Tonsillar carcinoma in dogs: Treatment outcome and potential prognostic factors in 123 cases

Abstract Background Tonsillar carcinomas are rarely reported in dogs. Information on outcome after treatment is sparse and prognosis is guarded to poor. Hypothesis/Objectives Assess treatment outcome and potential prognostic factors in a population of dogs with cytological or histopathological diagnosis of tonsillar carcinoma. Animals A total of 123 client‐owned dogs with diagnosis of tonsillar carcinoma confirmed by cytology or histopathology. Methods Retrospective, multi‐institutional study. Medical records of 12 institutions were reviewed from 2012 to 2021. Results Treatment included surgery, chemotherapy (conventional, tyrosine kinase inhibitors or metronomic chemotherapy), radiotherapy, nonsteroidal anti‐inflammatory drugs (NSAIDs) or a combination of these. Surgery was performed in 68 cases, chemotherapy was administered in association with NSAIDs in 64 cases, NSAIDs were used alone in 14 cases and in association with surgery in 21 cases, whereas radiotherapy was used alone or in combination with surgery or chemotherapy in 20 cases. Overall survival time (OST) was 126 days (95% confidence interval [CI], 88‐164). Significantly longer survival (P < .001) was seen in dogs without evidence of metastatic disease (median survival time, 381 days; 95% CI, 116‐646). Other significant positive prognostic factors included absence of clinicals signs at presentation, surgery (tonsillectomy), use of adjuvant chemotherapy and use of NSAIDs. Conclusion and Clinical Importance Asymptomatic dogs, those treated with surgery, those that received adjuvant chemotherapy, and those that received NSAIDs may have a better prognosis than previously expected, but overall survival remains short for dogs with tonsillar carcinoma.


| INTRODUCTION
Primary tonsillar carcinoma is reported rarely in dogs when compared to non-tonsillar oral carcinomas and has a higher potential for local metastatic disease (up to 73%), when compared to oral carcinomas in other locations. Metastatic disease usually involves regional lymph nodes (mandibular, retropharyngeal, cervical) and lungs. 1,2 Lymph node metastasis has been reported in 9/11 dogs (82%) that had their lymph nodes examined, and in the same study 2/40 patients (5%) had suspected pulmonary metastasis. 1 A more recent study found local lymph node metastasis in 7/15 dogs (47%), whereas 2/15 (13%) had nodal and distant (pulmonary) metastasis. 3 The most common histological subtype is squamous cell carcinoma (SCC, 55%), followed by lymphoma (17%) and melanoma (12%). 4,5 Carcinoma (CA) also has been reported. 6,7 The etiology of this tumor is unclear, but it initially was suggested that dogs living in urban areas were prone to develop tonsillar neoplasia. 8 A later epidemiologic study, however, did not find any difference in the incidence of tonsillar carcinoma in dogs living in smaller vs larger urban areas. 9 Tumor diagnosis often is challenging because of the small size of the primary tumors, vague or no clinical signs, or incidental ventrolateral neck masses, consistent with an enlarged retropharyngeal lymph node, noted by the owners. 1 1,6 The most common chemotherapy drugs administered to dogs with tonsillar carcinomas are platinum agents or anthracyclines. 1,3,10 Regardless of treatment, MST was short in early studies, ranging from 179 to 270 days. 1,3,10 Previously identified negative prognostic factors include presence of clinical signs (e.g., lethargy, anorexia). 1 Treatment with chemotherapy and RT was associated with longer MST, whereas clinical stage was not prognostic. 1 Conversely, a more recent study found that dogs with unilateral involvement and no evidence of metastatic disease had a MST of 637 days, compared to 134 and 75 days for those patients with local or distant metastasis, respectively. The MST for patients treated using surgery and adjuvant chemotherapy was 464 days. 3 Limited information is available with regard to patient outcome after various treatments, including surgery, chemotherapy, and radiation, for dogs with tonsillar carcinomas. Our aim was to report the outcome of dogs with tonsillar carcinoma that underwent different treatment modalities. A secondary aim was to assess prognostic factors influencing disease progression and overall survival.

| MATERIALS AND METHODS
Ours was a retrospective, multi-institutional study. Radiotherapy was delivered as palliative or definitive treatment, alone or in combination with surgery, chemotherapy or NSAIDs.
Various NSAIDs (carprofen, piroxicam, meloxicam, firocoxib, and robenacoxib) were used at standard dosages, q24h, alone or in association with surgery, chemotherapy and RT. Palliative-intent treatment included NSAIDs (q24h) or prednisolone (q24h), paracetamol or paracetamol and codeine (q12h) and tramadol (q12h or q8h), used PO alone at standard dosages, started at the time of diagnosis until disease progression or euthanasia.

| Treatment response
Response to treatment was determined by the primary clinician's assessment, including clinical examination and oral cavity examination under GA or sedation, or using cytology in combination with diagnostic imaging (CT or thoracic radiographs to assess for local and distant metastasis). Response was evaluated according to the Veterinary

Cooperative Oncology Group Response Evaluation Criteria in Solid
Tumors (VCOG RECIST, version 1) and classified as complete response (CR), partial response (PR), progressive disease (PD) and stable disease (SD). 11 Complete response was defined as the disappearance of all target lesions. Pathologic lymph nodes had to be <10 mm in the short axis. Partial response was defined as at least 30% reduction in the sum of diameters of target lesions, taking as reference the baseline sum. Progressive disease was defined as either the appearance of ≥1 new lesions or at least a 20% increase in the sum of diameters of target lesions, taking as reference the smallest sum on study.
The sum also must have shown an absolute increase of 5 mm. Stable disease was defined as <30% reduction (PR) or 20% increase (PD) in the sum of diameters of target lesions, taking as reference the smallest sum of diameters while on study.
Chemotherapy toxicity was graded according to the VCOG-Common Terminology Criteria for Adverse Events (VCOG-CTCAE, version 2). 12 Dose reductions when toxicity occurred were performed at the clinician's discretion. Classification of adverse events (AEs) was done retrospectively by reviewing medical records, as assessed by the clinician in charge of the case at the time of each visit.

| Clinical signs
Clinical signs at presentation were categorized as respiratory (e.g., cough, dyspnea, upper respiratory noises, stertorous breathing, stridor, reverse sneezing), gastrointestinal (GI; e.g., decreased appetite or anorexia, vomiting, diarrhea, ptyalism, dysphagia) or a combination of respiratory and GI. Additional constitutional (e.g., lethargy, weight loss) and neurological (e.g., neck pain, head tilt, facial nerve paralysis, behavioral changes, anisocoria, dysphonia) signs were documented if present. If the tumor was an incidental finding at presentation, clinical signs were not classified. If signs were not categorized as respiratory, GI, neurological or constitutional, they were classified as other, including halitosis, retching, oral pain, oral discharge, lymphadenomegaly, neck edema, presence of a mandibular mass, or pyrexia.

| Statistical analysis
Clinical data and follow-up information were acquired from medical records or phone conversations with the referring veterinarians. If patients were lost to follow-up, they were right censored at the time of their last visit or update.
The following characteristics were analyzed for statistical purposes and as potential prognostic factors: institution, breed, sex, age The MST was defined as the time from diagnosis until death because of any cause, whereas progression-free survival (PFS) was defined as the time from diagnosis until disease progression developed. A KM product-limit method was used to estimate survival and the log-rank test was applied to test differences between groups. All reported P-values were 2-sided, and a 5% significance level was set (P < .05). Multivariate Cox's proportional hazards regression analysis using a forward stepwise procedure was performed for factors significant on univariate analysis. Statistical analysis was performed using commercial software (IBM SPSS Statistics, version 23.0).
Patient demographics are summarized in Table 1.
Twenty patients (16%) additionally had constitutional and 13 (11%) neurological signs. In 3 cases (2%), the tumor was an incidental finding at the time of GA and intubation before a scheduled surgical procedure. One-hundred and fifteen dogs (94%) had laboratory blood tests performed: hematology and biochemistry were performed in 109 cases, whereas 2 cases had hematology only and 4 biochemistry only; urinalysis was performed in 12 cases and no abnormalities were found. Hematology and biochemistry abnormalities are summarized in Table 2.
The tonsillar carcinoma was localized in the left tonsil in 61 cases (50%), in the right tonsil in 54 cases (44%), unilateral in 115 cases and bilateral in 7 cases (6%). The location was not reported in 1 case (1%).
The remainder of patients were staged using a combination of head CT (n = 5), head magnetic resonance imaging (MRI, n = 2), or thoracic CT and neck ultrasound examination (n = 1). One patient did not undergo any staging procedures.
Regional lymph nodes (mandibular, superficial cervical, and retropharyngeal) were examined by means of cytology or histopathology in 94 dogs (76%), if they were found to be enlarged on physical examination or in case of abnormal findings on CT, and not sampled in 29 dogs (19%). Of those sampled, 77 (82%) had regional metastasis, 16 (13%) were negative for metastatic disease and, in 1 case, cytology was not diagnostic. In 40 dogs (33%), only 1 lymph node was involved; multiple lymph nodes were affected by metastases in 35 dogs (29%). Nineteen dogs (15%) had distant (pulmonary, n = 19 and splenic, n = 1) metastatic disease, presumed or confirmed by cytology. Overall, staging procedures were negative for local or distant metastatic disease in 29 dogs (24%). Abdominal imaging did not identify any abnormalities consistent with distant metastases in 30 cases, whereas in 1 case splenic metastases were identified.

| Survival analysis and prognostic factors
The median PFS (  Regarding treatment, tonsillectomy with or without regional lym-  Dogs were staged using a variety of different procedures, consistent with the retrospective and multicenter nature of our study, the time frame, and the owners' financial resources. The value of CT in assessing solid tumors and suspected metastasis has been reported for oral cancer, particularly with respect to regional lymph nodes, and the tomographic appearance of tonsillar tumors recently has been described. 5 Metastasis did seem to affect survival, particularly in cases of regional metastatic disease. This finding could be related to the severity of clinical signs associated with lymphadenomegaly and its effects on quality of life when compared to asymptomatic or mildly symptomatic dogs. Additionally, in some cases, local metastatic disease could be associated with more invasive inoperable tumors, rapid disease progression, owners declining surgical treatment or the decision to euthanize. Histopathological examination of bilateral mandibular and medial retropharyngeal lymph nodes previously has been recommended to confirm metastatic disease, but not all of our cases underwent bilateral lymphadenectomy or had cytology of the contralateral lymph node performed, especially if lymph nodes were normal on CT. 14 Sentinel lymph node mapping also has been shown to be helpful in identifying metastatic lymph nodes in dogs with oral tumors, and could similarly be considered when staging patients with tonsillar cancer. 4,15,16 The same applies to the contralateral tonsil, which was not evaluated in all cases. It is known that the contralateral tonsil can be affected, and a previous study found that 3 carboplatin also has been used in association with radiotherapy. 10,17,18 Response to toceranib phosphate has been reported in solid tumors of dogs including head and neck carcinomas, with CR and PR reported in 12.5% and 62.5% of patients, respectively. 19 In our cases, a few PR (n = 11) were recorded, although SD was most commonly reported (n = 16).
The use of RT was not significant, which could be a consequence of the low number of treated patients and the different protocols used (i.e., type II error). Therefore, its benefit remains unclear. A previous study also suggested that the addition of RT did not seem to influence prognosis. Similarly, in our study, RT was used palliatively and in association with antineoplastic agents in the majority of cases, and specifically as sole treatment in the gross disease setting in 2/20 cases and to treat residual disease in 8/20 cases. 3 If radiation is available, it is likely to be beneficial in treating residual disease and may prolong survival if used in combination with chemotherapy or as an adjuvant to surgery.
The use of NSAIDs generally was associated with significantly longer survival. In a previous study, the majority of dogs received either NSAIDs alone or in combination with other treatments (e.g., surgery, chemotherapy or RT) and a survival benefit was seen. 1 Cyclooxygenase (COX)-2 expression has been shown in different types of carcinomas in dogs, including urothelial, prostatic, nasal, thyroid, mammary and anal sac carcinomas, among others. [20][21][22][23][24][25][26][27][28][29] Oral carcinomas in dogs also express COX-2 receptors, but a specific study of tonsillar carcinomas has not yet been conducted. 30 These tumors may express specific receptors or respond to anti-inflammatory treatment because of the degree of inflammation associated with this tumor type, particularly in the gross disease setting. Previous measurable responses of non-tonsillar SCC to piroxicam have been reported. 31 In conclusion, our study suggests that asymptomatic patients, patients treated with surgery and adjuvant chemotherapy, and those receiving NSAIDs may have a better prognosis than untreated dogs or symptomatic patients, although overall survival remains short. Survival expectations for dogs with tonsillar carcinoma are limited, but the combination of multiple treatment modalities, whenever possible, is likely to be beneficial. Additional studies are needed to elucidate the role of surgery and different adjuvant treatments in the management of tonsillar carcinoma in dogs.