Impact of PCA3 on the number of repeat prostate biopsies and sensitivity to detect high-grade cancer in the placebo cohort of the REDUCE study.

Abstract

111 Background: 10-35% of men with a negative initial biopsy (Bx) have prostate cancer (PCa) found on repeat Bx (rBx). Fear of missing significant cancer frequently results in, often negative, rBx. As Bx is associated with pain, anxiety and complications, a better individualization of the decision for rBx to reduce the number of rBx and overdiagnosis of indolent PCa is needed. Recently, we used the RAND Appropriateness Method (RAM) to develop a model for rBx patient selection, based on expert recommendations and including the PROGENSA PCA3 Assay (Tombal B, et al. World J Urol 2011;doi:10.1007/s00345-011-0721-0). PCA3 has been shown to predict the probability that a rBx will be positive and may be indicative of PCa significance (Aubin SMJ, et al. J Urol 2010;184:1947-52). In the current analysis, we tested the expert recommendations model on the placebo cohort of the REDUCE study. Methods: The RAM expert recommendations were applied to the placebo cohort of the REDUCE study in which 1073 men with a baseline PSA 2.5-10 ng/mL and a prior negative Bx had a PCA3 test before the planned 2-year and 4-year rBx2. For each scenario (with and without PCA3), the number of rBx and the number of missed high-grade (Gleason sum ≥ 7) cancers were assessed. Results: Data from 1024 subjects were available for analysis. In the scenario without PCA3 consideration (reflecting the experts’ recommendations using PSA, DRE, number of previous negative Bx, prostate volume and life expectancy), 267 (26%) of study-mandated rBx were considered inappropriate and the number of missed high-grade cancers was 14. The scenario including also PCA3 consideration led to a reduction of the number of rBx (656 or 64%), while the number of missed high-grade cancers was decreased to 8. The diagnostic accuracy was superior in the model with PCA3: sensitivity for high-grade cancer was 85% vs. 75%, specificity 67% vs. 26%, positive predictive value PPV 13% vs. 5%, NPV 99% vs. 95%. Conclusions: Application of RAM expert recommendations for rBx that include the PCA3 Score can reduce the number of rBx for men with an elevated PSA while maintaining the sensitivity to detect high-grade cancer.