Criteria for Selection of Cases

Medical records of 139 dogs in which nasal carcinomas had been diagnosed at 15 veterinary institutions from January 1993 to December 2003 were reviewed. Medical records of dogs were eligible for inclusion in the study if a diagnosis of nasal carcinoma had been confirmed histologically; if the dog survived for ≥ 7 days from the date of diagnosis; and if the dog had not been treated surgically or with chemotherapy, radiation therapy, or immunotherapy. Medical records of dogs treated with palliative medications such as corticosteroids, NSAIDs, or antimicrobials were included.

Procedures

Information recorded from medical records included age, breed, weight, sex, clinical signs at the time of diagnosis, thoracic radiographic findings, and results of cytologic examination of regional lymph node aspirates. Medications administered as palliative treatments (eg, corticosteroids, NSAIDs, or antimicrobials) were recorded. Computed tomography images were reviewed by 1 radiologist (PVS), and clinical stage was classified according to 3 systems described by the WHO¹⁹, Théon et al,⁸ and Adams et al.¹⁸ Histologic sections from tumors were reviewed by 1 pathologist (BLN), and the type of carcinoma was histologically classified by use of the system described by the WHO.²⁰

Duration of disease was defined as the interval from the onset of clinical signs to a diagnosis of nasal carcinoma. Survival time was defined as the time from diagnosis of nasal carcinoma until death from any cause. Additional follow-up information was obtained via telephone communication with owners and referring veterinarians. Dogs alive during the last follow-up communication were included in analyses until the last day in which follow-up information was collected, and then censored.

Statistical analysis—Risk factors analyzed to determine their effect on survival time included age, weight, clinical signs at the time of diagnosis (eg, epistaxis, facial deformation, ocular discharge, sneezing, dyspnea, or neurologic abnormalities), duration of clinical signs, histologic subtype, metastases, individual CT findings (eg, unilateral nasal mass; bilateral nasal mass; exophthalmos; or involvement of the sinuses, cribriform plate, orbit, hard palate, or brain), clinical stage, palliative medical treatment, and improvement in clinical signs with use of palliative treatment. For analysis, histologic subtypes were categorized as adenocarcinoma, undifferentiated carcinoma, transitional carcinoma, and squamous cell carcinoma. Individual CT findings were scored as detected, not detected, or questionable, and for analyses, questionable lesions were categorized in 2 ways: first as detected and then as not detected. Palliative treatments were categorized as antimicrobials versus no antimicrobials, corticosteroids versus no corticosteroids (including piroxicam or other NSAIDs), and as piroxicam versus no piroxicam (including corticosteroids or other NSAIDs).

The Kaplan-Meier product-limit method was used to estimate survival curves for each potential categoric risk factors, and the log-rank test for censored data was used to compare survival curves. Wilcoxon rank sum tests were used for analyses of continuous risk factors that were not normally distributed. To evaluate the combined effects of potential risk factors on survival time, multivariable survival analysis was performed by use of the Cox proportional hazards regression model. Variables with values of P ≤ 0.20 in the univariable analyses were included in the multivariable survival analyses. Models were chosen by means of backward elimination; the variable with the highest P value (> 0.05) by use of the Wald test was removed, and nonsignificance of the removal was confirmed by the likelihood-ratio χ² test (the variable was returned to the model if the χ²-test result was significant). All statistical analyses were performed by use of a software package,^a and for the final analysis, 2-sided values of P ≤ 0.05 were considered significant.

Results

Median age of the 139 dogs was 11 years (range, 3 to 17 years). Seventy dogs were male (57 of which were castrated), and 69 were female (67 of which were spayed). Median body weight was 22 kg (48 lb; range, 4 to 68 kg [9 to 150 lb]). Purebred dogs accounted for 79% (110/139) of cases, and 21% (29/139) were mixedbreed dogs. Purebred dogs consisted of 45 breeds; the most frequently represented breeds were Golden Retrievers (n = 19), Shetland Sheepdogs (10), Labrador Retrievers (9), Beagles (6), Australian Shepherds (4), Siberian Huskies (4), and Miniature Poodles (4). The remaining purebred dogs were noted only 1 to 3 times.

Epistaxis was the most common clinical sign at the time of diagnosis, being detected in 107 dogs (77%). Other common clinical signs included sneezing (67%), nonhemorrhagic nasal discharge (56%), facial deformity (40%), dyspnea (27%), ocular discharge (26%), and neurologic abnormalities (6%). Duration of clinical signs prior to diagnosis of nasal neoplasia ranged from 1 to 300 weeks (median, 8 weeks).

Computed tomography of the skull and nasal cavity was performed in 112 dogs; 49 of these evaluations were available for review. The CT images were not sufficient to evaluate all anatomic areas for each dog.

Nasal carcinomas were unilateral in 29 of 49 dogs evaluated and bilateral in 11, and location was questionable in 9. Lesions were detected unilaterally in the frontal sinus in 32 of 43 dogs, bilaterally in 8, and there was questionable involvement of the frontal sinus in 3. Erosion of the cribriform plate was assessed in 45 dogs and was detected in 18, not detected in 22, and was questionable in 5. Erosion of the orbit was assessed in 46 dogs and was detected in 13, was not detected in 27, and was questionable in 6. Erosion of the hard palate was detected in 19 of 44 dogs and was not detected in 25. Five of 44 dogs had displacement of the brain, whereas this was not detected in 37 dogs and was questionable in 2.

Nasal carcinomas were staged on the basis of CT findings and were classified by use of the WHO¹⁹ criteria in 49 dogs; 5 (10%) had stage 1 disease (unilateral tumor with minimal or no bone involvement), 19 (39%) had stage 2 disease (bilateral tumor with moderate bone destruction), and 25 (51%) had stage 3 disease (extensive tumor with extranasal extension). In 48 dogs, tumor stage was classified by use of the system described by Théon et al⁸; 2 (4%) were classified as stage 1 (unilateral or bilateral tumor confined to the nasal passage without extension into the frontal sinuses), and 46 (96%) were classified as stage 2 (bilateral tumor extending into the frontal sinuses with erosion of any bone of the nasal passages). In 48 dogs, tumor stage was classified by use of a system described by Adams et al¹⁸; 10 (21%) were classified as T1 (tumor confined to 1 nasal passage, paranasal sinus, or frontal sinus, with no bony involvement), 15 (31%) were classified as T2 (any bony involvement, but with no evidence of orbit, subcutaneous, or submucosal mass), 5 (10%) were classified as T3 (involvement of orbit, or a subcutaneous or submucosal mass), and 18 (38%) were classified as T4 (tumor extension into nasopharynx or cribiform plate).

All dogs had histopathologic confirmation of nasal carcinoma, but histologic slides from only 40 tumors were available for review by the pathologist. Of the 40 slides that were reviewed, 18 (45%) were classified as adenocarcinomas, 13 (32%) were classified as undifferentiated carcinomas, 6 (15%) were classified as transitional carcinomas, 2 (5%) were classified as squamous cell carcinomas, and 1 (2.5%) was classified as a ductular carcinoma.

Results of cytologic examination of fine-needle aspirates of regional lymph nodes obtained at the time of initial diagnosis were available for 29 dogs. Seven (24%) of those dogs had evidence of regional lymph node metastases. Radiography of the thorax was performed at the time of initial diagnosis in 101 dogs, and only 2 had evidence of pulmonary metastases.

Thirty-one of the 139 (22%) dogs included in the study had been treated with corticosteroids to palliate clinical signs associated with nasal neoplasia, 53 (38%) had been treated with piroxicam, 19 (14%) had been treated with an NSAID other than piroxicam, and 36 (26%) did not receive corticosteroids or NSAIDs. One hundred four (75%) dogs were given antimicrobials. Information regarding improvement in clinical signs was obtained from 83 medical records; improvement in clinical signs was detected in 39 (47%) records, and no improvement was detected in 44 (53%) records. For dogs in which clinical signs had improved, it was not clear whether improvement in clinical signs was partial or complete.

One hundred twenty-six dogs died during the follow-up period, 11 were alive at the end of the study (range, 32 to 1,034 days; median, 141 days), and the remaining 2 were lost to follow-up at 17 and 38 days, respectively. One dog died because of congestive heart failure 55 days after nasal carcinoma had been diagnosed; the remaining 125 dogs died or were euthanatized because of nasal carcinomas. One dog's death 1,114 days after diagnosis was attributable to pulmonary metastases; this dog did not have evidence of metastases at the time of initial diagnosis.

The median survival time for all dogs in the study was 95 days (95% CI, 73 to 113 days), and the estimated probabilities of survival 6, 9, 12, and 24 months after diagnosis were 34%, 22%, 12%, and 2%, respectively (Figure 1).

Kaplan-Meier curve depicting overall survival time for 139 dogs with nasal carcinomas that did not receive treatment or received only palliative treatment for clinical signs associated with nasal carcinomas. The overall median survival time was 95 days (95% CI, 73 to 113 days), and the estimated probabilities of surviving 1 and 2 years after diagnosis were 12% and 2%, respectively.

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier curve depicting overall survival time for 139 dogs with nasal carcinomas that did not receive treatment or received only palliative treatment for clinical signs associated with nasal carcinomas. The overall median survival time was 95 days (95% CI, 73 to 113 days), and the estimated probabilities of surviving 1 and 2 years after diagnosis were 12% and 2%, respectively.

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curve depicting overall survival time for 139 dogs with nasal carcinomas that did not receive treatment or received only palliative treatment for clinical signs associated with nasal carcinomas. The overall median survival time was 95 days (95% CI, 73 to 113 days), and the estimated probabilities of surviving 1 and 2 years after diagnosis were 12% and 2%, respectively.

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

Risk factors significantly associated with survival time are depicted (Table 1). In multivariable analysis, only epistaxis continued to be significantly associated with survival time. In dogs that had epistaxis at the time of diagnosis, the hazard of dying was 2.3 times (95% CI, 1.4 to 3.8; P < 0.001) that of dogs that did not have epistaxis. For the 107 dogs with epistaxis, the median survival time was 88 days (95% CI, 65 to 106 days) and the estimated probabilities of survival 6, 9, 12, and 24 months after diagnosis were 27%, 14%, 6%, and 1%, respectively. For the 32 dogs without epistaxis, the median survival time was 224 days (95% CI, 54 to 467 days) and the estimated probabilities of survival 6, 9, 12, and 24 months after diagnosis were 53%, 46%, 42%, and 10%, respectively (Figure 2). The final multivariable Cox proportional hazards regression model also included age (dichotomized at the median as ≤ 11 years vs > 11 years; retained because of the likelihood-ratio criterion); however, for this variable, the hazard of dying was not significant (hazard ratio, 0.96; P = 0.83).

Kaplan-Meier curves depicting survival time for 107 dogs with nasal carcinomas in which epistaxis was detected at the time of diagnosis (solid line) and 32 dogs without epistaxis (broken line). Median survival time for dogs with epistaxis (88 days; 95% CI, 65 to 106 days) was significantly (P < 0.001) shorter than the median survival time for dogs without epistaxis (224 days; 95% CI, 54 to 467 days; in univariable log-rank test).

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier curves depicting survival time for 107 dogs with nasal carcinomas in which epistaxis was detected at the time of diagnosis (solid line) and 32 dogs without epistaxis (broken line). Median survival time for dogs with epistaxis (88 days; 95% CI, 65 to 106 days) was significantly (P < 0.001) shorter than the median survival time for dogs without epistaxis (224 days; 95% CI, 54 to 467 days; in univariable log-rank test).

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curves depicting survival time for 107 dogs with nasal carcinomas in which epistaxis was detected at the time of diagnosis (solid line) and 32 dogs without epistaxis (broken line). Median survival time for dogs with epistaxis (88 days; 95% CI, 65 to 106 days) was significantly (P < 0.001) shorter than the median survival time for dogs without epistaxis (224 days; 95% CI, 54 to 467 days; in univariable log-rank test).

Citation: Journal of the American Veterinary Medical Association 229, 3; 10.2460/javma.229.3.401

• Download Figure
• Download figure as PowerPoint slide

Discussion

On the basis of the results reported here, the prognosis for dogs with nasal carcinomas that did not receive treatment other than palliative medications is poor. The median survival time was 3.1 months, and the probability of surviving 1 and 2 years after diagnosis was only 12% and 2%, respectively. To the authors' knowledge, this is the largest study to evaluate survival of dogs with a single type of nasal neoplasia that was not treated. In 1 study,⁷ 5 dogs with nasal carcinomas that did not receive any treatments lived for 2 to 12 months. Most other studies^3,9,14,21,22 only report outcome for individual dogs or imprecise survival information.

Median survival times for dogs treated with radiation therapy range from 7.4 to 47.7 months.^8,9,15–18 Results of a study⁹ of 53 dogs with nasal adenocarcinomas that received radiation therapy indicate a median survival time of 14 months. This value was significantly longer than the median survival time (4.1 months) of 11 dogs that were not treated or treated by use of chemotherapy, surgical debulking, or both.⁹ Although statistical analyses were not performed, it appears that survival times reported in the present study for untreated dogs would be considerably shorter than those for dogs treated by use of radiation therapy.

In our study, most dogs that died reportedly died or were euthanatized because of progressive, local disease of the nasal cavity. Necropsy results were not available for most dogs; therefore, the risk of metastatic disease in dogs in the present study is not known. Surgical curettage can temporarily improve QOL in some patients with nasal tumors, but reported survival times have been short. Survival times for 42 dogs with various types of nasal tumors treated by use of cytoreductive surgery alone range from 1 month to 16 months, with an overall mean of approximately 4 months.^{3,7,13,14,23–25}

Although radiation therapy has become an integral part of treatment for dogs with nasal tumors, the risk of local recurrence is 60%, and survival is generally associated with failure to control the tumor locally rather than metastatic disease.^{8,10,11,15–18,26} Surgical exenteration of the nasal cavity after completion of radiotherapy has been described.¹⁷ Although the median survival time of 13 dogs undergoing surgical exenteration was long (47.7 months), the risk of local recurrence was not importantly different than that for 40 dogs treated by use of radiotherapy alone (60% vs 68%).¹⁷ Additionally, dogs undergoing surgery had a substantial risk of developing postoperative complications (including chronic rhinitis and osteomyelitis).¹⁷

Of the factors analyzed in the study reported here, epistaxis as an initial clinical sign was associated with a short survival time. For 107 dogs with epistaxis, the median survival time was 3 months, compared with 7.3 months for 32 dogs without epistaxis. Nasal hemorrhage is a major clinical sign, and many owners may have equated this with a poor QOL. Defining a pet's QOL is an enormously broad and complex concept.^27,28 In a study²⁹ of cats with neoplasia, all owners reported considering the cat's QOL in their decision for euthanasia. Other reports^30,31 have also discussed QOL as a factor contributing to the decision for euthanasia of an animal. In future studies, health-related questionnaires given to owners to assess QOL may be used to better understand the impact a diagnosis of nasal carcinoma is having on the life of their dog.

In our study, median age of dogs was 11 years. This finding is similar to that in other reports.^7–9,14-18 Increasing age at diagnosis was significantly associated with a short survival time, but only in the univariable analysis. LaDue et al¹⁵ also reported that age (> 10 years) was associated with a short survival time. The bond owners have with older pets has the potential to influence their interpretation of QOL, and as mentioned, this may contribute to the decision for euthanasia. Male and female dogs were equally represented, and in the univariable analysis, median survival time of males (87 days) was shorter than females (106 days). Of 504 dogs with carcinomas of the nasal cavity or paranasal sinuses, the risk in males was 1.3-fold greater than the hospital population at risk.²⁹ However, investigators of other studies^8,9,14,15 have reported similar proportions among sexes.

In the study reported here, 22% of dogs had been treated with corticosteroids to palliate clinical signs associated with nasal neoplasia (eg, sneezing, epistaxis, or nasal discharge) and 52% had been treated with NSAIDs. Nonsteroidal anti-inflammatory drugs inhibit COX; the isoform COX-2 and COX metabolites such as prostaglandin-E₂ contribute to tumorigenesis.³⁰ Overexpression of COX-2 has been reported in 80% of naturally occurring nasal carcinomas in dogs.^31,32 In the present study, 74% of dogs treated with NSAIDs received piroxicam. Piroxicam is a potent NSAID and has antitumor activity against transitional cell carcinoma of the urinary bladder³³ and squamous cell carcinoma of the oral cavity.³⁴ However, neither piroxicam nor NSAIDs generally had any association with survival of dogs in our study. Seventy-five percent of dogs received antimicrobials. Nasal tumors may be secondarily infected, but similar to the use of corticosteroids and NSAIDs, treatment with antimicrobials was not associated with survival.

Tumor stage was not prognostic for survival. Whether there is any association between clinical stage and outcome in dogs with intranasal neoplasia treated with radiation therapy is controversial.^8,9,15,18,21 Although a few individual CT findings were considered in the multivariable model, none were significantly associated with survival.

Only dogs with nasal tumors of epithelial origin were included in our study; therefore, survival of dogs with untreated mesenchymal or other types of tumors that can develop in the nasal cavity (transmissible venereal tumor, lymphoma, mast cell tumor, or melanoma) is not known. When treated with radiation therapy, the relationship between outcome and histologic type grouped as carcinoma versus sarcoma has yielded conflicting results, but recent reports^15–18 have not substantiated any such relationship. Authors of 2 studies^10,22 have suggested the prognosis is worse for dogs with either of 2 specific subtypes of epithelial tumors, squamous cell carcinoma or nonkeratinizing squamous cell carcinoma (also referred to as transitional carcinoma).³⁵ In the present study, histologic subtype of carcinoma was not a prognostic factor for survival.

In the study reported here, dogs with nasal carcinomas that did not receive treatment other than palliative medications had a median survival time of approximately 3 months. Radiation therapy has become the main method for treating dogs with nasal tumors. Numerous prognostic factors have been reported in dogs treated by use of radiation, including age,¹⁵ stage of disease, histologic type,^10,22,35 presence of metastatic disease,⁹ and facial deformation,¹⁶ and it has been suggested that dogs with factors negatively associated with outcome should not be treated.¹⁵ However, it is possible that radiation therapy will significantly extend survival time, compared with dogs that do receive treatment. On the contrary, results of the present study indicated that dogs without epistaxis survived for a median of 7.8 months and survival in this subset of patients may not necessarily be significantly improved with radiation therapy. The dataset generated in this study may be useful to test these hypotheses. Only 47% of dogs reportedly had improvement in clinical signs after treatment with palliative medications such as corticosteroids, NSAIDs, and antimicrobials. Mellanby et al³⁶ reported improvement in 95% of dogs with nasal tumors treated with radiotherapy on a hypofractionated schedule consisting of 4 large doses. Additional studies evaluating strategies to improve outcome of dogs with nasal tumors are warranted.