Effects of Palliative Radiation Therapy on Nonsplenic Hemangiosarcoma in Dogs

Introduction

Hemangiosarcoma (HSA) is a vascular endothelial cancer that most commonly affects the spleen in dogs.¹ Surgery is the treatment of choice for splenic HSA, followed by adjunctive chemotherapy.¹ Median survival times range from approximately 3 months for surgery alone to approximately 6 months for surgery plus chemotherapy, and most dogs die of metastatic disease.¹ In rare cases, HSA can arise from nonvisceral organs.² In 1985, Brown et al. found that 17% of HSAs in dogs involved soft tissues of the trunk and extremities.² In cases of nonsplenic HSA, surgery is the primary treatment modality, but complete excision is not possible in all animals.^3,4

In humans, vascular endothelial tumors similar to HSAs are referred to as angiosarcomas.⁵ Angiosarcomas are rare in humans and most commonly occur in previously radiated areas of the body or on the face or scalp of elderly patients.^5,6 Prognostic factors include size, location, and depth of invasion.^5–8 Although surgery is the primary therapy, various studies in humans have demonstrated that angiosarcomas may be sensitive to radiation therapy.^5,8–10 Similar reports are lacking regarding the efficacy of radiation therapy in dogs with nonsplenic HSAs.

The goal of the radiation therapy evaluated in this study was to improve quality of life by reducing pain, tumor burden, and/or dysfunction. Traditional (i.e., intent is to cure) radiation therapy was not employed because of the assumption that most dogs would die of metastatic disease before or shortly after local control of the tumor could be achieved. Palliative radiation therapy has been used effectively for other tumor types, such as osteosarcomas, mast cell tumors, and melanomas.^11–15 The purpose of this study was to determine the effect of palliative radiation therapy on tumor response and survival time in dogs with nonsplenic HSAs.

Materials and Methods

Selection criteria included any dog with histologically confirmed non-splenic HSA that was treated with radiation therapy. Dogs were included regardless of overt metastasis (i.e., radiographic changes consistent with pulmonary metastasis or ultrasonographic changes consistent with visceral organ metastasis). Dogs with microscopic (i.e., incompletely excised lesions) or dermal disease were excluded, and only dogs with grossly measurable disease were included. Twenty dogs treated from June 1995 to May 2004 fit the inclusion criteria.

Data collected and factors evaluated included signalment, tumor location, stage of disease, and treatments (including the modality used [e.g., radiation alone or in combination with surgery/chemotherapy], number of doses, and whether the treatment was done before or after radiation therapy). Tumor response, adverse effects from radiation therapy, and survival time were also evaluated. Survival time was defined as the time from initiation of radiation therapy to the time of death.

Complete response was defined as resolution of all gross disease. Because measurements were not consistently recorded in the medical records, reduction of tumor size could not be defined as true partial responses (i.e., ≥50% reduction). Tumor size was considered decreased if the mass was subjectively smaller following radiation therapy. Dogs in which a reduction in tumor size was not observed, or in which tumor size did not change following radiation therapy, were classified as having no response.

If the dog was lost to follow-up or another cause of death was documented, the dog was censored from the data at the date the dog was last known to be alive. If a necropsy was not performed, and there was a high level of suspicion that the dog died from metastatic disease based on clinical signs and recent history (e.g., difficulty breathing, hemoptysis, hemoabdomen), the dog was considered to have died from the HSA.

Survival data was evaluated using the Kaplan-Meier product limit method. Logrank analysis and Wilcoxon’s signed rank test were used to evaluate the effect of location, stage, treatment, and tumor response on outcome.^a Significance was set using a P value of ≤0.05. Multivariate analysis was not performed because of the small sample size.

Results

Ten breeds of dogs were affected, with most being mixed-breed dogs (n=7). Other breeds included the golden retriever (n=4), Labrador retriever (n=2), basset hound (n=1), bullmastiff (n=1), Dalmatian (n=1), German shepherd dog (n=1), Siberian husky (n=1), Chinese shar pei (n=1), and Shetland sheepdog (n=1). Twelve dogs were male (11 castrated), and eight were spayed females. Mean age was 10±2.5 years (median 11 years; range 2.5 to 14 years), and mean weight was 33±10 kg (median 30 kg; range 14 to 52 kg).

Most of the tumors involved soft tissues (n=17), such as the distal limbs, chest wall, and neck region [see Table]. The extent of tumor invasion (i.e., subcutaneous, muscular) could not be determined from all histopathology reports because of the incisional nature of most biopsies. Only six dogs with soft-tissue masses had advanced imaging with computed tomography, and bony destruction was not identified. Three dogs had subcutaneous masses that were intimately associated with the adjacent musculature. Three dogs had HSAs in the retroperitoneal space.

Four dogs were considered to have overt metastasis (i.e., stage III disease) based on thoracic radiography or abdominal ultrasonography at the time of radiation therapy [see Table]. Three dogs had evidence of pulmonary metastasis at the start of radiation therapy. Splenic lesions were identified on ultrasonography in three dogs (including one of the dogs with pulmonary metastasis), but further diagnostic tests were not performed. Two of the dogs with initial splenic lesions were presented with hemoabdomen at the time of euthanasia.

Seven dogs had palliative radiation therapy as the sole treatment modality, primarily because surgical resection was thought to have too high a morbidity [see Table]. One dog had surgery and radiation, three dogs received chemotherapy and radiation, and nine dogs were treated with all three treatment modalities.

In seven dogs, surgery was initially performed, the tumor recurred, and palliative radiation therapy was subsequently done. In three other cases, palliative radiation therapy was performed prior to surgery, either in the hope of decreasing the size of the tumor (n=2) or because the tumor regrew and amputation was subsequently elected (n=1).

Twelve dogs received chemotherapy [see Table]. Seven of these dogs received chemotherapy following radiation therapy. Most dogs received doxorubicin alone. One dog received a dose of cisplatin followed by one dose of doxorubicin. Another dog with retroperitoneal disease was treated with five doses of doxorubicin, four cycles of COP (i.e., cyclophosphamide, vincristine, prednisone), and one dose of carboplatin.

A 6-MV linear accelerator was used to administer radiation in the form of photons.^b Total radiation doses ranged from 6 to 24 Gray [see Table]. Thirteen of the dogs received 24 Gray. Ten dogs were treated with 8 Gray fractions on days 0, 7, and 21. Three dogs (case nos. 2, 14, 17) were treated with 6 Gray fractions on days 0, 7, 21, and 28 because of the tumor location and underlying normal tissue structures (e.g., tail base, retroperitoneum, shoulder area). Fifteen of the 20 dogs completed their prescribed dose.

Fourteen dogs had subjective decreases in tumor size following palliative radiation therapy, and four of these dogs had complete responses. A decrease in tumor size was generally seen after one treatment (n=9). Decreased pain was reported in one of the dogs that did not demonstrate tumor shrinkage.

Three of the four dogs that had a complete response had tumors involving soft tissues [see Table]. Of the dogs with complete responses, only one (case no. 2) had suspicious lesions suggestive of distant metastasis at the start of therapy. One dog with retroperitoneal disease (case no. 4) received concurrent chemotherapy. All dogs exhibiting a complete response received a total of 24 Gray. Survival times for the dogs that had a complete response ranged from 55 days to 408 days. Two of the dogs died from metastatic disease, one dog died of local disease, and the remaining dog was considered to have died from other causes (e.g., difficulty ambulating pelvic limbs, anorexia).

Acute effects from radiation were generally minimal and self-limiting, with the exception of one dog (case no. 1) that developed persistent oral mucositis and moist desquamation. Acute effects included diarrhea (n=1), alopecia in the radiation field (n=2), and lethargy (n=1). A few days’ duration of diarrhea was present in one dog (case no. 14) following the second dose of radiation for a retroperitoneal mass. Although the primary mass in the ventral mandibular area (case no. 1) completely resolved, the dog was euthanized 55 days after the start of treatment because of tumor necrosis in the radiation field. At necropsy, HSA was identified in the skin, but pulmonary metastasis was not identified in this dog.

Median survival time calculated by the Kaplan-Meier product limit method was 95 days (mean 178±42 days; range 6 to 500 days) [Figure 1]. Most dogs (n=15) died of their disease. Six dogs died of local disease, seven died of metastasis, and two (case nos. 13, 14) died from either local disease or metastasis. In case no. 13, evidence of pulmonary and rib metastasis was seen on thoracic radiographs at the last visit, and the dog died with signs of progressive anemia and hematuria; no necropsy was performed. Case no. 14 was presented with weak pulses and pale mucous membranes; this dog developed cardiopulmonary arrest. No necropsy was performed to determine if the primary mass (retroperitoneal) ruptured or if the clinical signs were secondary to metastatic disease. Two dogs died of other causes. The remaining three dogs were censored; two dogs were alive at the last known follow-up (6 and 112 days), and the cause of death could not be determined in the third dog (this dog had a history of atrioventricular block and was euthanized because of decreased quality of life; no necropsy was performed).

In univariate logrank and Wilcoxon analyses, the only significant prognostic factor for improved outcome was location of disease (logrank P value=0.03; Wilcoxon P value=0.05) [Figure 2]. The median survival time for dogs with retroperitoneal masses was 408 days (n=3; mean 389±71 days; range 258 to 500 days), while the dogs with soft-tissue masses at other locations had a median survival time of 89 days (n=17; mean 84±9 days; range 6 to 151 days).

Discussion

Results of this study support the use of palliative radiation therapy for nonsplenic HSAs. Fifteen (75%) of the 20 dogs benefited from radiation treatment, and four (20%) dogs had complete responses. Most dogs tolerated the palliative radiation therapy well.

Signalment data were consistent with previously published information on splenic HSAs, with the exception of lack of breed predisposition.¹ Signalment data were not compared to other nonsplenic cases of HSA, as larger case series have focused more on cutaneous HSA, which appears to have a different biological behavior than other nonvisceral locations.^3,16,17

Median survival time following radiation therapy was approximately 3 months, which was comparable to surgery alone for splenic HSA.¹ Unfortunately, because of the small sample size, multiple predictive factors for a favorable response were not determined. Stage of disease did not influence survival. The effect of adjunctive chemotherapy was not determined; however, multimodality treatment did not influence survival times. Only location was found to be statistically significant. Given the small number of dogs (n=3) with retroperitoneal masses, it is questionable whether location would still be a prognostic factor in a larger study.

Limitations of this study were related to its retrospective nature and small sample size. Additionally, dogs received various treatments and different doses of radiation. Based on the gross tumor response seen, different doses and the scheduling of palliative radiation therapy warrant investigation. Two of the six dogs that died of local disease received <24 Gray, which suggested that higher doses may result in better responses. A larger number of dogs would be needed to determine optimal dose and scheduling for nonsplenic HSAs treated with palliative radiation therapy.

Conclusion

Twenty dogs with histologically confirmed nonsplenic HSAs were treated with palliative radiation therapy. Fourteen dogs had a reduction in tumor size, and four of these dogs had complete resolution of their masses following radiation therapy. Acute effects from treatment were generally mild. Late effects from radiation therapy were not observed, although survival times were shorter than the expected onset time of late effects. While there are still many questions to be answered regarding radiation therapy for canine HSA, based on the results of this study, palliative radiation therapy is a well-tolerated treatment modality that may be considered for nonsplenic forms of the disease.

Acknowledgment

The authors thank Mr. James zumBrunnen for help with the statistical analysis.