Objectives: To evaluate the relation between age, prostate-specific antigen (PSA) level, and the probability of detecting curable prostate cancer.
Methods: A consecutive surgical series of radical prostatectomies was performed in 492 men with nonpalpable (Stage T1c) disease, who formed the study cohort (mean age 58 years). The cohort was systematically classified into three age groups ( 40 to 50 years [n = 69];  51 to 60 years [n = 227];  61 to 73 years [n = 196]) and five pretreatment PSA groups ( 2.5 to 4.0 ng/mL [n = 36];  4.1 to 6.0 ng/mL [n = 100];  6.1 to 8.0 ng/mL [n = 122];  8.1 to 10.0 ng/mL [n = 76];  greater than 10.0 ng/mL [n = 135]). The percent probability of curable cancer was determined by logistic regression analysis. Curable cancer was defined as organ-confined tumor or a tumor with capsular penetration of low grade (Gleason score less than 7), with negative margins and no involvement of seminal vesicles or lymph nodes.
Results: A comparison within age groups and within PSA groups revealed that the probability of curable cancer was more closely associated with age than PSA level. Although the probability of curable cancer was closely related to PSA level across a wide range of PSA values (0.4 to 52 ng/mL), there was only minimal difference (2% to 4%) in the probability of curable cancer within the range of PSA values from 2.5 to 6.0 ng/mL for all ages.
Conclusions: Age is a strong predictor of the probability of curable cancer. Thus, early detection efforts in younger men are more likely to lead to a decrease in prostate cancer mortality. These data suggest that the use of PSA thresholds below 4.0 ng/mL as an indicator of prostate cancer is unlikely to improve the probability of detecting curable cancer.