Endometrial carcinoma is the most common gynaecological malignancy. Progestins are widely used in the treatment of advanced and metastatic disease with about 20-40% response rate. Attempts to develop a predictive test for progestin-sensitivity of endometrial cancers are plagued by the problems of steroid receptor instability and the heterogeneous distribution of progesterone receptor. Successful development of a nude mouse model for human endometrial carcinoma has permitted a detailed investigation of the biological behaviour, hormonal modulation and resistance to progestin therapy. Use of this preclinical model should enhance our understanding of the hormonal mechanisms and lead to improved rational treatment strategies for this malignancy.