Feline Cutaneous Hemangiosarcoma: A Retrospective Study of 18 Cases (1998–2003)

Introduction

Hemangiosarcoma (HSA) is a malignant neoplasm arising from vascular endothelial cells.¹ Hemangiosarcoma is uncommon in dogs and cats, with a reported incidence of 0.3% to 2.0%.¹² Although the visceral form of HSA is the most common, primary cutaneous HSA has also been reported and accounted for 1.8% of all cutaneous tumors in one study of 340 cats.⁶

Previous studies suggest that the most common primary sites for cutaneous HSA in the cat are the pinna, lateral face, and inguinal or abdominal subcutis.⁷⁸ In dogs, the most common sites of origin are the ventral abdomen, prepuce, and pelvic limbs, with median survival times of 780 days reported when tumors were confined to the dermis alone.⁹ Published reports in cats have had limited clinical follow-up and have not specifically addressed the treatment, prognosis, and factors affecting survival.⁷⁸

The purposes of this study were to describe the clinical characteristics and survival times in cats with primary cutaneous HSA and to determine the significance of surgical excision on the disease. The hypothesis examined was that cutaneous HSA in cats has a good prognosis regardless of tumor location and size, if complete surgical excision is achieved.

Materials and Methods

Medical records of 18 cats diagnosed with cutaneous HSA at The Animal Medical Center from February 1998 to September 2003 were evaluated. The diagnosis of cutaneous HSA was obtained via either incisional (i.e., surgical wedge or punch biopsy) or excisional (i.e., aggressive surgical excision) biopsy. Only cats with a complete medical record and tumors limited to the skin, with or without underlying muscle involvement, were included in the study. Cats with cutaneous HSA that extended into bone or gingival tissues were excluded.

Data extracted from the medical records included age at time of diagnosis, breed, sex, tumor location, tumor size (calculated as the sum of the two largest measurements of the mass), treatment administered, survival time, disease-free interval, and cause of death. Results of three thoracic radiographs (n=18), abdominal ultrasonography (n=11), and echocardiography (n=5) performed at the time of initial diagnosis were also evaluated.

Follow-up information was obtained by review of the medical records and through interviews with the owners or referring veterinarians. If a cat was alive at diagnosis but no further information could be obtained from the medical record, referring veterinarian, or owner, the cat was excluded from the study. Follow-up data included results of recheck examinations (n=16), three-view thoracic radiographs (n=5), and abdominal ultrasonography (n=2) when available. Survival time was defined as the time from the date of diagnosis to euthanasia or death. The disease-free interval was defined as the time from complete surgical excision of the tumor to the time of local recurrence or the time to death.

Histopathological sections of 17 of the cats were retrospectively reviewed by a single pathologist (Newman) for tissue of origin, confirmation of tumor type, completeness of surgical excision, and presence or absence of peritumoral lymphocytic inflammation or vascular invasion. The histological slides from one case were not reviewed, because the slides were unavailable; but the original diagnosis was cutaneous HSA with incomplete margins and no evidence of vascular invasion. Tumor location was divided into four categories: tumors involving only the dermis; tumors involving the subcutaneous tissues of the trunk or proximal limbs; tumors involving the subcutaneous tissues of the paw; and tumors involving the subcutaneous tissues and extending into the underlying muscle of the body wall.

Nonparametric statistics of all parameters were evaluated using both Kendall rank and Kruskall Wallis/Mann-Whitney U tests. Chi-square analysis was utilized for categorical versus noncategorical variables. Kaplan-Meier analysis was used to analyze median survival times, and a level of significance of P<0.05 was used for all statistical tests.^a

Results

The study population consisted of 18 domestic shorthair cats (11 castrated males and seven spayed females). The mean age of the cats at the time of diagnosis was 11.5 years (median 12.2 years; range 5.1 to 15.6 years). All cats had solitary tumors with no evidence of metastatic disease at the time of diagnosis based on three thoracic radiographs (n=18) and abdominal ultrasonography (n=11). Twelve cats had tumors located in the subcutaneous tissues of the trunk or proximal pelvic limb; two cats had tumors located in the subcutaneous tissues of the paw; two cats had tumors that extended into underlying muscle of the body wall; and two cats had tumors that involved only the dermis [see Table].

The clinical appearances of the lesions varied with the location of the tumor. One of the dermal lesions (case no. 4) was located on the nasal planum and had a calculated size of 0.2 cm. The other dermal lesion (case no. 18) was located on the lower eyelid of the right eye, and the size was not recorded. Both dermal lesions were small, red, raised nodules on the skin surface that did not cause the cats any clinical problems at the time of diagnosis. The two tumors located in the subcutaneous regions of the paw (case nos. 10, 11) had a calculated size of 1.0 cm and 2.0 cm, respectively. These two lesions resembled a focal area of bruising on the ventral metacarpal or digital pad that bled easily when traumatized. The 12 tumors located on the trunk or proximal pelvic limbs had a mean calculated size of 11.2 cm (range 4.2 to 24.4 cm). These 12 tumors tended to be large, firm, red or purple lesions with occasional small areas of necrosis and ulceration extending into the subcutaneous tissues. These latter lesions were painful when palpated and often had a serosanguineous discharge associated with the ulcerated region. The lesions of the trunk were most commonly located on the ventral abdominal or inguinal regions. Two cats (case nos. 9, 12) had masses that extended into the underlying body wall musculature on the lateral abdomen, with calculated sizes of 12.0 cm and 17.0 cm, respectively. These two invasive tumors resembled the lesions confined to the trunk but tended to have more extensive red or purple discoloration, superficial ulceration, and necrosis. Despite the different clinical appearances of the tumors, there was no significant association between tumor location and median survival time (P>0.05).

Histopathologically, the majority of the tumors were comprised of pleomorphic endothelial cells that lined variably sized, blood-filled vascular spaces. A few tumors (n=2) had a more solid pattern. The two tumors that involved only the dermis were quite superficial, and one displayed marked coagulative necrosis. Most of the tumors (n=16) originated in the subcutis and were generally poorly circumscribed, infiltrative, and often composed of markedly anaplastic endothelial cells. No vascular or lymphocytic infiltration was noted in the surrounding tissues.

Surgical excision was performed in 10/18 (55%) cats. In each case, all of the grossly visible disease was removed. If the tumor was located in the subcutaneous regions of the trunk or proximal pelvic limb, the mass was excised with a surface margin extending 1 to 3 cm beyond the tumor and a deep margin of at least one fascial plane. Full-thickness body wall resection was performed in three cats, and the resulting defect was either reconstructed primarily (n=1) or with polypropylene mesh (n=2).^b Amputation of the digit was performed in one cat (case no. 10) with a mass located in the subcutaneous tissues of the right, front, third digit. Complete surgical margins were achieved in 5/10 (50%) cats. The tumors with complete surgical margins were located in the subcutaneous tissues of the proximal pelvic limb (case no. 2) and the dermis of the nasal planum (case no. 4), and those with full-thickness body wall resections were located in the subcutaneous tissues of the trunk (case nos. 1, 3, 5). The mean calculated tumor size of these five masses was 7.4 cm (range 0.2 to 14.0 cm). All of the cats that had complete surgical excisions were alive with no evidence of disease (local recurrence or metastasis) at their last evaluation. These five cats had a mean disease-free interval of 479 days (median 420 days; range 120 to 1186 days [see Table]). None of the cats with clean surgical margins received chemotherapy postoperatively.

Tumors that were incompletely excised involved either the subcutaneous tissues of the trunk (case nos. 6, 7, 8), subcutaneous tissues of the digit (case no. 10), or they extended deeper to the body wall musculature (case no. 9). The mean calculated tumor size of these incompletely excised masses was 6.8 cm (range 1.0 to 12 cm). The median survival time for the five cats with tumor at the surgical margins was 910 days (range 90 to 1460 days). One of the cats had local recurrence of disease 270 days postoperatively and was euthanized 912 days after the initial diagnosis. Three cats were euthanized for hypertrophic cardiomyopathy at 1460 days, 90 days, and 360 days postoperatively with no signs of local recurrence or metastasis. One cat was euthanized 1003 days postoperatively for chronic renal failure with no signs of local recurrence or metastasis. Although no necropsies were performed on the study cats, none of the cats had overt signs of metastatic disease at the time of death. Two intravenous doses of postoperative doxorubicin were given to two of the cats with incomplete surgical excisions (case nos. 8, 9). These two cats had survival times of 1003 days and 1460 days and were both euthanized for medical reasons unrelated to their cutaneous HSA.

Eight of the 18 cats did not have surgery for their tumors. Three cats died or were euthanized within 10 days of diagnosis because of disseminated intravascular coagulation (case no. 15), hypertrophic cardiomyopathy with anuric renal failure (case no. 12), and anemia and anorexia (case no. 17). Two cats were euthanized 60 days after diagnosis because of recurrent discharge from their tumors and increased tumor size (case nos. 16, 18). Two cats did not undergo surgical excision of their tumors because of their concurrent hypertrophic cardiomyopathy, and they were eventually euthanized at 60 and 180 days postdiagnosis. One cat with a mass originating in the subcutaneous region of the left forepaw (case no. 11) was euthanized 210 days postdiagnosis when the tumor finally caused a nonweight-bearing lameness. The cats that did not have surgery had masses located in subcutaneous tissues of the paw (n=1), lateral abdominal body wall (n=1), subcutaneous tissues of the shoulder (n=1), subcutaneous tissues of the lateral cervical region (n=1), subcutaneous tissues of the ventral abdomen (n=3), and the dermis of the right lower eyelid (n=1) [see Table]. The calculated tumor size of these masses was 13.1 cm (range 2.0 to 24 cm). The median survival time for the cats that were not operated was 60 days (range 4 to 210 days).

The Kaplan-Meier median survival time for all 18 cats was 912 days, and 40.6% of the cats were alive ≥3 years after diagnosis [Figure 1]. Comparisons of median survival times based on age, sex, tumor location, and tumor size were not statistically significant. Comparison of median survival times based on tumor size via Cox Proportional Hazards Analysis was not prognostic (P=0.0836). Cats that had surgery had a significantly longer median survival time [Figure 2]. The Kaplan-Meier median survival time for the cats that underwent surgery was >1460 days, while the cats that did not undergo surgery had a median survival time of 60 days (P=log rank 0.0004) [Figure 2].

Discussion

The mean age (11.5 years) of the cats reported here was consistent with previous reports on cutaneous HSA in cats.^6–9 While two previous studies found that male cats were more likely to be affected, there was no sex predilection in the present study.⁷⁸ Although the cats in this study were all domestic shorthair cats, previous reports on cutaneous HSA have shown no breed predilection.^6–8

Previous studies on feline HSA have documented a low metastatic potential for the cutaneous form of HSA.⁷⁹ However, a recent retrospective study of five cats with cutaneous HSA reported that they all died from complications associated with metastasis within 77 days of diagnosis, despite surgical intervention.⁹ All of the tumors in the study reported here were solitary, with no evidence of metastatic disease at the time of diagnosis. The authors do recognize the fact, however, that none of the cats in the present study underwent necropsies, and the potential for undetected metastatic disease at the time of death does exist. Three cats that were originally presented in poor clinical condition died or were euthanized soon after biopsy. These cats might have had metastatic disease that was not detected on thoracic radiography and abdominal ultrasonography.

It is uncommon to assign a histological grade to HSA. The tumors in this study were not graded histologically, but it is plausible that some tumors are inherently more aggressive (i.e., high grade) than others and that this study population had a higher proportion of low-grade lesions that responded favorably to surgical excision. Future studies are needed to explain the discrepancies in the metastatic propensity of feline cutaneous HSA and to identify cats that have a poor prognosis.

Lymphocytic infiltrates are associated with longer survival times in humans with cutaneous HSA.¹⁰ Only one prior study on feline cutaneous HSA commented on the presence or absence of lymphocytic infiltrates within the lesions and only speculated on the significance of this finding.⁷ None of the cats in the study reported here had lymphocytic infiltrates within the tumors or had tumors that extended into the surrounding vasculature or lymphatics. The significance of lymphocytic infiltrates in feline cutaneous HSA is unknown, and further studies are necessary to determine their significance.

Tumors in this study occurred most commonly in the subcutaneous regions of the trunk or proximal pelvic limbs. In comparison, a previous study found that the subcutaneous regions of the head and tips of the ears of fair-skinned cats were the most common regions affected.⁷ Numerous studies in dogs and humans have also documented that the cutaneous form of HSA tends to occur in nonhaired, nonpigmented skin.^10–12 Potential predisposing factors for cutaneous HSA in humans are vascular stasis, previous radiation therapy, trauma, and sun exposure.¹² Solar radiation has been proposed as a cause of cutaneous HSA in dogs, and cutaneous HSA has also been experimentally induced using ultraviolet radiation in dogs.¹³ Results of the study reported here showed only two tumors occurring on the cutaneous tissues of the head (i.e., nasal planum dermis and palpebral dermis), and affected cats were not fair skinned. Although it was not directly investigated, these results suggest that ultraviolet radiation and/or solar exposure were not causative factors in the disease.

Tumor location did not statistically affect median survival time in this study. In a previous study of 25 dogs with cutaneous HSA, surgical excision of superficial dermal HSA resulted in a median survival time of 780 days, whereas median survival times for subcuticular lesions were <1 year.¹¹ In the present study, cats with tumors located in the subcutaneous tissues and those with muscle involvement had the longest survival times. It is interesting to note that the four cats with tumors involving the ventral abdominal region either died or were euthanized within 90 days of diagnosis. Three of these masses were too large to attempt surgical excision, and it was difficult to determine if tumor location or size played a more important role in the survival times of these cats.

Previous studies in humans have indicated that tumor size was prognostic in cutaneous HSA, with tumors <5 cm in diameter having significantly longer survival times.¹⁰¹² None of the previous studies in cats evaluated tumor size, and, therefore, its clinical significance is unknown.⁶⁷⁹ Three of the cats in this study with the largest tumors were either euthanized or died shortly after diagnosis, and all of these cats were initially in poor clinical condition. This study failed to show a statistically significant relationship between tumor size and median survival time, but the relatively small sample size may have affected the statistical outcome. With a larger study population, tumor size may have been found to be a negative prognostic factor for cutaneous HSA.

The mean calculated tumor size for the cats with clean surgical margins was greater than for cats with incomplete surgical excisions. Although the significance of this data was unknown, it suggested that even large cutaneous HSA tumors can be resected completely if the lesion is located in an anatomical region that allows aggressive surgical excision (e.g., body wall).

Of the five cats with incomplete surgical excisions, only one had evidence of local tumor recurrence. The remaining four cats were euthanized for other chronic diseases (e.g., hypertrophic cardiomyopathy or renal failure) unrelated to their tumors. Even though these four animals were eventually euthanized, their survival times were >1 year. It was possible that some cats were euthanized before local recurrence could occur (as in case no. 7). This cat only lived 90 days after diagnosis and was euthanized for hypertrophic cardiomyopathy. Potentially, the cats that had incomplete surgical excisions benefited from the partial resections, because they provided alleviation of clinical signs such as bleeding, bruising, and pain, and thus allowed a better quality of life. Two cats with incomplete surgical excisions received additional doxorubicin therapy, and both of these cats were eventually euthanized for medical reasons unrelated to their cutaneous HSA. Too few cats were treated with doxorubicin to reach any conclusions on its efficacy, but chemotherapy has resulted in prolonged survival in dogs with cutaneous HSA.¹⁵ Vincristine, doxorubicin, and cyclophosphamide have been used in combination in six dogs with cutaneous HSA, with a median survival time of 425 days.¹⁵ Further studies are necessary to investigate the potential benefits of adjunctive chemotherapy in feline cutaneous HSA, especially when complete surgical excision cannot be achieved.

Previous studies of cats with cutaneous HSA reported a mean survival time of 385 days with surgical excision.⁸ In the present study, complete or partial surgical excision of cutaneous HSA in cats resulted in long survival times (mean survival time 622 days, range 90 to 1460 days) regardless of age, tumor size, or tumor location. In addition, all cats with complete surgical excisions were alive at the completion of this study (mean survival time 479 days, range 120 to 1186 days), and the median survival time was 60 days for unoperated cats, indicating that aggressive surgical excision of cutaneous HSA provided a better long-term prognosis than previously reported.

Conclusion

Cutaneous HSA in cats may be a more treatable disease than previously reported. Cats in this study that had complete surgical excision of their tumors had a favorable long-term prognosis. Based on the results of the current study, further studies investigating the use of surgery and adjunctive chemotherapy for this disease are warranted.