Serum prostate-specific antigen (PSA) levels and the biopsy Gleason sum are used along with clinical staging to predict prostatectomy pathology results for men with localized prostate cancer. The additional predictive value of perineural invasion (PNI) in pretreatment prostate needle biopsies for evaluating tumor stage in this setting is controversial. The current study evaluates the independent predictive value of PNI for tumor staging in a cohort of 632 men who underwent radical retropubic prostatectomies for clinically localized adenocarcinoma of the prostate between the years 1994 and 1998. None of these men received hormonal or radiation therapy before surgery. In addition to the Gleason sum, biopsy results contained detailed information regarding tumor burden: 1) total number of biopsy cores involved by adenocarcinoma, 2) greatest percentage of any single biopsy involved by prostate carcinoma (GPC), and 3) total percentage of cancer added over all cores (TPC). The presence or absence of any PNI was recorded. Pretreatment factors were analyzed in a univariate and multivariate fashion to determine their predictive value using the TNM tumor stage (pT2 vs pT3) and the modified tumor staging system, which includes surgical margin status (pT2 vs pT3 or positive surgical margin) as end points. Univariate analysis revealed a significant association between pT3 disease and several preoperative factors including age, Gleason sum, serum PSA, digital rectal examination, PNI, GPC, TPC, and the total number of positive cores (p <0.01). Multivariate analysis indicated that serum PSA, Gleason sum, age, and GPC contributed significantly to predicting pT3 disease with odds ratios of 2.7 (95% CI, 1.7-4.3), 2.3 (95% CI, 1.7-3.1), 1.7 (95% CI, 1.1-2.7), and 1.7 (95% CI, 1.4-2.1) respectively. PNI was significant in multivariate analysis only when GPC and TPC were not considered, due to a significant interaction between GPC and PNI (p <0.0001, Wilcoxon's rank sum test). These predictive factors showed a similar relationship to adverse pathology when an alternative definition of adverse pathology was used that included positive surgical margins (pT3 or any positive margin). In the interaction between GPC and PNI, GPC was more significant than PNI in predicting pT3 disease. However, PNI added additional information when adverse pathology was defined more broadly as pT3 or any positive margin.