5008 Introduction: Risk stratification in localized prostate cancer is based on PSA level, Gleason score and T stage. PSA derivatives may be of additional prognostic value. We have assessed the use of PSA derivatives in the prediction of histologic disease progression on repeat biopsy in patients on active surveillance. Methods: In a prospective cohort study of active surveillance, men with untreated, low- and intermediate-risk, clinically localized prostate cancer (T1/2a, PSA < 15 ng/ml, Gleason score ≤ 3 + 4, and % positive cores ≤ 50%) had repeat octant prostate biopsies 18–24 months after diagnosis. Histologic progression was defined as primary Gleason pattern ≥ 4, or % positive cores > 50 %, or an increase in Gleason score from ≤ 6 to ≥ 7. Standard clinical variables and PSA derivatives were analysed with respect to histologic progression. Multivariate analysis was performed including all standard variables both with each PSA derivative individually, and with all PSA derivatives. Results: 175 men had repeat biopsy. Median age was 67 yrs and median initial PSA, 6.5 ng/ml. Histologic disease progression was seen in 50 (29%) cases. Factors associated with histologic progression on univariate analysis were PSA velocity (p=0.0001), PSA density (p=0.0002), free-total PSA ratio (p=0.002), maximal % core involvement (p=0.009), age (p=0.003), and initial PSA level (p=0.02). On multivariate analysis, each PSA derivative was significant, independent of standard clinical variables. On combined multivariate analysis, free-total PSA ratio (p=0.0027), PSA velocity (p=0.009), age (p=0.0057) and maximal % core involvement (p=0.011) were significant determinants of histologic progression. Conclusions: Free-total PSA ratio and PSA velocity may be useful in the subclassification of favourable-risk localized disease. Larger validation studies in active surveillance patients are warranted. No significant financial relationships to disclose.