SELECTION OF OPTIMAL PROSTATE-SPECIFIC ANTIGEN CUTOFFS FOR EARLY DETECTION OF PROSTATE-CANCER - RECEIVER OPERATING CHARACTERISTIC CURVES

A prospective clinical trial of prostate cancer screening was conducte d at 6 university centers including 6,630 men 50 years old or older wh o underwent a serum prostate specific antigen (PSA) determination and digital rectal examination. Biopsies were performed if the PSA level w as greater than 4.0 ng./ml. (Hybritech Tandem assay) or digital rectal examination was suspicious for cancer. We evaluated the effect on bio psy rate and cancer detection if the cutoff value was shifted from 4.0 to age-specific reference ranges recommended in the literature. In me n 50 to 59 years old with normal digital rectal examination findings a decrease from 4.0 to 3.5 ng./ml. would have resulted in a 45% increas e in the number of biopsies (39 of 87) and a projected 15% increase in cancer detection. An increase from 4.0 to 4.5 ng./ml. in men 60 to 69 years old would result in 15% fewer biopsies (35 of 238) and would mi ss 8% of the organ confined tumors (2 of 25). Increasing the cutoff to 6.5 ng./ml. in men 70 years old or older would result in 44% fewer bi opsies (70 of 159) and would miss 47% of the organ confined cancers (7 of 15). The number of biopsies performed for each cancer detected wit h a PSA level of greater than 4.0 ng./ml. remains constant across age groupings, which suggests that the cutoff of 4.0 ng./ml. does not need to be altered in the older men, since it is apparently unaffected by the simultaneously increasing prevalence of benign prostatic hyperplas ia and cancer with age. We conclude that a serum PSA concentration of 4.0 ng./ml. should be used as a general guideline for biopsy in all ag e groups.