Objective: The grade of the prostate cancer is an important factor in defining prognosis and deciding on treatment. In this study, we compared the Gleason score determined by 18-gauge core needle biopsies with both the Gleason score and pathological staging of the radical prostatectomy specimens. Patients and Methods: Between July 1992 and September 1998, we performed 144 radical retropubic prostatectomies for clinically localized prostatic carcinoma, after a negative frozen section in bilateral pelvic lymphadenectomy in all cases. Ten patients with pathologic stage T1a and T1b were excluded. The final study group consisted of 134 patients, all of whom had been diagnosed with adenocarcinoma by transrectal needle biopsies with an 18-gauge automated spring-loaded biopsy gun. No patients received neoadjuvant therapy, including androgen deprivation and radiation therapy. All patients had a designated Gleason score on the needle biopsy and prostatectomy specimens. Results: We found that grading error was greatest with well-differentiated (Gleason score 2-4) tumors. The accuracy was 15% for Gleason score 2-4 on needle biopsy. Of the 113 evaluable patients with Gleason score 5-7 on needle biopsy, 110 (97%) were graded correctly. All of the Gleason score 8-10 on needle biopsy was graded correctly. But only 1 patient in our series had Gleason score 8 on needle biopsy. Twenty-seven (25%) of 110 patients with a biopsy grade of Gleason score <7 had the cancer upgraded to 7. Of patients with both Gleason score <7 in the needle biopsy and Gleason score 7 in the prostatectomy specimen, only 3 (11%) had tumor confined to the prostate. Conclusion: The potential for grading error is greatest with well-differentiated tumors and of patients with both Gleason scores <7 in the needle biopsy and Gleason score 7 in the prostatectomy specimen, only 11% had tumor confined to the prostate. This effects treatment policy, especially for watchful waiting criteria. Copyright (C) 2000 S. Karger AG, Basel.