Introduction

Hemangiosarcoma (HSA) is a malignant neoplasm that has been historically classified as a vascular endothelial tumor, but new evidence supports its origin as a pluripotent bone marrow progenitor cell, as the tumor cells can retain traits of cancer stem cells with self-renewal, chemoresistance, and aggressive tumorigenicity.¹ HSA has been commonly reported in dogs and cats but has also been reported in various other species. In dogs, visceral HSA is the most common form, accounting for 70% of all HSAs, with the spleen being the most common primary location.^2–4 Other common primary sites include the right atrium, subcutaneous tissue, liver, and the retroperitoneal space. As a neoplasm composed of endothelial cells, HSA can ultimately occur at any site with a vascular supply.⁵

Visceral canine HSA is typified by a very aggressive biologic behavior, with rapid and widespread metastasis.^2–4,6 Reported median survival time (MST) of canine splenic HSA treated with surgery alone is 48–86 days and 60–172 days with adjuvant chemotherapy.^3,4,6 Primary cardiac HSA has a similarly poor prognosis with an MST of 189 days with aggressive surgical resection and adjuvant chemotherapy in one report.⁷ Dogs with HSA arising in certain nonvisceral locations (retroperitoneal space, subcutaneous tissue, dermis, and tongue) may have a better outcome, with reported MST of 272–1189 days; however, in one study of dogs with subcutaneous and intramuscular HSAs, an MST of 172 days was achieved, and only 25% dogs lived >1 yr.^8–12

The falciform ligament of the liver (ligamentum falciforme hepatis) is a remnant of the primitive ventral mesentery.¹³ In adult dogs, the falciform ligament is expanded by adipose, and it persists only from the diaphragm to the umbilicus.¹³ Primary HSA of the falciform ligament is an extremely rare disease with only two cases reported in the veterinary literature^14,15 In both of those reports, the dogs lived considerably longer than dogs with the more common visceral form of HSA, with survivals of >4 yr and >11 mo. We describe the clinical presentation, treatment, and outcome of seven dogs with primary HSA of the falciform ligament.

Materials and Methods

The pathology databases of 14 veterinary schools (Colorado State University, Cornell University, Louisiana State University, Mississippi State University, North Carolina State University, The Ohio State University, Oklahoma State University, Texas A&M University, Tufts University, University of California Davis, University of Georgia, University of Michigan, University of Minnesota, and the University of Pennsylvania) and 1 private reference pathology laboratory (Animal Reference Pathology) were searched for cases of canine HSA involving the falciform ligament. Inclusion criteria included cases with a histopathologic diagnosis of HSA of the falciform only. Dogs who had HSA of other sites in addition to falciform involvement were excluded. One exception was made for a necropsy case (dog 4) in whom the primary HSA lesion was suspected to be in the falciform as a result of its size relative to the presumptive metastases. The search of these databases yielded five cases, provided by Cornell University, North Carolina State University, and the University of Pennsylvania, that met the criteria of inclusion. One private referral practice, Florida Veterinary Referral Center, also contributed two cases.

Histopathologic samples were reviewed by a board-certified pathologist (S.S.C.) to confirm the diagnosis of HSA and provide tumor grade. Tumor grade was determined based on a previously described grading system for HSA.³ In this grading scheme, neoplastic cell differentiation, nuclear pleomorphism, percent of tumor necrosis, and mitotic count (MC; mitoses in 10 high-powered fields) were individually scored and totaled to arrive at a final grade. Neoplastic cell differentiation was assessed based on the following criteria: 1 = well-differentiated with numerous, irregular vascular channels; 2 = moderately differentiated with at least 50% of the tumor revealing well-defined vascular channels; and 3 = poorly differentiated, mostly consisting of solid sheets of cells with few vascular channels. Nuclear pleomorphism was scored as 0 (none), 1 (mild), 2 (moderate), or 3 (marked). The amount of necrosis was assigned as 0 = none, 1 = <25%, 2 = 25–50%, and 3 = >50%. The MC was determined for each neoplasm, and the mitotic score was 0 = MC of 5–10, 1 = MC of 11–20, 2 = MC of 21–30, and 3 = MC >30. The total score for each neoplasm was calculated by adding the individual scores, and grades were assigned as 1 = total score of 0–5; 2 = total score of 6–9; and 3 = total score of 10–12.

Medical records were reviewed for signalment, presenting complaint, staging results, treatment, and outcome information. Dogs were staged according to the modified World Health Organization classification scheme. In this system, stage I indicates a nonruptured tumor in a primary site measuring <5 cm; stage II tumors have ruptured, have spread to local lymph nodes, or are >5 cm; and stage III tumors have evidence of distant metastasis or invade adjacent structures. Outcomes assessed included overall survival (OS) time and time to progression (metastasis and/or recurrence). OS was defined as the time from surgery to death or euthanasia. Variables evaluated for associations with overall and PFS time included the presence or absence of hemoabdomen, tumor size (≤5 versus >5 cm), histologic grade (1, 2, or 3), mitotic score (0, 1, 2, or 3), stage (I, II, or III), treatment with or without chemotherapy, and presence or absence of clinical signs. The Kaplan-Meier product limit method was used to estimate survival times for each potential risk factor. A P value <.05 was considered statistically significant. Dog 4, who was diagnosed at necropsy, was not included in survival analyses.

Results

Seven dogs were identified for inclusion in the study (Table 1). The signalment of these seven dogs were as follows: 6 yr old neutered male boxer (dog 1), 10 yr old spayed female golden retriever (dog 2), 14 yr old spayed female Labrador retriever (dog 3), 4 yr old intact female mixed-breed dog (dog 4), 13 yr old spayed female Australian cattle dog (dog 5), 9 yr old spayed female West Highland white terrier (dog 6), and a 6 yr old spayed female Siberian husky (dog 7). The median age was 9 yr. Body weight ranged from 9 to 34 kg, with a median of 28 kg.

TABLE 1 Summary of Patient Characteristics, Findings, Treatments, and Outcomes

View Fullscreen

The tumor was an incidental finding in three dogs (dogs 3, 5, and 6). Presenting complaints for the four clinical dogs included lethargy (n = 3), vomiting (n = 1), collapse (n = 1), polyuria and polydipsia (n = 1), and right pelvic limb lameness that progressed to paraplegia (n = 1). Two dogs (dogs 1 and 7) had hemoabdomen at presentation.

Complete blood counts and serum chemistry results at the time of diagnosis were available for all but one dog (dog 4). Two dogs were anemic (hematocrit of 19 and 21%; reference range: 40.3–60.3%), and two dogs were thrombocytopenic (36,000 and 60,000/uL; reference range: 177,000–398,000/uL). Two dogs had elevated alanine aminotransferase (131 and 400 U/L; reference range: 12–118 U/L), one dog had elevated aspartate aminotransferase (167 U/L; reference range: 15–66 U/L), one dog had elevated alkaline phosphatase (414 U/L; reference range: 14–120 U/L), one dog was hypocholesterolemic (122 mg/dL; reference range: 138–317 mg/dL), one dog was panhypoproteinemic (albumin 2.2 g/dL, reference range: 2.5–3.7 g/dL; globulin 2.1 g/dL, reference range: 2.4–4.0 g/dL), and one dog was azotemic (blood urea nitrogen 59 mg/dL, reference range: 5–30 mg/dL; creatinine 2.9 mg/dL, reference range: 0.7–1.8 mg/dL). Of the two dogs presenting with hemoabdomen, both were thrombocytopenic, one was hypoproteinemic, and one was anemic.

Tumor staging, consisting of abdominal ultrasound and thoracic radiographs, was performed prior to surgery in all dogs that had their tumor removed antemortem (n = 6). One of these dogs (dog 6) also had computed tomography (CT) of the abdomen. No dogs had evidence of pulmonary metastatic disease on imaging. On abdominal ultrasound, the falciform tumors were described as arising from the liver (n = 2), mesentery (n = 2), liver or falciform (n = 1), and mid-abdomen (n = 1). The CT radiology report described the falciform mass as arising from either the gastric wall or body wall.

Six dogs underwent exploratory laparotomy with surgical excision of a falciform mass. One dog (dog 4) was diagnosed on necropsy, and this case will be described in more detail separately. No dogs had gross evidence of abdominal metastasis at the time of surgery. One dog had splenic nodules that were not evaluated further. Two dogs had hepatic nodules that were biopsied. Histopathology of the hepatic nodules showed no evidence of neoplasia in either patient (hepatocellular necrosis and fibrosis in dog 6 and nodular hyperplasia in dog 3).

Histologically, the neoplasms were variably encapsulated, moderately cellular, infiltrative, dissected through and replaced the falciform adipose tissue, and contained regions of hemorrhage and necrosis. Neoplastic cells formed variably sized vascular channels and often lined or dissected between bands of fibrous connective tissue. The neoplastic cells were spindloid with variably distinct cell borders, a moderate amount of eosinophilic to lightly basophilic, fibrillar to homogeneous cytoplasm, and a central to paracentric ovoid nucleus with finely stippled chromatin and 1–3 prominent nucleoli. Nuclear and cellular pleomorphism were mild to marked. The MC ranged from 3 to 34 (median 7). Five tumors had a mitotic score of 0, and one each had a mitotic score of 2 and 3. There was a single histologic grade 1 tumor, five grade 2 tumors, and one grade 3 tumor (Table 2).

TABLE 2 Summary of Histologic Tumor Characteristics

View Fullscreen

Four dogs were treated with adjuvant chemotherapy. One dog (dog 7) was treated with six doses of weekly vinorelbine, followed by metronomic therapy (daily piroxicam given concurrently with daily etoposide alternating with daily cyclophosphamide in 3 wk cycles). Rescue chemotherapy was attempted with various drugs including doxorubicin, chlorambucil, vinorelbine, and temozolomide. One dog (dog 5) was treated with metronomic therapy (daily cyclophosphamide at 15 mg/m²) and an allogeneic tumor vaccine^a. One dog (dog 2) was treated with metronomic therapy (daily cyclophosphamide at 15 mg/m²) alone. The final dog (dog 3) was treated with four doses of doxorubicin given 2 wk apart.

No dogs that were diagnosed prior to necropsy were alive at the time of writing. Excluding the dog diagnosed at necropsy, the dogs in this series had a median OS of 339 days (range 69–496 days) and a 1 yr survival rate of 50%. Dogs treated with some form of chemotherapy (n = 4) had a median OS of 394 days compared with 83 days for those who did not (n = 2); this difference was significant (P = .018; Figure 1). Progression data were available for four dogs, all but one of whom were treated with chemotherapy. Median progression-free survival (PFS) was 181 days for dogs treated with chemotherapy compared with 41 days for the dog not treated with chemotherapy (P = .083). Median PFS was 154 days for all dogs. Excluding the dog diagnosed at necropsy, dogs who presented with clinical signs (n = 3) had a median OS of 288 days. Dogs who did not have clinical signs (n = 3) had a median OS of 389 days. There was no significant difference in OS between these groups (P = .49). Median PFS was not significantly different between these groups (111 days for dog with and 304 days dogs without clinical signs; P = .43). Tumor stage, tumor size, and the presence/absence of hemoabdomen did not significantly influence PFS or OS.

View Figure

• Download as Powerpoint
• Download Figure

Dogs with grade 2 tumors had a median PFS of 111 days and median OS of 192 days. The one dog with a grade 1 tumor had a PFS and OS of 126 and 389 days, respectively. Dogs whose tumors had a mitotic score of 0 had a median PFS and MST of 126 days and 339 days, respectively. Progression data was not available for the single dog whose tumor had a mitotic score of 2, and this dog had an OS of 96 days. The single dog whose tumor had a mitotic score of 3 (the only dog with a grade 3 tumor) had a PFS and OS of 482 and 496 days, respectively. Median OS for dogs with tumors having a mitotic score of 0 was 339 days, and 296 days for those having a score greater than 0. This difference was not significant (P = .44).

Four dogs were euthanized as a result of suspected progressive disease. Of these dogs, ultrasound confirmed a liver mass in one dog, a midabdominal mass in one dog, and liver, splenic, and pancreatic nodules in one dog. Tissue samples were not obtained for diagnosis. One of the euthanized dogs (dog 6) had also been diagnosed with end-stage hepatic disease based on histopathology (marked multifocal hepatocellular necrosis with bridging fibrosis) at the time of HSA diagnosis. It is unclear if the clinical signs that led to euthanasia in this dog were related to hepatic disease or HSA progression.

One dog in this case series was diagnosed on necropsy. A brief summary of the case is as follows: dog 4 was a 4 yr old intact female mixed-breed dog who was presented for a 1 wk history of right pelvic limb lameness that progressed to paraplegia. Neurologic examination revealed that deep pain was absent from her right pelvic limb and was questionable in her left pelvic limb. Pain was elicited on palpation of the lumbosacral region. Spinal radiographs showed no evidence of disc mineralization, trauma, or diskospondylitis. An MRI of the caudal spine was performed and revealed an extensive, primarily right-sided, intramedullary lesion at the level of the L3–L4 intervertebral disk where it made up >95% of the cross-sectional area of the spinal cord. The appearance of the lesion was most consistent with hemorrhage or a hematoma. A cerebrospinal fluid tap was performed multiple times, but the samples were heavily contaminated with blood. The dog’s owners elected euthanasia. On necropsy, there were multifocal pulmonary nodules, a mass at the fornix of the brain, a mass of the spinal cord adhered to the dura at the level of L4–L6, and a 4.5 × 3.35 × 3 cm mass within the falciform fat. Histopathology of all of the lesions was consistent with a malignant neoplasm of vascular origin. Immunohistochemistry for von Willebrand’s factor was performed and the neoplastic cells stained positively, confirming the diagnosis of HSA. The liver, spleen, and heart were unaffected. The size of the falciform tumor relative to the other sites was suggestive of it being the primary tumor.

Discussion

Primary HSA of the falciform adipose is an extremely rare anatomic variant of HSA in dogs. After searching the pathology databases of 14 veterinary schools, only five cases were identified for this study. An additional two cases were submitted from a private specialty practice. Falciform neoplasia has only been reported in the veterinary literature twice and both cases were HSA. Falciform neoplasia in humans is uncommon, and to the authors’ knowledge, there have been no reports of primary angiosarcoma of the falciform ligament or surrounding adipose in humans. The most commonly reported neoplasm of the falciform in humans is the perivascular epithelioid cell tumor (PEComa). PEComa encompasses a group of mesenchymal tumors of perivascular epithelioid cells including angiomyolipoma, lymphangioleiomyomatosis, renal capsuloma, clear-cell “sugar” tumor, and clear-cell myomelanocytic tumor.¹⁶ These tumors range from benign to malignant in their behavior. The histogenesis of PEComas is not entirely known, but they may arise from mesenchymal stem cells.¹⁶

In this study, patient characteristics were similar to previous reports of canine HSA of other sites with a median age at diagnosis of 9 yr and median weight of 28 kg.^10,17,18 Clinicopathologic findings commonly associated with HSA such as anemia, hypoproteinemia, and thrombocytopenia were present in three dogs in this study. Two of those three dogs initially presented with hemoabdomen and the other had a cavitated abdominal mass on ultrasound. None of the dogs in this series had evidence of metastasis at the time of diagnosis except for the dog diagnosed at necropsy. It is possible that the metastatic rate of this anatomic variant of HSA is lower when compared with visceral HSA, which is commonly associated with advanced-stage disease.¹⁹ However, the study design excluded dogs who had primary HSA of other sites with falciform metastasis. Dogs with large solitary or multifocal splenic, hepatic, or cardiac lesions in addition to a falciform lesion were presumed to be primary visceral HSA with secondary falciform involvement and were excluded from the study. This may have falsely lowered the reported metastatic rate. Additionally, staging for all dogs did not include echocardiograms or advanced imaging such as thoracic or abdominal CT to evaluate for metastatic lesions. No dogs in this series had biopsy confirmation of either local tumor recurrence or distant metastasis. In order to avoid underestimating these rates as much as possible, the development of any mass in the peritoneal cavity was reported as disease progression (recurrence or metastatic disease).

Histologic grading of HSA originating from various sites was first reported in 1996.³ In that study, tumor grade did not significantly affect outcome, but nuclear pleomorphism and mitotic scores appeared to predict survival time and disease-free interval for dogs treated with surgery and doxorubicin. A second study evaluating histologic features as a predictor of outcome for dogs with HSA found that a higher mitotic score was associated with a significantly increased risk of death.²⁰ More recently, a study of dogs with stage II splenic HSA treated by splenectomy and chemotherapy found that mitotic score was significantly associated with survival time.²¹ In this case series, there were not enough cases per grade to evaluate the effect of grade on outcome. Dogs with a mitotic score of 0 had a slightly longer median OS of 339 days compared with those with a mitotic score of 1–3 (296 days), but this difference was not significant (P = .44). The prognostic use of this histologic grading scheme is questionable, particularly given the fact that the one dog in our study with a mitotic score of 3 (and an overall tumor grade of 3) had the longest PFS and OS compared with the other dogs in this series, which had grade 1 or 2 HSA.

The prognosis for dogs with splenic HSA treated by surgery alone is poor, with MSTs of 1–3 mo, which is comparable to the dogs in this study treated by surgery alone (MST 82.5 days).^4,21–23 Surgery and adjuvant chemotherapy (anthracycline based) increases the MST to 141–179 days and the 1 yr survival rate to 11% in dogs with splenic HSA.^{3,6,19,20,24–28} In addition to anthracycline-based chemotherapy, metronomic chemotherapy (cyclophosphamide, etoposide, piroxicam) following splenectomy for dogs with stage II HSA has also yielded OS of 178 days.²⁹ In this case series, dogs treated with adjuvant chemotherapy (n = 4) had an MST of 394 days, and the 1 yr survival rate was 50%. Seventy-five percent of the dogs were treated with metronomic chemotherapy at some point in their protocols. As adjuvant chemotherapy has been shown to improve survival in other forms of canine HSA, it stands to reason that adjuvant chemotherapy would improve survival times for dogs with primary falciform HSA as well.

There have been several reports of secondary HSA of the spinal cord. One report describes a dog with intramedullary spinal cord metastasis that was initially paraplegic and became ambulatory for 17 wk with prednisone treatment alone.³⁰ The dog was ultimately euthanized as a result of disseminated HSA. The one dog diagnosed with HSA of the spinal cord in this study had a rapid progression from right pelvic limb lameness to paraplegia within 1 wk, which resulted in euthanasia. The spinal cord tumor in this patient is not definitively a secondary site; however, given the large size of the falciform tumor found in this patient, it is suspected that the falciform was the primary site. Primary spinal HSA has been reported.³¹

The results of this study are limited by issues inherent to most retrospective evaluations, including a small group of patients, incomplete medical records, and lack of standardized diagnostics, treatment, and follow-up. The lack of postmortem examination data was also a limitation of this study. Although this is a small group of dogs, these findings, along with those reported in the two previously published case reports, suggest that dogs with primary falciform HSA may have improved outcomes compared with the more common visceral sites. A prospective evaluation of dogs with falciform HSA would be needed to confirm this hypothesis.

Conclusion

This case series describes the clinical presentation, treatment, and outcome of seven dogs with primary HSA of the falciform. In this population, dogs treated with surgery and chemotherapy experienced longer survival times and higher 1 yr survival percentages compared with what is reported in the literature for dogs with splenic HSA treated with surgery and chemotherapy. This retrospective study was limited to dogs who underwent surgical excision of their primary tumor; therefore, it is unknown how surgical intervention influences survival times in dogs with HSA of the falciform. The findings of this case series suggest the possibility of a more favorable prognosis for dogs with HSA confined to the falciform compared with the more common visceral locations. Adjuvant chemotherapy likely improves outcome.