© 2002 Blackwell Science Ltd 15 Veterinary Dermatology 2002, 13, 15–21 sBlackwell Science Ltd Haemangiopericytoma: histological spectrum, immunohistochemical characterization and prognosis MAURIZIO MAZZEI,* FRANCESCA MILLANTA,* SIMONA CITI,† DAVIDE LORENZI* and ALESSANDRO POLI* *Department of Animal Pathology, and †Clinical Department, School of Veterinary Medicine, University of Pisa, Viale delle Piagge, I-56124, Pisa, Italy (Received 14 June 2001; accepted 17 October 2001) Abstract Canine haemangiopericytoma (CHP) is a vascular neoplasm thought to be derived from pericytes. The histological pattern and immunohistochemical profile were studied in 31 CHPs. Twenty-three subjects were followed for 2 years to evaluate the correlation among tumour location, histotype, immunostaining and outcome of the disease. Of the 31 CHPs examined, 20 exhibited a perivascular whorled pattern, 8 were storifom and 3 were epithelioid. All tumours were positive for vimentin and negative for cytokeratin, factor VIII-related antigen, glial fibrillary acidic protein and S-100 protein. Seventeen CHPs were positive for actin and nine co-expressed desmin. Six CHPs were also positive for CD34 antigen. The panel of immunohistochemical markers used confirmed the vascular lineage of CHP and aided in the exclusion of other mesenchymal tumours. Of the 23 dogs submitted to follow-up, 6 had recurrence or metastases of the primary tumour. The epithelioid pattern or a noncutaneous location were associated with a poorer prognosis. Keywords: canine haemangiopericytoma, CD34 antigen, immunohistochemistry, intermediate filaments, muscle actin, prognosis. INTRODUCTION Haemangiopericytoma (HP) is a soft tissue tumour first described in human beings by Stout and Murray in 1942 1 and recognized in dogs 4 years later. 2 Both human and canine HP (CHP) are cutaneous tumours of adult life and are more often found in the legs and hind limbs, respectively. Human HP has also been found in noncutaneous locations, whereas in dogs it is exclusively considered a tumour of skin and subcuta- neous tissue. 2,3 The putative cell of origin is the pericyte. In human beings, the origin of this tumour has been established on the basis of the tumour’s ultrastructure 4 and the immunohistochemical profile of tumour cells has been accurately investigated. 5,6 In the dog, however, the histogenesis of this tumour is not clear, because the tumour cell origin from pericytes has not been defini- tively confirmed. 7 Recently, an immunohistochemical study performed on 45 CHPs and other canine soft tissue tumours analysed the tissue marker expressed by this tumour. 8 The diagnosis of HP is known to confound even experienced pathologists, because the tumour cells lack readily identifiable morphological features. In human patients this tumour consists of tightly packed cells, with round to oval nuclei and moderate amounts of cytoplasm with ill-defined borders, around ramifying thin-walled, endothelium-lined vascular channels. Rarely, there are also focal spindle cell areas, but these cells are never arranged in long bundles or fascicles as in fibrosarcoma or synovial sarcoma. 9 HPs are often confused with other soft tissue tumours such as synovial sarcoma, malignant fibrous histiocytoma and angiosarcoma, all of which can display ramifying vessels. Microscopically, the CHP has a characteristic appearance with a typical ‘fingerprint’ pattern. The tumour often consists of layers of spindle cells arranged in a concentric fashion around a central vessel. 2,3,10 –12 In the dog, pericytes lack readily identifiable cellular features and histopathological diagnosis is therefore often based on the recognition of a characteristic pattern. This pattern may sometimes be lacking or may also be found in other soft tissue tumours such as Schwannomas, fibrosarcomas, cutaneous fibrous histi- ocytomas and leiomyosarcomas, so that differential diagnosis of these tumours may be very difficult. 8,13,14 The difficulties encountered in differentiating soft tissue tumours morphological or immunohistochemical grounds have recently led to the use of the term ‘spindle cell tumours of canine soft tissue’ to encompass this group of lesions. 15 CHP has a high recurrence rate, even if metastasis is quite uncommon. 11,13,16,17 The influence of histological patterns or the immunohistochemical expression of various tumour markers on tumour behaviour has not yet been investigated extensively. The purpose of our research was to analyse the histological spectrum and Correspondence: Alessandro Poli, Department of Animal Pathology, School of Veterinary Medicine, University of Pisa, Viale delle Piagge, 2 I-56124 Pisa, Italy. Fax: 39-50-540644; E-mail: apoli@vet.unipi.it