We explored the molecular correlates of the effect of finasteride on prostate tissue in patients undergoing radical prostatectomy. Patients undergoing radical prostatectomy for localized prostate cancer were eligible for study. After providing informed consent patients were randomized to receive 5 mg finasteride or placebo daily for at least 30 days before surgery. At surgery prostate tissue was harvested from the surgical specimen and sent for analysis. Tissue samples were analyzed for the pro-apoptotic factors caspase-3, caspase-7 and IGFBP-3. Samples were also analyzed for the tumor suppressor/proto-oncoproteins bcl-2, p53 and p21. Finally, tissues were analyzed for androgen receptor density and insulin growth factor-1. A total of 22 study and 20 placebo samples were collected and analyzed. Negligible staining for bcl-2 or caspase-3 was noted in each group. Statistical differences were not observed for bcl-2, p53, p21 or insulin growth factor-1 between the groups. There was a statistically significant difference in caspase-7 and IGFBP-3. A mean of 77% and 99.9% of cells stained for caspase-7 in the treatment and placebo groups, respectively (p = 0.007). In 3 patients caspase-7 staining disappeared completely and it was decreased by 70% and 50% in 1 patient each. Mean intensity staining for IGFBP-3 was 1.03 in the treatment group and 1.54 in the placebo group (p = 0.005). The staining intensity of nuclear androgen receptors on benign and cancerous cells was not significantly different between the treatment and placebo groups. However, there was a significant difference in androgen receptor staining between benign and cancer cells in the 2 populations. Mean nuclear androgen receptor staining intensity in all cancer and all benign tissue samples was 119.3 and 151.8, respectively (0.001). Finasteride administered 30 days before surgery appears to decrease the apoptotic factors caspase-7 and IGFBP-3 in cancer cells, while having little to no effect on caspase-3, insulin growth factor-1, bcl-2, p53 and p21. This short-term study may have interesting implications for interpreting Prostate Cancer Prevention Trial data on the molecular level. No differences were noted between the treatment and placebo groups in the expression of nuclear androgen receptor. However, decreased expression of androgen receptors was present in cancer cells compared to that in benign prostate cells in the 2 groups.