Objective: To compare the ability of loss of phosphatase and tensin homolog (PTEN) and Genomic prostate score assay (GPS) in predicting the biochemical-recurrence (BCR) and clinical-recurrence (CR) after radical prostatectomy (RP) for clinically localized prostate cancer (PCa). Methods: Three hundred seventy seven patients with and without CR were retrospectively selected by stratified cohort sampling design from RP database. PTEN status (by immunohistochemistry [IHC] and fluorescence in situ hybridization [FISH]) and GPS results were determined for RP specimens. BCR was defined as Prostate Specific Antigen (PSA) ≥ 0.2 ng/mL or initiation of salvage therapy for a rising PSA. CR was defined as local recurrence and/or distant metastases. Results: Baseline mean age, PSA, and GPS score for the cohort were 61.1 years, 8 ng/dL, and 32.8. PTEN loss was noted in 38% patients by FISH and 25% by IHC. The concordance between FISH and IHC for PTEN loss was 66% (Kappa coefficient 0.278; P <.001). On univariable analysis, loss of PTEN by FISH or IHC was associated with BCR and CR (P <.05). However, after adjusting for GPS results, PTEN loss was not a significant predictor for CR or BCR (P >.1). The GPS result remained strongly associated with CR and BCR after adjusting for PTEN status (P <.001). PTEN status and GPS results only weakly correlated. GPS was widely distributed regardless of PTEN status indicating the biological heterogeneity of PCa even in PTEN-deficient cases. Conclusion: GPS is a significant predictor of aggressive PCa, independent of PTEN status. After adjustment for GPS results, PTEN was not independently associated with recurrence for PCa.