Background: Finasteride, a 5-alpha reductase inhibitor may have effects on biomarkers such as prostate-specific antigen (PSA) that could be leveraged to improve screening.
Objective: To determine the predictive characteristics of biomarkers for prostate cancer for cancer on biopsy following 3 months of finasteride use compared with placebo.
Design, setting and participants: 383 men from multiple clinical sites with intermediate prostate cancer risk, without history of prostate cancer, were randomly allocated in a double-blinded manner, 4:1, to receive either finasteride or placebo for 90 days at which time a prostate biopsy was performed.
Outcome measurements and statistical analysis: The primary outcomes were associations of biomarkers with prostate cancer that were tested using multiple logistic regression and area under the receiver operating curves (AUC). Biomarkers for PCA risk (PCA3, TMPRSS2:ERG (T2:ERG) gene product, and PSA) were measured at baseline and at biopsy in a blinded fashion to assess the predictive performance of baseline levels, 90-day levels, and measures of change relative to standard predictors.
Results and limitations: A total of 292 (233 finasteride; 59 placebo) randomized patients underwent biopsy and were analyzed. On finasteride, baseline and 90-day measures of PCA3 and T2:ERG had similar moderate discrimination capacity with AUCs 62 to 65% (p-values < 0.001 and 0.001, respectively), but their rates of change had no discrimination ability (AUC 51%, (95% CI 43 to 60% p = 0.72) and 48% (95% CI 44 to 60%, p = 0.62), respectively).) Relative to baseline, the 90-day PCA3 and PSA decreased in the finasteride group by 25% and 50%, respectively (both p<0.001). T2:ERG had a smaller, non-significant change post finasteride treatment (p = 0.08).
Conclusions: Short-term finasteride therapy did not improve performance of the most commonly-employed prostate cancer biomarkers. Threshold values for new biomarkers of prostate cancer should be interpreted with caution in patients receiving finasteride until formal validation of test performance in these patients is conducted.
Patient summary: Three months of finasteride treatment did not increase the accuracy for predicting the outcome on prostate biopsy but did have a significant effect on biomarker values. Adjustments to thresholds for biopsy for men on finasteride are proposed.
Trial registration: ClinicalTrials.gov, NCT01296672.