Data on 2436 primary breast carcinomas diagnosed between 1992 and 2006 were collected to evaluate the rate of frozen section procedures performed over time. Frozen section procedures performed to evaluate resection margins for conservative surgery or sentinel node status were excluded. Over time, there was a decrease in the use of frozen sections indistinctly extended to all pT cancer categories. The rate of cancers diagnosed with frozen sections was 51.2% in 1999, and 0% in 2005-2006. In the same period, the adoption of cytology and core biopsy for breast cancer diagnosis increased from 40% in 1992 to more than 90% since 1999. In an audited diagnostic activity on breast pathology, the routine use of frozen sections on primary lesions was considered inappropriate, particularly in assessment of clinically non-palpable lesions, and should be limited to cases with inadequate pre-surgical sampling.