A 23-year-old woman presented to an emergency unit with severe anaemia. She had been taking oral contraceptives for 6 months because of intermittent vaginal bleeding which was attributed to benign menstrual disorders. Her haemoglobin concentration was 5 g/dL, which was corrected by a blood transfusion; coagulation and other blood parameters were normal. Further examination with transvaginal echography showed a mass in the uterus. Cytological analysis of cells taken from inside the cervix indicated the presence of a malignant growth, and 1 month later the patient had a hysterectomy with dissection of the pelvic and para-aortic lymph nodes. Examination of the excised tissue with histological and immunohistochemical techniques revealed that the patient had malignant melanoma of the endometrium with necrosis, parametrial infiltration, and para-aortic lymph-node metastases. We didn’t have access to images used for staging before surgery, so were unable to confirm whether the extensive pathological invasion corresponded with previous radiological assessments. However, adjuvant treatment with interferon, interleukin, and cisplatin was given on the basis of the initial histological diagnosis. Although the patient was told to expect tumoural progression within a few months, no suspicious lesions had been identified with computed tomography after a year. Therefore, despite feeling healthy, she sought a second opinion. During her initial clinical examination the patient said she had no history of malignant melanoma and, apart from laparotomy scars, a physical examination did not detect any skin lesions or other abnormalities. We then decided to do a full-body scan with fluorodeoxyglucose-18 positron emission tomography (FDG-PET), which showed no signs of metabolic tumour activity outside the pelvis. Consequently, we decided that samples taken from the hysterectomy should be systematically reviewed. Light microscopy (figure 1) showed an organoid growth pattern of large epithelioid cells with abundant eosinophilic cytoplasm. Immunohistochemical staining for the melanogenesisrelated marker HMB45 was positive, but stains for S100 protein, vimentin, and epithelial markers were all negative. We decided that electron microscopy was unnecessary because the histological evidence was very characteristic and our final diagnosis was clear-cell “sugar” tumour of the uterus. The patient agreed to a narrow clinicoradiological follow-up. She is still healthy 30 months after the hysterectomy. Gynaecological malignant melanoma is a rare tumour which is well described in young women and becomes rapidly