Feline ventral abdominal wall angiosarcoma: haemangiosarcoma or lymphangiosarcoma? Clinical and pathological characteristics in nine cases

Abstract

Objectives

Angiosarcomas are rare malignant mesenchymal neoplasms of endothelial cell origin with a predilection to the ventral abdominal wall in cats. Larger case series describing this entity are lacking.

Methods

Two referral centre laboratory databases were searched for angiosarcoma of the ventral abdominal wall. Nine cases with a histological diagnosis were included. Immunohistochemistry (factor VIII and PROX-1 antibodies) was used to phenotype them as haemangiosarcoma or lymphangiosarcoma.

Results

All cats presented with a ventral abdominal mass, five of which were producing a serosanguinous discharge. Eight underwent tumour staging and pulmonary metastases were suspected in one cat (but not histologically confirmed). With histopathology alone, a diagnosis of angiosarcoma and lymphangiosarcoma was made in four and five cases, respectively. After immunohistochemistry, five cases had a haemangiosarcoma phenotype and four had a lymphangiosarcoma phenotype, including two cases of lymphangiosarcoma that were reclassified as hemangiosarcoma. Eight cats received treatment (either surgery with or without adjuvant therapies or medical management alone). Six cats were euthanased due to local disease progression. The median survival time for haemangiosarcoma was 166 days (range 137–381), and for lymphangiosarcoma it was 197 days (range 67–208). Two cats with haemangiosarcoma remained alive for a follow-up period of 329 and 580 days, respectively.

Conclusions and relevance

Feline ventral abdominal angiosarcomas are rare locally aggressive neoplasms. While histology often provides a diagnosis of angiosarcoma, immunohistochemistry is ultimately required to differentiate between haemangiosarcoma and lymphangiosarcoma phenotypes. Further studies are required to evaluate whether the different phenotypes have an impact on treatment response and outcome.

Introduction

Angiosarcoma (AS) is a rare malignant mesenchymal neoplasm of endothelial cell origin. ¹ For ASs that demonstrate features of lymphatic differentiation, subclassification with the term lymphangiosarcoma (LAS) is favoured, whereas those of blood vessel origin are termed haemangiosarcoma (HSA). ² With histopathology alone, the subclassification of AS phenotypes may be challenging. ³ Both have the overlapping characteristic feature of irregular anastomosing vascular channels lined by spindle-shaped mesenchymal neoplastic cells; however, the absence of intraluminal erythrocytes, hobnail cell appearance, kaposiform architecture and lymphocytic infiltrates would favour a diagnosis of LAS. ³ In humans, confirmation of vascular endothelial origin for AS is achieved with immunohistochemical markers, such as factor VIII-related antigen, also known as von Willebrand factor, CD31, CD34, FLI1, ERG and Ki67. ³ Differentiation between HSA and LAS relies on immunohistochemistry, and for this, a number of antigens that are preferentially expressed by lymphatic endothelial cells, including podoplanin, prospero homeobox protein 1 (PROX-1), vascular endothelial growth factor receptor (VEGFR)-3 and lymphatic vessel hyaluronan receptor (LYVE)-1, are available. ²

In cats, AS is rare; however, cutaneous and visceral locations are the most commonly reported primary sites.^4–18 Metastasis at the time of presentation appears uncommon but has been documented in the lungs, liver, spleen, myocardium, pericardium, diaphragm, omentum, mesentery, mediastinum and regional lymph nodes.^6–9 Within the feline AS population, a predilection for the subcutis of the caudal ventral abdominal wall is reported and this has subsequently been termed feline ventral abdominal wall angiosarcoma (FVAA). ¹⁰ Affected animals typically present with solitary, poorly circumscribed, locally infiltrative masses with ecchymosis, lymphoedema with or without serosanguinous discharge.^{10 –18} Of these, an LAS phenotype appears favoured and from 18 cases previously reported in the literature, the perceived prognosis is guarded, with the majority of cats being euthanased due to disease progression within a period of days to months (range 3 days to 13 months), irrespective of treatment modalities pursued.^10–18 For cases of FVAA that favour a HSA phenotype, there is a comparative paucity of information available in the literature, often admixed among cutaneous and subcutaneous HSAs. Based on a limited number of cases that have experienced prolonged survival times with surgery alone or in combination with chemotherapy, it is possible that phenotypic subclassification may provide additional prognostication for AS in this location.^19–21 In cats, a diagnosis of neoplasia of endothelial cell origin is usually achieved based on histological features, as described above in humans, alongside immunohistochemistry (IHC) for CD34, CD31 and factor VIII. ²² Diagnostic tools previously described for confirmation of an LAS phenotype in cases of feline FVAA include PROX-1,^17,18 LYVE-1 ¹⁴ or a demonstration of a discontinuous basement membrane with methenamine silver–Periodic acid–Schiff stain (Jones’ stain) ¹¹ or electron microscopy.^12,14

The aim of this study was to collate the largest set of data on FVAA to date and thereby describe clinical presentation, tumour staging, diagnostic tests, potential treatment options and outcomes.

Materials and methods

Inclusion criteria and data collection

Histopathology reports from two laboratories at referral centres were searched for cats with AS between January 2000 and December 2018. Cats were included if they had a histological diagnosis of AS arising from the ventral abdominal wall. Patients without available medical records and formalin-fixed tissue samples were excluded from the study population. Cats with no follow-up available after a histological diagnosis of AS were also included. Ethical approval was gained before starting the study as per institution guidelines (approval number 06-2019).

Data retrieved from clinical records included the following: signalment; primary tumour description; previous treatments before referral; cytological or histological sampling of the primary mass; staging performed (ie, bloodwork, thoracic and abdominal imaging); and treatment. Where applicable for gross disease, the response to treatment was defined by the response evaluation criteria in solid tumours (RECIST) method: complete response (no residual macroscopic disease); partial response (tumour decreased by ⩾30%); stable disease (tumour decreased by <30% or increased by <20%); and progressive disease (tumour increased by ⩾20%). ²³ Follow-up information was obtained from the medical records of the referral institutes.

Histology and immunohistochemistry

Formalin-fixed tissue was retrieved, and histopathology samples were reviewed by a board-certified pathologist. Fresh-cut tissue sections, with a thickness of 4 µm, from formalin-fixed, paraffin-embedded tissue blocks were labelled with a mouse-monoclonal antibody against PROX-1 (Ab199359; Abcam) and a rabbit-polyclonal antibody against factor VIII (A0082; Dako). Heat-induced antigen retrieval was performed using a pH 6 buffer (Bond ER1; Leica) for 10 mins (factor VIII) and a pH 9 buffer (Bond ER2; Leica) for 10 mins (PROX-1) at 90°C. The Bond Polymer Refine Detection kit (Leica) was used for visualisation with a haematoxylin counterstain. Normal feline skin and lymph nodes were used as positive controls. When present, surrounding histologically normal blood and lymphatic vessels were used as internal positive controls. Negative controls were prepared by replacing the primary antibody with Leica Antibody Diluent (Leica) only. A positive immunolabelling was indicated by the presence of distinct brown cytoplasmic (factor VIII) or nuclear (PROX-1) staining of neoplastic cells. A diagnosis of HSA subtype was confirmed when the AS was factor VIII-positive and PROX-1-negative, and LAS subtype was confirmed when the AS was both factor VIII-positive and PROX-1-positive.

Statistical analysis

Descriptive statistics were applied to the data collected. In cases of complete response, the disease-free interval (DFI) was calculated from the time of surgery until local or distant recurrence was documented. For patients treated in a gross disease setting, the progression-free interval (PFI) was calculated from the time of initiating treatment until documented progressive disease. Overall survival time (OST) was calculated from the date of diagnosis to the time of death (of any cause). A Kaplan–Meier survival plot was used to estimate the survival of the overall population as well as that of cases of HSA compared with cases of LAS. This statistical analysis was performed using the commercial Prism software v8.4.3 (GraphPad Software).

Results

Signalment and clinical presentation

Nine cats met the inclusion criteria. Eight were domestic shorthair cats and one was a domestic longhair cat. Six were female spayed cats and three were male castrated cats. The median age was 10 years and 3 months (range 5–16 years). In all cases, there was no long-term pre-existing history of diseases or medication administrations reported.

All cats presented for a ventral abdominal wall mass. The median duration of clinical signs before presentation was 20 days (range 10–76 days). Six masses (67%) had evidence of ecchymosis and five (56%) appeared ulcerated, producing a serosanguinous discharge (Figure 1). Two cats (22%) had pitting oedema of the inguinal region, one which also involved the proximal right hindlimb. Five masses (56%) were described as poorly demarcated. The median primary tumour size was 4.5 cm in maximum width (range 2.0–10.0 cm). Before tumour biopsy, six cats received empirical treatment with antimicrobial therapy alone, with or without non-steroidal anti-inflammatory drugs (NSAIDs) or glucocorticoids. None of these treatments resulted in the resolution of the presenting clinical signs.

Figure 1.

Appearance of feline ventral abdominal angiosarcoma on presentation (case 4). The mass was ill defined and there was accompanying ecchymosis and serosanguinous discharge

Diagnostic investigations

Haematology was performed in all nine cases. Abnormalities were identified in six cats and included a mild to moderate neutrophilia (4/9; range 14.08–27.2 ×10⁹/l; reference interval [RI] 2.5–12.5), lymphopenia (2/9; 0.6 and 0.74 ×10⁹/l; RI 1.5–7.0) and a mild non-regenerative anaemia (haematocrit 0.23 l/l; RI 0.29–0.46). Serum biochemistry was performed in eight cases. Three cats had mild abnormalities, including elevated alanine aminotransferase (155 IU/l; RI 5–60), hypoalbuminemia (23 g/l; RI 24–40) and hypercalcaemia (total calcium 2.7 mmol/l; RI 2.1–2.6) with normal ionised calcium (1.37 mmol/l; RI 1.00–1.40). One cat had an assessment of prothrombin and activated partial thromboplastin clotting time. The former was within RI and the latter was mildly prolonged (11.4 s; RI 7.0–11.0).

Eight cats had bicavitary imaging performed: five with thoracic radiographs and abdominal ultrasound and three with CT. Pulmonary nodules were detected in one cat using CT. Fine-needle aspiration (FNA) cytology of the primary mass was performed in two cases for which the interpretation of one was non-diagnostic and the other was supportive of sarcoma.

Surgical biopsy of the primary tumour

Five cats underwent excisional biopsies and four cats had incisional biopsies. In case 5, primary wound closure was achieved with an inguinal fold advancement flap. For those undergoing excisional biopsy, histologically assessed tumour-free margins were complete in one (20%) case, with the narrowest horizontal and deep margins of 3 mm and 5 mm, respectively. For the remaining four cats, the margins were incomplete. Postoperative complications were described in 4/5 cats undergoing excisional surgery. Two cases had mild wound dehiscence that required no further surgical intervention and were allowed to heal via secondary intention. In case 5, complete healing was not observed, and recurrence of a mass lesion was grossly appreciated, followed by confirmation with incisional biopsies 67 days postoperatively (Figure 2). Case 6 developed herniation of its abdominal viscera into the subcutis, which required revision surgery. In cases 5 and 6, the wounds were swabbed for culture and sensitivity. A moderate mixed growth of beta-haemolytic Streptococcus species and coagulase positive Staphylococcus species was detected in case 5, prompting treatment with enrofloxacin at 5 mg/kg PO q24h for 7 days (Baytril; Bayer). Pseudomonas aeruginosa was cultured in case 6; thereafter, the patient received marbofloxacin at 2 mg/kg PO q24h for 10 days (Marbocyl; Vetoquinol). Of the four cats that underwent an incisional biopsy, mild wound dehiscence requiring no further intervention was described in two cases.

Figure 2.

(a) Day 0 postoperative appearance after incomplete excision of a feline ventral abdominal wall angiosarcoma (lymphangiosarcoma-phenotype) using an inguinal fold advancement flap (case 5). (b) Local regrowth (green arrow) was subsequently confirmed on histopathology after an incisional biopsy at day 67. (c) The patient was euthanased at day 97 owing to progressive local disease and wound dehiscence

Histopathology and immunohistochemistry

A histological diagnosis of AS and LAS was made in four and five cases, respectively. Four cases had evidence of intraluminal erythrocytes within channels lined by spindle-shaped neoplastic cells and five cases had anastomosing channels without intraluminal erythrocytes. In eight cats, the neoplastic cells had positive cytoplasmic expression for factor VIII (two strongly, five moderately and one weakly positive), while the remaining cat was negative. Four cats had strong positive nuclear staining for PROX-1 and five were negative. With the addition of IHC, five cats were assigned an HSA phenotype and four cats an LAS phenotype (Figures 3 and 4).

Figure 3.

Haemangiosarcoma (HSA) phenotype (case 1). (a) Photomicrograph with haematoxylin and eosin stain (×200 magnification). Spindle-shaped neoplastic cells formed blood filled vascular channels. (b) Neoplastic cells demonstrated moderate positive cytoplasmic staining for factor VIII (×400 magnification) confirming endothelial cell origin. (c) Nuclei of neoplastic cells were negative for the PROX-1 (×400 magnification) consistent with an HSA phenotype

Figure 4.

Lymphangiosarcoma phenotype (case 4). (a) Photomicrograph with haematoxylin and eosin stain (×200 magnification). Spindle-shaped neoplastic cells formed vascular channels with the absence of intraluminal erythrocytes. (b) Neoplastic cells demonstrated moderate positive cytoplasmic staining for factor VIII (×400 magnification) confirming endothelial cell origin. (c) Nuclei of neoplastic cells were diffusely and strongly positive for PROX-1 (×400 magnification) consistent with an LAS phenotype

Treatment and outcome

After excisional surgery, four cats received adjuvant treatment (including systemic chemotherapy or tyrosine kinase inhibitors). Of the remaining four cats that underwent incisional biopsies, one cat was lost to follow-up after diagnosis, one cat was palliated with an NSAID and two received chemotherapy. Table 1 summarises the study population.

Table 1.

Characteristics of the study population

Case	Signalment	Description of primary lesion	Metastasis	Histological diagnosis	Immunohistochemistry		Phenotype after IHC	Treatment(s)	Outcome
					Factor VIII	PROX-1
1	9 yo FS DSH	Non-ulcerated mass (5 × 3 cm) with ecchymosis	N	AS	+	–	HSA	S(CE) and adjuvant treatment with doxorubicin (1 mg/kg IV q3wks) and cyclophosphamide (50 mg total dose PO q3wks) for 5 cycles	DFI >580 days with routine restaging performed
2	5 yo MN DSH	Ulcerated mass (10 × 5 cm) discharging serosanguinous-like fluid	N	AS	+	–	HSA	S(IE)	DFI >329 days. Remained clinically well with no evidence of local recurrence
3	7 yo MN DSH	Non-ulcerated mass (4.1 × 3.6 × 8.4 cm); inguinal and right hindlimb oedema	NA	AS	+	+	LAS	IB; meloxicam (0.1 mg/kg PO q24h)	Euthanased at 208 days owing to advanced local disease
4	10 yo FS DSH	Ulcerated mass (4.5 cm) with ecchymosis; discharging serosanguinous-like fluid	N	AS – LAS	+	+	LAS	IB	NA
5	10 yo FS DSH	Ulcerated mass (2.0 cm); discharging serosanguinous-like fluid	N	AS – LAS	+	+	LAS	S(IE) with inguinal fold advancement flap; adjuvant doxorubicin (1 mg/kg IV q3wks) for 1 dose	Local recurrence at 67 days; euthanased at 97 days owing to progressive local disease and wound dehiscence
6	12 yo FS DSH	Ulcerated mass (7.0 × 2.3 × 3.0 cm) with ecchymosis; discharging serosanguinous-like fluid	N	AS – LAS	–	+	LAS	S(IE); adjuvant toceranib phosphate (2.6 mg/kg PO q24h on a Monday–Wednesday–Friday schedule)	Herniation of abdominal viscera postoperatively requiring revisional surgery; local recurrence at 120 days; arrested under anaesthesia at 197 days at the time of assessing progressive local disease
7	10 yo FS DSH	Non-ulcerated mass (2.5 × 2.0 cm) with ecchymosis	N	AS – LAS	+	–	HSA	S(IE) and adjuvant doxorubicin (1 mg/kg IV q3wks) for 5 doses	Local recurrence of a mass at 311 days; euthanased at 381 days owing to progressive local disease
8	14 yo MN DLH	Ulcerated mass (2.0 × 3.0 × 1.5 cm) with ecchymosis; discharging serosanguinous-like fluid	Y	AS – LAS	+	–	HSA	IB; chlorambucil (4 mg/m2/day) and meloxicam (0.1 mg/kg PO q24h)	Stable disease; PFI of 71 days; euthanased at 137 days owing to progressive local disease
9	16 yo FS DSH	Non-ulcerated mass (2.0 × 3.0 cm); inguinal oedema	N	AS	+	–	HSA	IB; prednisolone (0.5 mg/kg PO q24h) and carboplatin (240 mg/m2 IV q3–4wks) for 3 doses	Partial response; PFI of 105 days; euthanased at 166 days owing to progressive local disease

+ = positive; – = negative; AS = angiosarcoma; DFI = disease-free interval; DLH = domestic longhair; DSH = domestic shorthair; FS = female spayed; HSA = haemangiosarcoma; IB = incisional biopsy; IHC = immunohistochemistry; IV = intravenous; LAS = lymphangiosarcoma; MN = male neutered; N = no; NA = not available; PFI = progression-free interval; PO = per os; S(CE) = surgery with complete excision; S(IE) = surgery with incomplete excision; Y = yes; yo = years old

Five cats died due to local disease progression with a median survival time of 197 days (range 137–381). One cat died from cardiopulmonary arrest under anaesthesia at the time of investigating local mass regrowth at 97 days. Of the five cats that underwent primary tumour excision, three (60%) developed local recurrence with a median DFI of 120 days (range 53–311). Of these three cats, two with LAS phenotype experienced local recurrence at 67 and 120 days vs 311 days for the HSA phenotype. The median overall survival time for HSA cases was 166 days (range 137–381) and for LAS was 197 days (range 67–208). Two cats with HSA that underwent surgical excision were alive and disease-free at 329 and 580 days after diagnosis (Figure 5).

Figure 5.

Kaplan–Meier plot of overall survival time for cats with haemangiosarcoma and lymphangiosarcoma phenotypes. Marks indicate time of censoring

Discussion

FVAA is an uncommon malignancy with a limited number of cases reported in the literature. A common feature in the described cases is their clinical presentation of an ill-defined caudal abdominal wall mass with or without ecchymosis, serosanguinous discharge and oedema. Leading differential diagnoses for a solitary ventral abdominal wall mass include inflammatory (eg, cellulitis, eosinophilic granuloma complex, mastitis, panniculitis and steatitis), infectious (eg, mycobacterium species, nocardia species or actinomyces species), trauma and other neoplastic conditions (eg, mammary gland adenocarcinoma, cutaneous lymphoma). ¹⁰ Within our small number of cases, the distribution between HSA (55%) and LAS (45%) phenotypes arising from the ventral abdominal wall was similar and distinguishing between the two on clinical presentation alone was not possible. A definitive diagnosis of AS is usually achieved with histopathology, as was the case for all the patients in our study. Conversely, FNA cytology was only supportive of sarcoma in 1/2 of our cases, which reflects the typical behaviour of sarcomas being poorly exfoliative when minimally invasive diagnostics are attempted. ⁵

In the present cases, no pathognomonic clinicopathological abnormalities were detected on routine blood sampling. The most common haematological abnormality, documented in 44% of cases, was a neutrophilia. Possible causes for this include tumour-associated inflammation, a paraneoplastic syndrome or in response to secondary opportunistic infections. Paraneoplastic immune-mediated haemolytic anaemia has previously been reported in two cases of feline LAS, one of which was FVAA; however, this was not observed in any of our cases.^9,11

Tumour staging in our study revealed no detection of regional metastasis and a low detection of distant metastasis (12.5%) at presentation. Given the lack of regional lymph node sampling or subsequent extirpation, the true incidence of regional metastasis may have been underrepresented. In addition, in the one case (HSA phenotype) in which pulmonary metastases were suspected, these were not confirmed cytologically or histologically. Comparatively, the metastatic rate of previously reported feline cutaneous/subcutaneous HSA appears low and the most frequent location is the lungs. ²⁰ In a study by McAbee et al, ²¹ no regional or distant metastasis was detected in 18 cats with cutaneous/subcutaneous HSA for which 11 underwent both thoracic radiographs and abdominal ultrasound at initial presentation. Four of these cases were described as arising from the ventral abdominal wall. Pulmonary metastasis was reported by Johannes et al ²⁰ in 3/13 cats with HSA that were known to have complete tumour staging, with two of these cases arising from cutaneous or subcutaneous tissue of an unspecified location. For the previously reported cases of FVAA with a favoured LAS phenotype, metastasis was documented in three cats, one to the inguinal lymph nodes and two had distant metastasis (pleural and pericardial in one and cardiac, pericardial and mediastinal in the other).^{10 –12} Appendix 1 in the supplementary material summarises the previously reported FVAA cases in which an LAS phenotype has been assigned in the literature.

In our study, with histopathology alone, four cases were diagnosed as AS, while in five cases, an LAS phenotype was assigned. The use of IHC in our cohort was an important tool for confirmation of the tumours being of endothelial cell origin but also allowed differentiation between LAS and HSA phenotypes. In particular, two cases of LAS were reclassified as HSA, suggesting that based on histopathology alone, differentiation between phenotypes is not definitive. Based on this, it is possible that some of the previous cases of FVAA that had been assigned an LAS phenotype may be reclassified as HSA, with the addition of lymphatic specific IHC markers. Of particular interest was that one of our cases of LAS had negative expression for factor VIII-related antigen but positive expression of PROX-1. Negative factor VIII expression has not been previously described in cases of FVAA with an LAS phenotype (see Appendix 1 in the supplementary material); however, in human literature, this has been previously reported for AS believed to have at least partial lymphatic differentiation. ²²

The predominant challenge in the management of sarcomas in cats is local tumour control and surgical resection with wide margins, where possible, is considered the mainstay treatment. ⁵ This challenge was also apparent in our study where a high rate of incomplete excision (75%) and subsequent local tumour recurrence (60%) was observed. This is likely secondary to the type of surgery performed in most cases (ie, excisional biopsy), infiltrative behaviour of sarcomas, challenging anatomical location, primary tumour size and commonly ill-defined gross appearance of the mass. Therefore, before surgery, incisional biopsies would be recommended for improved treatment planning. Similar to our study, reported cases of feline subcutaneous HSA were also associated with high rates of incomplete excision (50–94%) and local recurrence (50–80%),^19–21,24 and in one study, the median time for local recurrence was 208 days. ¹⁹ However, for a subset of cases treated with aggressive surgical excision (with or without adjuvant therapies), median survival time ranged from 9 months to 4 years.^21,24 Interestingly, two cases of HSA in our study experienced prolonged survival but this was not observed in any cats with an LAS phenotype. This may suggest that a subset of cats with FVAA with the HSA phenotype may have improved outcomes when treated with surgery; however, larger studies are required. Of note, for one of our cases with a HSA phenotype in which excision was incomplete and no adjuvant treatment was pursued, a good survival outcome was still observed (DFI >329 days). Possible explanations may include challenges to accurately assess histological margins due to tissue shrinkage after excision and/or because of formalin fixation vs the destruction of residual tumour cells owing to the inflammatory response in the surgical wound or the subsequent development of tumour dormancy. ²⁵ For FVAA with an LAS phenotype, local tumour control also posed the most significant challenge in our study with both cases, where excisional surgery was performed experiencing local recurrence at 67 and 120 days. This was also evident in 50% of cases previously reported in the literature undergoing surgical excision and with follow-up available (see Appendix 1 in the supplementary material).

Regarding surgical complications, six of our cases experienced delays in wound healing; however, the majority were considered minor, with the exception of the two patients that developed visceral herniation and incomplete wound healing. Possible causes of delayed wound healing include incomplete excision and thus the possible effects of the ongoing tumour microenvironment and/or secondary infection (as detected in two cases).

In humans, AS makes up less than 2% of all soft tissue sarcoma and principally affects adults (median age 60–71 years). ²⁶ Cutaneous lesions represent approximately 60% of cases and are often located on the head and neck. ³ Reported rates of advanced or metastatic disease on presentation are in the range of 16–44% and the OST is 6–16 months. ²⁷ Radical surgery followed by adjuvant radiotherapy is currently the best treatment to achieve local control. ²⁶ The status of surgical margins and its association with outcome remains to be fully defined; however, overall clear margins tend to improve patient survival. ²⁷ The subclassification of AS based on phenotypes had previously been abandoned, as most studies had concluded that AS frequently showed mixed lymphatic and blood vascular endothelial differentiation, although a subset of tumours can undergo at least partial differentiation along the lymphatic endothelial lineage. ² Furthermore, there was questionable clinical relevance as, previously, there were no specific therapies directed to tumours with blood vessel or lymphatic differentiation. ² More recently, tyrosine kinase inhibitors, such as sorafenib and pazopanib, have been found to be beneficial in the treatment of AS in humans with LAS phenotype through inhibition of VEGF/VEGFR and/or platelet-derived growth factor receptor signalling pathways.^28,29 Comparatively, in cats, the role of systemic treatments for both HSA and LAS is yet to be fully defined. Adjuvant anthracycline-based treatments have significantly improved survival of canine patients with HSA; however, systematic evaluations are lacking in cats. ³⁰ In addition, further studies are required to assess whether LAS phenotypes would warrant alternative medical therapies to HSA. Recently, toceranib phosphate (Palladia; Zoetis) was used in a recurrent LAS at the level of the caudal mammary gland in a 5-year-old female neutered Boxer, which resulted in a partial response lasting beyond 3 months. ³¹ Only one of our patients received toceranib phosphate but the cat died as a consequence of local disease progression before the efficacy of this treatment could be assessed.

The limitations of this study are primarily due to its retrospective nature and limited case numbers, which is partially reflective of the rarity of this condition. In addition, there was heterogeneity in treatment and patient management, which was clinician-dependent and likely acts as a confounding bias. Finally, the treatment landscape would have differed over the 15-year period in which cases were retrieved; for example, the introduction of small molecule inhibitors such as toceranib phosphate (Palladia; Zoetis).

Conclusions

The aim of this retrospective study was to describe the clinical characteristics, treatments and outcome of cases of FVAA. This is a rare, locally aggressive neoplasm associated with a guarded prognosis in most cases. The main challenge is local disease control with a high risk for incomplete excision, local tumour recurrence and postoperative complications. Histopathology is a cornerstone in achieving a definitive diagnosis; however, IHC may be used to differentiate between HSA and LAS phenotypes. It remains to be defined whether confirming this differentiation in feline AS is prognostic alongside whether different treatment modalities may be pursued in the future, depending on the phenotype.