To perform the first evaluation of the effects of the 5-alpha-reductase inhibitor class of drugs on cancer histopathologic features at radical prostatectomy in a placebo-controlled multicenter trial. We analyzed prostatectomy slides in a blinded manner from 17 men treated with dutasteride, an inhibitor of types 1 and 2 isoenzymes of 5-alpha-reductase, and 18 men treated with placebo for 5 to 11 weeks before radical prostatectomy. The histopathologic features of benign epithelium, high-grade prostatic intraepithelial neoplasia, and cancer were recorded, and the treatment effect was scored. Digital imaging analysis was used to measure the stroma/epithelium ratio and epithelial height, as well as the nuclear area in cancer. In benign epithelium, treatment caused distinctive cytoarchitectural changes of atrophy and a decrease in the epithelial height (P = 0.053). The peripheral zone showed the most marked response to treatment. In cancer tissue, the tumor volume was significantly lower in the dutasteride-treated men than in the placebo-treated men (mean 15% versus 24%, respectively, P = 0.025), the percentage of atrophic epithelium was increased (P = 0.041), and the stroma/gland ratio was doubled (P = 0.046). The treatment alteration effect score was doubled (P = 0.055) and did not correlate with any Gleason score changes. After short-term dutasteride treatment, benign epithelium showed involution and epithelial shrinkage, and prostate cancer tissue demonstrated a decrease in epithelium relative to stroma. These findings indicate that dutasteride induces significant phenotypic alterations in both the benign and the neoplastic prostate, supportive of a chemopreventive or chemoactive role.