The automated spring-loaded 18-gauge gun recently introduced for prostatic needle biopsy provides less than half the tissue of the traditional 14-gauge biopsy, possibly influencing the accuracy and predictive ability of biopsy tumor grade. In order to determine the value of tumor grade in contemporary needle biopsy specimens, we compared grade in 316 biopsies with matched whole-mounted radical retropubic prostatectomy specimens according to Gleason primary pattern, Gleason secondary pattern, Gleason score, percent of Gleason patterns 4 and 5, and nuclear grade. Biopsy grading accuracy was correlated with rates of capsular perforation, seminal vesicle invasion, pelvic lymph node metastasis, serum prostate specific antigen level, prostatic volume, prostatic weight, cancer volume, perineural invasion, DNA ploidy, and pathologic stage. The greatest grading error was encountered with low-grade tumors; there was no correlation of grading error with clinical staging error or other pathologic factors. Significant differences were noted between biopsy and prostatectomy for Gleason primary pattern, secondary pattern, and score. The percent of poorly differentiated carcinoma (Gleason patterns 4 and 5) in biopsies and prostatectomies showed a moderate positive correlation. The results indicate that the accuracy of 18-gauge needle biopsy in predicting tumor grade in the prostatectomy is similar to that reported with 14-gauge biopsies. Based on these findings, we recommend that the Gleason score (sum of primary and secondary patterns) be employed in all needle biopsies, recognizing that the accuracy of grade is decreased in cases with low-grade cancer and small amounts of cancer.