BACKGROUND: HER-2/neu gene amplification, established as a prognostic factor in breast carcinoma and other cancers, has not been correlated with outcome in prostate carcinomas (PCs). METHODS: HER-2/neu gene amplification was determined by automated fluorescence in situ hybridization (FISH) using a unique sequence cosmid probe on 113 formalin fixed, paraffin embedded 4-microns tissue sections and the results compared with tumor grade, DNA ploidy, HER-2/neu protein expression by immunohistochemistry (IHC), serum prostate specific antigen, pathologic stage, and postoperative disease recurrence (mean follow-up of 44 months). RESULTS: HER-2/neu gene amplification by FISH (41% of PCs) correlated with tumor grade (P = 0.001) and DNA ploidy status (P = 0.0003). HER-2/neu protein overexpression by IHC (29% of PCs) correlated with grade (P = 0.03), but not with DNA ploidy. A trend for similar HER-2/neu status in each PC by IHC and FISH did not reach statistical significance (P = 0.25). On univariate analysis, HER-2/neu amplification by FISH (P = 0.029), tumor grade (P = 0.013), and DNA ploidy (P = 0.016) correlated with postoperative disease recurrence. HER-2/neu expression by IHC did not correlate with outcome. On multivariate analysis, grade (P = 0.0001) and ploidy (P = 0.001) were independent outcome predictors; HER-2/neu amplification by FISH reached near-independent significance (P = 0.125). CONCLUSIONS: HER-2/neu gene amplification by FISH on archival PCs significantly correlates with grade and DNA ploidy status, is more sensitive than IHC in detecting HER-2/neu gene abnormalities, predicts postoperative disease recurrence, and may prove important in planning therapy for patients with prostate carcinoma.