Because prostate cancer has a long latency and high incidence, it is a good
target for chemoprevention by agents such as retinoids, antiandrogens, ant
iestrogens, and vitamin D analogs. Phase II chemoprevention trials are freq
uently conducted on cohorts of patients with previous cancers or premaligna
nt lesions who are scheduled for prostate cancer surgery; such trials are c
urrently in progress with several agents. Prostatic intraepithelial neoplas
ia (PIN) can be used as a surrogate endpoint biomarker for prostate cancer
incidence. Studies of men with high-grade PIN (HGPIN) are particularly usef
ul in that they require a much smaller cohort of 200-400 patients instead o
f the 18 000 patients required for typical Phase III trials. Even with a sm
aller sample size, statistically significant evidence of cancer prevention
is achieved due to the high probability of HGPIN progressing to cancer (35-
55%). A Bayesian sequential monitoring system allows interim analysis of bi
omarker modulation as early as the completion of 30 patients. Putting all t
hese strategies together will help inhibit, delay, or modulate the natural
history of prostate carcinogenesis.