Surg 2011; 2: 225-7.

8. Neville B, Damm DD, Allen CM, Bouquot J. Oral and maxillofacial pathology. Philadelphia: WB Saunders, 2009.

9. Krolls SO, Pindborg JJ. Calcifying epithelial odontogenic tumor: a survey of 23 cases and discussion of histomorphologic variations. Arch Pathol 1974; 98: 206-10.

10. Sharma U, Gulati A, Batra H, Singh D. Calcifying epithelial odon- togenic tumor in anterior maxilla associated with a supernumer- ary tooth: a case report. J Dent Res Dent Clin Dent Prospects 2013; 7: 51-4.

11. Ai-Ru L, Zhen L, Jian S. Calcifying epithelial odontogenic tumors: a clinicopathologic study of nine cases. J Oral Pathol 1982; 11: 399- 406.

12. Philipsen HP, Reichart PA. Calcifying epithelial odontogenic tu- mour: biological profile based on 181 cases from the literature. Oral Oncol 2000; 36: 17-26.

13. Channappa NK, Krishnapillai R, Rao JB. Cystic variant of calcify- ing epithelial odontogenic tumor. J Investig Clin Dent 2012; 3: 152- 6.

14. Ng KH, Siar CH. A clinicopathological and immunohistochemical study of the calcifying epithelial odontogenic tumour (Pindborg tumour) in Malaysians. J Laryngol Otol 1996; 110: 757-62.

15. Habibi A, Saghravanian N, Zare R, Jafarzadeh H. Clear cell vari- ant of extraosseous calcifying epithelial odontogenic tumor: a case report. J Oral Sci 2009; 51: 485-8.

16. Kaushal S, Mathur SR, Vijay M, Rustagi A. Calcifying epithelial odontogenic tumor (Pindborg tumor) without calcification: a rare entity. J Oral Maxillofac Pathol 2012; 16: 110-2.

17. Wang YP, Lee JJ, Wang JT, et al. Non-calcifying variant of calcify- ing epithelial odontogenic tumor with Langerhans cells. J Oral Pathol Med 2007; 36: 436-9.

18. Ide F, Obara K, Mishima K, et al. Peripheral odontogenic tumor: a clinicopathologic study of 30 cases. General features and hamarto- matous lesions. J Oral Pathol Med 2005; 34: 552-7.

19. Montes Ledesma C, Mosqueda Taylor A, Romero de Leon E, de la Piedra Garza M, Goldberg Jaukin P, Portilla Robertson J. Adeno- matoid odontogenic tumour with features of calcifying epithelial odontogenic tumour (The so-called combined epithelial odonto- genic tumour): clinico-pathological report of 12 cases. Eur J Can- cer B Oral Oncol 1993; 29B: 221-4.

20. Abrahão AC, Camisasca DR, Bonelli BR, et al. Recurrent bilateral gingival peripheral calcifying epithelial odontogenic tumor (Pin- dborg tumor): a case report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009; 108: e66-71.

21. Poomsawat S, Punyasingh J. Calcifying epithelial odontogenic tu- mor: an immunohistochemical case study. J Mol Histol 2007; 38: 103-9.

22. Mori M, Tatemoto Y, Yamamoto N, Okada Y. Immunohistochemi- cal localization of intermediate filament proteins in calcifying epi- thelial odontogenic tumors. J Oral Pathol 1988; 17: 236-40.

23. Morimoto C, Tsujimoto M, Shimaoka S, Shirasu R, Takasu J. Ultra- structural localization of alkaline phosphatase in the calcifying ep- ithelial odontogenic tumor. Oral Surg Oral Med Oral Pathol 1983; 56: 409-14.

24. Chomette G, Auriol M, Guilbert F. Histoenzymological and ultra- structural study of a bifocal calcifying epithelial odontogenic tu- mor: characteristics of epithelial cells and histogenesis of amyloid- like material. Virchows Arch A Pathol Anat Histopathol 1984; 403: 67-76.

25. El-Labban NG, Lee KW, Kramer IR. The duality of the cell popu- lation in a calcifying epithelial odontogenic tumour (CEOT). His- topathology 1984; 8: 679-91.

26. El-Labban NG, Lee KW, Kramer IR, Harris M. The nature of the amyloid-like material in a calcifying epithelial odontogenic tu- mour: an ultrastructural study. J Oral Pathol 1983; 12: 366-74.

27. Murphy CL, Kestler DP, Foster JS, et al. Odontogenic ameloblast- associated protein nature of the amyloid found in calcifying epi- thelial odontogenic tumors and unerupted tooth follicles. Amy- loid 2008; 15: 89-95.

28. Takata T, Ogawa I, Miyauchi M, Ijuhin N, Nikai H, Fujita M. Non- calcifying Pindborg tumor with Langerhans cells. J Oral Pathol Med 1993; 22: 378-83.

29. Perdigão PF, Carvalho VM, De Marco L, Gomez RS. Mutation of ameloblastin gene in calcifying epithelial odontogenic tumor. An- ticancer Res 2009; 29: 3065-7.

30. Paine ML, Wang HJ, Luo W, Krebsbach PH, Snead ML. A trans- genic animal model resembling amelogenesis imperfecta related to ameloblastin overexpression. J Biol Chem 2003; 278: 19447-52.

31. Franklin CD, Pindborg JJ. The calcifying epithelial odontogenic tumor: a review and analysis of 113 cases. Oral Surg Oral Med Oral Pathol 1976; 42: 753-65.

32. Kamath G, Abraham R. Recurrent CEOT of the maxilla. Dent Res J (Isfahan) 2012; 9: 233-6.

33. Veness MJ, Morgan G, Collins AP, Walker DM. Calcifying epithe- lial odontogenic (Pindborg) tumor with malignant transformation and metastatic spread. Head Neck 2001; 23: 692-6.

34. Kawano K, Ono K, Yada N, et al. Malignant calcifying epithelial odontogenic tumor of the mandible: report of a case with pulmo- nary metastasis showing remarkable response to platinum deriva- tives. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2007; 104: 76-81.

35. Somayaji G, Rajeshwary A, Ramesh S, Dinesh S. Recurrent Pin- dborg tumor of the maxilla: a case report and review of the litera- ture. Ear Nose Throat J 2013; 92: 84-7.

36. Gorlin RJ, Pindborg JJ, Odont, Clausen FP, Vickers RA. The calci- fying odontogenic cyst: a possible analogue of the cutaneous calci- fying epithelioma of Malherbe: an analysis of fifteen cases. Oral Surg Oral Med Oral Pathol 1962; 15: 1235-43.

37. Freedman PD, Lumerman H, Gee JK. Calcifying odontogenic cyst: a review and analysis of seventy cases. Oral Surg Oral Med Oral Pathol 1975; 40: 93-106.

38. Siar CH, Ng KH. Histological variation of the lining epithelium of the calcifying odontogenic cyst: a light microscopic study. Dent J Malays 1988; 10: 42-7.

39. Kwon HI, Lim WB, Kim JS, et al. Odontogenic keratocyst associat- ed with an ectopic tooth in the maxillary sinus: a report of two cas- es and a review of the literature. Korean J Pathol 2011; 45(Suppl 1): S5-10.

40. Kim YS, Lee SK. Aberrant ossification in basophilic ghost cells of calcifying odontogenic cyst resembling to pilomatricoma. Korean J Oral Maxillofac Pathol 2012;36:287-92.

41. Mori M, Yamada K, Kasai T, Yamada T, Shimokawa H, Sasaki S. Immunohistochemical expression of amelogenins in odontogenic epithelial tumours and cysts. Virchows Arch A Pathol Anat Histo- pathol 1991; 418: 319-25.

42. Yoshida M, Kumamoto H, Ooya K, Mayanagi H. Histopathologi- cal and immunohistochemical analysis of calcifying odontogenic cysts. J Oral Pathol Med 2001; 30: 582-8.

43. Johnson A 3rd, Fletcher M, Gold L, Chen SY. Calcifying odonto- genic cyst: a clinicopathologic study of 57 cases with immunohis- tochemical evaluation for cytokeratin. J Oral Maxillofac Surg 1997; 55: 679-83.

44. Li TJ, Yu SF. Clinicopathologic spectrum of the so-called calcifying odontogenic cysts: a study of 21 intraosseous cases with reconsid- eration of the terminology and classification. Am J Surg Pathol 2003; 27: 372-84.

45. Lucchese A, Petruzzi M, Scivetti M, et al. Calcifying odontogenic cysts associated with odontomas: confocal laser scanning micros- copy analysis of 13 cases. Ultrastruct Pathol 2011; 35: 146-50.

46. Hong SP, Ellis GL, Hartman KS. Calcifying odontogenic cyst: a re- view of ninety-two cases with reevaluation of their nature as cysts or neoplasms, the nature of ghost cells, and subclassification. Oral Surg Oral Med Oral Pathol 1991; 72: 56-64.

47. Gallana-Alvarez S, Mayorga-Jimenez F, Torres-Gómez FJ, Avellá- Vecino FJ, Salazar-Fernandez C. Calcifying odontogenic cyst asso- ciated with complex odontoma: case report and review of the lit- erature. Med Oral Patol Oral Cir Bucal 2005; 10: 243-7.

48. Basile JR, Klene C, Lin YL. Calcifying odontogenic cyst with odon- togenic keratocyst: a case report and review of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010; 109: e40-5.

49. Philipsen HP, Reichart PA. Classification of odontogenic tumours: a historical review. J Oral Pathol Med 2006; 35: 525-9.

50. Barnes L, Eveson JW, Reichart P, Sidransky D. World Health Or- ganization classification of tumours: pathology and genetics of head and neck tumours. Lyon: IARC Press, 2005; 1-41.

51. Prætorius F, Ledesma-Montes C. Calcifying cystic odontogenic tumour. In: Barnes L, Eveson JW, Reichart P, Sidransky D, eds. World Health Organization classification of tumours: pathology and genetics of head and neck tumours. Lyon: IARC Press, 2005; 313.

52. Prætorius F, Ledesma-Montes C. Dentinogenic ghost cell tumour. In: Barnes L, Eveson JW, Reichart P, Sidransky D, eds. World Health Organization classification of tumours: pathology and ge- netics of head and neck tumours. Lyon: IARC Press, 2005; 314.

53. Takata T, Lu Y. Ghost cell odontogenic carcinoma. In: Barnes L, Eveson JW, Reichart P, Sidransky D, eds. World Health Organiza- tion classification of tumours: pathology and genetics of head and neck tumours. Lyon: IARC Press, 2005; 293.

54. Ledesma-Montes C, Gorlin RJ, Shear M, et al. International collab- orative study on ghost cell odontogenic tumours: calcifying cystic odontogenic tumour, dentinogenic ghost cell tumour and ghost cell odontogenic carcinoma. J Oral Pathol Med 2008; 37: 302-8.

55. Lombardi T, Kuffer R, Di Felice R, Samson J. Epithelial odonto- genic ghost cell tumour of the mandibular gingiva. Oral Oncol 1999; 35: 439-42.

56. Ellis GL. Odontogenic ghost cell tumor. Semin Diagn Pathol 1999; 16: 288-92.

57. Hogge M, Velez I, Kaltman S, Movahed R, Yeh F. Ghost cell odon- togenic tumor associated with odontoma: report of two rare cases. J Clin Pediatr Dent 2012; 36: 373-6.

58. Yoon JH, Ahn SG, Kim SG, Kim J. Odontogenic ghost cell tumour with clear cell components: clear cell odontogenic ghost cell tu- mour? J Oral Pathol Med 2004; 33: 376-9.

59. Fejerskov O, Krogh J. The calcifying ghost cell odontogenic tumor - or the calcifying odontogenic cyst. J Oral Pathol 1972; 1: 273-87.

60. de Lima AP, Kitakawa D, Almeida JD, Brandão AA, Anbinder AL. Peripheral calcifying cystic odontogenic tumour of the maxillary gingiva. BMC Res Notes 2012; 5: 455.

61. Kler S, Palaskar S, Shetty VP, Bhushan A. Intraosseous calcifying cystic odontogenic tumor. J Oral Maxillofac Pathol 2009; 13: 27-9.

62. Habibi A, Saghravanian N, Salehinejad J, Jafarzadeh H. Thirty years clinicopathological study of 60 calcifying cystic odontogenic tumors in Iranian population. J Contemp Dent Pract 2011; 12: 171- 3.

63. Buchner A, Merrell PW, Carpenter WM. Relative frequency of pe- ripheral odontogenic tumors: a study of 45 new cases and com- parison with studies from the literature. J Oral Pathol Med 2006; 35: 385-91.

64. Chindasombatjaroen J, Poomsawat S, Kakimoto N, Shimamoto H. Calcifying cystic odontogenic tumor and adenomatoid odonto- genic tumor: radiographic evaluation. Oral Surg Oral Med Oral Pathol Oral Radiol 2012; 114: 796-803.

65. Chindasombatjaroen J, Poomsawat S, Klongnoi B. Calcifying cys- tic odontogenic tumor associated with other lesions: case report with cone-beam computed tomography findings. Oral Surg Oral Med Oral Pathol Oral Radiol 2012; 113: 414-20.

66. Poleti ML, Duarte BG, Lara VS, Freitas-Faria P, Rubira-Bullen IR, Goncales ES. Odontoma associated with calcifying cystic odonto- genic tumor in deciduous dentition: case report. Oral Maxillofac Surg 2013; 17: 77-80.

67. Soares EC, Costa FW, Neto IC, Bezerra TP, do Socorro Vidal Pa- trocínio RM, Alves AP. Rare hybrid odontogenic tumor in a 2-year-old child. J Craniofac Surg 2011; 22: 554-8.

68. Chaubey SS, Mishra SS, Degwekar SS, Chaubey S. A rare presen- tation of hybrid odontogenic tumor involving calcifying cystic odontogenic tumor and plexiform ameloblastoma. Contemp Clin Dent 2013; 4: 406-8.

69. Crivelini MM, Felipini RC, Coclete GA, Soubhia AM. Immunoex- pression of keratins in the calcifying cystic odontogenic tumor epi- thelium. J Oral Pathol Med 2009; 38: 393-6.

70. Poomsawat S, Punyasingh J, Vejchapipat P. Expression of base- ment membrane components in odontogenic tumors. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2007; 104: 666-75.

71. Chen J, Aufdemorte TB, Jiang H, Liu AR, Zhang W, Thomas HF. Neoplastic odontogenic epithelial cells express bone sialoprotein. Histochem J 1998; 30: 1-6.

72. Siar CH, Kawakami T, Buery RR, et al. Notch signaling and ghost cell fate in the calcifying cystic odontogenic tumor. Eur J Med Res 2011; 16: 501-6.

73. Andrade FR, Sousa DP, Mendonça EF, Silva TA, Lara VS, Batista AC. Expression of bone resorption regulators (RANK, RANKL, and OPG) in odontogenic tumors. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008; 106: 548-55.

74. Karathanasi V, Tosios KI, Nikitakis NG, et al. TGF-beta1, Smad-2/- 3, Smad-1/-5/-8, and Smad-4 signaling factors are expressed in ameloblastomas, adenomatoid odontogenic tumors, and calcify- ing cystic odontogenic tumors: an immunohistochemical study. J Oral Pathol Med 2013; 42: 415-23.

75. Ribeiro AL, Nobre RM, Alves-Junior SM, et al. Matrix metallopro- teinases, tissue inhibitors of metalloproteinases, and growth fac- tors regulate the aggressiveness and proliferative activity of kera- tocystic odontogenic tumors. Oral Surg Oral Med Oral Pathol Oral Radiol 2012; 114: 487-96.

76. Siqueira AS, Carvalho MR, Monteiro AC, Freitas VM, Jaeger RG, Pinheiro JJ. Matrix metalloproteinases, TIMPs and growth factors regulating ameloblastoma behaviour. Histopathology 2010; 57: 128-37.

77. Gong Y, Wang L, Wang H, Li T, Chen X. The expression of NF- kappaB, Ki-67 and MMP-9 in CCOT, DGCT and GCOC. Oral On- col 2009; 45: 515-20.

78. Tsuneki M, Maruyama S, Yamazaki M, Cheng J, Saku T. Podo- planin expression profiles characteristic of odontogenic tumor- specific tissue architectures. Pathol Res Pract 2012; 208: 140-6.

79. Wong YK, Chiu SC, Pang SW, Cheng JC. Peripheral dentinogenic ghost cell tumour presenting as a gingival mass. Br J Oral Maxillo- fac Surg 2004; 42: 173-5.

80. Ramaglia L, Esposito D, Bruno MP, Siano M. Dentinogenic ghost cell tumor: histopathology and clinical aspects in pediatric age. Minerva Stomatol 2012; 61: 509-17.

81. Kumar U, Vij H, Vij R, Kharbanda J, Aparna I, Radhakrishnan R. Dentinogenic ghost cell tumor of the peripheral variant mimick- ing epulis. Int J Dent 2010; 2010: 519494.

82. Mori M, Kasai T, Nakai M, et al. Dentinogenic ghost cell tumor: histologic aspects, immunohistochemistry, lectin binding profiles, and biophysical studies. Oral Oncol 2000; 36: 134-43.

83. Candido GA, Viana KA, Watanabe S, Vencio EF. Peripheral den- tinogenic ghost cell tumor: a case report and review of the litera- ture. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009; 108: e86-90.

84. Kim TJ, Lee YS, Kim BK, Lee KY. Ameloblastoma associated with dentinogenic ghost cell tumor: a case report. Korean J Pathol 2006; 40: 297-302.

85. Tanaka A, Okamoto M, Yoshizawa D, et al. Presence of ghost cells and the Wnt signaling pathway in odontomas. J Oral Pathol Med 2007; 36: 400-4.

86. Kim SA, Ahn SG, Kim SG, et al. Investigation of the beta-catenin gene in a case of dentinogenic ghost cell tumor. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2007; 103: 97-101.

87. Singhaniya SB, Barpande SR, Bhavthankar JD. Dentinogenic ghost cell tumor. J Oral Maxillofac Pathol 2009; 13: 97-100.

88. Sun G, Huang X, Hu Q, Yang X, Tang E. The diagnosis and treat- ment of dentinogenic ghost cell tumor. Int J Oral Maxillofac Surg 2009; 38: 1179-83.

89. Park HR, Min JH, Huh KH, et al. Distant metastasis of intraosse- ous dentinogenic ghost cell tumour to the donor site of a bone graft. Dentomaxillofac Radiol 2013; 42: 20120172.

90. Li BB, Gao Y. Ghost cell odontogenic carcinoma transformed from a dentinogenic ghost cell tumor of maxilla after multiple recur- rences. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009; 107: 691-5.

91. Li BH, Cho YA, Kim SM, Kim MJ, Hong SP, Lee JH. Recurrent odontogenic ghost cell carcinoma (OGCC) at a reconstructed fibu- lar flap: a case report with immunohistochemical findings. Med Oral Patol Oral Cir Bucal 2011; 16: e651-6.

92. Arashiyama T, Kodama Y, Kobayashi T, et al. Ghost cell odonto- genic carcinoma arising in the background of a benign calcifying cystic odontogenic tumor of the mandible. Oral Surg Oral Med Oral Pathol Oral Radiol 2012; 114: e35-40.

93. Lu Y, Mock D, Takata T, Jordan RC. Odontogenic ghost cell carci- noma: report of four new cases and review of the literature. J Oral Pathol Med 1999; 28: 323-9.

94. Cheng Y, Long X, Li X, Bian Z, Chen X, Yang X. Clinical and ra- diological features of odontogenic ghost cell carcinoma: review of the literature and report of four new cases. Dentomaxillofac Radi- ol 2004; 33: 152-7.

95. Kim HJ, Choi SK, Lee CJ, Suh CH. Aggressive epithelial odonto- genic ghost cell tumor in the mandible: CT and MR imaging find- ings. AJNR Am J Neuroradiol 2001; 22: 175-9.

96. Nazaretian SP, Schenberg ME, Simpson I, Slootweg PJ. Ghost cell odontogenic carcinoma. Int J Oral Maxillofac Surg 2007; 36: 455-8.

97. Gong YL, Wang L, Chen XM, Wang HK, Wang XH. Expression of nuclear factor-kappaB, Ki-67 and matrix metalloproteinase-9 in calcifying odontogenic cyst. Zhonghua Kou Qiang Yi Xue Za Zhi 2006; 41: 627-30.

98. Kim J, Lee EH, Yook JI, Han JY, Yoon JH, Ellis GL. Odontogenic ghost cell carcinoma: a case report with reference to the relation between apoptosis and ghost cells. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2000; 90: 630-5.

99. Roh GS, Jeon BT, Park BW, et al. Ghost cell odontogenic carcinoma of the mandible: a case report demonstrating expression of tar- trate-resistant acid phosphatase (TRAP) and vitronectin receptor. J Craniomaxillofac Surg 2008; 36: 419-23.

100. Goldenberg D, Sciubba J, Tufano RP. Odontogenic ghost cell carci- noma. Head Neck 2004; 26: 378-81.