The correct identification of HER2-positive cases is a key point to provide the most appropriate therapy to breast cancer (BC) patients. We aimed at investigating the reproducibility and accuracy of HER2 expression by immunohistochemistry (IHC) in a selected series of 35 invasive BC cases across the pathological anatomy laboratories in Tuscany, Italy. Unstained sections of each BC case were sent to 12 participating laboratories. Pathologists were required to score according to the Food and Drug Administration (FDA) four-tier scoring system (0, 1+, 2+, 3+). Sixteen and nineteen cases were HER2 non-amplified and amplified respectively on fluorescence in situ hybridization. Among 192 readings of the 16 HER2 non-amplified samples, 153 (79.7%) were coded as 0 or 1+, 39 (20.3%) were 2+, and none was 3+ (false positive rate 0%). Among 228 readings of the 19 HER2 amplified samples, 56 (24.6%) were scored 0 or 1+, 79 (34.6%) were 2+, and 93 (40.8%) were 3+. The average sensitivity was 75.4%, ranging between 47% and 100%, and the overall false negative rate was 24.6%. Participation of pathological anatomy laboratories performing HER2 testing by IHC in external quality assurance programs should be made mandatory, as the system is able to identify laboratories with suboptimal performance that may need technical advice. Updated 2013 ASCO/CAP recommendations should be adopted as the widening of IHC 2+ "equivocal" category would improve overall accuracy of HER2 testing, as more cases would be classified in this category and, consequently, tested with an in situ hybridisation method.