OBJECTIVE: To evaluate the concordance between the Gleason scores of prostate biopsies and radical prostatectomy specimens, thereby highlighting the importance of the prostate-specific antigen (PSA) level as a predictive factor of concordance. METHODS: We retrospectively analyzed 253 radical prostatectomy cases performed between 2006 and 2011. The patients were divided into 4 groups for the data analysis and dichotomized according to the preoperative PSA, <10 ng/mL and ≥10 ng/mL. A p-score <0.05 was considered significant. RESULTS: The average patient age was 63.3±7.8 years. The median PSA level was 9.3±4.9 ng/mL. The overall concordance between the Gleason scores was 52%. Patients presented preoperative PSA levels <10 ng/mL in 153 of 235 cases (65%) and ≥10 ng/mL in 82 of 235 cases (35%). The Gleason scores were identical in 86 of 153 cases (56%) in the <10 ng/mL group and 36 of 82 (44%) cases in the ≥10 ng/mL group (p = 0.017). The biopsy underestimated the Gleason score in 45 (30%) patients in the <10 ng/mL group and 38 (46%) patients in the ≥10 ng/mL (p = 0.243). Specifically, the patients with Gleason 3 + 3 scores according to the biopsies demonstrated global concordance in 56 of 110 cases (51%). In this group, the patients with preoperative PSA levels <10 ng/dL had higher concordance than those with preoperative PSA levels ≥10 ng/dL (61% x 23%, p = 0.023), which resulted in 77% upgrading after surgery in those patients with PSA levels ≥10 ng/dl. CONCLUSION: The Gleason scores of needle prostate biopsies and those of the surgical specimens were concordant in approximately half of the global sample. The preoperative PSA level was a strong predictor of discrepancy and might improve the identification of those patients who tended to be upgraded after surgery, particularly in patients with Gleason scores of 3 + 3 in the prostate biopsy and preoperative PSA levels ≥10 ng/mL.