The prognostic significance of perineural invasion by prostate cancer is debated. We have evaluated the association between biochemical failure and measurements of perineural invasion in radical prostatectomy specimens. Perineural invasion was identified in sections using S-100 protein immunostaining. For nerves showing invasion, the involved nerve closest to the edge of the prostate and to the surgical excision margin, as well as the diameter of these nerves, the largest nerve showing perineural invasion and its proximity to the excision margin, and the percentage of nerves showing perineural invasion up to 1.75 mm from the excision margin was determined and tested against time to prostate-specific antigen failure, along with preoperative prostate-specific antigen levels, highest Gleason primary grade, Gleason score and TNM T category. Perineural invasion was present in 90% of cases, with extraprostatic perineural invasion in 25% of tumors. Diameter of nerves showing perineural invasion ranged from 11 to 680 microm and the shortest distance to the surgical excision margin ranged from 33 to 2.57 mm. Perineural invasion density ranged from 6 to 96%. Gleason scores were six in 58 cases, seven in 43 cases, eight in three cases and nine in one case. Clinical T categories were T1c in 75 cases, T2a in 22 cases, T2b in five cases, T2c in two cases, T3 in one case. During a follow-up period of 11 to 55 months (median 26 months), 27 patients showed prostate-specific antigen failure. On univariate analysis only presence of extraprostatic perineural invasion, among parameters of perineural invasion, showed a weak correlation with outcome, while on multivariate analysis this lost significance and preoperative prostate-specific antigen levels, Gleason score and excision margin status were independently associated with biochemical failure. We conclude that the investigated parameters of perineural invasion do not predict prostate-specific antigen recurrence in radical prostatectomy specimens.