Abstract Background: Prostate cancer (PC) is an ideal target for chemoprevention, with high-grade prostatic intraepithelial neoplasia (HGPIN) as its most likely precursor finasteride, a 5α-reductase inhibitor, reduced prostate cancer incidence by 25% in average-risk men. However, a higher proportion of high-grade PC has been noted whose biological significance is unclear. An alternative targeted approach is the use of androgen receptor antagonists such as flutamide. Presurgical window-of-opportunity trials have been proposed as a model to assess the activity of preventive interventions in a cost-effective manner in PC. In this study we test the effect of minimal low-dose flutamide exposure or finasteride in comparison to placebo on biomarkers associated with PC, HGPIN, and normal prostate tissue. Materials and Methods: This is a prospective randomized phase IIB placebo-controlled double-blind presurgical trial of finasteride 5 mg/day (FIN) vs. flutamide 250 mg/day (FLU) vs. placebo (PLA) in men with PC. Patients with biopsy proven PC, clinical stage ≤T2, candidates to radical retropubic prostatectomy were enrolled. Drugs and placebo were administered for 4 to 6 weeks before surgery. A sample size of 120 patients (40 per arm) was calculated to detect significant differences on change in nuclear area (NA) of epithelial cells between treatment and placebo arms (primary endpoint). To establish a direction of NA change from normal to malignancy a discriminant function derived from differences in morphometric parameters in the pretreatment biopsies between the normal epithelium and the PC tissue was applied to the posttreatment groups. Secondary endpoints include change in proliferation markers (Ki67 labeling index, topoisomerase-IIα), in serum PSA and hormones, safety, and toxicity. Results: 125 patients were randomized at four cooperating institutions (41 FIN, 42 FLU, and 42 PLA). Mean age and PSA at baseline were 64 years and 8.7 ng/mL, respectively, without differences among arms. There were significant differences in NA and Ki67LI of normal and neoplastic cells both before and after treatment (NA interquartile range: 27-40µ2 in normal tissue, 30-50µ2 in HGPIN and 36-55µ2 in cancer tissue) but no differences were detected among arms. Image and discriminant analysis performed in a subgroup of patients (16 FIN, 24 FLU, and 20 PLA) showed in the PLA and FLU arms a higher discriminant function score after treatment, with a significant increase of the mean score by 90% and 50%, respectively, meaning that cases treated with placebo or flutamide undergo a statistically significant further progression of karyometric features. Conversely, FIN-treated PC exhibited only a small further progressive change, with a net increase of the mean score by only 8%. LH significantly increased in the FLU arm; PSA and fPSA decreased while T increased significantly in both treatment arms compared to placebo. Treatment was well tolerated. Conclusion: NA and KI67LI were significantly different among normal, HGPIN and PC tissue and between pre- and posttreatment biopsies. Discriminant analysis detected an inhibition of progressive karyometric pattern in patients treated with finasteride relative to low-dose flutamide and placebo. Our results support the use of nuclear changes as measured by karyometry in presurgical models to assess the activity of preventive agents. This talk is also presented as Poster A68. Citation Information: Cancer Prev Res 2010;3(12 Suppl):PR-03.