Aim-Orbital granulocytic sarcoma is a localised tumour composed of cells of myeloid origin. Histological diagnosis can be diYcult in patients with poorly diVerentiated orbital tumours and no evidence of systemic leukaemia. The naphthol AS-D chloracetate esterase (Leder stain) and immunohistochemical stains for lysozyme and MAC387 were used to determine the staining characteristics of these tumours. A case series of seven patients with orbital granulocytic sarcoma is presented. Methods-Seven patients with orbital granulocytic sarcoma were studied. Haematoxylin and eosin, Leder, and lysozyme stained sections were available in seven cases. Unstained formalin fixed paraYn embedded sections of seven cases were available for immunohistochemical evaluation using the avidin-biotin-complex technique for MAC387. Results-The mean age of presentation of the orbital tumour was 8.8 years. Four patients presented with an orbital tumour before any systemic manifestations of leukaemia. In two cases the diagnosis of the orbital tumour and systemic leukaemia was made simultaneously. There was one case of established systemic myeloid leukaemia in remission with the subsequent development of orbital granulocytic sarcoma. Six of seven cases (86%) were positive for the Leder stain. Five of seven cases (71%) showed positive immunoreactivity with lysozyme. The immunohistochemical stain for MAC387 was positive in all seven cases (100%) including one case that was negative for both lysozyme and Leder stains. Conclusions-Orbital granulocytic sarcoma is a tumour that aVects children and can present with rapidly progressive proptosis. This tumour may develop before, during, or after the occurrence of systemic leukaemia. The combination of Leder and lysozyme stains is useful in the diagnosis of orbital granulocytic sarcoma. MAC387 may be a more reliable marker for orbital granulocytic sarcoma. (Br J Ophthalmol

Human uveal melanoma cell lines OCM-1, SP 6.5, and MKT-BR and the ocular melanocyte cell line UW-1 exhibited maintenance of some structural and ultrastructural characteristics of melanocytic cells. All four MoAbs, PAL-M2, NK1/C3, IND-1, and MAAMA against cutaneous melanoma-associated antigens stained positively all melanoma cell lines as well as the melanocytic cell line, suggesting that in vitro proliferation of melanocytes could modify their antigenic expression.

We investigated the need for continuous immunosuppression to maintain experimental tumours derived from human uveal melanoma cells implanted in the choroid of pigmented rabbits. Two groups of pigmented rabbits immunosuppressed with cyclosporin A (CsA) were implanted with human uveal melanoma cells in the suprachoroidal space. After 5 weeks, CsA was discontinued in group 2. Animals were treated with prophylactic antibiotics and examined weekly for tumour growth, weight and secondary effects; blood urea nitrogen levels were measured every two weeks. Autopsies and histopathological studies were performed after death or euthanasia at the end of week 12. The difference between the groups in the development of ophthalmoscopic tumours was not statistically significant 5 weeks after implantation. Tumours in group 1 grew progressively throughout the experiment, whereas group 2 tumours showed marked regression 3-4 weeks after discontinuing CsA. Tumours in group 1 were significantly larger and had greater mitotic activity and showed more ciliary body, optic nerve and extrascleral invasion than tumours in group 2, which showed massive fibrosis, minimal mitotic activity and marked inflammatory cell infiltration. Continuous immunosuppression with CsA seems to be necessary to maintain tumour growth in this experimental model of uveal melanoma.