The grading system for prostate carcinoma devised by Gleason is a strong prognostic indicator. The primary and secondary patterns are combined to give a tumor score, referred to as Gleason score or sum. Gleason scores on biopsy correlate with the prostatectomy Gleason scores, and in combination with pretreatment serum prostate-specific antigen and digital rectal examination results, predict tumor stage and lymph node status. However, when only a minute focus of tumor is present on biopsy, the Gleason score is assigned by doubling the Gleason pattern. The goal of this study was to determine if a Gleason score assigned to a minimal focus of adenocarcinoma had predictive value. Paired biopsies and prostatectomy specimens from 963 cases of men with clinically localized prostate cancer were examined. Minimal tumor on biopsy was defined as less than 1 mm or 5% involvement of one biopsy core; excluded from this definition were biopsies where two Gleason patterns could be identified and/or tumor was seen on more than one biopsy core. Terms often used to describe these lesions include "single minute focus of carcinoma" or "adenocarcinoma, too small to give a Gleason grade." One hundred five cases (10.9%) met the above criteria for minimal carcinoma. The correlation of Gleason scores between biopsies and prostatectomy specimens overall was good with exact agreement for 57% of cases and a difference of +/-1 unit in 92% of cases. The correlation for the minimal tumors on biopsy and prostatectomy was slightly worse with exact agreement in 52.4% (55 of 105) and a difference of +/-1 unit in 87.6% (92 of 105). The majority of minimal tumors (83.8% or 88 of 105) were assigned a Gleason score of 6. A total of 31.8% of these 88 cases were upgraded and 5.7% were downgraded. Multivariate analysis on all cases looking for predictors of tumor stage found biopsy Gleason score, perineural invasion, pretreatment prostatic-specific antigen, and digital rectal examination all predicted higher tumor stage with odds ratios of 1.86 (95% confidence interval [CI], 1.53-2.27; p = 0.0001), 2.06 (95% CI, 1.43-2.95; p = 0.0001), 1.08 (95% CI, 1.05-1.11; p = 0.0001), and 1.41 (95% CI, 1.04-1.91; p = 0.0289), respectively. In a model restricted to the 105 cases with minimal carcinoma, pretreatment prostatic-specific antigen was the only independent predictor of higher tumor stage with an odds ratio of 1.15 (95% CI, 1.01-1.31; p = 0.0380); Gleason score was not found to significantly predict higher tumor stage (odds ratio, 1.156; p = 0.6680). The results of this study confirm that biopsy Gleason score in most cases predicts prostatectomy Gleason score and tumor stage. However, for cases with minimal tumor on biopsy, the assigned Gleason score did not predict tumor stage. To properly convey this uncertainty to clinicians, a cautionary note should accompany Gleason scores derived from a minimal focus of carcinoma.