Use of Single-Agent Carboplatin as Adjuvant or Neoadjuvant Therapy in Conjunction With Amputation for Appendicular Osteosarcoma in Dogs
Survival following amputation and administration of single-agent carboplatin for treatment of appendicular osteosarcoma (OSA) in dogs was retrospectively examined. Records of 155 dogs with appendicular OSA treated with amputation and single-agent carboplatin were included from 14 centers. Any carboplatin dosage, number of doses, and protocol schedule were eligible for inclusion. The median disease-free interval (DFI) was 256 days. The median overall survival time was 307 days. Similar prognostic survival factors were identified in this study as reported in prior studies of canine appendicular OSA. Median DFI and survival were comparable to those reported in the original Bergman et al publication. Carboplatin treatment improves the survival probability in dogs with appendicular OSA compared to amputation alone and remains an acceptable alternative to adjuvant treatment with cisplatin.
Introduction
Osteosarcoma (OSA) of the appendicular skeleton is a biologically aggressive neoplasm in dogs, representing approximately 85% of skeletal malignancies in the species. Appendicular OSA is associated with significant pain, decreased quality of life, and short survival without treatment. Surgical excision (typically via amputation of the affected limb) is the standard of care for palliation of pain; however, median survival time following amputation alone has been reported to be 14.5 to 25 weeks, with 1-year survival rates of 11.5% to 21%.1–5 In the majority of cases treated with amputation alone, death is a result of development of metastatic disease.2
Various chemotherapy protocols have been reported to improve survival following amputation in dogs, with reported median survival times for cisplatin-based protocols ranging from 262 to 413 days.1,4–6 The first reported use of carboplatin as a single adjuvant chemotherapy following amputation for canine appendicular OSA revealed similar efficacy to that of cisplatin protocols. Median disease-free interval (DFI) and overall survival time were 257 days and 321 days, respectively, and a 35.4% 1-year survival rate was reported.7
Carboplatin has advantages over cisplatin in dogs with OSA because of its similar reported efficacy, ease of administration compared to cisplatin (i.e., no need for concurrent diuresis), and generally mild and reversible toxicity (usually grade 2 or lower based on reported Veterinary Cooperative Oncology Group [VCOG] toxicity grading scheme).8 However, in two recent investigations, the median overall survival durations following carboplatin as single- agent adjuvant treatment in conjunction with amputation were less favorable (230 days9 and 20710 days) than previously reported.8 These findings prompted this retrospective study by the VCOG, which includes most cases from the animal populations of the three aforementioned reports in addition to cases not previously reported.7,9,10 Chemotherapy was initiated in a neoadjuvant fashion in the latter two studies; however, chemotherapy treatments were administered in an adjuvant manner in the initial report. In this study, therefore, timing of chemotherapy administration in relation to amputation was a variable that was evaluated. 7,9,10 This study was designed to determine the survival outcomes for dogs treated with amputation and carboplatin chemotherapy in a far greater number of animals than reported in the three previous studies and to support the Bergman et al study conclusion that carboplatin has efficacy similar to cisplatin.7,9,10
Materials and Methods
Case Characteristics
The VCOG membership submitted cases for this investigation. Cases were identified by a chronological computer search, a chronological manual search of medical records, and a review of cases from previous studies (n=50).7,9,10 For animals included from prior studies, records were reviewed and information was reabstracted using case report forms identical to those used for dogs not previously reported. Animals from the Khanna et al publication that were treated with both carboplatin chemotherapy and Oncolar (a long-acting analogue of somatostatin) were not included in this report. Inclusion criteria included dogs with histologically confirmed OSA of the appendicular skeleton, no evidence of metastatic disease at the time of amputation, and treatment with only amputation and single-agent carboplatin chemotherapy. All carboplatin dosage and schedule protocols were eligible for inclusion.
Case variables included age, gender, neuter status, breed, and weight. For purposes of statistical analysis, ages of dogs were divided into ≤5 years versus >5 years at diagnosis. Gender was examined using the variables of intact male, intact female, neutered male, and neutered female, as well as all intact versus neutered dogs. Breed analysis was divided between golden retriever versus rottweiler (the two most prevalent breeds) versus greyhounda versus other pure breeds versus mixed-breed dogs. Weights were divided into variables of <25 kg and ≥25 kg.
Tumor Characteristics
Tumor-related variables examined included histological grade (low, intermediate, high); total serum alkaline phosphatase (SAP) level (within normal limits versus elevated serum levels); and location of primary tumor (e.g., humerus, distal femur, radius, proximal tibia, ulna, other appendicular location). Humeral location versus all other locations was also examined.
Histological specimens from 71 cases were provided for review by one board-certified veterinary pathologist (Powers). No specimens were provided from cases treated neoadjuvantly. The subtype of OSA based on matrix produced was described as osteoblastic (bone matrix), fibroblastic (collagenous matrix), chondroblastic (cartilage matrix), telangiectatic (large, blood-filled spaces formed), and osteoclastic (little matrix with numerous osteoclasts).
Histological grades for 69 of the tumors were scored with the following four parameters:
Amount of matrix was described as abundant matrix, moderate amount of matrix, or scant matrix. Percent necrosis values were divided into the following categories: 0% to 10%, 11% to 49%, and ≥50%. Nuclear pleomorphisms were described as mild, moderate, or marked. Mitoses counts per 10 high-power fields were divided as follows: 1 to 9, 10 to 25, and >25. Final histological grade (low, intermediate, high) was determined using a combination of the four parameters described above.
Treatment Protocols
Protocol-related variables evaluated included duration of clinical signs prior to amputation (0 to 45 days, 46 to 90 days, 91 to 180 days, >181 days); number of intended chemotherapy doses (three versus four doses); timing between amputation and chemotherapy (neoadjuvant therapy was defined as day of amputation or prior, versus postamputation administration); average chemotherapy dosage (<270 mg/m2 versus 270+ mg/m2); and average chemotherapy dosing interval (<25 days versus ≥25 days).
Outcomes and Statistical Considerations
Disease-free interval (DFI) was defined as the interval between amputation and tumor recurrence. Overall survival was defined as the interval between amputation and death. Disease-free interval and overall survival values were determined using Kaplan-Meier life table analysis. Dogs were censored in DFI analysis for the following reasons: recurrence or metastasis had not occurred before the end of the study period; they were lost to follow-up; or they died before relapse. Dogs were censored in survival analysis for the following reasons: they were lost to follow-up; death was not caused by OSA; or they were alive at the end of the study period. Differences in DFIs and overall survival between dogs grouped by several variables were assessed by univariate analysis using the logrank test. Multivariate analysis was performed for variables with a P value <0.25 on univariate analysis using forward stepwise multiple logistic regression to fit the Cox proportional hazards model, producing risk ratios, 95% confidence intervals (CIs), and significance levels. Statistical significance was set at P≤0.05.
Results
Case Characteristics
A total of 157 case submission forms were provided for review. Of these, 155 dogs from 14 different facilities met inclusion criteria. Animals included in the review were diagnosed with appendicular OSA between December 2, 1991, and July 25, 2001. The final attempt at determination of status for those alive and not considered lost to follow-up was made on June 1, 2005. Included in the study were 123 purebred and 31 mixed-breed dogs (one breed was not stated). Breed, gender, weight, and age characteristics are presented in the Table. Considerable variation existed in duration of clinical signs before amputation [see Table].
Tumor Characteristics
The breakdown of tumor location, SAP concentrations, and tumor histology and grade are presented in the Table. Seventy-one tumor specimens were available for evaluation of subtype, and 69 tumor specimens were evaluable for histological grade.
Treatment Protocols
Protocols differed in the number of intended treatments, the timing of chemotherapy in relationship to amputation, administered dosage, and dosing intervals. These are described in the Table. One case did not have intended number of treatments stated, and for one case only two doses were intended. The mean, median, and range of carboplatin prescribed doses were 292 mg/m2, 300 mg/m2, and 130 to 351 mg/m2. In dogs receiving neoadjuvant carboplatin, only the first dose was given neoadjuvantly, 1 week prior to amputation, with two or three subsequent treatments given adjuvantly after amputation. Use of nonsteroidal antiinflammatory drugs (NSAIDs) varied considerably regarding drug type, dosing, and perioperative timing.
Outcomes
Median DFI for all dogs was 256 days. Rates for DFIs of 1, 2, and 3 years were 30.3%, 18%, and 10.3%, respectively. The median overall survival time was 307 days. Rates for survival times of 1, 2, and 3 years were 36.8%, 18.7%, and 11%, respectively. Results of univariate analysis are presented in the Table. Of the variables analyzed, only proximal humeral location, elevated SAP, and age >5 years were associated with shorter DFI and overall survival.
Following multivariate analysis, three variables retained negative prognostic significance for DFI: age >5 years (P=0.038; risk ratio of 2.1 [95% CI of 1.04 to 4.25]); proximal humeral location (P=0.043; risk ratio of 1.6 [95% CI of 1.02 to 2.58]); and elevated SAP (P=0.0006; risk ratio of 0.45 [95% CI of 0.29 to 0.71]). Following multivariate analysis, two variables maintained negative prognostic significance with respect to overall survival: proximal humeral location (P=0.046; risk ratio of 1.6 [95% CI of 1.01 to 2.63]) and elevated SAP (P=0.0005; risk ratio of 0.44 [95% CI of 0.28 to 0.70]).
While evaluation of toxicity was not a goal of this study, and bone marrow and gastrointestinal toxicity data were inconsistently reported, toxicity was generally grade 2 or lower based on the reported VCOG toxicity grading scheme.8 Worth noting is that 37 dogs experienced at least one episode of gastrointestinal toxicity during the course of their treatment, and three dogs died from nontumor-related causes during their treatment periods (one death was strongly attributed to chemotherapy, and two other deaths were possibly treatment related). Dogs treated with carboplatin should be carefully monitored for gastrointestinal and bone marrow toxicity.
Discussion
The animal population in the current report has similar signalment characteristics to those of previous reports.7,11 A prognostic impact of age was noted in the study, with younger dogs having improved DFIs and overall survival times on univariate analysis. This contrasts to a previous report where a negative impact was found with younger age at diagnosis.2 However, age was not significant for overall survival when evaluated using multivariate analysis. While speculative, younger dogs (representing the smaller of two known incidence peaks for OSA) may represent a form of the disease in this animal population that is more susceptible to initial chemotherapy or is biologically less aggressive because of some unknown factor.
The proximal humeral location was identified in this investigation as having a negative prognostic impact, and this has been noted in a previous report.7 This may represent a biologically more aggressive location, or it may simply represent a site that becomes more advanced before detection because of an abundance of surrounding soft tissues not present at other common sites. Due to the retrospective nature of this study, measurement of tumor burden was not possible, but this variable should be critically evaluated in future prospective studies. Histological grade was not evaluated with respect to tumor location; evaluation of this combination of variables should be performed in future studies. The study population in the current investigation reiterates the negative prognostic impact of elevated SAP documented in other reports.10,12 It has been theorized that SAP may be elevated due to increased osteoblastic activity, and increased levels may indicate a more biologically aggressive osseous neoplasm.
This study did not reveal outcome differences between various carboplatin dosages nor dosing intervals <25 and ≥25 days. Chemotherapeutic principles dictate that increased dosage and decreased dosing intervals should improve therapeutic outcome; however, the differences between the dosage and dosing interval protocols used in this study may not have been enough to impact prognosis. Additionally, so few dogs received the lower dose that the analysis would likely be underpowered to detect any clinically relevant difference.
No outcome effect was attributable to short-term or chronic NSAID use in the current population. Nonsteroidal antiinflammatory drugs have several potential antineoplastic effects, and the reader is referred to several reviews on the subject.13–15 The heterogenous type, dose, and duration of NSAID use in this population do not allow any conclusions to be made regarding potential activity.
Tumor grade and histological subtype had no impact on prognosis for either DFI or overall survival, as has been reported previously.16 However, less than half of tumor samples were available for analysis; therefore, nonrepresentative sampling may have contributed to the lack of significance, and the low numbers in some histological subtypes lack sufficient power for comparison.
The impetus behind this retrospective study was to confirm or dispel reports that carboplatin was not as efficacious for canine OSA as cisplatin. While this is not a prospective, randomized study, the DFI and overall survival reported in this current population mimic previous reports using cisplatin and seem to dispel concerns of lesser activity with carboplatin.9,10
Conclusion
In this study, amputation followed by adjuvant or neoadjuvant carboplatin administered as a single agent was determined to provide DFIs and overall survival times similar to those in the original report by Bergman et al. In that original report, similar efficacy had been identified between carboplatin and cisplatin for treatment of canine appendicular OSA. Though study population and treatment protocols are very similar, it is difficult to make a meaningful comparison between a retrospective study and previously published prospective studies. However, these response durations appear to dispel concerns generated by two smaller reports that claim carboplatin may be inferior to cisplatin when used for canine appendicular OSA.
Age of animal, SAP level at time of diagnosis, and tumor location were prognostic for disease-free survival. Pretreatment SAP level and tumor location were also prognostic for overall survival. Although treatment with carboplatin improved survival compared to historical reports of amputation alone, the minority of dogs survived ≥1 year, further demonstrating the need for the development of novel therapeutic agents or combinations to improve survival for the majority of dogs diagnosed with appendicular OSA.
Greyhound breed evaluated as a separate variable due to reporting of the greyhound breed as a negative prognostic factor in research abstract reported by Couto et al, at the 2005 midyear conference of the Veterinary Cancer Society.
Acknowledgments
Cases comprising <10% of overall reported cases were submitted by the following members of the Veterinary Cooperative Oncology Group (VCOG), in descending order of case submissions: M Kent, MK Klein, R Gamblin, K Rassnick, L Fox, A Moore, M O’Brien, P Woods, A Smith, and CL Harris.
Contributor Notes
