Outcome and Metastatic Behavior of Canine Sinonasal Osteosarcoma (2005–2015)
ABSTRACT
Sinonasal neoplasms in dogs behave locally aggressively, and metastatic disease has not been a common cause of death. The metastatic rate of sinonasal osteosarcoma (OSA) is not well characterized, and reported outcomes of these patients are variable. The purpose of this study is to evaluate the outcome and metastatic behavior of canine sinonasal OSA. Medical records of canine patients diagnosed with sinonasal OSA via histopathology between January 2005 and December 2015 were reviewed. Patients with any form of treatment or no treatment were included. Time to local progression, time to metastasis, and overall survival data were evaluated. Variables that may impact outcome, such as tumor stage and treatment type, were evaluated. Twenty-seven dogs were identified that fit the inclusion criteria. Overall, 30.0% of dogs developed metastasis over the disease course, with a median time to metastasis of 458 days (95% confidence interval [CI] 318–758 days). The median time to local progression was 335 days (95% CI 264–544 days). The overall median survival time was 410 days (95% CI 341–627 days). Regarding metastasis, sinonasal OSA behaves similarly to sinonasal neoplasms of other histologies and dissimilarly to appendicular OSA. The outcome of treated patients appears similar to that of sinonasal tumor patients with other histologies.
Introduction
Canine sinonasal tumors are rare, encompassing ∼1% of all canine cancers.1 Two-thirds of these tumors are carcinomas, whereas the majority of the remaining one-third are sarcomas, with osteosarcoma (OSA) being a rare subtype.2–4 Overall, canine sinonasal tumors are locally aggressive. The metastatic rate is typically low at diagnosis yet can be higher at the time of death (40–50%), although dogs usually die of local disease.1,2,5–7 OSA of the appendicular skeleton is much more common than sinonasal OSA, with chondrosarcoma and fibrosarcoma being more common in the nasal cavity.2,3
Although there is literature regarding canine sinonasal tumors, the literature specifically addressing survival times of dogs with sinonasal OSA is lacking. There have been many studies investigating the survival times of sinonasal tumors treated with different radiation therapy (RT) protocols and radiation techniques. At this time, RT is the recommended treatment for local control of nasal tumors versus surgery.1,2,6,8–15 Kubicek et al. investigated outcomes and prognostic factors associated with canine sinonasal tumors treated with curative-intent stereotactic radiosurgery.5 In this study, 7 dogs with nasal OSA had an overall median survival time (MST) of 94 days, which is shorter when compared with the dogs with nasal carcinoma or other nasal sarcomas (MST = 256 days). In another study, Sones et al. found an MST of 624 days in 6 dogs with nasal OSA treated with RT. The metastatic rate of canine sinonasal OSA has been reported to be 12.5% in one retrospective study from 1984.3 Another study, from 1992, investigated canine axial skeletal OSA.15 In that study, 10 (9%) of the 116 cases were OSA of the nasal cavity or paranasal sinuses. None of those 10 dogs were found to have any evidence of pulmonary metastasis. In that study, the overall pulmonary metastatic rate for dogs with axial skeletal OSA was 11%, although only 54 of 116 dogs had monitoring pulmonary radiographs and none had necropsies. In comparison, metastasis is very common for patients with appendicular OSA, with ∼90% of dogs succumbing to metastatic disease within 1 yr.16 These results suggest that OSA in different locations may have varying biological behavior, with respect to metastasis.
As a result of divergent results reported in the literature regarding outcome for sinonasal OSA as well as outdated and limited information available about metastatic rate, the purpose of this study was to evaluate the overall survival time and metastatic behavior of canine sinonasal OSA.2,5 A secondary objective of the study was to evaluate variables that may impact outcome, such as tumor stage and treatment type.
Materials and Methods
Case Collection and Medical Record Review
Cases were retrospectively identified from Colorado State University Veterinary Teaching Hospital’s medical record database between December 2005 and December 2015. Dogs who were included had a histopathologic diagnosis of OSA in the sinonasal location. The tumors were defined as sinonasal, arising from the nasal and/or sinus cavities, based on imaging studies. Dogs were excluded if they lacked adequate follow-up, which was defined by not returning for diagnostics or treatment after initial diagnosis.
Information that was retrospectively gathered included age, breed, sex, and neuter status at the time of diagnosis. The results of diagnostic tests from the time of diagnosis were also collected, when available. These results included reports from computed tomography (CT) scan, histopathology, cytology of the primary mass, regional lymph node cytology, thoracic radiographs, thoracic CT scan, and abdominal ultrasound. When CT scans were available for review, patients were retrospectively staged based on the Modified Adams Staging System.17 This staging system is summarized in Table 1. All CT scans and reports were reviewed by the same radiation oncologist (author not identified in the reviewed text).
Time of diagnosis was defined as the time of CT imaging or biopsy confirming sinonasal OSA, whichever came first. Time to local disease progression and time to metastasis were evaluated. Local disease progression was based on CT imaging or recurrence or progression of the patient’s original clinical signs—specifically, clinical signs that were consistent with local disease progression including a mass effect, visible swelling, and/or epistaxis.
Other information collected included whether the patient received treatment, type of treatment, radiation technique used, chemotherapy protocols, follow-up diagnostics that were performed, if the patient had any additional treatment for local disease recurrence or metastasis, and location of metastasis.
Statistical Analysis
Survival time was defined from the time of diagnosis to the time of death, as a result of any cause. The progression-free interval (PFI) was defined as the time of diagnosis to time of first event, whether local disease progression or metastasis. The time to metastasis was defined as the time of diagnosis of local disease to time of detection of metastatic disease. The time to local progression was defined as the time of diagnosis to the time of local tumor progression based on CT findings or clinical signs. The MST and median PFI were estimated using the Kaplan-Meier method. Tumor stage and treatment type (radiation with or without chemotherapy, stereotactic RT [SRT] versus conventionally fractionated RT) were evaluated for effect on outcome via Log-rank (Mantel-Cox) analysis. Statistical analysis was performed using Prism v 7.0 softwarea. Of the factors analyzed, P < .05 was considered statistically significant. All deaths were attributed to the disease being investigated. Patients were censored if still alive at last follow-up.
Results
Patient Population
Thirty patient medical records were found to be reviewed for potential inclusion. One patient was excluded as a result of a final diagnosis of multilobular osteochondrosarcoma based on necropsy report. One was excluded as a result of no follow-up after initial appointment. Another patient was excluded as a result of findings most consistent with maxillary OSA. Twenty-seven dogs met the inclusion criteria.
Of the 27 included dogs, the mean and median ages were both 9.0 yr (range 2–14 yr). There were 17 neutered males, 9 spayed females, and 1 intact male. Mixed-breed dogs (n = 6) and Labrador retrievers (n = 6) were the most commonly represented breeds, followed by golden retrievers (n = 3) and German shepherd dogs (n = 3). There was also 1 each of Great Pyrenees, bloodhound, cocker spaniel, Boston terrier, rottweiler, Rhodesian ridgeback, greyhound, miniature schnauzer, and Cairn terrier.
All 27 dogs had histopathology performed of their primary tumor, confirming OSA (n = 24), likely OSA (n = 2), or probable OSA (n = 1). Five of 27 dogs also had local cytology of their primary tumor with results of normal airway epithelium (n = 1), suppurative inflammation (n = 1), suspect carcinoma (n = 1), and sarcoma (n = 2).
Twenty-four of the 27 dogs had a CT of their nasal cavity or skull performed. Eighteen of these CT scans were able to be reviewed to assess tumor stage in accordance with the Modified Adams Staging System.17 Tumor stages included stage I (n = 1), stage II(n = 1), stage III(n = 11), and stage IV(n = 5).
Regional lymph node enlargement was noted in 10 cases according to the CT scan reports. Regional lymph node cytology was performed in 13 cases. Out of these, none were positive for lymph node metastasis. Regional lymph node cytology of 2 cases was consistent with normal lymph node, whereas the remaining 11 were consistent with reactive lymph nodes. Seven of the cases noted to have lymphadenopathy on CT reports had cytology performed, whereas 3 did not.
Thoracic imaging was performed at the time of diagnosis in 26 cases. Twenty-three had thoracic radiographs only, 1 case had thoracic CT scan only, and 2 cases had both thoracic radiographs and thoracic CT scan. No evidence of pulmonary metastatic disease was documented at diagnosis in these cases. One case had a soft tissue nodule seen on thoracic radiographs, yet this was suspected to be an extrathoracic structure. No further imaging or monitoring was performed in this case. The one case that did not have thoracic imaging at the time of diagnosis did not have treatment and was euthanized shortly after diagnosis.
Twenty of the 27 patients had repeat or serial thoracic imaging including thoracic radiographs (n = 18), thoracic CT scans (n = 1), or both (n = 1) following their diagnosis. Twelve of the 20 dogs had two or greater imaging series, whereas 8 had thoracic imaging repeated once following the imaging at the time of their initial diagnosis. The frequency of imaging of the dogs who had two or greater monitoring imaging series ranged from every 2 to 6 mo. Of the 8 dogs with only one repeat imaging after time of diagnosis, 5 had thoracic radiographs performed that showed no evidence of metastasis at 112, 214, 134, and 445 days after diagnosis, 1 was euthanized at the time of repeat radiographs that discovered pulmonary and rib metastasis at 234 days after diagnosis, 1 was diagnosed with pulmonary metastasis (two soft tissue pulmonary nodules) at 804 days, and 2 had thoracic radiographs performed at the time of euthanasia, at 62 days and 284 days after diagnosis, respectively (as a result of continued epistaxis after treatment in the former case and obvious local recurrence with significant mass effect, swelling, and epistaxis in the latter case), that did not show metastasis. Three cases had necropsies performed, which found disseminated OSA metastasis (n = 2) and no evidence of metastasis (n = 1).
Abdominal imaging was performed in five cases at the time of initial diagnosis. Three out of the five had some type of abnormality found, including a nodular liver and splenic nodule in one case, enlarged mesenteric lymph node in one case, and hepatic nodules and splenomegaly in one case. Aspirates were performed in all of these cases, with no evidence of metastasis and findings typically consistent with other unrelated benign changes.
Treatment
Five dogs had no treatment. One dog was treated with chemotherapy as a primary treatment with an alternating carboplatin (five doses) and doxorubicin (four doses) protocol. One patient had surgery as a primary treatment, followed by adjuvant chemotherapy. This dog received a dorsal rhinotomy followed by five doses of carboplatin chemotherapy.
Fourteen dogs received RT alone as a primary treatment, without adjuvant chemotherapy following RT. Of these 14 dogs, 9 received SRT with 3 fractions of 10 Gy for a total of 30 Gy. Five of the 14 dogs received a conventionally fractionated protocol. These protocols included 18 fractions of 3 Gy for a total of 54 Gy (n = 2), 18 fractions of 3.15 Gy for a total of 56.7 Gy (n = 1), 19 fractions of 3 Gy for a total of 57 Gy (n = 1), and 10 fractions of 4.2 Gy for a total of 42 Gy (n = 1).
Six cases received RT followed with adjuvant chemotherapy. Four dogs received SRT (three fractions of 10 Gy for a total of 30 Gy) followed by four doses of carboplatin (n = 2), toceranib phosphate (n = 1), and metronomic cyclophosphamide chemotherapy (n = 1). One case received a fractionated RT protocol of 18 fractions of 3 Gy for a total of 54 Gy, followed by six doses of carboplatin chemotherapy followed by metronomic chlorambucil chemotherapy. One case was started on a fractionated RT protocol and received four fractions of 3 Gy and was then converted into a palliative protocol of five fractions of 6 Gy. The total radiation dose for this dog was 42 Gy for the combined protocols. This dog was then treated with four doses of carboplatin for adjuvant therapy. Table 2 summarizes treatment and outcome information for the reviewed cases.
Analysis of Survival Time and Progression-Free Interval
One dog of the 27 was alive at the time of data analysis, and this dog was censored from survival data at 883 days. At this time, this patient had no evidence of metastatic disease on recent thoracic radiographs and did not have any clinical signs consistent with a nasal tumor or recurrence. All other patients were confirmed to be deceased.
Most patients in this study died from local disease (70%, 19/27 cases). Three patients (11.1%, 3/27) died from metastatic disease. One dog had clinical signs associated with metastatic disease and local disease progression at the time of euthanasia. One dog was alive at the time of data collection and analysis. Three dogs died of the following: one with a hemoabdomen, one with sudden death after seizure-like activity and hypoglycemic episodes, and one with sudden death of unknown cause. These three dogs did not have necropsies; therefore, the cause of their deaths was assumed to be a result of their sinonasal OSA.
The MST was 410 (range 2–1,459) days. The median PFI from the time of diagnosis was 284 (range 2–969) days (Figure 1). The survival times of the five untreated dogs were 2, 13, 78, 175, and 284 days, with an MST of 78 days.
Citation: Journal of the American Animal Hospital Association 56, 2; 10.5326/JAAHA-MS-6972
Analysis of Overall Metastatic Rate, Median Time to Metastasis, and Median Time to Local Progression
No dogs were confirmed to have metastasis at the time of diagnosis. Six of 27 dogs developed metastasis over the course of their disease. Yet, only 20 dogs received monitoring thoracic imaging and/or had a necropsy performed. In those 20 cases, the overall metastatic rate was 30.0%. The median time to metastasis in these 6 dogs was 458 days (234–935 days). Locations of metastasis included pulmonary (n = 4), disseminated to multifocal ribs, left proximal humerus, and right distal tibia (n = 1), and disseminated to brain, lung, pleura, and diaphragm (n = 1).
In the 19 dogs experiencing progression, the median time to local progression was 335 days. Local recurrence was determined by CT scan (n = 7) or suspected based on clinical signs consistent with local disease progression (n = 12), including exophthalmos (n = 1), epistaxis (n = 8), visible mass effect or swelling (n = 6), and visible mass effect with epistaxis (n = 1).
Effect of Tumor Stage on Outcome
Eighteen of the 27 canine patients with sinonasal tumors were staged according to the Modified Adams Staging System.17 PFI and MST were compared between stage I–III and stage IV. The median PFI was 431 days (86–969 days) for stage I–III dogs and 284 days (2–883 days) for stage IV dogs. The MST was 526 days (126–971 days) for stage I–III dogs and 284 days (2–883 days) for stage IV dogs (Figure 2). These differences were not statistically significant.
Citation: Journal of the American Animal Hospital Association 56, 2; 10.5326/JAAHA-MS-6972
Effect of Treatment on Outcome
Both MST and median PFI were compared between dogs who received RT alone (n = 14) versus RT in combination with any adjuvant chemotherapy (n = 6). The median PFI was 419 days (86–969 days) for RT alone and 430 days (2–883 days) for RT with chemotherapy. The MST was 550 days (126–971 days) with RT alone and 444 days (2–883 days) for RT with chemotherapy. Neither MST nor median PFI between these groups were statistically different (Figure 3). Of the six dogs who did have detectable metastasis, only one of them received chemotherapy at any point throughout their treatment.
Citation: Journal of the American Animal Hospital Association 56, 2; 10.5326/JAAHA-MS-6972
The MST and median PFI were also compared between patients treated with SRT (n = 13) versus conventionally fractionated RT (n = 7). The MST was 526 days and 729 days for dogs treated with SRT and conventionally fractionated RT, respectively (P = .42). The median PFI was 431 days and 405 days for dogs treated with SRT and conventionally fractionated RT, respectively (P = .80). This comparison was not statistically significant. The MST for dogs receiving any form of RT with or without chemotherapy was 531.5 (110–971) days.
Discussion
Of the patients included in this study, males were overrepresented when compared with females (2:1). This is reported in previous literature, but the significance of this finding is unknown.15
In our patient population, all 27 dogs had histopathology performed of their primary intranasal tumor, confirming OSA (n = 24), likely OSA (n = 2), or probable OSA (n = 1). The 1 patient’s result reported as probable OSA was confirmed to be OSA on necropsy. The other 2 cases were included as a result of very high suspicion on the histopathology report (both reported as likely OSA, as a result of identification of an eosinophilic bone matrix and osteoid production), yet other differential diagnoses for these tumors could include chondrosarcoma, undifferentiated sarcoma, or fibrosarcoma.
In our study evaluating the biological behavior of sinonasal OSA, we found an MST of 410 days and a median PFI of 284 days. In a previous study, a shorter MST of 93 days was found in seven dogs with sinonasal OSA treated with curative-intent RT.5 In another study, a longer MST of 624 days was found in six cases of intranasal OSA treated with curative-intent RT.2 Our population of 27 patients had an outcome more consistent with the later report and is similar to outcomes of sinonasal tumors of all histologies receiving treatment including sarcomas. Sones et al. found an MST of 444 days in dogs with intranasal sarcomas treated with either curative or palliative intent RT. These patients had intranasal chondrosarcoma (n = 42), undifferentiated sarcoma (n = 22), fibrosarcoma (n = 12), and OSA (n = 6).2 Kubicek et al. included dogs with sinonasal tumors treated with curative-intent cone-based stereotactic radiosurgery. Median survival times were 316, 326, and 94 days for carcinoma, sarcoma, and OSA, respectively.5 Gieger et al. reported an MST of 586 days for dogs with nonlymphomatous nasal tumors treated with SRT. The majority of these dogs had nasal carcinomas (n = 20) followed by chondrosarcomas (n = 16).6
The metastatic rate of canine sinonasal OSA was found to be 30%, which is lower than that of appendicular OSA and closer to that of OSA of the axial skeleton.15,18,19 Metastasis was also found to occur late in the disease process, with a median time to metastasis of 458 days. Overall, the metastatic behavior is more similar to that of sinonasal neoplasms of other histologies.2,5,6,12
Two studies from 1984 and 1992 did report the metastatic rate of sinonasal OSA to be low, at 12.5% and 0%, respectively.3,15 The dogs in the first study were not reported to have treatment. Out of the seven nasal or paranasal OSA patients in the second study, three were treated with attempted surgical excision followed with no adjuvant therapy (n = 2) or orthovoltage RT of 48 Gy total with concurrent cisplatin chemotherapy (n = 1). This low rate of metastasis in these studies is consistent with our findings yet slightly lower. A slightly higher rate of metastasis may have been found in the present study as a result of better local and systemic treatments. This could have allowed dogs to live long enough to develop detectable metastasis.
The median time to metastasis in the current study was 458 days. The median time to local progression was 335 days, which is consistent with other nasal tumors having local disease recurrence as the most common cause of death versus metastatic disease. 2,5–7,13,15
Consistent with other studies evaluating sinonasal neoplasms, most patients in this study died from local disease (70%, 19/27 cases). 2,5,6,13 Three patients (11.1%, 3/27) died from metastatic disease. One dog had clinical signs associated with metastatic disease and local disease progression at the time of euthanasia. One dog was alive at the time of data collection and analysis. Three dogs died of the following: one with a hemoabdomen, one with sudden death after seizure-like activity and hypoglycemic episodes, and one with sudden death of unknown cause. These three dogs did not have necropsies. It is possible that the seizure-like activity leading to death in one dog could be attributed to local disease progression into the forebrain or late radiation side effects, but this is not known. These three dogs were included in data analysis and were assumed to die of their sinonasal OSA because their death was not confirmed to be from other causes.
As a result of low numbers of dogs in stages I and II, stages I–III (n = 13) were grouped and compared with stage IV dogs (n = 5). Comparing stages I–III with stage IV was also performed as a result of previous reports of dogs with sinonasal tumors with cribriform plate involvement having a worse outcome.17 Tumor stage did not significantly affect outcome, most likely because of low numbers in the two compared groups of T1–T3 (n = 13) and T4 (n = 5). Although not statistically significant, dogs with stage IV tumors did have shorter median PFI of 284 days and MST of 284 days compared with stage I–III dogs with a median PFI of 431 days and MST of 526 days. More cases could be collected in the future to evaluate this to confirm the impact of stage IV disease on outcome.
In this study, there was no evidence of lymph node metastasis at the time of diagnosis in the 13 dogs that had regional lymph node cytology performed. Seven of these dogs were noted to have lymphadenopathy on CT imaging. Three dogs who were noted to have lymphadenopathy did not have regional lymph node cytology performed; therefore, it is possible metastatic disease was missed in these patients. This low incidence of regional lymph node metastasis is consistent with other sinonasal neoplasms and appendicular OSA.7,12,20,21
Although not found to be statistically significant, treatment with chemotherapy following RT did not extend median PFI or MST compared with dogs treated with RT alone in this population of dogs with intranasal OSA. As a result of the majority of these patients dying from local disease progression and having a prolonged time to develop metastasis (458 days), the role of adjuvant chemotherapy is currently unknown in dogs with sinonasal OSA.
One variable in this patient group is the change of RT technique over the 10 yr period of the retrospective study. Literature, even over the past 30 yr, involving RT for treatment of canine nasal tumors has shown varying results in regard to outcomes, and technology has evolved to incorporate intensity-modulated radiation techniques. There is also a paucity of literature directly comparing the different radiation techniques’ impact on outcome for dogs with nasal tumors.2,5,6,8–11,13,14,20,22,23 At our institution, SRT has become a more common treatment modality compared with conventionally fractionated RT. However, in our study, we did not find a difference in outcome between dogs treated with SRT or conventionally fractionated RT, yet the numbers in each group are small.
There are other limitations to this study. The retrospective nature of the study did not allow for standardized diagnostics, imaging, and follow-up criteria. A major limitation of the study is that patients received a multitude of different treatment modalities or no treatment at all. Although including dogs with any or no treatment is a limitation, this represents that clients choose different treatment paths for different patients that may ultimately lead to differences in their outcome. It would be possible to only include dogs with sinonasal OSA who received one specific treatment type, yet this was not the primary purpose of this study. Moreover, follow-up diagnostics to assess local disease recurrence or metastatic disease were also variable. Return of clinical signs was used as a marker for recurrence of local disease in some cases as a result of a lack of repeat CT scans. Although clinical signs may occur secondary to radiation adverse events (e.g., rhinitis), the clinical signs noted in these cases were more consistent with local disease progression.
Although standardization of diagnostics and treatment would allow for evaluation of different variables, the data gathered were able to provide confirmation of outcome and metastatic behavior, which was the intent of the study. Higher patient numbers would also allow for better evaluation of different variables on outcome including stage of disease, the use of adjuvant chemotherapy, and radiation technique. Another limitation of this study is that a low number of patients had necropsies performed (n = 3), possibly missing some metastatic lesions.
Conclusion
The purpose of this study was to evaluate the outcome and metastatic rate associated with sinonasal OSA. The overall metastatic rate developed over the course of disease was 30.0%. The MST and median PFI were 410 days and 284 days, respectively. The median time to metastasis was later in the disease course (458 days) compared with the median time to local disease progression (335 days), and most of these patients died from progression of local disease. In this population, it was found that sinonasal OSA behaved similarly to sinonasal neoplasms of other histologies and dissimilarly to appendicular OSA based on metastatic rate and outcome.
Kaplan-Meier curves of progression-free interval and survival time from time of diagnosis. Median PFI was 284 days and median survival time was 410 days. PFI, progression-free interval; ST, survival time.
Kaplan-Meier curves of median progression-free interval (PFI; left) and median survival time (MST; right) of dogs with stage I–III nasal tumors (n = 13) and stage IV nasal tumors (n = 5). The PFI for dogs with stage I–III tumors was 431 days, whereas the median PFI for dogs with stage IV tumors was 284 days. MST for dogs with stage I–III tumors was 526 days, whereas MST for dogs with stage IV tumors was 284 days. When comparing tumor stage on outcome, dogs with stage IV nasal tumors had a shorter median PFI and MST, yet this was not statistically significant.
Kaplan-Meier curves of median progression-free interval (PFI; left) and median survival time (MST; right) of patients treated with radiation therapy with and without adjuvant chemotherapy. PFI was 419 days in patients treated with radiation therapy alone (n = 14) and was 430 days in patients treated with radiation therapy followed by chemotherapy (n = 6). MST was 550 days in patients treated with radiation therapy alone and 444 days in patients treated with radiation therapy followed by chemotherapy. Treatment with adjuvant chemotherapy following radiation therapy did not seem to extend median PFI or MST, yet these numbers are not statistically significant. RT, radiation therapy.
Contributor Notes
CI (confidence interval); CT (computed tomography); MST (median survival time); OSA (osteosarcoma); PFI (progression-free interval); RT (radiation therapy); SRT (stereotactic radiation therapy)
