The treatment of complex atypical hyperplasia (CAH) and well-differentiated endometrioid carcinoma (WDC) by progestin therapy has been shown to be a safe alternative to hysterectomy. Accurate assessment for regressive changes induced by the progestins is critical to successful treatment. However, there are few studies detailing the histopathologic changes associated with progestin therapy. A total of 44 patients with CAH or WDC, treated with oral progestins or a progesterone or levonorgestrel-releasing intrauterine device, were followed by endometrial biopsy and/or curettage at 3 to 6 month intervals for a maximum of 25 months. The pretreatment and posttreatment endometrial samples were evaluated for response to treatment and for the histologic features of gland-to-stroma ratio, architectural abnormalities [back-to-back glands and confluency (cribriform and/or papillary patterns)], glandular cellularity, mitotic activity, cytologic atypia, and cytoplasmic changes. Histologic changes seen in progestin-treated endometria included a decreased gland-to-stroma ratio, decreased glandular cellularity, decreased to absent mitotic activity, loss of cytologic atypia, and a variety of cytoplasmic changes including mucinous, secretory, squamous, and eosinophilic metaplasia. Architectural changes tended to resolve later in the course of treatment. Some architectural abnormalities, specifically cribriform and papillary patterns, were induced by progestins mimicking progression. Twelve (67%) of 18 women with CAH had complete resolution, 2 (11%) regressed to complex hyperplasia without atypia, and 4 (22%) demonstrated persistent disease over a median follow-up period of 11 months. Eleven (42%) of 26 women with WDC had complete resolution and 15 (58%) had persistent disease over a median follow-up period of 12 months. Three instances of disease progression occurred, presumably only after discontinuing progestin treatment. Only persistent architectural abnormalities and/or cytologic atypia in the 7 to 9-month biopsies were predictive of treatment failure, with a trend for cytologic atypia to be the most powerful predictor. These findings indicate that progestin therapy should be continued for no less than 6 months to accurately assess treatment response. A modified classification for progestin-treated lesions of the endometrium is proposed.