Parotid salivary gland carcinoma in a geriatric horse



This case report describes a geriatric gelding with a 2 month history of fast-growing masses within the parotid and submandibular regions. The horse was dyspnoeic on presentation and upper airway endoscopy revealed partial airway obstruction at the level of the pharynx, secondary to a space-occupying mass. Cytological evaluation of multiple fine-needle aspirates obtained from the masses were suggestive of salivary gland neoplasia, therefore the horse was subjected to humane euthanasia. A computed tomographical scan was obtained post mortem and revealed a large multi-lobulated mass involving both guttural pouches, resulting in 80% occlusion of the naso- and oropharynx. Histopathology confirmed a parotid salivary gland carcinoma (papillary-cystic type). Salivary gland tumours are extremely rare in horses, with only a few cases reported in the literature. Treatment of these malignancies proves to be challenging, requiring complete parotidectomy or surgical debulkment, in combination with adjunctive therapy.


The number of geriatric horses and ponies (age ≥15 years) in the equine population is steadily rising, probably due to improvements in nutrition and an increasing willingness by owners to seek veterinary care and prolong life (Brosnahan and Paradis 2003; Ireland et al. 2011). Ageing has been identified as a risk factor for the development of numerous conditions including chronic laminitis and neoplasia (Ireland et al. 2011), resulting in a growing number of reports involving equine tumours. A study of disease prevalence in 467 geriatric horses, revealed an incidence of neoplasia (excluding pituitary adenomas and lipomas) of 10%, with the most common neoplasia being squamous cell carcinoma (Brosnahan and Paradis 2003). No mention of neoplasia affecting the salivary glands was made, highlighting the infrequent occurrence of such a condition.

Salivary gland tumours are rare in domestic species, with an overall incidence of 0.17% in dogs and cats (Carberry et al. 1988). Currently, the incidence of salivary gland neoplasia in horses is unknown; however, given the scant number of reports available in the equine literature, it is deemed to be low. The horse has 3 pairs of major salivary glands (parotid, submandibular and sublingual) and 4 minor salivary glands (buccal, labial, lingual and palatine), responsible for the production of serous and mucous salivary secretions. Of these, the parotid salivary glands are the largest and the most frequently affected by disease (Schumacher and Schumacher 1995). The most common neoplasia associated with the major salivary glands in horses is melanoma, although lymphoma, adenocarcinoma and acinic cell carcinoma have also been reported (Stackhouse et al. 1978; Schumacher and Schumacher 1995; Cole MacGillivray et al. 2002). To our knowledge, this is the first case of computed tomographic (CT) imaging of a salivary gland tumour in a horse.

Case history

A 25-year-old crossbreed gelding weighing 423 kg was presented to the Onderstepoort Veterinary Academic Hospital of the Faculty of Veterinary Science, University of Pretoria for further investigation of fast-growing masses within the left parotid and submandibular regions. The lesions were first noted 2 months prior to presentation and were reported to grow steadily throughout this period. The horse was observed to have a decreased appetite for roughage and was reluctant to graze. No abnormalities were detected on skull and pharyngeal radiographs, obtained by the attending veterinarian prior to referral.

Clinical findings

Clinical examination revealed multiple firm, irregular masses extending ventrally from the base of the left ear along the ramus of the mandible, and continuing approximately 15 cm rostrally to occupy the intermandibular space (Fig 1). The masses appeared nonpainful as the horse did not resent palpation and no heat was detected over the affected area. Mild tachypnoea was noted (22 breaths/min) with moderate bilateral nostril flare and increased inspiratory and expiratory effort. Airflow from the nostrils was significantly decreased bilaterally, and inspiratory stridor was noted. Mild mucopurulent bilateral nasal discharge was also present and the gelding was observed to stand in an orthopnoeic position. No other abnormalities were detected following completion of the physical examination.

Figure 1.

Photograph demonstrating the extensive nature of the parotid and submandibular masses. Note the presence of a suture at the site of the initial single core biopsy.

Upper airway endoscopy revealed mild dorsoventral pharyngeal collapse along with marked dorsomedial displacement of the soft palate, resulting in partial occlusion of the airway (Fig 2). Endoscopy of both guttural pouches demonstrated marked collapse of both the lateral and medial pouches; the left guttural pouch was more severely affected.

Figure 2.

Still endoscopic image of the pharyngeal region demonstrating partial occlusion of the airway caused by marked displacement of the soft palate.

Haematological evaluation revealed haemoconcentration (packed cell volume 0.45 l/l, normal reference range [rr] 0.24–0.44 l/l; haemoglobin 161 g/l, rr 80–140 g/l; red cell count 9.27 × 1012/l, rr 5.5–9.5 × 1012/l), mature neutrophilia (8.19 × 109/l, rr 3.54–7.08 × 109/l) and monocytosis (0.83 × 109/l, rr 0.0–0.72 × 109/l). Serum biochemistry revealed a hyperglobulinaemia (44.5 g/l, rr 18–38 g/l) and hyperfibrinogenaemia (6 g/l, rr <4 g/l). Ionised calcium was 1.72 mmol/l (rr 1.5–1.75 mmol/l).

A single core-biopsy was taken from the mass, approximately 10 cm distal to the base of the left ear; however, histopathological examination was consistent with a normal parotid salivary gland. Fine-needle aspirates were obtained at the same time, from the left parotid salivary gland and submandibular lymph node. Cytological evaluation of both samples revealed similar morphological findings, with a moderate number of anaplastic cells being identified. These cells were of glandular origin and displayed marked basophilic cytoplasm and macrokaryosis, with prominent and multiple nucleoli (Fig 3). Based on these cytological findings, salivary gland neoplasia was presumed.

Figure 3.

500 × magnification of fine needle aspiration (FNA) from the left parotid salivary gland. Note the basophilic cytoplasm, macrokaryosis and prominent and multiple nucleoli (arrows) suggestive of malignancy.


Due to the fast-growing nature of the masses, development of dyspnoea secondary to airway obstruction and the likelihood of a neoplastic process, the horse was subjected to humane euthanasia. At the owner's request a post mortem examination was not performed, although permission was granted for computed tomography and further histological examination of the parotid salivary gland and submandibular masses.

Computed tomography

Computed tomographic evaluation followed disarticulation of the head. The examination was performed with a Siemens Somatom Emotion dual slice CT unit1. Slices were set at 3 mm thickness, with a pitch of 1.5. Helical scanning commenced in a soft tissue window (window level of 58 and window width of 414). Reconstruction in a bone window optimised bone detail (window level of 560 and window width of 2500). The study revealed a large multi-lobulated soft tissue mass (45–149 Hounsfield units) involving the left guttural pouch. The mass crossed the midline to involve and partially occlude the right guttural pouch resulting in 80% occlusion of the oro- and nasopharynx (Fig 4). Ventrally, extending caudally from the submandibular region, another 6 × 3 cm multi-lobulated soft tissue mass was seen. This mass contained wispy, irregular mineralised opacities that were more prominent centrally and rostrally, probably due to necrosis and dystrophic calcification in that region (Fig 5).

Figure 4.

A transverse soft tissue (window level 58, window width 414) image at the level of the guttural pouches: the near total occlusion of the oro-and nasopharynx by a fairly homogenous soft tissue mass (arrows) is clearly evident, the larynx and trachea (T) is markedly displaced to the right.

Figure 5.

A sagittal image of the skull in the same window as Figure 4 shows another multi-lobulated soft tissue mass (arrows) on the ventral aspect of the mandible with irregular, wispy mineralisation seen at the rostral aspect (block arrow).


Samples obtained from the left parotid salivary gland and submandibular lymph nodes were examined histologically, revealing the presence of normal salivary gland tissue, and normal lymphoid tissue embedded within areas of chronic active inflammation. At the periphery of the parotid gland biopsy were aggregates of neoplastic cells with large round to oval nuclei, finely granular chromatin and usually just a single prominent nucleolus. These cells also had moderate amounts of pinkish cytoplasm within which appeared to be secretory vesicles. Mitotic index was 8/high-powered field. Similar cells were also found within the subcapsular sinus and medulla of the submandibular lymph node (Fig 6). Of particular interest was the manner in which the epithelial carcinoma cells formed papillary projections and tortuous cystic spaces; along with the associated desmoplastic reaction, characterised by an abundance of fibrous connective tissue (Fig 7). Based on the cellular morphology, their propensity to form sheets and the suspected presence of secretory product within the cytoplasm, a diagnosis of parotid salivary gland carcinoma (papillary-cystic type), with metastasis to the regional submandibular lymph nodes was made.

Figure 6.

Histopathological examination (haematoxylin-eosin staining) of a submandibular lymph node. Note the large neoplastic cells with prominent eosinophilic cytoplasm (arrows) within the subcapsular sinus of the lymph node.

Figure 7.

Histological section (haematoxylin-eosin staining) of the parotid salivary gland (SG) demonstrating the formation of papillary projections (arrows) and tortuous cystic spaces (C) by the epithelial carcinoma cells. This tumour is therefore classified as a papillary-cystic type carcinoma. Also note the significant desmoplasia (D) associated with the neoplastic cells.


Whilst salivary gland tumours have been mentioned in a small number of papers in the veterinary literature, to the authors' knowledge only 2 reports describing the clinical details associated with salivary gland neoplasia exist; one of which was a salivary gland adenocarcinoma (Stackhouse et al. 1978) and the other a parotid gland lymphoma (Coumbe 1994). Consequently, this is the first detailed case report describing a parotid gland carcinoma in a geriatric horse.

As previously stated, salivary gland tumours are rare, with an overall annual incidence of 0.003% being reported in man (Speight and Barrett 2002). Salivary gland malignancies are reported more often in men and older patients, with 66% of tumours occurring in patients aged >55 years (Ho et al. 2011). Currently, the World Health Organisation recognises 24 malignant types of primary salivary gland tumours, classified based on recognisable morphological patterns (Speight and Barrett 2002; Jones et al. 2008). The current classification of salivary gland tumours involves simply listing the different tumour types based on microscopic appearance, with no indication of tumour grade or behaviour, and as a result such a system is frequently challenged (Speight and Barrett 2002). Similar classification systems have been described in veterinary medicine by Koestner and Buerger (1965) and Head (1976). Collectively, 75 salivary gland tumours, of which 6 were identified in horses, were examined and subsequently classified as either: 1) adenomas (either pleomorphic or monomorphic); 2) mucoepidermoid tumour; 3) acinic cell tumour; 4) carcinomas (adenocarcinoma, squamous cell carcinoma, undifferentiated carcinoma or carcinoma in pleomorphic adenoma); or 5) unclassified tumours (Koestner and Buerger 1965; Head 1976). The histological classification of the carcinoma in the current case report was of a papillary-cystic type, based on its papillated appearance and cyst formation (Fig 6). This type of growth pattern is often described in human medicine, particularly associated with acinic cell carcinomas and adenocarcinomas (Ali 2002; Speight and Barrett 2002), but has not been described in the equine literature to date.

Veterinary clinicians are often required to make decisions regarding the management of cases in the absence of a final diagnosis. In this case, cytological evaluation of the fine-needle aspirates proved very useful in allowing more informed decision-making with regards to potential management strategies. The results were consistent with a neoplastic process, although the type could not be determined, and given the rapid deterioration in the patient's condition, as described by the owner, the decision to subject the horse to euthanasia was made in a timely manner. Although cytological evaluation of a fine-needle aspirate is not as sensitive or specific as a histopathological examination, it is both noninvasive and cost-effective, and should therefore be considered in all cases where visible and accessible masses are present.

In this case, histopathological examination was crucial for making a definitive diagnosis and this report clearly illustrates the importance of taking multiple biopsies from different sites within a mass. The first biopsy, taken from the left parotid salivary gland at the same time as the fine-needle aspirates, identified normal salivary gland tissue which ultimately was not consistent with the final diagnosis. Following euthanasia of the patient, repeat biopsies were obtained from the parotid salivary gland, as well as the submandibular lymph nodes, resulting in a post mortem definitive diagnosis.

Due to the relatively superficial location of the parotid salivary gland, abnormalities of these can readily be evaluated using ultrasound. This particular patient was evaluated ultrasonographically but involvement of the underlying structures and the full extent could not be accurately assessed. Due to its complexity and overlap of adjacent structures, CT is the most commonly used imaging modality to assess abnormalities involving the equine head, especially caudally in the region of the temporomandibular and temporohyoid joints (Warmerdam et al. 1997; David et al. 2008; Hilton et al. 2009; Naylor et al. 2010), the nasal passages (Tietje et al. 1996; Tucker and Farrell 2001; Solano and Brawer 2004) and the cranial cavity (Tietje et al. 1996). The air-filled nature of the guttural pouches prohibits deeper and accurate ultrasonographic assessment but make them most amenable to CT examination due to their inherent contrast. Computed tomography generally requires general anaesthesia (Solano and Brawer 2004), so the risk-gain ratio must be carefully assessed. In our patient, CT images provided information regarding the full extent of the mass and the involvement of adjacent structures; should surgery have been contemplated, this information would have been vital. Currently reports of CT images of salivary gland neoplasia are lacking.

Haemoconcentration was identified on initial presentation, and given the history of inappetance was considered most likely to be a result of hypovolaemia and dehydration. Relative polycythaemia due to dehydration is frequently accompanied by an increase in total plasma protein concentration (Sellon and Wise 2010), which, however, was not observed in this case (total protein concentration 73.7 g/l, rr 66–83 g/l). In the face of inflammation, the liver is stimulated by inflammatory cytokines to increase production of acute phase proteins, many of which are α- and β-globulins, and limit the production of albumin (Eckersall 1995), resulting in a respective hyperglobulinaemia and hypoalbuminaemia, as is seen in this instance. Although the albumin concentration reported here is within normal limits (29.2 g/l, rr 22–38 g/l), it is highly possible that this value is falsely increased secondary to decreased plasma volume as a result of dehydration.

Hyperglobulinaemia is often observed in cases of equine neoplasia and has previously been reported in association with a salivary gland adenocarcinoma (Stackhouse et al. 1978). In a study of gastric neoplasia, hyperglobulinaemia was identified in 11 out of 24 horses, and in these cases the increased globulin concentration was deemed to be associated with chronic inflammation (Taylor et al. 2009). The hyperglobulinaemia reported in this particular case is similarly most likely to be due to inflammation, given the extensive and infiltrative nature of the tumour, and this notion is highly supported by the presence of a mature neutrophilia, monocytosis and hyperfibrinogenaemia.

The high-normal ionised calcium concentration identified in this horse may be attributable to humoural hypercalcaemia of malignancy (HHM), an uncommon manifestation of neoplasia. Currently there are no estimates of the prevalence of HHM in the equine population; however, this condition has been reported in association with gastric, vulvar and preputial squamous cell carcinoma, adrenocortical carcinoma, ameloblastoma, and lymphosarcoma (Marr 1994; Rosol et al. 1994; Toribio 2010). This paraneoplastic endocrinopathy is thought to develop due to the production and secretion of biologically active parathyroid hormone-related protein (PTHrP) by certain tumour cells (Rosol et al. 1994). This protein subsequently interacts with PTH-1 receptors, resulting in the resorption of calcium from bone and inhibition of renal calcium excretion (Toribio 2010). Inflammatory cytokines secreted by neoplastic cells have also been implicated in the development of hypercalcaemia by stimulating osteoclastic bone resorption (Rosol et al. 1994). Unfortunately, neither PTHrP nor PTH concentrations were measured in this case, so the presence of HHM cannot be confirmed.

Historically, complete parotidectomy and tumour resection is considered the mainstay of treatment in human patients with salivary gland tumours (Noh et al. 2010; Ho et al. 2011). Due to the complicated anatomy, however, it is often not possible to achieve adequate margins; thus adjunctive treatments such as radiation therapy and systemic chemotherapeutics are also instituted.

In horses, the parotid salivary gland lies within the retromandibular fossa, situated between the vertical ramus of the mandible and the wing of the atlas, with the base of the gland lying adjacent to the lingual facial vein. The rostral border attaches to the ramus of the mandible and the masseter muscle, whilst the caudal border is attached to the underlying muscle. The medial surface lies against the maxillary vein, in close approximation with the facial, glossopharyngeal and hypoglossal nerves, the internal and external carotid arteries and the vagosympathetic trunk (Schumacher and Schumacher 1995; Dixon and Gerard 2006). Due to the number of vital structures within the region of the parotid salivary gland, complete parotidectomy in the horse is considered both challenging and impractical.

Numerous human studies have demonstrated the benefit of adjunctive radiation therapy (either intra- or post operative) in improving local salivary gland tumour control and patient survival, compared to surgical resection alone (Chen et al. 2008; Noh et al. 2010). Unfortunately, the costs and risks associated with multiple general anaesthetics, combined with the limited number of institutions with the required equipment, makes teletherapy an infrequently used treatment option for equine patients.

Interstitial brachytherapy involves the implantation of a metal seed or wire containing high doses of a radioactive isotope such as iridium, cobalt or radium directly into the tumour itself (Pennock 1967; Burks et al. 2009). As a result, the delivery of radiation is focal, with limited exposure of the surrounding tissues. The disadvantage of brachytherapy, however, is the small therapeutic radius (approximately 1 cm) surrounding the implant, thus surgical debulkment is often performed prior to implantation (Witte and Perkins 2011).

The prognosis for successful treatment in the case described in this report was considered hopeless, due to the extent of neoplastic infiltration and the development of dyspnoea, therefore any treatment provided would have been purely palliative. The involvement of the parotid gland, along with the extensive and infiltrative nature of the tumour prevented complete surgical resection and due to the anatomical location, surgical debulkment was deemed too risky. External beam radiation therapy is not currently available in South Africa and, given the size of the tumour, interstitial brachytherapy would have been cost-prohibitive. Local metastasis to the regional lymph nodes was confirmed by histopathology; however, as a complete necropsy was declined by the owners it remains unknown whether metastasis to other organs had occurred. If other organs were implicated, systemic chemotherapy would have been indicated. Few chemotherapeutic drugs, all of which have adverse side effects, are readily available for use in horses, and the cost of these drugs and the requirement for long-term treatment means their use is greatly limited (Burks et al. 2009).

Authors' declaration of interests

No conflicts of interest have been declared.

Manufacturer's address

1 Siemens, Forchheim, Germany.