General considerations

SCCs are the most common oral neoplasm of cats, horses, and the production animal species. They are the second most common malignant oral neoplasm of dogs. In humans, SCCs represent 95% of all oral neoplasia. Tobacco and alcohol consumption is considered the most common cause of oral SCCs in developed countries. However, poor oral hygiene and an inadequate diet also increase neoplasia risk.¹ Infection by papillomaviruses has been reported to cause up to a quarter of human oral SCCs.² People with an affected close relative are at increased risk, possibly due to genetically inherited differences in the activation and detoxification of carcinogens.¹

Gross appearance and cytological diagnosis

In all species, oral SCCs initially appear as pale plaques or irregular roughened raised areas in the oral mucosa. As the neoplasm increases in size, it can either protrude, forming a fleshy mass, or appear as a raised plaque with a central area of ulceration. Invasion by the neoplastic cells can distort surrounding structures and ulceration and secondary infection is likely. As most oral SCCs in domestic animals are advanced when first diagnosed, they most often present as ulcerated raised masses that infiltrate surrounding tissues.

Cytology is useful in evaluating focal ulcerative lesions in the mouth of dogs and cats and allows differentiation between a SCC and an inflammatory lesion such as an eosinophilic granuloma or bacterial stomatitis. However, due to the presence of ulceration and secondary infection it can be difficult to obtain diagnostic samples of an oral SCC. To avoid sampling the surface inflammation, deep scrapes or fine‐needle aspirates are recommended. The cytological appearance of an oral SCC is dependent on the differentiation of the neoplastic cells. Cytology from a poorly differentiated neoplasm reveals a highly anaplastic population of cells with few showing evidence of keratinization (angular cells with increased quantities of pale cytoplasm). In contrast, evidence of malignancy can be difficult to find in well‐differentiated SCCs. Most oral SCCs are intermediate between these two extremes and diagnosis is usually made by the identification of clumps of large angular cells showing variable keratinization. Features of malignancy within the cell population such as high variability in cell and nuclear size and appearance can be used to differentiate neoplastic from dysplastic epithelial cells. Neutrophil emperipolesis is common within oral SCCs and can often be detected cytologically. A cytological diagnosis of SCC is supported by other clinical findings such as radiographic evidence of osteolysis.

Histological diagnosis and subtyping of oral SCCs

Histology of an oral SCC reveals a proliferation of epithelial cells. Depending on the appearance of the cells, oral SCCs can be divided into subtypes. Although few pathologists currently routinely subtype oral SCCs from domestic animals, a study of canine oral SCCs revealed 69 conventional, 5 papillary, 5 basiloid, 3 adenosquamous, and 2 spindle cell SCCs.³ An awareness of the subtypes of SCCs is important for two reasons. First, while histological diagnosis of a conventional oral SCC is generally straightforward, other oral SCC subtypes may represent more of a diagnostic challenge and may require IHC for definitive diagnosis. Second, in humans different oral SCC subtypes show different clinical behavior resulting in differences in prognosis. Currently no studies in the domestic species have investigated whether the subtypes show differences in behavior. However, considering the evidence from human studies, is it probable that subtyping will be found to be of prognostic significance in the domestic species.

Conventional SCCs are the most common subtype of oral SCC in all domestic species and appear as trabeculae and nests of epithelial cells extending into the submucosa (Figure 13.3A) Cells at the periphery of the nests are smaller and more basiloid with cells in central areas larger with eosinophilic cytoplasm. Disorderly maturation of neoplastic squamous cells is visible with keratinized cells interspersed between basilar cells and variable numbers of keratin pearls. Areas that contain solid masses of neoplastic cells are often present. Intercellular bridging is present and is most easily identified using ultrathin sections that are stained with toluidine blue. Most conventional SCCs are highly infiltrative with small clusters or individual cells visible separate from the main neoplasm mass. Neoplasms are often ulcerated with necrosis, neutrophils, and bacteria visible within superficial aspects of the SCC. Often the neoplastic cells can be seen merging with the overlying epithelium; however, as the tumors are infiltrative, cells can be present infiltrating under intact gingival epithelium and a lack of connection to the overlying epithelium does not exclude a diagnosis of SCC. In SCCs that invade bone, osteolysis is visible with nests of cells and increased connective tissue present within the remaining spicules of bone (Figure 13.3B). Scattered clusters of neoplastic cells, often surrounded by lymphocytes, are typically present within surrounding muscle and connective tissue. Rarely, the epithelial cells induce a marked fibrous response. Although cellular pleomorphism can be marked, neoplastic cells are generally large and polygonal with eosinophilic cytoplasm. Giant cells and multinucleate cells can be found scattered within some SCCs. Large neoplastic cells often contain intracytoplasmic neutrophils (emperipolesis).

Histology usually allows a definitive diagnosis of a conventional SCC. However, differentiation between an inflammatory lesion and a SCC can be difficult in samples that contain predominantly inflammation with only small numbers of anaplastic epithelial cells. This is often encountered when superficial samples of alveolar bone are taken during tooth extraction. Definitive diagnosis can require deeper biopsies so that epithelial cells that are not associated with the inflammation can be adequately evaluated. IHC using cytokeratin can be useful to demonstrate small numbers of epithelial cells infiltrating within small samples that do not contain an identifiable tumor mass. Differentiation between a conventional SCC and an acanthomatous ameloblastoma (formerly acanthomatous epulis) is necessary. Although both contain cords or nodules of epithelial cells, the presence of keratin is more suggestive of a diagnosis of SCC. Poorly differentiated conventional SCCs can contain little keratin. However, the cells within an acanthomatous ameloblastoma remain well differentiated and show features of odontogenic epithelium such as regular palisading. In contrast, the neoplastic cells within a poorly differentiated SCC are irregularly arranged and have marked anaplasia present within the cell population.

Carcinoma in situ appears as a thickening of the oral epithelium due to a proliferation of dysplastic cells that remain confined by the basement membrane. Although all SCCs are thought to initially develop from this subtype, oral in situ carcinomas are rarely recognized in domestic animals.

Verrucous SCCs appear grossly as exophytic cauliflower‐like tumors. Histologically, they are characterized by the presence of broad tongues of mature squamous epithelium that “push” into underlying tissue rather than infiltrating as in conventional SCCs (Figure 13.4A).⁴ The neoplastic cells typically have little evidence of atypia or other features of malignancy. In humans, verrucous SCCs have a more favorable prognosis than other oral SCC subtypes and are considered to have little, if any, metastatic potential. Verrucous SCCs have been reported in the mouth of a young dog as well as in two older dogs.⁵ Although surgical excision of one canine verrucous SCC was curative, rapid neoplasm recurrence followed by euthanasia was reported in the two older dogs and it cannot be confirmed that oral verrucous SCCs of dogs and cats have a less aggressive clinical behavior than other SCC subtypes.

Papillary SCCs have been reported in the oral cavity in dogs and appear as proliferative exophytic friable masses. The neoplastic cells are predominantly confined to the epithelium resulting in marked papillomatous folding (Figure 13.4B). The neoplastic epithelial cells are supported by a fibrovascular stalk and can be basilar in appearance or undergo keratinization resulting in the development of keratin pearls. The key diagnostic feature is the presence of invasion of cells into the underlying submucosa or supporting fibrovascular stalk. Infiltration by the neoplastic cells can be difficult to identify due to the complex filiform architecture of the SCC. However, the presence of infiltration is critical to differentiate a papillary SCC from a squamous papilloma.⁴ Due to the exophytic nature of this subtype, complete excision is more likely to be possible and this subtype has a more favorable prognosis in people. Currently there is no evidence to suggest that a papillary SCC develops as progression from either a squamous or a viral oral papilloma.

Basiloid SCCs have also been reported in the canine oral cavity and cells within this subtype remain undifferentiated and resemble cells within the basal layer of the epithelium. Keratinization is not prominent although is present in most neoplasms (Figure 13.4C). Neoplastic cells can form solid lobules infiltrating into the underlying submucosa. Cells can palisade around the periphery of the lobules and gland‐like spaces can be present.⁴ Although the palisading of the cells can appear histologically similar to an acanthomatous ameloblastoma, a basiloid SCC can be differentiated by the predominance of small dark basilar neoplastic cells rather than the larger, more eosinophilic cells present within an acanthomatous ameloblastoma. In addition, cells within a basiloid SCC will generally be more anaplastic than those of an acanthomatous ameloblastoma. Basiloid SCCs have a worse prognosis than conventional oral SCCs in people.

Spindle cell SCCs consist of a proliferation of elongated cells arranged within whorls or bundles (Figure 13.4D). Histological differentiation between a spindle cell SCC and a sarcoma or melanoma can be difficult; however, examination of cells within the infiltrating border of the SCC will often reveal foci of squamous differentiation. Although the IHC detection of cytokeratin expression confirms a diagnosis of SCC, up to a third of human oral spindle cell SCCs do not contain cytokeratin immunostaining.⁶ In these cases IHC (as described elsewhere in this chapter) to exclude a fibroblastic, neural, or melanocytic origin is required with a failure to detect the expression of any tumor marker considered supportive of a spindle cell SCC. This subtype is associated with a worse prognosis than conventional SCCs in people.⁴ Oral spindle cell SCCs have been reported in dogs.

Other histological subtypes that have been described in domestic animals include adenosquamous and adenoid (also known as acantholytic) SCCs. Both subtypes are best considered variants of conventional SCCs and are comprised predominantly of infiltrative trabeculae of anaplastic keratinizing epithelial cells. In humans, these two subtypes have a similar biological behavior to conventional SCCs. Adenosquamous SCCs have the additional feature of containing foci of glandular differentiation (Figure 13.4E). Adenoid SCCs contain large spaces that are lined by neoplastic squamous cells and contain acantholytic cells (Figure 13.4F).⁴ Both subtypes could be mistaken for an oral or salivary adenocarcinoma. Staining with alcian blue to detect the presence of mucosubstances within an adenocarcinoma may help with differentiation. A pseudoangiomatous variant of an adenoid SCC contains prominent blood‐filled clefts and may histologically resemble a hemangiosarcoma. The presence of keratinizing epithelial cells will allow diagnosis of a SCC; however, IHC may be required to definitively exclude a hemangiosarcoma in neoplasms that lack this feature.⁷

Prognostic indicators of oral SCCs

In humans, the clinical stage of an oral SCC is considered to be the strongest prognostic indicator. Other factors that have been reported to be prognostic include the histological subtype of the SCC, the location in the mouth (rostral neoplasms have a better prognosis), the maximum thickness of the neoplasm, the pattern of invasion of the neoplastic cells, lymphocyte infiltration of the neoplasm, and perineural and vascular invasion by the neoplastic cells.⁸

Histological evaluation of the degree of keratinization, cellular and nuclear pleomorphism, and mitotic rate can be used to grade conventional oral SCCs into well, moderately, or poorly differentiated. Limitations of histological grading include the heterogeneity present within SCCs (especially important when only small samples are available) and the subjective nature of the assessment. Due to these limitations, the histological grade of a human conventional oral SCC is has not been consistently found to be prognostic.

Numerous antibodies have been investigated as IHC markers of prognosis in human oral SCCs.⁹ Of these, IHC to detect increased cellular p16^CDKN2A protein (p16) is well established to indicate a papillomaviral etiology and a better prognosis.² Although p16 immunostaining was found to predict prognosis in feline cutaneous SCCs,¹⁰ there is little evidence that papillomaviruses cause oral SCCs in domestic animals, suggesting p16 immunostaining will not be prognostic. Antibodies against epidermal growth factor receptor, p53, and matrix metalloproteinases have also been extensively studied in human oral SCCs, although none has been found to consistently predict prognosis.

Prognostic indicators for oral SCCs in domestic animals are less established and are discussed in more detail in the descriptions of canine and feline SCCs. In contrast to human oral SCCs, staging has only inconsistently been associated with survival times in canine and feline oral SCCs. When examining histology samples of oral SCCs it should be determined whether or not the neoplasm has been completely excised. The subtype of the SCC should also be determined. If the SCC is a conventional subtype, the neoplasm can be graded histologically. When a series of 69 conventional canine oral SCCs were graded, 33 were classified as well differentiated, 31 as moderately differentiated, and 5 as poorly differentiated.³ Although there have been few studies to determine whether the histological grade is prognostic, it is possible that the subtype and histological grade may be useful to predict tumor behavior, especially in cases in which complete clinical staging is not possible.

Location, signalment, and etiology

Canine oral SCCs develop most frequently on the gingiva with the maxillary and mandibular gingiva involved roughly equally (Figure 13.5). Less common locations for canine oral SCCs include the lips (6%), hard or soft palate (3%), and pharynx (2%).¹¹ SCCs of the tongue and tonsil are discussed separately. The average age of dogs with oral SCCs is 9 years with a range of 1.2–14 years.¹² Larger dogs develop oral SCCs more frequently than small dogs, and English springer spaniels, Shetland sheepdogs, and German shepherds may be predisposed.^3,13 The cause of most canine oral SCCs is unknown, although intraoral radiation therapy may rarely predispose to neoplasm development. As discussed with oral papillomatosis, papillomaviruses do not appear to be a significant cause of canine oral SCCs.

Biological behavior, treatment, and prognosis

Canine oral SCCs are invasive, with evidence of bone invasion detectable at diagnosis in around 70% of cases. Neoplasms are slow to metastasize, with fewer than 15% of dogs having detectable nodal metastases and fewer than 5% of dogs having radiographic evidence of pulmonary metastases at diagnosis. Due to this slow metastasis, many canine oral SCCs are considered curable. If euthanasia is necessary, this is usually due to pain and dysfunction caused by SCC invasion rather than metastatic disease.^11–13 Currently, no associations between the biological behavior and prognosis of a canine oral SCC has been established with SCC subtype, histological grade, or a molecular marker.

There are numerous treatment options for canine oral SCCs, including surgical excision, radiation therapy, chemotherapy, and photodynamic therapy. Surgical excision of the SCC without removing the adjacent bone is often not curative and local recurrence was observed in 5 (63%) of 8 dogs treated this way. These dogs had a median survival time (MST) of 9 months and a 1‐year survival of 38%.¹² In comparison, local recurrence was observed in just 2 (8%) of 24 dogs with oral SCCs excised by partial mandibulectomy. In this study 3 of the 24 dogs died of oral SCCs, 2 dogs had local recurrence and metastases, while 1 dog developed metastases without recurrence of the primary neoplasm. Overall, a disease‐free interval of 26 months and a 91% 1‐year survival was reported. Younger dogs had a better prognosis; however, neither the size of the neoplasm nor the presence of bone invasion was significantly associated with survival time.¹⁴ Hemimaxillectomy was used to treat oral SCCs in 7 dogs. Two (29%) SCCs recurred and the dogs had an MST of 19.2 months and a 57% 1‐year survival. Neither the size of the SCC nor the location of the SCC within the maxilla had prognostic significance in this study.¹⁵

Radiation therapy can be used as a sole treatment or as an adjunct with surgery. When used as a sole treatment MSTs of 9–15 months have been reported.^12,16,17 Two studies reported that the age of the dog was prognostic and a rostral location of the tumor was associated with longer survival in one study, but not in another. A study of 39 irradiated oral SCCs reported a median progression‐free time of 36 months with the size of the SCC, but not with the anatomic site of the SCC, the age of the dog, or the presence of bone involvement associated with progression‐free time.¹⁸

Small numbers of dogs with oral SCCs have been treated using chemotherapy. Tumor remission was observed in 5 of 9 dogs in one study and 4 of 7 in another.^19,20 Photodynamic therapy has also been used to treat canine oral SCCs and no tumor recurrence was observed in 8 of 11 dogs receiving this treatment. One of the dogs treated with photodynamic therapy subsequently developed metastatic disease.²¹

Overall, these studies suggest that canine oral SCCs are locally aggressive neoplasms that frequently invade bone, but metastasize late in the clinical course. Surgical excision is curative with a potential for long survival times with appropriate treatment of nonresectable SCCs. Due to conflicting data, it is currently unknown whether the age of the dog, size of the tumor, or location of the SCC in the mouth significantly influences prognosis.

Location, signalment, and etiology

SCCs comprise the overwhelming majority of oral neoplasia in cats and oral SCCs are the fourth most common feline neoplasm.^22,23 Feline oral SCCs most commonly develop within the mandibular, maxillary, and sublingual regions and neoplasia develops within these three locations at approximately equal rates.^23–25 Sublingual SCCs often develop on the floor of the oral cavity close to the frenulum and can invade the ventral surface of the tongue (Figure 13.6).²³ Less commonly, SCCs develop on the hard palate, soft palate, larynx, pharynx, and lip, although each location represents less than 2% of feline oral SCCs.²⁵ The average age for neoplasm development is around 13 years with a range of 1.5–22 years.²⁵ No sex, coat color, hair length, or breed predispositions have been identified.^25,26

The cause of feline oral SCCs is unknown. Despite the strong association between tobacco use and oral SCCs in humans, exposure to environmental tobacco smoke did not significantly increase oral SCC risk in an epidemiological study of 36 cats.²⁶ Likewise, although papillomaviruses cause human oral SCCs, these viruses are rarely detectable in feline oral SCCs and there is no evidence of a causative association.^27,28 Neither feline immunodeficiency virus (FIV) nor feline leukemia virus infection (FeLV) has been shown to significantly increase the risk of oral SCC. Dental disease is a risk factor for human oral SCCs, with tartar, caries and tooth loss all associated with increased neoplasia development.¹ As dental disease and tooth loss is ubiquitous in older cats this could be a factor in feline oral SCC development. Canned food has been associated with increased risk of feline oral SCCs.²⁶ As canned food is associated with higher rates of dental disease in cats, this could be the link between canned food and oral SCCs in cats.

Biological behavior

Feline oral SCCs are highly invasive but have a low metastatic potential. In a study of 52 cats with oral SCCs, bone invasion was present at diagnosis in 38 (73%).²⁹ Regional lymphadenopathy was detected in 15 cats, although the presence of nodal metastasis was confirmed microscopically in just 7 (13%) cats. Detectable distant metastases were not present in any cat at diagnosis.²⁹ In a study of 18 feline oral SCCs, bone invasion was present at diagnosis in 9 (50%) of the cats with cytologically confirmed nodal metastases in 5 (28%).³⁰ Feline oral SCCs cause necrosis of surrounding tissues which is often worsened by ulceration and secondary bacterial infection. Cats are almost invariably euthanized due to local disease resulting in anorexia, dysphagia, or dyspnea rather than the development of clinically significant metastases. Hypercalcemia was reported in two cats with SCCs of the mandibular gingiva.

Treatment and prognosis

Survival times in cats with oral SCCs are short with most studies reporting MSTs of 44–60 days and 1‐year survival rates of 5–10%.^24,29–31 However, two studies in which oral SCCs were excised by mandibulectomy reported MSTs of 217³² and 420 days³³ with SCC recurrence observed in 8 of 21 (38%) cats and a 43% 1‐year survival rate.³² Feline oral SCCs have historically been thought to be poorly responsive to both radiation therapy and chemotherapy. However, there is some evidence that survival times can be extended (MST of 163 days in a group of 31 cats) using accelerated radiation therapy and carboplatin, although no controlled studies have been performed.³⁴

Due to the generally short survival times observed in cats with oral SCCs, few prognostic indicators have been established. Cats with maxillary SCCs were reported in one study to survive longer than cats with SCCs in other locations within the mouth.³⁰ In another study, 9 cats with metastases to multiple lymph nodes or distant metastases had an MST of 24 days, while 35 cats without such metastases had an MST of 90 days, although whether survival times were significantly different was not reported. Several studies reported that neither the size of the SCC nor the location of the SCC in the mouth was associated with survival times.^29,30,32 Thirty‐six cats with oral SCCs that had diffuse cyclooxygenase (COX)‐1 immunostaining survived significantly longer than 5 cats with SCCs that had patchy COX‐1 immunostaining.²⁴ A study of 67 cats with oral SCCs revealed that cats with neoplasms with a low Ki67 proliferation index survived significantly longer than those with SCCs that had a high index.³¹ However, this association was not observed in a subsequent study of 21 cats with oral SCCs.³⁵ No significant association between epidermal growth factor receptor expression and survival was detected in either study.^31,35

Neuroendocrine carcinomas

Neuroendocrine carcinomas within the oral cavity are rare in the domestic animals. These tumors have previously been referred to as oral carcinoids and oral Merkel cell tumors. As with neuroendocrine carcinomas elsewhere, neoplastic cells are arranged in cell nests supported by a delicate fibrovascular stroma. Individual cells tend to be round with pale basophilic cytoplasm and variably sized oval nuclei. Definitive diagnosis is by IHC (see section on Neuroendocrine carcinomas (intestinal carcinoids)) or by the demonstration of neuroendocrine granules using electron microscopy. In a series of 4 canine oral neuroendocrine carcinomas, surgical excision was curative for 3 tumors with the remaining tumor successfully treated using radiation therapy.³⁷

General considerations

Fibrosarcomas are the third most common oral malignancy in dogs and the second most common in cats. Fibrosarcomas are rare in the oral cavity of horses and these neoplasms represented just 2 of 29 equine oral malignancies in one survey.² No cause of fibrosarcoma development can be identified in most cases in companion animals. In contrast, 9 (12%) oral fibrosarcomas were detected in a survey of 75 tumors from sheep grazing bracken fern. Oral fibrosarcomas have been rarely reported in other surveys of sheep tumors, supporting a causative association between bracken fern and ovine oral fibrosarcoma.³ In sheep, oral fibrosarcomas frequently infiltrate adjacent bone, but metastases are rare and limited to regional lymph nodes. No oral fibrosarcomas were identified in a retrospective study of tumors in cattle that were grazing bracken fern.¹

Oral fibrosarcomas are usually solitary, pale lesions that, compared to other oral malignancies, are less likely to be ulcerated (Figure 13.7). In contrast to fibroma, fibrosarcomas are not well delineated and are typically infiltrative with destruction of surrounding tissue present at the time of diagnosis. The resulting osteolysis can be assessed using radiography, ultrasonography, or newer techniques such as computerized tomography, and this is useful for accurate clinical staging of the neoplasm as well as determining necessary surgical margins. If adequate cells can be harvested, cytology may reveal spindle‐shaped cells with variable degrees of cellular atypia. The identification of multinucleate cells and numerous mitoses is more supportive of a diagnosis of fibrosarcoma than fibroma although definitive differentiation is dependent on the presence of an infiltrative growth pattern visible on histology.

Histology reveals moderately to poorly differentiated large spindle‐shaped cells that are arranged in interlacing bundles separated by small amounts of collagenous matrix (Figure 13.8A). Compared to a fibroma, fibrosarcomas are expected to contain less cellular differentiation and more frequent mitotic figures. Additionally, necrosis is often present within the neoplasm and the presence of an infiltrative growth pattern within the deeper aspects of the mass will confirm a diagnosis of fibrosarcoma. Differentiation between an odontogenic tumor and a fibrosarcoma is usually easy if odontogenic epithelium is visible. However, if no odontogenic epithelium is visible, the location of the mass away from the dental arcade is useful to exclude odontogenic origin. Presence of ossification favors diagnosis of osteosarcoma. Differentiation from an amelanotic or poorly melanotic melanoma dominated by spindle cells can be problematic and may require IHC using melanocytic markers.

Oral tumors of smooth muscle

Oropharyngeal tumors of smooth muscle are rare in domestic animal, but may be underreported as differentiation from fibroma and fibrosarcoma may prove impossible using H&E‐stained slides alone. Histologically, smooth muscle tumors comprise spindle‐shaped cells arranged in streams and interlacing bundles. Extracellular matrix and cross‐striation are absent. The presence of a uniform population of neoplastic cells that have small nuclei and few mitotic figures is suggestive of a leiomyoma rather than a leiomyosarcoma. IHC to detect the expression of smooth muscle actin can be used to confirm smooth muscle differentiation. Although desmin expression suggests muscle origin, immunostaining can be present in both smooth and striated muscle neoplasms. Absence of actin–myosin bundles (cross‐striations) or collagen fibers on electron microscopy can also be used to confirm a smooth muscle origin of the neoplasm, although this technique is generally not used in a diagnostic setting.

Successful surgical excision of a canine oropharyngeal leiomyoma has been described.²¹ Recurrence after initial resection of an angioleiomyoma of the palate of a young dog was reported, but this tumor did not recur after subsequent complete excision.

Oral leiomyosarcomas are rarely reported in domestic animals. In dogs, leiomyosarcomas occur in the tongue, gingiva of the rostral maxilla or mandible, upper lip, and soft palate.²¹

Oral tumors of striated muscle

Oral tumors of striated muscle are uncommon but have been reported in dogs, cats, and horses. They are most frequently located in the tongue and less commonly the laryngopharynx, cheek and gingiva.

Oral rhabdomyomas have only been described in the tongue and laryngopharynx of dogs.²⁰ They are well demarcated and contain few mitotic figures. Rhabdomyomas may appear histologically similar to granular cell tumors or oncocytomas on routine histology,²⁰ but can be differentiated using IHC for myoglobin or other muscle‐specific markers or by the demonstration of actin–myosin bundles by electron microscopy.

As discussed with tumors of muscle, rhabdomyosarcomas can be histologically subdivided into types, and embryonal, alveolar, and pleomorphic rhabdomyosarcoma types have all been reported in the oral cavity of domestic animals. Recognition of the histological subtype aids in making a diagnosis, but has not been shown to influence prognosis. Poorly differentiated rhabdomyosarcomas may lack histologically identifiable cross‐striations and therefore IHC or electron microscopy may be required for diagnosis. Expression of markers of striated muscle varies between and within tumors, recapitulating the progression of differentiation during normal muscle development. Antibodies that can be useful include MyoD, an early myogenic regulating factor, myogenin, myoglobin, and alpha‐sarcomeric actin. However, IHC evidence of a skeletal muscle origin may be difficult to detect in some poorly differentiated rhabdomyosarcomas.

In dogs, oral rhabdomyosarcomas are generally solitary, although one dog developed multiple neoplasms within the oral cavity.²² Rhabdomyosarcomas have been reported in the gingiva, tongue, pharynx, and hard palate. Affected dogs are generally younger than those with other oral neoplasms with many less than 2 years old at diagnosis.²³ Most canine oral rhabdomyosarcomas are of the embryonal subtype with alveolar rhabdomyosarcomas less frequently reported.²³ Surgical resection is the recommended therapy, although both tumor recurrence and metastases occur.

In a study of 8 equine rhabdomyosarcomas, 2 were in the tongue and 2 in the cheek.²⁴ Surgical excision was curative in all three horses in which it was attempted. Most rhabdomyosarcomas in this study were of the embryonal subtype. A pleomorphic rhabdomyosarcoma was diagnosed in the masseter area of one horse and this subtype has been reported also in the equine tongue. An embryonal rhabdomyosarcoma involving the maxilla and hard palate was described in a goat (Figure 13.10).²⁵

General considerations

Melanocytic neoplasms are the most common malignant neoplasm in the oral cavity of dogs.^1–7 Despite being so common, some can present a diagnostic challenge due to their poor pigmentation and cellular variability.^6,8,9 In general, melanocytic neoplasms of the oral cavity are more aggressive and have shorter MSTs than cutaneous melanocytic neoplasms.^8,9 Historically, all lip and oral melanocytic neoplasms were considered to be highly malignant. However, this dogma has recently been challenged with the identification of a subset of canine oral melanocytic tumors that have a more favorable prognosis.^7,8,10,11 Although no canine oral melanocytic neoplasm should ever be unequivocally classified as benign, they can be subclassified as lowly or highly malignant melanomas. In one study, 8% of canine oral melanomas were classified as low malignancy and 92% as highly malignant. However, the true incidence of low malignancy neoplasms may be higher, as only 59% of the tumors classified as highly malignant subsequently recurred or metastasized after surgical excision.⁹ A study of 71 dogs with histologically well‐differentiated melanocytic neoplasms (melanocytic neoplasms of low malignant potential) located in the mucous membranes of the lips and oral cavity revealed that surgical excision resulted in an MST of approximately 3 years with some dogs surviving almost 4 years after diagnosis.¹⁰

Incidence and signalment

There is no sex or well‐established breed predisposition for the development of oral melanocytic neoplasms.¹² However doberman pinschers and miniature and giant schnauzers have been suggested to have a higher incidence of oral melanocytic tumors and these tumors may be more likely to be of low malignancy.² Other breeds that have been overrepresented in studies of oral melanocytic neoplasms include golden retrievers, Labrador retrievers, and cocker spaniels, although this may be due to a higher frequency of these breeds in the study population.¹² Most melanocytic neoplasms occur in older dogs. Dogs with low malignancy oral melanomas are reported to have a mean age of approximately 8 years, whereas dogs with highly malignant oral melanomas have a mean age of approximately 12 years.^1–4,7,12 The relative risk of developing oral melanocytic neoplasia increases with age more than with other oral malignancies. Dogs with oral melanocytic neoplasms are usually older than dogs that develop cutaneous melanomas.^2,13

Canine oral melanocytic tumors most frequently develop on the gums and lips (Figure 13.12A).^1,3,5,7,12 Occasionally, asymptomatic nodules less than 1 cm in size are found during dentistry. Many of these smaller lesions are pedunculated and are consistent with low malignancy tumors.¹⁰ Larger lesions (often 3–4 cm diameter) cause clinical signs of oral disease, are sessile, and often have an ulcerated surface (Figure 13.12B). Gingival tumors tend to be oval, molded by the anatomy of the jaw, and immobile. The surface may be black (Figure 13.12C), but white or pink mucosa can overlie pigmented tumors, and a red granulation tissue reaction secondary to ulceration may mask melanin pigment. The tumor consistency is firm unless necrosis and secondary infection have caused softening. Some tumors are uniformly black on cut surface, but more often there are foci with less pigment, and these areas are brown, gray, or white (amelanotic). Amelanotic neoplasms may give rise to pigmented secondaries and vice versa. The majority of oral melanocytic neoplasms grow as a sessile mass, but some may be pedunculated and grow from a relatively narrow stalk. Such pedunculated tumors tend to behave less aggressively although it is uncertain whether this is because of the biologic nature of the neoplasms or because they are easier to completely excise.

Microscopic diagnosis

The identification of neoplastic cells containing melanin pigment will allow a cytological diagnosis of oral melanocytic neoplasm in many cases. However, due to the intratumor variation in neoplastic cell appearance, cytology should not be used to predict the behavior of an oral melanocytic neoplasm. Cytology of a poorly pigmented melanocytic tumor will often allow a diagnosis of a malignant neoplasm; however, additional histological studies are required for definitive diagnosis of a melanoma and more accurate prognostication. Aspiration of regional lymph nodes can be used to detect early metastatic spread. Approximately half of oral melanomas first metastasize to lymph nodes other than the mandibular lymph nodes and multiple nodes should be evaluated for nodal metastasis.¹⁴

There is currently some confusion regarding the use of terms used to describe melanocytic tumors in animals. In human pathology, melanocytic tumors have been extensively studied and the terminology has been well defined. Therefore, we propose that the terms used to describe the appearance and behavior of melanocytic neoplasms in domestic animal species should be equally well defined. In human pathology, a dermal melanocytic tumor is confined to the dermis with no epidermal involvement. A compound melanocytic tumor is defined as a benign melanocytic neoplasm with both epidermal and dermal components. In contrast to human medicine, the designation “compound” has been used in veterinary medicine for benign and malignant melanocytic tumors. Preferably, a malignant melanocytic neoplasm with both epidermal and dermal components should be described as an invasive melanoma. Dermal invasion is also often described as the vertical growth phase. In mucocutaneous melanomas the terms epithelial and submucosal or parenchymal are used instead of epidermal and dermal. Other features often described in oral melanocytic neoplasms include extension of the neoplastic cells to the dermo‐epidermal junction, junctional activity, lentiginous spread, and pagetoid growth. Extension of neoplastic cells to the dermo‐epidermal junction appears as the presence of neoplastic melanocytes within the superficial dermis, without these cells penetrating the basement membrane (Figure 13.13A). Junctional activity is defined as the presence of aggregates, nests or individual neoplastic melanocytes in the basal layer of the epidermis without a dermal component (Figure 13.13B). Previously junctional activity has been loosely used in descriptions of veterinary melanocytic tumors and we believe that the rigid application of the human definition of junctional activity will decrease confusion about this term in veterinary pathology. Lentiginous spread is defined as intraepidermal proliferation of migrating melanocytes which preferentially reside among basal keratinocytes. Theoretically, junctional activity may be differentiated from lentigenous spread by the cohesive nature of the neoplastic melanocytes preventing lateral spread. However, as migration cannot be assessed in histologic sections, both terms are often used interchangeably in human medicine, with lentiginous spread being used more commonly to clinically characterize a melanocytic neoplasm. Intraepidermal growth in melanomas is also often referred to as radial growth. Although junctional activity is a strong diagnostic criterion to differentiate melanocytic neoplasms from other tumor entities, it may be difficult to identify individual intraepidermal neoplastic melanocytes in H&E‐stained slides, especially if the neoplastic cells are not pigmented. Additionally, differentiation between hyperplastic and neoplastic intraepidermal melanocytes can be problematic. Features that indicate neoplasia include retraction artifacts around intraepithelial melanocytes as well as cellular atypia, especially variation in nuclear size, altered chromatin patterns, and nuclear pleomorphism. Lentiginous lateral spread is also an important feature to evaluate for prognostic purposes in order to accurately determine tumor margins, since it occurs away from the primary mass. Pagetoid spread is used in human melanocytic pathology to describe “upward spreading” of melanocytes into the epidermis; neoplastic melanocytes display pagetoid growth when they invade the upper epidermis from the stratum basale (Figure 13.13C). Although the term nevus was originally defined as a birthmark in humans, it has been used to describe benign melanocytic lesions in both human and veterinary pathology. Since this term lacks a good definition, it should be avoided when describing melanocytic neoplasms.

Canine oral melanocytic neoplasms of low malignant potential most closely resemble the common variant of blue nevi in people and have been described as histologically well‐differentiated melanocytic neoplasms.^10,15 They tend to be heavily pigmented, non‐ulcerated, typically have a diameter of 1.0 cm or less, and are often raised or pedunculated (Figure 13.13D).¹⁰ Neoplastic cells are located in the mid and upper subepithelial parenchyma and commonly form symmetric, wedge‐shaped lesions with variably fibrotic stroma. Pigment granules may obscure the nucleus, requiring bleaching with 1% potassium permanganate. The neoplastic cells tend to be quite uniform in size and are round or elongated and contain a small round nucleus that often has a small, single, centrally placed nucleolus (Figure 13.13E).¹⁰ There is little nuclear atypia and the mitotic count is low, with less than 4 mitoses per 10 HPFs. Discrete intraepithelial nests of neoplastic melanocytes are often not visible in these neoplasms. However, junctional activity or lentiginous spread can be present in some melanocytic neoplasms of low malignant potential and may indicate an early stage of malignant transformation.

Highly malignant oral melanomas are most often compound (invasive) melanomas that are characterized by junctional activity and deep nodular expansion into the underlying parenchyma (Figure 13.13F). The intraepithelial component is believed to be the primary site of neoplastic cell proliferation and lentiginous growth is also responsible for lateral spread. This pattern closely resembles human mucocutaneous melanomas of the rectum or nasopharynx.^6,7,15 Although these tumors appear grossly as well‐circumscribed solitary nodules, histology often reveals intraepithelial proliferation of discrete nests of often heavily pigmented neoplastic cells that may extend laterally for more than double the diameter of the grossly recognizable tumor. In invasive malignant melanomas, intraepithelial neoplastic cells are often better differentiated than the submucosal neoplastic melanocytes and may be the only cells that contain melanin pigment or show immunolabeling for Melan‐A or PNL‐2.⁸ Even with the submucosal component being negative for Melan‐A or PNL‐2, if the intraepithelial neoplastic cells are positive, then the diagnosis should still be melanoma. Due to the presence of these more differentiated, lentiginous spreading neoplastic cells, submitting surgical biopsy samples from the edge of lesions that include intact mucosal epithelium, rather than only an ulcerated surface over the center of a nodule, will increase the likelihood of differentiating a melanoma from a spindle cell sarcoma.⁸ Unlike the fairly uniform intraepithelial neoplastic cells, cells in the deeper portion of the tumors often show greater variation in the size and shape of both cytoplasm and nuclei. The deeper portion of the tumor is often divided into lobules, and the cells are supported by minimal collagenous stroma. Oral melanomas can be subdivided into subtypes using differences in cell morphology. The epithelioid or polygonal type consists of closely packed round or polyhedral cells with abundant cytoplasm, well‐defined borders, and large central nuclei with one or more prominent nucleoli (Figure 13.14A). In the spindloid or fibromatous type the nuclei are ovoid or elongate and have small nucleoli (Figure 13.14B). The third type is mixed and is characterized by both epithelioid and spindloid areas. A melanocytic neoplasm should be included in the list of differentials for any oral tumor with mixed epithelioid and spindloid features, regardless of the degree of pigmentation.

Additional rare types, such as whorled type (dendritic), balloon cell type, signet ring cell type, clear cell type and adenoid/papillary cell types have also been reported. Melanocytic neoplasms with osteocartilaginous differentiation occur rarely in the oral cavity and have to be differentiated from osteo‐ or chondrosarcomas (Figure 13.15A).¹⁶

Additional diagnostic tests

Pigmentation is an important diagnostic feature of oral melanocytic neoplasms and highly pigmented tumors are easily recognized grossly and microscopically. However, diagnosis of tumors with little or no pigment can be more problematic. Although histological features such as junctional activity or pagetoid spread support a diagnosis of a melanocytic neoplasm; ultimate confirmation requires identification of cells containing melanin in H&E‐stained slides or IHC.^6–9 Antibodies that are highly specific for the diagnosis of oral melanocytic neoplasms include Melan‐A (Figure 13.15B), PNL‐2 (Figure 13.15C), and tyrosine reactive proteins 1 and 2. A “cocktail” containing all four antibodies provides the highest diagnostic sensitivity of canine melanocytic neoplasms.^5,8,16–20 In contrast, antibodies against S100 and MITF‐1 are highly sensitive (almost all melanomas are positive) but have a low specificity (many other tumors are positive).^5,8

Antibodies commonly used for the detection of human melanocytic lesions, such as HMB‐45 and tyrosinase, have a high specificity but very low sensitivity to detect canine oral melanocytic neoplasms.⁸ In contrast to melanocytic neoplasms, fibrosarcomas, histocytic sarcomas, peripheral nerve sheath tumors, and spindle cell SCCs do not contain immunostaining against Melan‐A, PNL‐2, or tyrosine reactive proteins, but, as described with the individual neoplasms, can be differentiated using antibodies such as CD34, laminin, CD18, or pancytokeratin. Melan‐A will also detect canine steroid‐producing tissues and their tumors such as adrenal cortical tumors and sex‐cord stromal tumors, but neither represents a histologic differential for oral melanocytic neoplasms. Masson Fontana silver stain has previously been used to highlight small quantities of melanin in poorly pigmented melanocytic neoplasms, but it is less specific and sensitive than IHC and will also react with lipofuscin and argentaffin granules. Electron microscopy has been used to identify premelanosomes in amelanotic tumors, but is rarely useful for routine diagnosis.

Prognostic indicators of canine oral melanomas

Canine oral melanocytic tumors can be tentatively subdivided into melanocytic neoplasms of low malignant potential and highly malignant melanomas by their gross and histological appearance. Melanocytic neoplasms of low malignant potential are less commonly ulcerated and tend to be small (55/71 melanocytic neoplasms of low malignant potential were <1.0 cm with only one tumor being >2 cm in one study¹⁰) and pedunculated. The degree of pigmentation of the neoplastic cells has been reported to be prognostic in human oral melanomas. However, using the pigmentation of canine oral melanocytic neoplasms to predict outcome has resulted in conflicting data.^{6,11,12,20,21} Overall, it appears that high pigment is an indicator of a good prognosis, but low or no pigmentation does not necessarily indicate a poor prognosis.^6,10,11 Additionally, variability in the pigmentation of cells within different areas of the neoplasm makes it hard to assess the overall degree of pigmentation, especially if only a portion of the mass is evaluated.

A recent review of the literature revealed that nuclear atypia, mitotic count, and the Ki67 labeling index most consistently separated canine oral melanocytic neoplasms with low malignancy from those with high malignancy. In contrast, ulceration, inflammation, necrosis, or junctional activity did not consistently predict behavior.^6,9,11

Nuclear atypia

Neoplastic melanocytes with little nuclear atypia contain a small nucleus with minimal clumping of chromatin.^9,11 These cells tend to have a single, central nucleolus and potentially some condensed strands of nuclear chromatin extending from the nucleolus to the nuclear membrane with condensation along the inner surface of the membrane. Some cells may lack a nucleolus and have fine and evenly dispersed chromatin at the periphery of the nucleus.^9,11 Compared to melanocytic neoplasms with low malignancy, highly malignant neoplasms contain cells with atypical nuclei characterized by large irregularly shaped nucleoli that are eccentrically located (Figure 13.16).^9,11 They often have multiple nucleoli that, in some cases, are haphazardly connected to the inner surface of the nuclear membrane by thin strands of chromatin to give the appearance of a coarsely vacuolated nucleus.^9,11 The use of nuclear atypia using a threshold value of ≥30% atypical nuclei per 100 cells evaluated predicted whether or not a dog would die of oral melanocytic tumor within 1 year in 86.3% of oral and lip melanocytic neoplasms with a sensitivity of 90.3% and a specificity of 84.4%.¹¹ However, not all studies detected a correlation between nuclear atypia and outcome,^13,16 possibly as the subjective nature of the measurement increases inter‐observer variation. Additional disadvantages to using nuclear atypia to predict prognosis include the long time required for assessment, the requirement of bleaching of pigmented tumors, and the inability to accurately assess the nuclear morphology of cells within spindloid tumors or other tumors with little nuclear detail.^6,9,11

Ki67 immunolabeling

Ki67 is a marker of the growth fraction of neoplastic cells and more accurately predicts cell proliferation than a phase index such as mitotic index. Other advantages of Ki67 immunolabeling over the mitotic count include the easier assessment in heavily pigmented neoplasms (by using a red chromogen) and the prevention of mis‐identifying apoptotic and mitotic cells. In addition, it is easier to identify an area of the tumor with high Ki67 immunolabeling than an area with high mitoses. Furthermore, as measuring Ki67 immunolabeling is objective, it avoids the subjective assessment of determining the degree of nuclear atypia present within a neoplasm. Different studies have used different methods to measure Ki67 index, however, they all demonstrated the prognostic significance of Ki67 index for canine melanocytic neoplasms (Figure 13.17).^11,13,16,22 In a study in which the average number of positively labeled neoplastic cell nuclei within 5 optical grid reticle 1 mm² areas was determined at 40×, dogs with oral melanomas with a Ki67 index above the statistically ROC determined threshold value 19.5 had a significantly higher death rate within the first year after diagnosis than those with tumors below the threshold (p < 0.0005). The sensitivity and specificity of this threshold value was 87.1% and 85.7%, respectively.¹¹

Other prognostic indicators

Other methods to predict the behavior of a canine oral melanoma have been investigated. In a recent study, a decrease of expression of microRNA‐203 in canine oral melanocytic tumors was significantly associated with a shorter survival time.²³ In contrast, expression of E‐cadherin, β‐catenin, KIT, p53, PTEN, Rb, p21, or p16 did not predict clinical behavior or survival in a series of canine oral melanocytic neoplasms.^23–27

Incidence, signalment, and etiology

As these lesions develop as a hyperplastic response to a viral infection and almost invariably spontaneously resolve, oral viral papillomas are best considered to be a hyperplastic disease rather than a true neoplasm. Canine oral papillomatosis is considered to be common; however, the true incidence is unknown as few dogs receive treatment and most cases are unrecorded. The disease is caused by canine papillomavirus (CPV)‐1 (formerly canine oral papillomavirus) and is mostly seen dogs that are less than 1‐year‐old. Aspects of the epidemiology of the disease are poorly understood with large numbers of CPV‐1 particles produced during clinical infection, but CPV‐1 DNA also detectable in swabs of clinically normal oral mucosa and haired skin.^1,2 Outbreaks of disease have been reported in colonies of research dogs with up to 25% of dogs developing oral papillomas.²

Gross and microscopic diagnosis

Papillomaviruses increase both the division of infected basal cells and the commitment of basal cells to squamous maturation. This increases the number of cells within all layers of the epithelium resulting in epithelial thickening and folding and the formation of a papilloma.

Examination reveals exophytic lesions that are typically cauliflower‐shaped. They develop most frequently on the oral mucosa of the lips and pharynx, although they can also involve the tongue and esophagus. Most dogs will develop multiple papillomas and the presence of multiple large papillomas can result in bulging of the lips (Figure 13.19).

Canine oral papillomatosis is usually diagnosed when multiple exophytic oral tumors are detected in a young dog. If confirmation of the clinical diagnosis is required, cytology may allow definitive diagnosis if well‐differentiated squamous cells that contain evidence of viral infection (smudged blue‐gray cytoplasm) are visible. However, such cells are infrequently present within aspirates and histology is often required.

Histologically, viral papillomas appear as polypoid lesions supported by a central fibrous stalk. Examination of the base of the papilloma reveals a sudden transition to hyperplastic epithelium covered by increased keratin. The surface of the papilloma is composed of many finger‐like projections of epithelium that are supported by a thin core of fibrous tissue. Replicating viral particles are assembled within the superficial layers of epidermis and virally induced cytopathologic changes may be visible in these cells. The most easily recognizable histological evidence of viral infection is the presence of cells containing increased quantities of finely granular gray‐blue cytoplasm. Cells containing a pyknotic nucleus that is surrounded by a clear halo (koilocytes or bird’s eye cells) may also be present. Intranuclear deeply eosinophilic viral inclusions tend to be rare and hard to detect. IHC detection of papillomaviruses is only possible when viral replication is present. Prominent lymphocytic inflammation without histological evidence of virally induced cell changes characterizes a regressing papilloma.

A viral papilloma can be differentiated from a nonviral squamous papilloma by the presence of virally induced cell changes or lymphocytic inflammation in the underlying dermis. Viral papillomas are strictly exophytic and any evidence of invasion of the basement membrane prompts a diagnosis of papillary SCC.

Biological behavior and prognosis

After infection by CPV‐1, there is an incubation period of 4–8 weeks before papilloma development. In most dogs, papillomas regress after 4–8 weeks and the dog is protected from further papilloma development.³ Although extensive papillomatosis can develop, dogs rarely show other clinical signs of disease and oral papillomatosis is a transient non‐life‐threatening disease for the overwhelming majority of dogs.

There are rare reports of dogs with persistent oral papillomatosis.^4,5 This is presumed to be the result of a host immune dysfunction, although an underlying immunodeficiency is not always identifiable.^4,5 Unless a normal host immune response can be restored, treatment of nonregressing oral papillomas is difficult and extensive involvement of the esophagus and haired skin can necessitate euthanasia.⁴ Vaccination against CPV‐1 prevents oral papillomatosis,⁶ but vaccines do not influence lesion resolution.⁴

Although an oral SCC has been reported to develop within an area of papillomatosis, there is no evidence that canine oral viral papillomas are predisposed to malignant transformation. CPV DNA can be detected in a small proportion of canine oral SCCs. However, as CPV‐1 DNA can also be detected in swabs of normal oral mucosa,¹ it is difficult to prove a causative association.

Location, gross appearance, and etiology

The feline eosinophilic granuloma complex (also referred to as feline eosinophilic dermatosis) is thought to be due to either a hypersensitivity reaction or dysregulation in eosinophil migration or function.⁹ Two manifestations of the feline eosinophilic granuloma complex can develop within the oral cavity. Feline indolent ulcers (also referred to as eosinophilic ulcers and rodent ulcers) most often develop at the mucocutaneous junction of the upper lip. Lesions appear as well‐demarcated areas of ulceration with an elevated margin, but without the formation of a mass. Ulcers can be up to 3 cm in diameter and result in lip distortion and exposure of the teeth. Unilateral lesions are most common, although bilateral and confluent lesions are reported. Feline oral eosinophilic granulomas (also referred to a feline collagenolytic granuloma) appear as solitary or multiple firm nodules that can affect the lingual, sublingual, tonsillar, or palate areas (Figure 13.21A). Ulceration, secondary infection, and necrosis is common within the lesions and, unlike indolent ulcers, oral eosinophilic granulomas are often painful and cause dysphagia.

Microscopic diagnosis

Gross differentiation between a feline eosinophilic granuloma and an oral SCC is not possible and microscopic diagnosis is always recommended. The presence of numerous eosinophils on cytology strongly suggests a diagnosis of indolent ulcer (mucocutaneous junction) or eosinophilic granuloma. Histologically, both indolent ulcers and eosinophilic granulomas are characterized by eosinophils and eosinophil degranulation results in the presence of brightly eosinophilic cellular debris surrounding collagen fibers (flame figures) within the submucosa (Figure 13.21B). Flame figures are often surrounded by macrophages and giant cells with mast cells and smaller numbers of lymphocytes also present. Ulcerated lesions have prominent neutrophilic inflammation and superficial necrosis within the lesions. The presence of large numbers of eosinophils and macrophages usually allows differentiation from an oral mast cell tumor. Differentiation between an eosinophilic granuloma and an oral mast cell tumor is discussed with canine eosinophilic granulomas.