Introduction

Osteosarcoma is the most common primary bone tumor in dogs and is an important clinical problem in veterinary medicine. The natural behavior of this disease has been well defined.^1–3 The primary tumor usually occurs at the metaphysis of long bones, particularly the proximal humerus, distal radius, distal femur, and proximal tibia, in that order. It affects primarily large and giant dogs, and usually presents initially with lameness and a painful bone swelling at the site of the primary tumor. Osteosarcoma can affect dogs of any age, although most are middle-aged to elderly animals with a median age of 8 yr.

A number of prognostic factors have been identified for dogs with appendicular osteosarcoma. Several studies have shown an association between elevated serum alkaline phosphatase (ALP) levels at diagnosis and worse survival after amputation and chemotherapy for dogs with appendicular osteosarcoma. Both total serum ALP and bone ALP are independently predictive.^4,5 For example, in one study, median survival times for dogs with normal or increased total ALP activities before treatment were 12.5 and 5.5 mo, respectively, and median survival times for dogs with normal or increased bone ALP activities before treatment were 16.6 and 9.5 mo, respectively. A recent meta-analysis of prognostic factors for osteosarcoma identified serum ALP level as one of the strongest prognostic factors for dogs with appendicular osteosarcoma.⁶ Histologically high-grade tumors are associated with a worse prognosis for appendicular osteosarcomas. In one study, dogs with grade 3 tumors accounted for more than 75% of tumors; those with low grade tumors had a better prognosis even with less aggressive treatment.⁷ Another study found that the presence of regional lymph node metastasis is rare, but it imparts a poor prognosis, with median survival time of 59 days.⁸ Higher numbers of circulating monocytes (>0.4 × 10³/μL) and lymphocytes (>1.0 × 10³/μL) before treatment have been significantly associated with shorter disease-free interval.⁹

The most pressing clinical problem for dogs with appendicular osteosarcoma is pain control, and the quickest and best pain relief measure is amputation. Although surgical treatment of osteosarcoma by amputation dramatically improves quality of life, it only increases survival time slightly and, therefore, adjunctive chemotherapy is usually recommended. The median survival time for dogs treated by amputation alone is about 4 mo, only 10% of dogs are alive one yr after surgery, and dogs usually die as a result of pulmonary metastases.¹⁰ Alternative approaches to management of the primary tumor are limb-sparing surgery, external beam radiotherapy, intensity modulated radiotherapy, and samarium-153 therapy.^11–15

For dogs that can achieve pain control either through surgical or nonsurgical means, survival can be significantly prolonged by adjuvant chemotherapy. Various chemotherapeutic approaches have been reported. Cisplatin monotherapy appears to improve median survival times to between approximately 6 and 13 mo, and 1-yr survival rates to between 30 and 62%, while 2-yr survival rates are between 7 and 21%.^16–20 The main disadvantages of cisplatin are the need for diuresis to prevent nephrotoxicity and its emetogenicity. Carboplatin monotherapy is associated with median survival times of between approximately 10 and 11 mo.^21–23 Subcutaneous infusion of carboplatin was recently reported to yield outcomes similar to those previously reported for intravenous infusion of carboplatin, although it is possible this group of dogs may have been earlier stage than dogs in previous studies since 16 of the 17 dogs were screened for pulmonary metastases using computerized tomography scan rather than radiographs as predominantly used in the older studies.²⁴ Lobaplatin monotherapy has been reported to be associated with a 1-yr survival proportion of 31.8%, similar to that seen with other platinum monotherapy, and also did not require pretreatment diuresis.²⁵ Doxorubicin monotherapy prolonged survival in dogs treated by amputation and five bi-weekly doses, with similar results.²⁶ The main disadvantage of doxorubicin monotherapy is the risk of cumulative cardiotoxicity. Combinations of cisplatin and doxorubicin, alternating every 21 days for two cycles, gave a median survival comparable to cisplatin monotherapy.^27–29 However, of 102 dogs treated with cisplatin and doxorubicin combined on the same day, 47% survived 1 yr after surgery, 28% were alive 2 yr after surgery, and 17% were alive 3 yr after amputation and chemotherapy.³⁰ This observation suggests that the intensity of scheduling appears to play a role in the outcomes in this combination (Table 1). Combinations of carboplatin and doxorubicin have been reported to yield median survival times of approximately 8 to 11 mo.^31–34 Carboplatin and gemcitabine combination has been reported to be associated with a median survival time of approximately 9 mo.³⁵ Most recently, the only randomized phase 3 trial of two different chemotherapy protocols yet published for dogs with appendicular osteosarcoma compared 6 doses of carboplatin only against 6 total doses of alternating carboplatin and doxorubicin on a 21-day schedule.³⁶ This study evaluated disease-free interval (DFI) rather than survival time and found a significant difference between the two protocols, with a median DFI of 135 days for alternating carboplatin and doxorubicin and a median DFI of 425 days for carboplatin monotherapy.

TABLE 1 Review of Relative Dose Intensity and Outcomes for Selected Published Chemotherapy Trials for Canine Appendicular Osteosarcoma

View Fullscreen

Taken together, these results show that chemotherapy appears to improve survival times over surgery alone for dogs with appendicular osteosarcoma. However, most reported chemotherapeutic approaches yield similar median survival times, so the choice of protocol for the individual patient is based primarily on the best fit of the toxicity profile and practical factors, such as cost and time commitment. Moreover, there appears to be a “ceiling” beyond which present conventional chemotherapy does not appear to be able to improve median survival times. On the other hand, while median survival times are similar between all of these published protocols, there are some differences in outcomes when results are compared in terms of 2-yr (and beyond) survival fraction. Table 1 makes a comparison of selected published chemotherapy protocols in terms of the relative dose intensity and tumor control outcomes, and suggests that the intensity of scheduling, while not able to break the aforementioned “ceiling,” can help to maximize the proportion potentially cured within these apparent limitations.

Based on the observation that the best published outcomes to date for dogs with appendicular osteosarcoma has been with more dose-intense chemotherapy, we sought to increase the relative dose intensity of a sequential doxorubicin—carboplatin combination protocol by using accelerated scheduling for dogs with appendicular osteosarcoma after amputation.³⁰ Here we present toxicity outcomes with this protocol, survival outcomes for treated dogs, and prognostic factors.

Materials and Methods

Medical records from 2007 to 2013 from Veterinary Oncology Consultants and the Animal Referral Hospital were retrospectively reviewed for dogs with a confirmed histologic diagnosis of appendicular osteosarcoma. To be included, dogs were required to have undergone physical examination, three-view thoracic radiographs, complete blood cell count and serum biochemistry panel pre-operatively, to have no macroscopic evidence of pulmonary metastases on thoracic radiographs, to have undergone amputation followed by histopathology of the surgical specimen, and to have started on the prescribed chemotherapy protocol. Chemotherapy consisted of three doses of doxorubicin at 30 mg/m² for dogs greater than 15 kg or 1 mg/kg for dogs less than 15 kg administered 14 days apart, followed by three doses of carboplatin at 300 mg/m² for dogs greater than 15 kg or 250 mg/m² for dogs less than 15 kg 21 days apart. Recommended rechecks included physical examination and three-view thoracic radiographs upon completion of chemotherapy and every 2 mo for the first 6 mo, then every 3 mo for the next 6 mo, then every 6 mo thereafter.

Data recorded included age; breed; sex and neuter status; body weight; date of diagnosis; pre-amputation ALP (both level and whether it was normal or high); pre-amputation lymphocyte count; pre-amputation monocyte count; tumor site (proximal humerus, distal radius, distal femur, proximal tibia, or other); duration of clinical signs prior to amputation; amputation date; histopathologic grade of osteosarcoma; mitotic index (mitoses per 10 high power fields) of the tumor; chemotherapy start date; chemotherapy finish date; whether or not the prescribed protocol was completed and the reason why if it was not completed; whether or not dose reductions were performed and, if so, for which drug; whether or not dose delays were performed and, if so, for which drug; the highest toxicity score overall and the nature of the toxicity and with which drug it was associated; whether the patient was still alive and, if not, the date and cause of death; whether or not there was local recurrence of primary tumor; and whether or not there was distant metastasis and the location of any metastasis. Hematological and gastrointestinal (GI) toxicities were scored according to the Veterinary Co-Operative Oncology Group— Common Terminology Criteria for Adverse Events v1.0.³⁷ The DFI was calculated from the date of amputation to the date of local recurrence or metastasis, whichever came first. Survival time was calculated from the date of amputation to the date of death due to any cause.

A commercial software statistics package^a was used for analyses. The Kaplan-Meier product limit method was used to calculate the median survival time and the 1-yr and 2-yr survival proportion. In addition, univariate analysis using Cox regression analysis was performed to evaluate the effect on survival time of age, body weight, tumor site, ALP (normal or high), monocyte count (<0.4 × 10³/μL versus ≥0.4 × 10³/μL), lymphocyte count (<1.0 × 10³/μL versus ≥1.0 × 10³/μL), the duration of clinical signs, histologic grade, mitotic index, the interval from amputation to starting chemotherapy, whether or not chemotherapy was completed, and whether or not dose reductions or delays were performed. To evaluate the combined effects of potential risk factors on survival, multivariate survival analysis was done using forward conditional Cox regression analysis. Variables with P values ≤0.1 in the univariate analysis were offered to the multivariate analysis. For the final analysis, values of P < 0.05 were considered significant.

Results

Thirty-eight dogs met the criteria for inclusion and had follow-up information available regarding whether or not they finished the protocol and, if so, for how long they were followed after finishing chemotherapy. Nine dogs were censored from the survival analysis; eight were still alive and one dog was lost to follow-up while free of disease and was censored at the date of last known follow-up. Of the 29 dogs that died, 14 were known to have died of osteosarcoma, none died from chemotherapy complications, 3 were known to have died of causes other than the cancer or the chemotherapy (2 euthanized for progressive, unmanageable degenerative joint disease and 1 after a rupture of the anterior cruciate ligament), and for 12 dogs the cause of death could not be determined.

The median survival time was 317 days and 1- and 2-yr survival percentages were 43.2 and 13.9%, respectively.

Twenty-seven dogs (71%) completed the prescribed chemotherapy protocol and 11 dogs (29%) did not. Of the 11 dogs that did not complete it, the reasons were progressive osteosarcoma (metastases) in 3 dogs, unrelated illness in 2 dogs (histiocytic sarcoma, renal failure), chemotherapy toxicity or possible toxicity in 2 dogs, unknown in 2 dogs, owner financial constraints in 1 dog, and owner preference for other therapy in 1 dog.

Toxicity was common, but grade 4 events were uncommon, with eight of the 38 dogs (21.0%) experiencing grade 3 adverse events and 2 dogs (5.2%) experiencing grade 4 adverse events, and 2 dogs (5.2%) requiring hospitalization for complications of myelosuppression. Altogether, 19 dogs had hematological toxicity (grade 1 = 1, grade 2 = 9, grade 3 = 7 [18.4%], grade 4 = 2 [5.2%]), 6 dogs experienced gastrointestinal toxicity (grade 1 = 3, grade 2 = 2, grade 3 = 1 [2.6%]), and 4 dogs had both gastrointestinal and hematological toxicity (grade 1 = 4). Of the 19 dogs that experienced hematological toxicity, 16 experienced it following carboplatin. Two dogs (5.2%) had doxorubicin dose reductions, 11 dogs (28.9%) had carboplatin dose reductions, and no dog had dose reductions in both drugs (total 13 dose reductions). One dog (2.6%) had a doxorubicin dose delay, eight dogs (21.1%) had a dose delay dose in carboplatin, and one dog (2.6%) had a delay in both drugs (total 10 dose delays).

Of the possible prognostic factors evaluated, ALP level (normal or high) (p = 0.020), duration of clinical signs prior to amputation (p = 0.099), whether or not chemotherapy was completed (p = 0.001), and whether or not dose reductions or dose delays were performed (p = 0.099), met the criteria on univariate analysis to enter into multivariate analysis. Of those, only ALP level (normal or high) (p = 0.004) and whether or not chemotherapy was completed (p = 0.001) were found to significantly impact survival time on multivariate analysis.

Twenty-five dogs (seven censored) with a normal ALP level had a median survival of 378 days (12.6 mo) (range, 102– 1446 days) with a 1-yr survival of 52.3% and 2-yr survival of 13.4%. Ten dogs (two censored) with a high ALP level had a median survival time of 163 days (5.4 months) (range 92–487 days) with a 1-yr survival of 11.7% and 2-yr survival of 0% (Figure 1) (p = 0.004).

View Figure

• Download as Powerpoint
• Download Figure

For 27 dogs (six censored) that completed chemotherapy, the median survival time was 378 days (range 138–1446 days) (12.6 mo). One- and 2-yr survival percentage was 53.9 and 17.3%, respectively. For 11 dogs (3 censored) that did not complete chemotherapy, the median survival time was 148 days (range 92–309 days) (4.9 mo). One-yr survival percentage was 0% (Figure 2) (p = 0.001).

View Figure

• Download as Powerpoint
• Download Figure

The difference between low and high histologic grade was not statistically significant (p = 0.150); however, this data is presented as being possibly of interest. For 33 dogs (6 censored) with high histologic grade, the median survival time was 317 days (10.6 mo) (range 92–769 days), and 1- and 2-yr survival percentage was 41.1 and 10.0%. Four dogs (three censored) had low histologic grade, the median survival could not be calculated and their mean survival time was 1120 days (37.3 mo) (three dogs alive at 305, 385, and 1446 days) (Figure 3).

View Figure

• Download as Powerpoint
• Download Figure

Discussion

The improvement in median survival times and proportion of long-term survivors brought about by adding chemotherapy to amputation for dogs with appendicular osteosarcoma demonstrates that this is a chemosensitive disease in its early stages. However, despite many attempts at improved chemotherapy approaches, the proportion of dogs cured or with very long survival remains intractably low at around 20%. Moreover, the outcomes with a variety of different chemotherapy approaches are very similar, with median survival times on the order of 10.5 mo for cisplatin monotherapy, 10.1 mo for carboplatin monotherapy, 9.5 mo for doxorubicin monotherapy, 10.5 mo for doxorubicin/carboplatin alternating combination therapy, and 11 mo for doxorubicin/cisplatin same-day combination therapy. One- and 2-yr survival proportions are also similar, at 46 and 11%, respectively, for cisplatin monotherapy; 37 and 19%, respectively, for carboplatin monotherapy; 40 and 21%, respectively, for doxorubicin monotherapy; 48 and 18%, respectively, for doxorubicin/carboplatin alternating combination therapy; and 48 and 28%, respectively, for doxorubicin/cisplatin same-day combination therapy.^16–35 A recent randomized phase 3 trial comparing alternating carboplatin and doxorubicin to 6 cycles against 6 cycles of carboplatin monotherapy reported a statistically significant difference between median DFI of 25 dogs each receiving the two protocols (135 days for combination therapy versus 425 days for monotherapy). The reason for the apparently worse outcomes (DFI 135 days) for the 25 dogs in the combination therapy arm of that study compared to previously reported dogs receiving the same protocol in non-randomized studies (50 dogs with a median DFI of 202 days and 32 dogs with a median DFI of 227 days) is unclear, however.^32,33,36

Chemotherapy dose intensity is a measure of dose of drug per unit of time, and it is a general rule of chemotherapy biology that increased conventional chemotherapy dose intensity increases anti-tumor response. The two ways to increase dose intensity are by increasing drug dose and by reducing the time over which the chemotherapy protocol is given (reducing the intertreatment intervals). Because of an apparently somewhat better proportion of long-term survival seen with a dose-intense same-day combination therapy (28% 2-yr survival and 17% 3-yr survival) (Table 1), we hoped to render some advantage to dogs by increasing the dose intensity of therapy using accelerated scheduling.³⁰ In the protocol examined here, reducing the intertreatment intervals increased the dose intensity of the doxorubicin component by 50% (or relative dose intensity of 1.5) and the dose intensity of the carboplatin component was left unchanged (or relative dose intensity of 1.0), resulting in an overall protocol dose intensity of 1.25 relative to most other published chemotherapy protocols for appendicular osteosarcoma in dogs. However, the outcomes of the dogs treated with this protocol were very similar to those previously published and noted above, with median survival time of 10.6 mo and 1- and 2-yr survival proportions of 43 and 14%, respectively. This suggests that either the amount of dose intensification achieved by the schedule acceleration we used is insufficient to provide a clinically meaningful increase in antitumor efficacy or that the disease is inherently resistant to further increase in antitumor efficacy based on intensification of these drugs. If inherent resistance to these drugs is the limiting factor in treating dogs with osteosarcoma then it may be that a new class of drugs or an entirely new treatment approach, such as, for example, a cancer stem-cell targeted approach, is needed to make a clinically meaningful further improvement in median survival time and proportion of long-term survivors for dogs with appendicular osteosarcoma.

We found the toxicity of this protocol to be similar to most other published chemotherapy protocols for this disease. Only 5.2% of dogs experienced grade 4 toxicities and the same number required hospitalization for supportive care for complications of myelosuppression and recovered well. Most pet owners are particularly concerned about GI toxicities with regard to impact on quality of life, and this was a minor toxicity with this protocol, with only 1 dog experiencing grade 3 GI toxicity and no grade 4 GI toxicities. Both dose reductions and dose delays were more common with carboplatin than with doxorubicin, and, since the accelerated (dose intensified) component of the protocol was the doxorubicin component, we conclude that the acceleration did not cause clinically significant increased toxicity.

Three dogs (7.9%) died of causes unrelated to the cancer or the chemotherapy; however, an argument could certainly be made that the cause of death was not entirely unrelated to the treatment. These dogs were all euthanized with no evidence of osteosarcoma (for other orthopedic problems [two with progressive degenerative joint disease and one with anterior cruciate ligament rupture]) and it is certainly possible that these problems were accelerated, were more clinically significant, or were more difficult to manage because of the amputation.

Previous publications have identified a variety of prognostic factors for dogs with appendicular osteosarcoma, and we examined many of these factors in our population of dogs. One of the most strongly predictive factors to have been identified is ALP level at diagnosis.^4–6 In our population, ALP level was the only pre-treatment factor found to have statistically significant predictive value (Figure 1). Dogs with low grade tumors have been reported to have a better prognosis.⁷ This was not statistically significant in our population of dogs; however, the number of dogs with low grade tumors was very low (four dogs, 10.5% of the study population) and these four dogs did clinically do very well (Figure 3), so we propose this likely represents a type I error. Higher lymphocyte and monocyte counts have been shown to worsen the prognosis in studies but were not found to be significantly prognostic in our population of dogs.^6,9 Other factors found to be prognostic in other studies that were not found to be statistically significant in our study include age and tumor location.

A frequent question in osteosarcoma chemotherapy is how long chemotherapy needs to be continued, or how many cycles of chemotherapy should be administered, to achieve ideal outcomes. The only factor other than ALP level found to be significantly prognostic in our population of dogs was whether or not they finished the prescribed chemotherapy protocol (six treatments altogether) (Figure 2). This is obviously a potentially confounded factor because dogs may stop treatment because of disease progression. However, of the 11 dogs in this study that did not complete therapy, progressive osteosarcoma was the stated reason in only three of the dogs (the other reasons were as listed above). In our opinion, this certainly leaves the notion that dogs that stopped therapy early for financial or other reasons were more likely to experience disease progression as a real possibility. Of course, this does not answer the question as to how the other dogs would have done if they had received even more cycles of therapy, or if there is a “threshold” number of cycles of therapy less than six that would achieve the same outcome. A recent publication would appear to shed some light on this question by reporting a longer DFI for dogs treated with 6 cycles of carboplatin (425 days) than previously reported for dogs that received shorter courses of carboplatin monotherapy (3 to 4 cycles, 256 days; and 4 cycles, 257 days).^21,22,36 However, the published results cannot be directly compared to each other and, further, in the 6-cycle protocol study, only 16 of the 25 dogs in the carboplatin monotherapy group received the prescribed 6 cycles. A very recent study compared a large number of dogs with appendicular osteosarcoma treated with one of several different chemotherapy protocols after amputation: 6 cycles of carboplatin, 4 cycles of carboplatin, 3 cycles of alternating carboplatin and doxorubicin, and 5 cycles of doxorubicin given at either 2-wk or 3-wk intervals.³⁸ Again, there may have been some biasing in the stage of dogs in this study because the majority of them underwent pretreatment whole-body scintigraphy and, therefore, may have been lower stage than dogs in other studies. This study found no statistically significant difference in the DFI or survival times between dogs assigned to these different protocols; however, only 45% of the dogs assigned to the 6-cycle carboplatin protocol completed the protocol, although it was associated with the lowest frequency of adverse events (again, this was mostly for owner reasons, such as finance, rather than because of disease progression). Therefore, while this most recent report certainly represents an interesting addition to this discussion, the underlying question of the ideal number of cycles, duration of chemotherapy, and intertreatment interval, remains not yet conclusively answered. Given all this, we believe that the data reported here support completing the intended chemotherapy whenever possible.

Although the chemotherapy protocol was designed with a specific goal in mind, the accrual of patients and collection of data in this study were retrospective and so this study is subject to all the typical limitations of a retrospective study. These include patient and treatment selection biases and potential for flaws in the accuracy of the information collected. However, we do not feel that these limitations in this study invalidate the conclusions. Further, the number of cases in this study, 38 dogs, is not large, so there is certainly potential for sampling error (particularly noticeable in the lack of finding histologic grade statistically significant as a prognostic factor). However, the number is similar in scale to many other published studies of chemotherapy for dogs with appendicular osteosarcoma. Our primary purpose was to substantially improve outcomes for these dogs and the similarity in survival times found here relative to other published studies suggests that we did not achieve our aim.

The results presented here add to the robustness of the existing published data regarding expected survival times for dogs treated with conventional chemotherapy after amputation for appendicular osteosarcoma. More importantly, the results support the view that relatively minor alterations to schedule and dose of conventional chemotherapy are unlikely to result in substantially improved outcomes for dogs with appendicular osteosarcoma. Accordingly, we suggest that therapy using an entirely new class of drugs or an altogether different treatment approach will most likely be needed to significantly further improve outcomes for dogs with this disease.

Conclusion

Median survival time and 1- and 2- yr survival percentage for our accelerated protocol were similar to those previously reported with other chemotherapy protocols for dogs with similar disease. Adverse effects with this protocol were frequent, but they were predominantly low-grade and easily clinically manageable. ALP level (normal or high) (p = 0.004) and whether or not chemotherapy was completed (p = 0.001) were significantly prognostic for survival time on multivariate analysis. We conclude that while conventional chemotherapy is tremendously helpful to dogs with appendicular osteosarcoma compared to surgery alone, further improvements to survival time for these dogs are unlikely to result from conventional chemotherapy approaches and may require a new treatment paradigm.