BACKGROUND.
There is debate over whether the recent increases seen in prostate cancer are due to lead-time bias from screening, or identification of clinically insignificant lesions. METHODS. Population-based incidence rates for 1973-1996 were calculated, based on the Surveillance, Epidemiology, and End Results Program (SEER) tumor registries. Relative incidence rates for prostate cancer by stage, fatal incidence, and lymph nodes were calculated, adjusted for age. RESULTS. Localized and regional stage prostate cancer increased through 1992 and then dropped. The rate of distant-stage disease was relatively stable from 1973-1991, with a decrease in distant stage starting in 1992. The 2-year mortality rates were constant for 1973-1989. A decline in the 2-year mortality among cases (fatal incidence) also began in 1992. CONCLUSIONS. These data show large increases in early disease, followed by a drop and leveling off along with a decrease in advanced disease (distant stage, 2-year mortality, positive lymph nodes). This indicates that the increasing incidence rates for prostate cancer are largely due to lead-time bias from increased early detection and treatment of prostate cancer. However, since incidence rates have not declined to rates seen in the 1970s, the additional cases may also reflect length bias from insignificant lesions or a true increase in incidence over time.