Pathogenesis and Nomenclature
The question of what to call these lesions warrants deliberate consideration and discussion of epidermal inclusion cysts, epidermoid cysts, and various types of keratinizing odontogenic cysts in humans. In dogs, the nailbed epidermal inclusion cyst is the only other type of cyst that contains hard keratin and occurs with some regularity.13 The lining epithelium is most often orthokeratinized and produces densely laminated keratin that can become mineralized. Unlike the KOC, there is a logical source of keratinizing epithelium, i.e., the germinative nailbed epithelium. For these oral keratinizing cysts, the term epidermal inclusion cyst is anatomically inaccurate. The term epidermoid cyst would imply that the lesions are a type of developmental anomaly, which seems unlikely since the KOC occurs most commonly in adult dogs and probably represents a type of acquired odontogenic cyst, not a developmental cyst. There is not a legitimate precedent for the use of odontogenic keratocyst in animals. The two keratin-filled cysts that have been previously published as odontogenic keratocysts in dogs6,7 have features of KOC as described in this case series. One other odontogenic keratocyst was reported in a dog.14 This cyst had fluid contents and the diagnosis of odontogenic keratocyst was previously determined to have been inappropriately classified2 and we agree with this assessment. Veterinary medicine borrowed the term odontogenic keratocyst from human oral pathology with little regard for clear diagnostic guidelines. Over time, the veterinary use of this term has proven to be inappropriate. Odontogenic keratocyst is a specific entity in humans that is often associated with mutation or inactivation of the PTCH1 gene.15 Unless lesions in animals are proven to be analogous, including genetic evaluation, the diagnosis is best avoided in veterinary medicine. Compared to odontogenic keratocysts, the KOC of dogs are more analogous to the orthokeratinized odontogenic cyst (OOC) in humans. Changes in nomenclature over the past several decades have complicated efforts to establish the true prevalence of this OOC in humans, which is considered to be rare. The pathogenesis is unknown and, like KOC in dogs, OOC can mimic other cyst types.16 The posterior mandible is the most common location, which is certainly not true of KOC in dogs. Radiographically, OOC in humans are usually well demarcated and may be either unilocular or multilocular. They are treated by enucleation and recurrence is considered rare.15,17 We strongly favor omitting the ortho prefix and using the term keratinized odontogenic cyst in animals. One reason to diverge from the human nomenclature is to avoid the implication that these lesions in dogs are truly analogous to OOC in humans, which has not been proven and is not our claim. Another important reason is to avoid diagnostic confusion since the KOC in dogs can have orthokeratinized and/or parakeratinized epithelium. We feel that it is more important to recognize production of keratin rather than the pattern of keratinization. This case series demonstrates that, as a group, KOC are more heterogeneous than most odontogenic cysts. It is simply not clear if KOC is a distinct entity or a keratinized variant of another cyst type. Cyst position with respect to the affected tooth/teeth is often one of the most useful features for determining the specific type of an odontogenic cyst. Based on position with respect to the associated teeth, 3 KOCs in this study may have originated as a canine furcation cyst, there is compelling evidence that 1 KOC started as a dentigerous cyst, 5 KOCs had a definite lateral periodontal location, and several had a periapical location. Considering the variable location and presentation of cysts in this case series, we expect that keratinized odontogenic cysts could represent another type of odontogenic cyst that acquired the ability to produce abundant keratin. It is not known why the epithelial lining of some odontogenic cysts might acquire the ability to keratinize. Keratin metaplasia has been reported in various types of odontogenic cysts in humans.18 In additional to several cytokines, calcium can play a role in keratinization of epithelium.19

Limitations
The design of retrospective studies presents certain limitations. In this study, obtaining follow up information and inconsistent diagnostic image quantity and quality were specific challenges. We were unable to obtain any patient information beyond the first post-operative exam for approximately one third of our cases. For the cases we received information about, the majority of these patients did not pursue follow up imaging at the specialty center that performed their surgery, or with their regular veterinarian. Our follow up data therefore includes subjective oral examination findings by the referring veterinarian or by the pet owners. Visual examination does not reveal the subtle changes of an initial recurrence that can be detected with imaging. Without imaging and consistent follow up visits, we cannot make conclusions regarding the rate of growth of KOCs, if aggressive surgery is warranted in all cases of KOC, or if enucleation can either be curative or delay recurrence. While imaging was available for all cases, the type, quantity, and quality varied widely. This variation necessitated that almost all the KOC features evaluated had to include a category for inconclusive cases. A small minority of our cases included three-dimensional imaging. Those cases allowed us to provide more accurate evaluations regarding tooth displacement and margination. Full mouth radiographs were not available in for all cases. Our evaluation of tooth resorption bordering and distant from the KOC would have been more accurate with images of complete patient dentition. Lastly, the amount of tissue available for histopathology was not consistent and often did not include the entire lesion. Samples from enucleated lesions do not have intact tissue architecture and do not allow for evaluation of the relationship of the cyst to surrounding tissues. However, this is a challenge when evaluating all types of odontogenic cysts since en bloc excision is rare. Inability to evaluate the associated teeth was a major limitation, particularly with respect to determining how many of the KOCs may have been a type of keratinized periapical (radicular) cyst.

Conclusions
In this case series, 15 cases (51.7%) had been biopsied previously and evaluated by pathologists at various laboratories. In these biopsy reports, a specific diagnosis was rarely given although the pathologist often recognized features of a keratin-producing cyst. By defining KOC as a specific entity, even with limitations regarding its pathoetiology, we are providing veterinary pathologists and clinicians with a concise and specific diagnosis that will guide diagnostic confirmation, treatment, and client education. The keratinized odontogenic cyst may be either a unique type of odontogenic cyst or a keratinized variant of another recognized type of odontogenic cyst. KOCs occur in the adult canine population and the clinical and radiographic features are variable. For patients where an en bloc resection is not possible or permitted, aggressive enucleation may be appropriate with a cautiously optimistic prognosis. Looking forward, the etiology of the KOC may be elucidated through excellent clinical history, appropriate histopathology of tissues and teeth, and careful documentation of all dental abnormalities and treatments prior to the development of these lesions. References available upon request.