The purpose of this study was to determine the accuracy of Gleason scores in prostate needle biopsy diagnosis and to investigate factors affecting the accuracy of the tumor grade. A single pathologist reviewed 116 sets of prostate cancer biopsies and radical prostatectomy specimens. The following factors were examined to determine their effect on the accuracy of the biopsy Gleason scores: (i) relative tumor differentiation; (ii) pathological stage; (iii) amount of tissue in the biopsy specimen; (iv) amount of cancer tissue in the biopsy specimen; (v) tumor heterogeneity; (vi) clinical findings (prostate specific antigen value and digital rectal examination); and (vii) interobserver variability. In 53 cases the Gleason score of biopsy specimens was identical to the score of prostatectomy specimens (45.7%). Fifty-four cases (46.6%) of biopsy specimens were undergraded. The most common discrepancy was diagnosis of well-differentiated carcinoma in the biopsy but diagnosis of moderately differentiated tumor in the corresponding prostatectomy specimen. This discrepancy occurred when the amount of tumor in the biopsy was 3 mm or less. Biopsy and prostatectomy results showed less agreement when the original biopsy tumor grade rendered by nine different pathologists was used, suggesting that interobserver variability can adversely affect the accuracy of tumor grade. Clarifying the histologic criteria for distinguishing each grade, especially between Gleason grades 2 and 3, is important for accurate grading.