Objective. The cyclin-dependent kinase inhibitor p27 has been shown to mediate cell growth arrest in response to various environmental stimuli. p27 protein levels have shown prognostic value in several different types of cancer. We examined the prognostic value of p27 protein expression in endometrial cancer, the most common gynecologic malignancy. Methods. A total of 95 paraffin-embedded tumor blocks were obtained and stained via immunohistochemical techniques with a monoclonal antibody against p27. Ten high-power fields were evaluated per slide with at least 1000 cells per slide and two slides per specimen evaluated by two reviewers for nuclear and cytoplasmic staining. The specimens were evaluated for associations with age, stage, grade, and histology. Statistical analysis was performed using the Student t test, X2 Kaplan-Meier, and likelihood ratios to assess the data and to generate P values. Results. A total of 91 patients met inclusion criteria for statistical analysis. Fifty-three patients were stage I, 13 stage II, 14 stage III and 11 stage IV with a positive stain (>50% of cells) for p27 obtained in 32.1, 23.1, 35.7, and 36.4%, respectively (Student t test P = 0.77). Survival data were available on 24 advanced stage patients, p27 protein immunostaining showed no association with patient survival. We also found no association of p27 staining with age or histology. Notably, we found a trend in increasing staining with increase in grade, particularly with stage I patients. Also, there was an association of the nuclear and cytoplasmic staining and stage (P = 0.05), but it had no correlation with patient survival. Conclusion. Our study showed decreased p27 protein staining in endometrial cancers compared to normal endometrial cells. We found that p27 protein staining shows no association with stage, age, or histology and is not prognostic for survival in advanced endometrial cancers. However, there may be a trend associated with increased p27 protein staining with advanced grades of tumors. © 2001 Academic Press.