Purpose: Although showing a 25% reduction in the biopsy prevalence of cancer compared with placebo in the Prostate Cancer Prevention Trial, finasteride was associated with a higher prevalence of high-grade disease. This observation was driven by "for-cause" biopsies. We sought to understand how volume-dependent changes in prostate-specific antigen test performance characteristics may have contributed. Experimental Design: A retrospective review was done on 1,304 men referred for initial biopsy with a prostate-specific antigen between 4 and 10 ng/mL or an abnormal digital rectal examination. Receiver-operator curves and positive predictive values were ascertained for prostate-specific antigen stratified by diagnosis and prostate volume. Results: The performance of prostate-specific antigen changed for any and high-grade (Gleason, >= 3 + 4) cancer in a volume-specific manner. For any cancer, the area under the curve (AUC) decreased from 0.758 to 0.629 to 0.520 as prostate volume increased (<30, 30-50, >50 cm(3), respectively). For high-grade cancer, a similar trend was shown (AUC, 0.712, 0.639, and 0.497, respectively). The positive predictive value of a prostate-specific antigen of >= 4 ng/mL was also affected by prostate volume. Trends for Gleason <= 6 decreased as prostate volume increased (positive predictive value for <30 cm(3), 25.0%; positive predictive value for 30-50 cm(3), 23.8%; and positive predictive value for) >50 cm(3), 17.3%). A more significant trend was seen for high-grade cancer (positive predictive value for <30 cm(3), 39.0%; positive predictive value for 30-50 cm(3), 22.3%; and positive predictive value for) >50 cm(3), 10.7%). Conclusion: Decreases in prostate volume over time and the resultant change in prostate-specific antigen performance characteristics may have contributed a bias toward the detection of high-grade disease in the finasteride arm of the Prostate Cancer Prevention Trial.