OBJECTIVE To determine both how the diagnosis of an atypical biopsy influences a urologist's decision to repeat the biopsy and the outcome of rebiopsy. METHODS Of 200 atypical biopsies that we confirmed from outside consultations to the Johns Hopkins Hospital from 1992 to 1993, we were able to retrieve follow-up information for 144 cases. Each atypical biopsy was evaluated for the reason for atypia (atrophic glands, rule out [r/o] adenosis, atypical not otherwise specified [NOS; insufficient cytologic and/or architectural atypia], r/o prostatic intraepithelial neoplasia [PIN], inflammation, crush artifact) and a favored diagnosis (cancerous, benign, and undetermined). RESULTS Of the 144 atypical biopsies, 92 were rebiopsied (63.9%). The time from the initial atypical biopsy to rebiopsy ranged from 0.5 months to 3 years (63% less than 6 months; 39% less than 3 months). Rebiopsy revealed carcinoma in 48.9%, benign in 38%, atypical in 8.7%, and PIN in 4.4%. The median prostate-specific antigen (PSA) value was lower in men who did not undergo a repeat biopsy (6 versus 7.8) (rank sum analysis, P = 0.04). No correlation was found between PSA level and results of the rebiopsy. Of the atypical biopsies in which cancer was favored, 61% were cancerous on rebiopsy versus 33% where a benign process was favored. The three reasons for atypical biopsies that seemed to correlate with outcome of rebiopsy were atypical NOS (68% cancer on rebiopsy); inflammation (63% cancer on rebiopsy); and r/o adenosis (36% cancer on rebiopsy). CONCLUSIONS Although 48.9% of the rebiopsied cases were cancerous, only 63% of men underwent rebiopsy, raising a concern that cancers are being missed in those cases not rebiopsied after an atypical diagnosis. Although there was a trend for serum PSA to correlate with outcome of rebiopsy, this correlation was not significant, and even men with serum PSA less than 4 ng/mL had a 33% risk of cancer on rebiopsy. Although histologic features of the atypical foci may be useful as factors in determining the urgency for rebiopsy, they also were not statistically significant in predicting outcome. Men with atypical diagnoses should undergo rebiopsy regardless of serum PSA levels and regardless of why the lesions were atypical.