Objetive To assess the safety and efficacy of FDG-PET in breast cancer in the diagnostic of primary tumours, lymph node staging, the detection of recurrent disease/metastases, and the assessment of chemotherapy treatment. Methods A systematic review was undertaken. A search was made for primary studies, other systematic reviews, and health technology assessment reports in different databases. Results A total of 73 reports were included. FDG-PET does not appear to be sufficiently accurate to be used in isolation for ruling out the presence of a primary tumour. In lymph gland staging, FDG-PET does not appear to be accurate enough to detect occult axillary metastases or micrometastases (sensitivity 20 and 50%, respectively); sentinel node biopsy is required for confirmation. In the detection of bone metastases, FDG-PET should be complemented with other tests such as bone gammagraphy or SPECT. The assessment of response to chemotherapy, there seems to be no uniform criterion for establishing a standardized uptake value (SUV) for FDG that would allow responders and non-responders to be distinguished. Conclusions FDG-PET is insufficiently sensitive to rule out small primary tumours. Due to the high number of false positives returned, it cannot replace axillary dissection in axillary lymph gland staging. A complete biochemical response identified by FDG-PET should not be relied upon to mean an absence of disease since the technique cannot detect residual microscopic elements.