ABSTRACT
BACKGROUND: Controversial results have shown a significant association with either low or high total testosterone (tT) levels and high risk prostate cancer (PCa). We tested the relationship between circulating tT and grade group 5 (G5) PCa at radical prostatectomy (RP) in patients with preoperative low- to intermediate-risk disease. METHODS: Serum sex hormones were assessed the day before RP in a cohort of 846 patients with low- to intermediate-risk PCa. Patients were segregated using the new 5-tiered Gleason grade groups. Restricted cubic spline functions and logistic regression analyses tested the association between sex hormones and G5 PCa. Differences in potential predictive accuracy (PA) were assessed for tT and prostate-specific antigen (PSA) levels. RESULTS: Overall, 27 men (3.2%) had G5 PCa at RP, and this group had higher PSA values than patients with G1-G4 PCa (P = 0.02). The groups did not differ in terms of preoperative mean hormonal values. Both low and high circulating tT values depicted a nonlinear U-shaped correlation with G5 PCa at RP. The lowest and highest (10th and 90th percentiles) tT values and biopsy PCa grade emerged as multivariable independent predictors of G5 PCa at RP (all P < 0.05). PA for G5 PCa did not differ between tT (area under the curve [AUC] 0.631) and PSA (AUC 0.636). CONCLUSIONS: Circulating tT was a significant predictor of G5 PCa at RP in patients with preoperative low- to intermediate-risk disease. Preoperative tT and PSA values showed similar PA for the most aggressive disease, confirming a potential role for circulating androgens in preoperative risk assessment of PCa patients. Prostate 77:234-241, 2017. © 2016 Wiley Periodicals, Inc.
