We analyzed concordance between immunohistochemical analysis and fluorescence in situ hybridization (FISH) in HER-2 status and studied the effect of dual-color (D-FISH) vs single-color FISH (S-FISH) scoring on the assignment of tumors to amplified or nonamplified categories. The assays were performed on formalin-fixed, paraffin-embedded sections of 2,279 invasive breast carcinomas. Immunohistochemical results were interpreted as negative (0, 1+) or positive (2+, 3+). For FISH analyses, a ratio for HER-2/chromosome 17 of 2.0 or more (D-FISH) or an absolute HER-2 copy number per nucleus of more than 4.0 (S-FISH) were interpreted as positive gene amplification. We found 547 (24.0%) cases positive immunohistochemically, 326 (14.3%) by D-FISH, and 351 (15.4%) by S-FISH. Overall concordance in HER-2 status with immunohistochemical analysis was 87% for D-FISH and 86% for S-FISH. Excellent concordance was found among groups scored immunohistochemically as 0, 1+, and 3+ (with D-FISH, 97%; with S-FISH, 96%). The most discordant category was the group scored 2+ immunohistochemically, in which only a quarter of the 2+ tumors were FISH(+). D-FISH and S-FISH scoring results were discordant in 89 tumors (4%), of which 8 (9%) had 3+ immunohistochemical staining and none showed high-level HER-2 amplification. Among all FISH(+) tumors, 10% were negative by immunohistochemical analysis, and notably almost half (47%) showed borderline to low HER-2 amplification (D-FISH score, 2.0-3.9); the clinical significance of these findings warrants further investigation.